[time-nuts] A/D specs (was EFC tracking)
jim...@earthlink.net said: > So, over the, say, 100 Hz and up range, they're probably pretty good. I like to think I'm pretty good at reading data sheets, but when it comes to modern A/D chips, I'm not so sure. 20 years ago, the specs were for DC and you hoped it did something sensible at high frequencies. Now the specs for many chips are for AC (typically they show a FFT plot) and you are in trouble if you want to use it for DC. Typical audio A/Ds are oversampling with a low pass filter implemented in DSP. This makes the anti-aliasing filter a lot simpler/cheaper. Has anybody tested modern audio A/Ds at very low frequencies? Is there some simple way I should be thinking about this tangle? Has anybody tried adding a signal, say 1 KHz, and filtering it out? I'm fishing for something along the lines of the old chopper amplifiers - shift things to a mode that is known to work well and then filter out the junk you don't like. -- These are my opinions, not necessarily my employer's. I hate spam. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Didier Juges wrote: I am not sure how to translate the IMD specs into integral or differential non-linearity, but from what I have seen, IMD specs are not significantly better for 24 bit sound cards than for the older high-end 16 bit models, when high-end 16 bit models were available. Noise specs are better, and that's about it. In sound cards as in many other consumer products, customers equate more bits to higher quality and today you cannot buy anything but 24 bit cards, regardless of the actual improvement obtained. I am not sure the actual specs reflect the higher number of bits. I believe Bruce has quite a bit of data on that. Some of the high end ADC parts (not necessarily cards, which could greatly compromise the performance) are quite good... they're sold into applications for professional recording as well as the consumer gear. The (amateur) software defined radio folks have looked into this, in connection with direct conversion type receivers, because it affects the performance on things like instantantaneous dynamic range: being able to demodulate a small signal next to a big one, for instance. So, over the, say, 100 Hz and up range, they're probably pretty good. What I know is that the standalone A/D converters like the AD parts referenced in the other emails have outstanding linearity specs (at low bandwidth of course) at a price well below that of a "high quality" sound card, so even if the specs were the same, it would still make more sense to use the external device to measure a slow moving DC voltage. I agree there.. Didier ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Didier Juges wrote: I can only guess at what a sound card with linearity specs approaching those of the AD kit would cost, and it's still AC coupled, and as Poul pointed out, has no long term stability spec. This *is* an issue.. the audio ADCs have great performance in the "sub 1 second" sort of regime.. small variations in gain or offset would have no impact on audio applications, but would be a problem with the EFC measurement.. To me, it sounds like a no brainer, pun intended :) Alternately, my last HP3456A cost me about $50, and that's a 6 digit DVM with decent specs. You will need an HPIB controller though, if you do not already have one. There is that... ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] yet another GPSDO design, or so
Attilia What you want is basically a Shera Board. That design has been around for quite some time and has served me very well. I have a total of six running including two controlling Rubidium. There are in my opinion a couple of problems: not every 4066 works on the design the 18 bit D/A is very hard to find and now expensive and the single step of the D/A is intended for a 1.7 E-13 frequency step. I have build a input section that counts 100 MHz in stead of 24 MHZ making the unit create steps of 4.3 E-14 which works better on my Rubidium's and Datum FTS 1000. Also it eliminates the 4066's. Since I do not know how to write code that was my solution. I have also designed a later version Shera, with less IC's and a low cost dual D/A but I do not have the programming skill. If you contact me directly I will send you a copy of the QST Shera article, my design and the D/A data sheet.. I am sure you can replace the PIC with an Atmel device. In a message dated 6/26/2010 1:16:09 P.M. Eastern Daylight Time, att...@kinali.ch writes: Moin, I recently had a look at the data sheet of the LEA6-T GPS module from ublox, which now features a second time pulse output that is capable of delivering a 10MHz signal, synchronized to GPS. After thinking quite some time quite some time about building my own GPSDO and struggling with the question how to synchronize a 10MHz signal to a 1Hz signal that has some substantial phase noise, the new LEA6-T module seems like to make things a lot easier. Although the LEA6 specs do not say anything about how the timepulse output is generated or how it is synchronized to GPS, i assume that it will either have some jumps or phase/frequency noise due to oszillator and synchronization imperfections. But, it should be possible to use the LEA6-T together with some OCXO and a PLL setup to stabilize the OCXO to get a high quality frequency standard. Unfortunately, my knowledge in that field is rather limited, thus before starting to make wrong design decisions i'd like to ask for some advice here. My basic idea is to feed the 10MHz output of the LEA6-T and the 10MHz OCXO into a current output PFD, do some low-order filtering of the output signal. Feed that into an ADC which is read by a uC which in turn controls an DAC that sets over some amplifier stage the EFC input of the OCXO. As PFD i thought about using a ADF4002 from Analog, which is actually an PLL, but allows to bypass the input dividers, so that it can be used as pure current output PFD. I'm not yet sure what kind of output filter i want to use. I probably have to add at least one low noise opamp there, to isolate the PFD output/filter from the ADC. I'm also not sure what filter frequency i should use here. It will have to be below 10MHz for sure, probably in the lower kHz range, but how low is the question. The lower the easier gets the ADC stage and the less work has to be done in the uC, but using a low frequency filter either means using an active filter (noise) or high value R or L (again noise, especially the L might couple in 50Hz noise from the enviroment or show microphone effects). The ADC will be either a low-noise 16bit type or a 24bit type. This will largely depend on the sample rate to be used and the availabilty of the ADCs. Any good advices on what to use here? Should there be some form of signal conditioning done? If, what form of conditioning would you advise me to use? As a uC i thought about using a AT91SAM7 variant from Atmel. I know these beasts (and their bugs) pretty well by now and already have some code ready for those. I thought about clocking the uC with a 40MHz crystal that is synchronized to the 10MHz OCXO using a PLL. This would allow me to generate quite precise+accurate digital signals. Unfortunately, there doesnt seem to be VCXOs at 40MHz available so that means that i'd have to build one by hand. The loopfilter is going to end up in the uC as it is easier to build such low frequency filters digitally than in analog. I havent put much thought into how that filter should look like, as this can be easily changed later. The DAC will probably be a 16bit type (there does not seem any higher resolution DAC with sane specs and still reasonable availability). The amplifier for the DAC output will be a two stage amplifier. One stage that adds an (adjustable) offset and one stage that adds the (again adjustable) amplification. This approach is choosen as the needed EFC range will probably much lower than the full range. Hence the resolution of the DAC can be enhanced by producing only values within that range. The disadvantage here is that it requires calibration. A rough guestimate is that the whole thing will probably cost less than 500CHF (including PCB production, but excluding OCXO). Yes, i know, i could get a Rb frequency standard for that money on ebay. But where is the fun in that? ;-) Beside whether this
Re: [time-nuts] EFC tracking
I am not sure how to translate the IMD specs into integral or differential non-linearity, but from what I have seen, IMD specs are not significantly better for 24 bit sound cards than for the older high-end 16 bit models, when high-end 16 bit models were available. Noise specs are better, and that's about it. In sound cards as in many other consumer products, customers equate more bits to higher quality and today you cannot buy anything but 24 bit cards, regardless of the actual improvement obtained. I am not sure the actual specs reflect the higher number of bits. I believe Bruce has quite a bit of data on that. What I know is that the standalone A/D converters like the AD parts referenced in the other emails have outstanding linearity specs (at low bandwidth of course) at a price well below that of a "high quality" sound card, so even if the specs were the same, it would still make more sense to use the external device to measure a slow moving DC voltage. Didier Sent from my BlackBerry Wireless thingy while I do other things... -Original Message- From: "John Miles" Sender: time-nuts-boun...@febo.com Date: Sat, 26 Jun 2010 15:25:28 To: Discussion of precise time and frequency measurement Reply-To: Discussion of precise time and frequency measurement Subject: Re: [time-nuts] EFC tracking > >sound card ADCs, the high end 24 bit ones, are pretty darn linear [...] > > That is actually a very debatable proposition, a lot of them are > tracking types that conveniently cover up any lack of linearity > on the analog side of the fence. Can you elaborate on that? Linearity is linearity - wouldn't it ultimately show up in the card's IMD specs? > The major problem with using sound ADC's is that their references > has absolutely no long term stability, so you will see your EFC > wiggle and wander all over the place, even when it stays perfectly > still. One idea: the sound card gives you at least two input channels to work with, and it isn't much of a stretch to assume they share a common reference. Perhaps you could use one channel to digitize the oscillator's supply voltage or (ideally) EFC reference voltage output, using the same chopper/mixer/VFC/whatever approach, and have the software take the difference out. > I would find one of those cheap-ish DVM's with a serial or USB port... I'd be surprised if a cheap DVM outperformed a decent sound-card ADC in any respect, other than perhaps DC baseline drift. -- john, KE5FX ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Hi Steve, Let me build on the VFC idea for a minute. If I am understanding your interest properly, you are wanting to monitor and track the EFC changes then process that information as a form of studying the stability to some degree. Is that correct ? Well, it occurred to me, admittedly after the fact, that there are several programs written for Amateur radio use to receive and observe very slow CW on the really long wave radio frequencies, like 160 to 190 KHz. These programs provide the waterfall type displays among other types and, I think, they store the sessions as a file for later review. As I am only familiar with the fact they exist, I have never used them, you would have to peruse their abilities to see if there is a fit. The primary point is they use a computer sound card to look at the audio out of a receiver. That audio is somewhere around 1KHz. I am not sure what kind of sensitivity the software has as to resolution, but I believe a number of Amateurs, including some on the list, have used such for the "FMT's" (frequency monitoring contests) with fair, if not better then fair, results. The software may do all you want or at least a major portion that being the file creation and storage, hopefully in a raw PCM format (i.e., no compression. I am not sure what is required for the circuit of the VFC these days but it shouldn't be anymore complex then that of building up a CPU and so forth. Good luck with the project, 73BillWB6BNQ Steve Rooke wrote: > Bill, > > On 27 June 2010 01:21, WB6BNQ wrote: > > Steve, > > > > I think using a voltage-to-frequency converter would solve that problem. > > They > > are not too expensive and there are several flavors from Amalog devices and > > some > > others. > > I just had one of those duh! moments :) It does make the sampling a > little more complicated as I'd have to take a number of samples and do > something like a FFT to get the frequency, I think. My idea seemed > just a simple way to just make a single, or limited number, of > sample(s) and I have the data. Still, this is a good idea and I'll > have a good thunk about it. > > > Just set it for a 1KHz start point or maybe 10KHz. > > Thanks for your advice, > Steve > > > BillWB6BNQ > > > > > > Steve Rooke wrote: > > > >> I would like to track the EFC voltage in hardware using something > >> cheap and ready to hand. I was thinking of using a sound card as it > >> has good resolution but it's obviously only AC coupled so it would not > >> measure the DC of the EFC. I thought about modifying a sound card to > >> make it DC coupled but most of them seem to reference the 0V point to > >> some internal reference voltage hence there is a DC shift there. I > >> next thought about turning the DC into AC by chopping it, IE. > >> inverting 50% of the voltage via an oscillator. This way I could pass > >> the square wave directly into an unmodified sound card, take > >> measurements and then do an RMS calculation on them (really just need > >> to flip the sign on, say, the negative readings). > >> > >> I wonder if anyone has done something like this before and could share > >> their experiences. I've attached a diagram image (hope it is accepted > >> by the list) which is my first go with Eagle so I'm not exactly very > >> familiar with it, sorry. The R's and C's in the astable would be set > >> to a clock frequency that enables this to work without bias given the > >> sampling frequency. I'm not sure if this clock should be slower than > >> the sampling frequency or higher, just haven't got my head around that > >> yet. The R's around the op-amp would need to be set in a ratio that > >> transforms the EFC voltage into the range that the sound card can > >> handle (that is yet to be calculated by measuring the limits). If you > >> have any suggestions or ways of doing this in a better way, I'd be > >> very grateful for the advice. > >> > >> Thanks, > >> Steve > >> -- > >> Steve Rooke - ZL3TUV ? G8KVD > >> The only reason for time is so that everything doesn't happen at once. > >> - Einstein > >> > >> > >> Name: DCchop.png > >>DCchop.pngType: PNG image (image/png) > >> Encoding: base64 > >> > >> > >> ___ > >> time-nuts mailing list -- time-nuts@febo.com > >> To unsubscribe, go to > >> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > >> and follow the instructions there. > > > > > > ___ > > time-nuts mailing list -- time-nuts@febo.com > > To unsubscribe, go to > > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > > and follow the instructions there. > > > > -- > Steve Rooke - ZL3TUV & G8KVD > The only reason for time is so that everything doesn't happen at once. > - Einstein > > __
[time-nuts] FE-405 DOCXO info request
I have a FEI Communications FE-405B DOCXO looking for info. Sales flyer and pictures of the unit here: www.n4iqt.com/fei-fe405b The Sales flyer listed the 405A and I have the 405B don't know the difference. A manual that shows the digital error correction via the i/o port would be great, even better would be info on changing the frequency of the synthesizer. Thank You, Stanley ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
> >sound card ADCs, the high end 24 bit ones, are pretty darn linear [...] > > That is actually a very debatable proposition, a lot of them are > tracking types that conveniently cover up any lack of linearity > on the analog side of the fence. Can you elaborate on that? Linearity is linearity - wouldn't it ultimately show up in the card's IMD specs? > The major problem with using sound ADC's is that their references > has absolutely no long term stability, so you will see your EFC > wiggle and wander all over the place, even when it stays perfectly > still. One idea: the sound card gives you at least two input channels to work with, and it isn't much of a stretch to assume they share a common reference. Perhaps you could use one channel to digitize the oscillator's supply voltage or (ideally) EFC reference voltage output, using the same chopper/mixer/VFC/whatever approach, and have the software take the difference out. > I would find one of those cheap-ish DVM's with a serial or USB port... I'd be surprised if a cheap DVM outperformed a decent sound-card ADC in any respect, other than perhaps DC baseline drift. -- john, KE5FX ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
I can only guess at what a sound card with linearity specs approaching those of the AD kit would cost, and it's still AC coupled, and as Poul pointed out, has no long term stability spec. To me, it sounds like a no brainer, pun intended :) Alternately, my last HP3456A cost me about $50, and that's a 6 digit DVM with decent specs. You will need an HPIB controller though, if you do not already have one. Didier Sent from my BlackBerry Wireless thingy while I do other things... -Original Message- From: jimlux Sender: time-nuts-boun...@febo.com Date: Sat, 26 Jun 2010 12:57:37 To: Discussion of precise time and frequency measurement Reply-To: Discussion of precise time and frequency measurement Subject: Re: [time-nuts] EFC tracking Didier Juges wrote: > Steve, > > You may want to check the "Analog Devices MiniKit for ADuC702x-series". > > http://www.google.com/search?q=Analog+Devices+MiniKit+for+ADuC702x-series > > This kit includes a 24 bit ADC and integrated ARM processor in a small PWB > with all the tools and sample code to do what you want with very little code > to write (you can probably use the sample code as-is). > > The kit is $30 (or $35, depending on where you look...) and you will easily > spend that much building something that will not work as well using your > sound card. > > Sound card ADCs are intended for audio, and I'll bet their linearity does not > come close to that of the ADuC702x series, if you can even get the spec for > it. > sound card ADCs, the high end 24 bit ones, are pretty darn linear, and have pretty good SNR as well. 120-130dB S/(N+Distortion+etc) wouldn't be unusual. The surrounding circuitry is likely to be a bigger problem. 1 LSB on a 1V p-p 24bit converter is about 60nV. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Here we go again C -Original Message- From: time-nuts-boun...@febo.com [mailto:time-nuts-boun...@febo.com] On Behalf Of Poul-Henning Kamp Sent: 26 June 2010 21:06 To: Discussion of precise time and frequency measurement Subject: Re: [time-nuts] EFC tracking In message <4c265bb1.8090...@earthlink.net>, jimlux writes: >sound card ADCs, the high end 24 bit ones, are pretty darn linear [...] That is actually a very debatable proposition, a lot of them are tracking types that conveniently cover up any lack of linearity on the analog side of the fence. The major problem with using sound ADC's is that their references has absolutely no long term stability, so you will see your EFC wiggle and wander all over the place, even when it stays perfectly still. I would find one of those cheap-ish DVM's with a serial or USB port... -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 p...@freebsd.org | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
In message <4c265bb1.8090...@earthlink.net>, jimlux writes: >sound card ADCs, the high end 24 bit ones, are pretty darn linear [...] That is actually a very debatable proposition, a lot of them are tracking types that conveniently cover up any lack of linearity on the analog side of the fence. The major problem with using sound ADC's is that their references has absolutely no long term stability, so you will see your EFC wiggle and wander all over the place, even when it stays perfectly still. I would find one of those cheap-ish DVM's with a serial or USB port... -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 p...@freebsd.org | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Didier Juges wrote: Steve, You may want to check the "Analog Devices MiniKit for ADuC702x-series". http://www.google.com/search?q=Analog+Devices+MiniKit+for+ADuC702x-series This kit includes a 24 bit ADC and integrated ARM processor in a small PWB with all the tools and sample code to do what you want with very little code to write (you can probably use the sample code as-is). The kit is $30 (or $35, depending on where you look...) and you will easily spend that much building something that will not work as well using your sound card. Sound card ADCs are intended for audio, and I'll bet their linearity does not come close to that of the ADuC702x series, if you can even get the spec for it. sound card ADCs, the high end 24 bit ones, are pretty darn linear, and have pretty good SNR as well. 120-130dB S/(N+Distortion+etc) wouldn't be unusual. The surrounding circuitry is likely to be a bigger problem. 1 LSB on a 1V p-p 24bit converter is about 60nV. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] yet another GPSDO design, or so
Moin, On Sat, 26 Jun 2010 12:38:29 -0600 Ed Palmer wrote: > Another GPS board with a 10 MHz output is the Navsync CW-12 module > (price ~US$85-90). Hmm.. i didn't know about this one. Both seem to be comparable from the specs, the LEA6-T being slightly better (well, the design is newer, so that's to be expected). As for the price. ublox is here, literally, sitting in the next town and i'm in the fortunate position to get a LEA-6T out of a bigger batch which makes it considerably cheaper than the CW-12. > I measured the 1 PPS output and found a Standard > Deviation of < 5 ns with a range of < 30 ns. The 10 MHz output is kept > on frequency by occasionally adjusting the period of the 10 MHz output > by one cycle of the CPU clock (~8.3 ns = 120 MHz). How often this > happens depends on the exact clock frequency of the particular unit. On > mine it happens about 200 times per second. Interesting information, thanks. I know that the LEA6 has a TXCO that can be adjusted and the timepulse output is generated from this source. What i dont know is whether the TXCO is adjusted tracking the GPS signal or whether it uses a similar "skip a cycle" technique like the CW-12. Maybe i should drop ublox a mail and ask about that. Attila Kinali -- If you want to walk fast, walk alone. If you want to walk far, walk together. -- African proverb ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] yet another GPSDO design, or so
Hello Attila, Another GPS board with a 10 MHz output is the Navsync CW-12 module (price ~US$85-90). I measured the 1 PPS output and found a Standard Deviation of < 5 ns with a range of < 30 ns. The 10 MHz output is kept on frequency by occasionally adjusting the period of the 10 MHz output by one cycle of the CPU clock (~8.3 ns = 120 MHz). How often this happens depends on the exact clock frequency of the particular unit. On mine it happens about 200 times per second. I don't know if this board is better or cheaper than the LEA-6T, but you might want to look into it. Ed Attila Kinali wrote: Moin, I recently had a look at the data sheet of the LEA6-T GPS module from ublox, which now features a second time pulse output that is capable of delivering a 10MHz signal, synchronized to GPS. After thinking quite some time quite some time about building my own GPSDO and struggling with the question how to synchronize a 10MHz signal to a 1Hz signal that has some substantial phase noise, the new LEA6-T module seems like to make things a lot easier. Although the LEA6 specs do not say anything about how the timepulse output is generated or how it is synchronized to GPS, i assume that it will either have some jumps or phase/frequency noise due to oszillator and synchronization imperfections. But, it should be possible to use the LEA6-T together with some OCXO and a PLL setup to stabilize the OCXO to get a high quality frequency standard. Unfortunately, my knowledge in that field is rather limited, thus before starting to make wrong design decisions i'd like to ask for some advice here. My basic idea is to feed the 10MHz output of the LEA6-T and the 10MHz OCXO into a current output PFD, do some low-order filtering of the output signal. Feed that into an ADC which is read by a uC which in turn controls an DAC that sets over some amplifier stage the EFC input of the OCXO. As PFD i thought about using a ADF4002 from Analog, which is actually an PLL, but allows to bypass the input dividers, so that it can be used as pure current output PFD. I'm not yet sure what kind of output filter i want to use. I probably have to add at least one low noise opamp there, to isolate the PFD output/filter from the ADC. I'm also not sure what filter frequency i should use here. It will have to be below 10MHz for sure, probably in the lower kHz range, but how low is the question. The lower the easier gets the ADC stage and the less work has to be done in the uC, but using a low frequency filter either means using an active filter (noise) or high value R or L (again noise, especially the L might couple in 50Hz noise from the enviroment or show microphone effects). The ADC will be either a low-noise 16bit type or a 24bit type. This will largely depend on the sample rate to be used and the availabilty of the ADCs. Any good advices on what to use here? Should there be some form of signal conditioning done? If, what form of conditioning would you advise me to use? As a uC i thought about using a AT91SAM7 variant from Atmel. I know these beasts (and their bugs) pretty well by now and already have some code ready for those. I thought about clocking the uC with a 40MHz crystal that is synchronized to the 10MHz OCXO using a PLL. This would allow me to generate quite precise+accurate digital signals. Unfortunately, there doesnt seem to be VCXOs at 40MHz available so that means that i'd have to build one by hand. The loopfilter is going to end up in the uC as it is easier to build such low frequency filters digitally than in analog. I havent put much thought into how that filter should look like, as this can be easily changed later. The DAC will probably be a 16bit type (there does not seem any higher resolution DAC with sane specs and still reasonable availability). The amplifier for the DAC output will be a two stage amplifier. One stage that adds an (adjustable) offset and one stage that adds the (again adjustable) amplification. This approach is choosen as the needed EFC range will probably much lower than the full range. Hence the resolution of the DAC can be enhanced by producing only values within that range. The disadvantage here is that it requires calibration. A rough guestimate is that the whole thing will probably cost less than 500CHF (including PCB production, but excluding OCXO). Yes, i know, i could get a Rb frequency standard for that money on ebay. But where is the fun in that? ;-) Beside whether this setup makes sense, the two biggest questions i have are, what OCXO to use. Are the ISOTEMP 134-10 that are available on ebay "good enough" for such an application? Or shall i look for something better/different? And the other is, how do i amplify and distribute the 10MHz signal i get out of the OCXO to be used by other devices with minimal phase noise? Thanks for your help Attila Kinali ___ time-nuts mailing list -- time-nuts@febo
Re: [time-nuts] EFC tracking
At 01:43 +1200 27-06-2010, Steve Rooke wrote: > I don't know if it qualifies as simple/cheap, but Analog Devices and others have single chip low-rate sigma/delta converters with good to excellent > properties; these were meant for strain gauges but should be able to track > slow-moving control voltages just fine. [...] Do you have any part numbers to hand that I could go and look up please? (You didn't specify runtime or resolution; I am assuming that you plan to monitor an OCXO CV line over several days; this would imply that you're interested in sub-mV wiggles on a constantish offset of a few V. Do set me straight if I'm wrong) I was looking at the Microchip MCP355x-series and Cirrus CS551x-series. Both are available in DIY-friendly SOIC-packages; if you can handle SOT-23-6 and TSSOPs your choices widen (ie the AD778x that Jim Lux mentioned). Both families are S/D converters with internal digital filtering, offering 20+ bit resolution (although not all bits are usable, depending on your config). Most of these parts are available from Farnell and Digi-Key (and if it were me, I'd order from the latter). At 10:20 -0500 26-06-2010, Didier Juges wrote: You may want to check the "Analog Devices MiniKit for ADuC702x-series". http://www.google.com/search?q=Analog+Devices+MiniKit+for+ADuC702x-series This kit includes a 24 bit ADC and integrated ARM processor in a small PWB with all the tools and sample code to do what you want with very little code to write (you can probably use the sample code as-is). Seconded, but note that according to the datasheet it only offers an internal 12-bit ADC (although the docs from AD seem to contradict eachother here). This may be Good Enough. The advantage of the ADuC7xxx series over most other microcontrollers with built-in ADCs is that the ADuC7xxx ADC has a differential input mode. For more cheap boards, have a look at Olimex (http://www.olimex.com/dev/). At 06:21 -0700 26-06-2010, WB6BNQ wrote: I think using a voltage-to-frequency converter would solve that problem. They are not too expensive and there are several flavors from Amalog devices and some others. Good idea, but might be hard to keep stable enough if my "sub-mV over days"-assumption is right. JDB. -- Years from now, if you are doing something quick and dirty, you imagine that I am looking over your shoulder and say to yourself, "Dijkstra would not like this," well that would be immortality for me. -- Edsger Dijkstra, 1930 - 2002 ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
[time-nuts] yet another GPSDO design, or so
Moin, I recently had a look at the data sheet of the LEA6-T GPS module from ublox, which now features a second time pulse output that is capable of delivering a 10MHz signal, synchronized to GPS. After thinking quite some time quite some time about building my own GPSDO and struggling with the question how to synchronize a 10MHz signal to a 1Hz signal that has some substantial phase noise, the new LEA6-T module seems like to make things a lot easier. Although the LEA6 specs do not say anything about how the timepulse output is generated or how it is synchronized to GPS, i assume that it will either have some jumps or phase/frequency noise due to oszillator and synchronization imperfections. But, it should be possible to use the LEA6-T together with some OCXO and a PLL setup to stabilize the OCXO to get a high quality frequency standard. Unfortunately, my knowledge in that field is rather limited, thus before starting to make wrong design decisions i'd like to ask for some advice here. My basic idea is to feed the 10MHz output of the LEA6-T and the 10MHz OCXO into a current output PFD, do some low-order filtering of the output signal. Feed that into an ADC which is read by a uC which in turn controls an DAC that sets over some amplifier stage the EFC input of the OCXO. As PFD i thought about using a ADF4002 from Analog, which is actually an PLL, but allows to bypass the input dividers, so that it can be used as pure current output PFD. I'm not yet sure what kind of output filter i want to use. I probably have to add at least one low noise opamp there, to isolate the PFD output/filter from the ADC. I'm also not sure what filter frequency i should use here. It will have to be below 10MHz for sure, probably in the lower kHz range, but how low is the question. The lower the easier gets the ADC stage and the less work has to be done in the uC, but using a low frequency filter either means using an active filter (noise) or high value R or L (again noise, especially the L might couple in 50Hz noise from the enviroment or show microphone effects). The ADC will be either a low-noise 16bit type or a 24bit type. This will largely depend on the sample rate to be used and the availabilty of the ADCs. Any good advices on what to use here? Should there be some form of signal conditioning done? If, what form of conditioning would you advise me to use? As a uC i thought about using a AT91SAM7 variant from Atmel. I know these beasts (and their bugs) pretty well by now and already have some code ready for those. I thought about clocking the uC with a 40MHz crystal that is synchronized to the 10MHz OCXO using a PLL. This would allow me to generate quite precise+accurate digital signals. Unfortunately, there doesnt seem to be VCXOs at 40MHz available so that means that i'd have to build one by hand. The loopfilter is going to end up in the uC as it is easier to build such low frequency filters digitally than in analog. I havent put much thought into how that filter should look like, as this can be easily changed later. The DAC will probably be a 16bit type (there does not seem any higher resolution DAC with sane specs and still reasonable availability). The amplifier for the DAC output will be a two stage amplifier. One stage that adds an (adjustable) offset and one stage that adds the (again adjustable) amplification. This approach is choosen as the needed EFC range will probably much lower than the full range. Hence the resolution of the DAC can be enhanced by producing only values within that range. The disadvantage here is that it requires calibration. A rough guestimate is that the whole thing will probably cost less than 500CHF (including PCB production, but excluding OCXO). Yes, i know, i could get a Rb frequency standard for that money on ebay. But where is the fun in that? ;-) Beside whether this setup makes sense, the two biggest questions i have are, what OCXO to use. Are the ISOTEMP 134-10 that are available on ebay "good enough" for such an application? Or shall i look for something better/different? And the other is, how do i amplify and distribute the 10MHz signal i get out of the OCXO to be used by other devices with minimal phase noise? Thanks for your help Attila Kinali -- If you want to walk fast, walk alone. If you want to walk far, walk together. -- African proverb ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] PICTIC II
Moin Richard, On Fri, 25 Jun 2010 23:11:39 -0800 (AKDT) "Richard H McCorkle" wrote: > Over the last 18 months I have developed a new diode switched > interpolator based on the comments made on line and have thoroughly > tested it. Some suggestions made improvements in the performance and > some resulted in poorer performance Would it be possible to put the schematics of the board in PDF form onto your website? For all those poor souls that dont have windows at home and cannot install ExpressPCB? Thanks in advance Attila Kinali -- If you want to walk fast, walk alone. If you want to walk far, walk together. -- African proverb ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
At 06:40 -0700 26-06-2010, jimlux wrote: J.D. Bakker wrote: Charge injection is a bit on the high side on a 4066; a more expensive (A)DG4xx-series chip may improve on that. Or the traditional chopper approach of a mercury wetted reed relay? Good luck with that in this day and age, especially if you live in and/or sell to RoHS-land. JDB. [I would *love* to use mercury wetted relays in the transistor tester I'm working on, but sourcing them here in Europe proved impractical, to say the least. They do show up from time to time on eBay, but not in quantities large enough to make what amounts to a lifetime buy. There are some newer low-signal relays that are guaranteed down to low-low wetting currents, but in my tests Hg-wetted relays beat them every time] -- LART. 250 MIPS under one Watt. Free hardware design files. http://www.lartmaker.nl/ ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
[voltage-to-frequency converter] > I just had one of those duh! moments :) It does make the sampling a little > more complicated as I'd have to take a number of samples and do something > like a FFT to get the frequency, I think. My idea seemed just a simple way > to just make a single, or limited number, of sample(s) and I have the data. > Still, this is a good idea and I'll have a good thunk about it. How many bits of resolution do you want? A FFT chops the energy into buckets. To get smaller (narrower) buckets you need more samples. But 16 bits is only 64K samples... No big deal if you are working on a real PC but it probably won't fit into a tiny micro. There is probably some simple way to move things around but I can't work it out on the fly. It's something like use every Nth sample. Your signal will get aliased down to the wrong place but since you know (roughly) where it is you can correct for that. (Radar geeks to this sort of thing all the time.) I'd look into something like scanning for zero crossings. -- These are my opinions, not necessarily my employer's. I hate spam. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
I am commenting on my own comment :) I have been using a similar kit from TI for the MSC1210Y5 processor (originally Burr-Brown, acquired by TI since). It was $50 and worked great. I have since made several projects at work using the chip. Unfortunately, the MSC1210 kit is no longer available, but the new kit is even cheaper. Didier > -Original Message- > From: time-nuts-boun...@febo.com > [mailto:time-nuts-boun...@febo.com] On Behalf Of Didier Juges > Sent: Saturday, June 26, 2010 10:20 AM > To: 'Discussion of precise time and frequency measurement' > Subject: Re: [time-nuts] EFC tracking > > Steve, > > You may want to check the "Analog Devices MiniKit for > ADuC702x-series". > > http://www.google.com/search?q=Analog+Devices+MiniKit+for+ADuC > 702x-series > > This kit includes a 24 bit ADC and integrated ARM processor > in a small PWB with all the tools and sample code to do what > you want with very little code to write (you can probably use > the sample code as-is). > > The kit is $30 (or $35, depending on where you look...) and > you will easily spend that much building something that will > not work as well using your sound card. > > Sound card ADCs are intended for audio, and I'll bet their > linearity does not come close to that of the ADuC702x series, > if you can even get the spec for it. > > Didier > > > -Original Message- > > From: time-nuts-boun...@febo.com > > [mailto:time-nuts-boun...@febo.com] On Behalf Of Steve Rooke > > Sent: Saturday, June 26, 2010 7:13 AM > > To: Discussion of precise time and frequency measurement > > Subject: [time-nuts] EFC tracking > > > > I would like to track the EFC voltage in hardware using something > > cheap and ready to hand. I was thinking of using a sound card as it > > has good resolution but it's obviously only AC coupled so > it would not > > measure the DC of the EFC. I thought about modifying a > sound card to > > make it DC coupled but most of them seem to reference the > 0V point to > > some internal reference voltage hence there is a DC shift there. I > > next thought about turning the DC into AC by chopping it, IE. > > inverting 50% of the voltage via an oscillator. This way I > could pass > > the square wave directly into an unmodified sound card, take > > measurements and then do an RMS calculation on them (really > just need > > to flip the sign on, say, the negative readings). > > > > I wonder if anyone has done something like this before and > could share > > their experiences. I've attached a diagram image (hope it > is accepted > > by the list) which is my first go with Eagle so I'm not > exactly very > > familiar with it, sorry. The R's and C's in the astable > would be set > > to a clock frequency that enables this to work without bias > given the > > sampling frequency. I'm not sure if this clock should be > slower than > > the sampling frequency or higher, just haven't got my head > around that > > yet. The R's around the op-amp would need to be set in a ratio that > > transforms the EFC voltage into the range that the sound card can > > handle (that is yet to be calculated by measuring the > limits). If you > > have any suggestions or ways of doing this in a better way, I'd be > > very grateful for the advice. > > > > Thanks, > > Steve > > -- > > Steve Rooke - ZL3TUV & G8KVD > > The only reason for time is so that everything doesn't > happen at once. > > - Einstein > > > > > ___ > time-nuts mailing list -- time-nuts@febo.com To unsubscribe, > go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Steve, You may want to check the "Analog Devices MiniKit for ADuC702x-series". http://www.google.com/search?q=Analog+Devices+MiniKit+for+ADuC702x-series This kit includes a 24 bit ADC and integrated ARM processor in a small PWB with all the tools and sample code to do what you want with very little code to write (you can probably use the sample code as-is). The kit is $30 (or $35, depending on where you look...) and you will easily spend that much building something that will not work as well using your sound card. Sound card ADCs are intended for audio, and I'll bet their linearity does not come close to that of the ADuC702x series, if you can even get the spec for it. Didier > -Original Message- > From: time-nuts-boun...@febo.com > [mailto:time-nuts-boun...@febo.com] On Behalf Of Steve Rooke > Sent: Saturday, June 26, 2010 7:13 AM > To: Discussion of precise time and frequency measurement > Subject: [time-nuts] EFC tracking > > I would like to track the EFC voltage in hardware using > something cheap and ready to hand. I was thinking of using a > sound card as it has good resolution but it's obviously only > AC coupled so it would not measure the DC of the EFC. I > thought about modifying a sound card to make it DC coupled > but most of them seem to reference the 0V point to some > internal reference voltage hence there is a DC shift there. I > next thought about turning the DC into AC by chopping it, IE. > inverting 50% of the voltage via an oscillator. This way I > could pass the square wave directly into an unmodified sound > card, take measurements and then do an RMS calculation on > them (really just need to flip the sign on, say, the negative > readings). > > I wonder if anyone has done something like this before and > could share their experiences. I've attached a diagram image > (hope it is accepted by the list) which is my first go with > Eagle so I'm not exactly very familiar with it, sorry. The > R's and C's in the astable would be set to a clock frequency > that enables this to work without bias given the sampling > frequency. I'm not sure if this clock should be slower than > the sampling frequency or higher, just haven't got my head > around that yet. The R's around the op-amp would need to be > set in a ratio that transforms the EFC voltage into the range > that the sound card can handle (that is yet to be calculated > by measuring the limits). If you have any suggestions or ways > of doing this in a better way, I'd be very grateful for the advice. > > Thanks, > Steve > -- > Steve Rooke - ZL3TUV & G8KVD > The only reason for time is so that everything doesn't happen at once. > - Einstein > ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Steve, Do a simulation and see how your card behaves. I've had some unexpected results in similar attempts years ago. Nothing like an add on card suitable for the job. These cards are not cheap. >I would like to track the EFC voltage in hardware using something >cheap and ready to hand. I was thinking of using a sound card as it >has good resolution but it's obviously only AC coupled so it would not >measure the DC of the EFC. I thought about modifying a sound card to >make it DC coupled but most of them seem to reference the 0V point to >some internal reference voltage hence there is a DC shift there. I >next thought about turning the DC into AC by chopping it, IE. >inverting 50% of the voltage via an oscillator. This way I could pass >the square wave directly into an unmodified sound card, take >measurements and then do an RMS calculation on them (really just need >to flip the sign on, say, the negative readings). -- Raj, VU2ZAP Bangalore, India. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Steve, Large amplitude signals on the input ports tend to leak into the inputs on sound cards. Isolation testing required! >> Soundcards have inputs and outputs. You can feed the output with a >> series of samples that represent your control waveform. The PC becomes >> the oscillator and you know it's frequency and relative phase track. > -- Raj, VU2ZAP Bangalore, India. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Steve Rooke wrote: The eval board for the part may have a computer interface built into it. So I need to locate this if possible, any pointer please? Go to Analog Devices website (http://www.analog.com/) and find your ADC.. typically there's an eval board with a USB interface available. For an example, look at the AD7785.. it might not be what you're looking for, but it's an example of what's available. Thanks for your help, Steve ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Oz-in-DFW wrote: Chopping is used to cancel DC offsets in imperfect amplifiers, it adds no gain. If there is a DC component and you filter with a cutoff frequency below the chop rate, the offsets of the amplifier can be effectively canceled. Chopping isn't quite that magical. You chop the DC signal before the first amplifier, and synchronously (usually) detect the amplified signal after the last amplifier. The key benefit to chopped amplifiers is the high gain stages can be AC coupled, which eliminates any DC drifting in the high gain amplifiers from affecting the results. Chopping doesn't cancel any DC offsets that might exist in the input signal, or in the input, or output, chopper mechanism. -Chuck Harris ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] crystal oscillators & TPLL
Steve, It's for people like you that makes this whole thing worthwhile. Keep hanging on the line and I'm certain something will transpire. Best regards, Steve On 27 June 2010 01:24, Steve Roberts wrote: > I imagine that there are people on this list like myself with limited > engineering experience > but a healthy appetite for knowledge and an interest in time/frequency > measurement. > I have little interest in complex math - an inherent brain block methinks, > but enjoy > building hardware and incorporating such 'blocks' into my projects. > > I am not necessarily unintelligent - being a Consultant Anaesthetist with 3 > degrees, > 2 diplomas and currently doing a Masters. > My interest in this list comes form having a HP106, 2 rubidium standards and > 2 z3801a > gps units. > > Why am I posting on this thread? - simple - I wish to build this beastie > described by > Warren and try it out on my equipment. I need more guidance than a block > diagram. > Shame on me for being so ignorant but I respectfully request more info > (circuit diagram) > emailed privately if necessary. > I am not here to criticise, but enjoy reading the discussions of people so > knowledgeable > in this area. I have my own areas of expertise. > > Thanks for reading. > > Steve > st...@borisone.demon.co.uk > > > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > -- Steve Rooke - ZL3TUV & G8KVD The only reason for time is so that everything doesn't happen at once. - Einstein ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Oz, On 27 June 2010 01:49, Oz-in-DFW wrote: > Soundcards have inputs and outputs. You can feed the output with a > series of samples that represent your control waveform. The PC becomes > the oscillator and you know it's frequency and relative phase track. Brilliant! > Chopping is used to cancel DC offsets in imperfect amplifiers, it adds > no gain. If there is a DC component and you filter with a cutoff > frequency below the chop rate, the offsets of the amplifier can be > effectively canceled. OK, I understand that. I'm not so sure this is the original problem I would have but thinking about it now I wonder if the chopped signal would end up being symmetrical with my original proposal as there is likely to be some small offsets thinking about it now. I could resolve this in software by sampling over a full cycle period of the chopper and then do averaging of the RMS readings. Thanks, Steve > -- > mailto:o...@ozindfw.net > Oz > POB 93167 > Southlake, TX 76092 (Near DFW Airport) > > > > > > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > -- Steve Rooke - ZL3TUV & G8KVD The only reason for time is so that everything doesn't happen at once. - Einstein ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
On 27 June 2010 01:40, jimlux wrote: > J.D. Bakker wrote: >>> >>> I wonder if anyone has done something like this before and could share >>> their experiences. >> >> The general principle should work. However, as you're interested in slow >> changes, there are some error sources that might be unacceptable, including >> the drift of (differential) channel resistances for the 4066 over >> temperature, voltage and time. As shown the scheme is also sensitive to >> impedance mismatch/drift on the two inputs. Charge injection is a bit on the >> high side on a 4066; a more expensive (A)DG4xx-series chip may improve on >> that. > > Or the traditional chopper approach of a mercury wetted reed relay? > If you're processing with a sound card, you have the advantage that you > don't need to process the samples coming from the time of transition (unlike > a traditional analog chopper with synchronous detection), so a fairly crummy > relay would probably work. The key is that it can toggle at, say, 100Hz, > forever. I really hadn't thought of factoring in the resistance of the 4066 or environmental effects on it I must admit. I wonder how long a mercury wetted reed relay would last at 100Hz though. The sampling rate would be much higher than this to discard the transition readings. >> >> I don't know if it qualifies as simple/cheap, but Analog Devices and >> others have single chip low-rate sigma/delta converters with good to >> excellent properties; these were meant for strain gauges but should be able >> to track slow-moving control voltages just fine. Interfacing them to a >> parallel port (or USB PP adapter) should be close to trivial. Do have a >> close look at the data sheet: some parts have unbuffered inputs, and present >> a fluctuating input impedance which might couple onto EFC lines. A simple >> isolation amp with one or two precision op-amps should fix that. > > > The eval board for the part may have a computer interface built into it. So I need to locate this if possible, any pointer please? Thanks for your help, Steve > > > > >> >> JDB. >> [had just been looking into this for a transistor matcher/noise test rig >> I'm working on] > > > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > -- Steve Rooke - ZL3TUV & G8KVD The only reason for time is so that everything doesn't happen at once. - Einstein ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
On 6/26/2010 8:36 AM, Steve Rooke wrote: > Oz, > > On 27 June 2010 01:09, Oz-in-DFW wrote: > > On 6/26/2010 7:12 AM, Steve Rooke wrote: > >>> >>> >>> >> I've done similar stuff in work projects, but never written code. I've >> thought about this some as well. I'd consider a few things; >> >> 1. Use the sound card output as the chopper control signal instead of >> the discrete unit. You'll have more control and phase sync will >> be easier. >> * I'd be temped to take the sound card output and run it >>through a comparator to square it up, but I'm almost certain >>this isn't needed. >> > Sorry, not sure what you mean here. Are you saying that I should > derive the chopper frequency directly from a connection to the sound > card? I was hoping not to modify the sound card in any way so as to > keep it simple. > Soundcards have inputs and outputs. You can feed the output with a series of samples that represent your control waveform. The PC becomes the oscillator and you know it's frequency and relative phase track. > > >> 5. The probelm with chopping is that signal levels around zero don't >> have much amplitude and are a challenge to extract from noise. >> > I was under the impression that this was the idea that is used to > amplify very low level signals like the output from the likes of > strain-gauges. It would surely seem to me to be a problem to amplify > small signals around zero due to offsets in the amp unless you do this > sort of thing. > Chopping is used to cancel DC offsets in imperfect amplifiers, it adds no gain. If there is a DC component and you filter with a cutoff frequency below the chop rate, the offsets of the amplifier can be effectively canceled. -- mailto:o...@ozindfw.net Oz POB 93167 Southlake, TX 76092 (Near DFW Airport) ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Bill, On 27 June 2010 01:21, WB6BNQ wrote: > Steve, > > I think using a voltage-to-frequency converter would solve that problem. They > are not too expensive and there are several flavors from Amalog devices and > some > others. I just had one of those duh! moments :) It does make the sampling a little more complicated as I'd have to take a number of samples and do something like a FFT to get the frequency, I think. My idea seemed just a simple way to just make a single, or limited number, of sample(s) and I have the data. Still, this is a good idea and I'll have a good thunk about it. > Just set it for a 1KHz start point or maybe 10KHz. Thanks for your advice, Steve > BillWB6BNQ > > > Steve Rooke wrote: > >> I would like to track the EFC voltage in hardware using something >> cheap and ready to hand. I was thinking of using a sound card as it >> has good resolution but it's obviously only AC coupled so it would not >> measure the DC of the EFC. I thought about modifying a sound card to >> make it DC coupled but most of them seem to reference the 0V point to >> some internal reference voltage hence there is a DC shift there. I >> next thought about turning the DC into AC by chopping it, IE. >> inverting 50% of the voltage via an oscillator. This way I could pass >> the square wave directly into an unmodified sound card, take >> measurements and then do an RMS calculation on them (really just need >> to flip the sign on, say, the negative readings). >> >> I wonder if anyone has done something like this before and could share >> their experiences. I've attached a diagram image (hope it is accepted >> by the list) which is my first go with Eagle so I'm not exactly very >> familiar with it, sorry. The R's and C's in the astable would be set >> to a clock frequency that enables this to work without bias given the >> sampling frequency. I'm not sure if this clock should be slower than >> the sampling frequency or higher, just haven't got my head around that >> yet. The R's around the op-amp would need to be set in a ratio that >> transforms the EFC voltage into the range that the sound card can >> handle (that is yet to be calculated by measuring the limits). If you >> have any suggestions or ways of doing this in a better way, I'd be >> very grateful for the advice. >> >> Thanks, >> Steve >> -- >> Steve Rooke - ZL3TUV ? G8KVD >> The only reason for time is so that everything doesn't happen at once. >> - Einstein >> >> >> Name: DCchop.png >> DCchop.png Type: PNG image (image/png) >> Encoding: base64 >> >> >> ___ >> time-nuts mailing list -- time-nuts@febo.com >> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts >> and follow the instructions there. > > > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > -- Steve Rooke - ZL3TUV & G8KVD The only reason for time is so that everything doesn't happen at once. - Einstein ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
JDB, On 27 June 2010 01:19, J.D. Bakker wrote: >> I wonder if anyone has done something like this before and could share >> their experiences. > > The general principle should work. However, as you're interested in slow > changes, there are some error sources that might be unacceptable, including > the drift of (differential) channel resistances for the 4066 over > temperature, voltage and time. As shown the scheme is also sensitive to > impedance mismatch/drift on the two inputs. Charge injection is a bit on the > high side on a 4066; a more expensive (A)DG4xx-series chip may improve on > that. I'd be looking at buffering the inputs to provide a low impedance there and choose values of the R's and C's around the op-amp to try and alleviate problems caused by the 4066's contact resistance. I am aware of the higher spec devices but will have to see if it's easy to source them over here in New Zealand without having to order them overseas. > I don't know if it qualifies as simple/cheap, but Analog Devices and others > have single chip low-rate sigma/delta converters with good to excellent > properties; these were meant for strain gauges but should be able to track > slow-moving control voltages just fine. Interfacing them to a parallel port > (or USB PP adapter) should be close to trivial. Do have a close look at the > data sheet: some parts have unbuffered inputs, and present a fluctuating > input impedance which might couple onto EFC lines. A simple isolation amp > with one or two precision op-amps should fix that. Do you have any part numbers to hand that I could go and look up please? Certainly if there is a decent custom solution it may be better but I'd have to bear in mind the cost of probably shipping this in from abroad. Thanks for your help, Steve > > JDB. > [had just been looking into this for a transistor matcher/noise test rig I'm > working on] > -- > LART. 250 MIPS under one Watt. Free hardware design files. > http://www.lartmaker.nl/ > > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > -- Steve Rooke - ZL3TUV & G8KVD The only reason for time is so that everything doesn't happen at once. - Einstein ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
J.D. Bakker wrote: I wonder if anyone has done something like this before and could share their experiences. The general principle should work. However, as you're interested in slow changes, there are some error sources that might be unacceptable, including the drift of (differential) channel resistances for the 4066 over temperature, voltage and time. As shown the scheme is also sensitive to impedance mismatch/drift on the two inputs. Charge injection is a bit on the high side on a 4066; a more expensive (A)DG4xx-series chip may improve on that. Or the traditional chopper approach of a mercury wetted reed relay? If you're processing with a sound card, you have the advantage that you don't need to process the samples coming from the time of transition (unlike a traditional analog chopper with synchronous detection), so a fairly crummy relay would probably work. The key is that it can toggle at, say, 100Hz, forever. I don't know if it qualifies as simple/cheap, but Analog Devices and others have single chip low-rate sigma/delta converters with good to excellent properties; these were meant for strain gauges but should be able to track slow-moving control voltages just fine. Interfacing them to a parallel port (or USB PP adapter) should be close to trivial. Do have a close look at the data sheet: some parts have unbuffered inputs, and present a fluctuating input impedance which might couple onto EFC lines. A simple isolation amp with one or two precision op-amps should fix that. The eval board for the part may have a computer interface built into it. JDB. [had just been looking into this for a transistor matcher/noise test rig I'm working on] ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Oz, On 27 June 2010 01:09, Oz-in-DFW wrote: > > > On 6/26/2010 7:12 AM, Steve Rooke wrote: >> >> I next thought about turning the DC into AC by chopping it, IE. >> inverting 50% of the voltage via an oscillator. This way I could pass >> the square wave directly into an unmodified sound card, take >> measurements and then do an RMS calculation on them (really just need >> to flip the sign on, say, the negative readings). >> > I've done similar stuff in work projects, but never written code. I've > thought about this some as well. I'd consider a few things; > > 1. Use the sound card output as the chopper control signal instead of > the discrete unit. You'll have more control and phase sync will > be easier. > * I'd be temped to take the sound card output and run it > through a comparator to square it up, but I'm almost certain > this isn't needed. Sorry, not sure what you mean here. Are you saying that I should derive the chopper frequency directly from a connection to the sound card? I was hoping not to modify the sound card in any way so as to keep it simple. > 2. Buffer the input so that your waveform is not so dependent on > source impedance. Good idea, thanks. > 3. Make the input buffer differential so that you can get some small > amount of ground isolation and CMRR If you look closely at it you can see that it is a differential input. > 4. look at the 4053 mux, it might make your interconnect life easier. Thanks, will do that. > 5. The probelm with chopping is that signal levels around zero don't > have much amplitude and are a challenge to extract from noise. I was under the impression that this was the idea that is used to amplify very low level signals like the output from the likes of strain-gauges. It would surely seem to me to be a problem to amplify small signals around zero due to offsets in the amp unless you do this sort of thing. > 6. If you mix (in the RF receiver sense, not sum in the audio studio > sense) rather than chop the DC offset becomes a phase shift, > generally pretty easy to calibrate for and decode from the output > samples of a sound card. See > http://en.wikipedia.org/wiki/Frequency_mixer Interesting idea, hadn't thought of doing it that way but it's a good idea, thanks. >> I wonder if anyone has done something like this before and could share >> their experiences. I've attached a diagram image (hope it is accepted >> by the list) which is my first go with Eagle so I'm not exactly very >> familiar with it, sorry. The R's and C's in the astable would be set >> to a clock frequency that enables this to work without bias given the >> sampling frequency. I'm not sure if this clock should be slower than >> the sampling frequency or higher, just haven't got my head around that >> yet. > The clock needs to be much higher than the highest frequency of the > input waveform to keep Nyquist happy and things simple. You can do this > inband, but you don't want to. > > If you chop very close to half the soundcard sample rate I suspect > you'll get no output because you'll be in the roofing filter cutoff and > your waveform will integrate to zero. I suspect you want to be 5 - 10X > below that to make waveform recovery easier, and even lower is better. > > So, if you use a 44.1 ksps default rate, Nyquist is 22.05. I'd run the > chopper at less than 1 kHz. The good news is that your input waveform > period is hours (maybe ~100 microhertz) and chopping at 1 Khz will make > 100 Hz response easy and 500 Hz possible with great care and some effort. Right, that makes sense, thanks. >> The R's around the op-amp would need to be set in a ratio that >> transforms the EFC voltage into the range that the sound card can >> handle (that is yet to be calculated by measuring the limits). > Most sound cards I've seen are ~ 1V pk to peak, though some are MUCH > higher. Gives me a ball-park to start work with, thanks. >> If you >> have any suggestions or ways of doing this in a better way, I'd be >> very grateful for the advice. >> > It's worth exactly what you've paid for it... And worth every penny :) Thanks, Steve >> Thanks, >> Steve >> >> > Oz > > -- > mailto:o...@ozindfw.net > Oz > POB 93167 > Southlake, TX 76092 (Near DFW Airport) > > > > > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > -- Steve Rooke - ZL3TUV & G8KVD The only reason for time is so that everything doesn't happen at once. - Einstein ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] crystal oscillators & TPLL
Hi Now, that's one I can heartily agree with. Bob On Jun 26, 2010, at 9:08 AM, Steve Rooke wrote: > Warren, advice from a friend, stop doing this please. > > Everyone else, please stop feeding the troll. > > Steve > > On 26 June 2010 08:32, WarrenS wrote: >> >> Charles Posted a bunch of stuff (below), >> >> Most think I should just ignore him, but I can not help myself, >> he has after all made this one just too easy and silly not to respond to. >> >> I hope Charles did not consider this to be just another "good example" of >> all the 'constructive helpful criticism' I've received. >> >>> His childish tantrums, insults, and outlandish claims are his and his >>> alone. >> >> Funny, I have to wonder if maybe Charles was just reading and referring to >> his own posting. >> Charles's past and latest posting does show that he has several problems, >> both technical as well as emotionally. >> Don't we all?, It is just that most have the good sense and taste not to >> make them so public. >> Sounds to me that he is someone that does need a lot of help, but certainly >> not the kind of help I can give. >> >> Some of the more ironic things, I find in the latest 'attack facts' is his >> statement: [paraphrased] >> >>> [Warren does not understand that all the name calling and insults and >>> attacks have been fair attempts by professional engineers to understand >>> Warren's TPLL implementation] >>> so that they can TRY to ascertain to what degree the TPLL is likely to >>> provide useful results over a broader range of conditions than those that >>> have been publicly demonstrated. >> >> 1) why would a professional engineer have to resort to attacks in a fair >> attempt to understand something so simple or so old and basic? >> >> 2) why would the professional engineers need to have more information, when >> everything that is needed is already on John's site? >> especially if AS CLAMED over and over, they are asking to get this >> information from someone that does not even know what he is doing. >> >> 3) Just how much broader range do they want or need than has already been >> publicly demonstrated, that it works good enough from near DC to 100Hz for >> every device and noise type it has been tested WITH NO exceptions (limited >> only by the controlled OCXO). >> >> There are some things that I do not Understand, Such as: >> >> I do not understand, Nor do I really care, what part of this Charles does >> not understand. >> I will not let his or others shortcoming and non-understandings be MY >> problem >> >> Also posted: >>> >>> I know [some] have said more than once that we should just ignore "the >>> femtosecond thing," but why? >>> (Not that anything turns on this one claim anyway) >> >> Just because of Charles's and others own non-understandings and limitations, >> why one would then feel it is MY reasonability to try and educate someone >> like that is way beyond my understanding. >> but >> >> I'll try again to comment on the femto second thing, since some seem to be >> hung up on that part most of all. >> In order to work good (which no one seems to be denying any more), the TPLL >> method has to hold the two Oscillator's phase differences real, real close >> OVER the Bandwidth of interest. >> Anyone that can understand what limits a noise floor plot, can see that the >> phase differences are being held to about 10 fs at 100 Hz, from the data >> posted on John's site. >> >> Anyone that can do simple math and has a vary basic understanding of the >> TPLLs could calculate for their self with the BW information given in John's >> site, >> that the TPLL is "trying" to hold the phase difference over the Bandwidth of >> interest from DC to 1 KHz down to single digit Femtoseconds varation for >> low noise oscillators. >> If you do not have a favorite Phase detector to use, can use the >> mini-Circuits SYPD-1 for you calculations, (or any other), >> >> A little less obvious but still very easy to calculate with simple math (OK, >> just a little harder than 2+2, but not by too much), >> is that the noise floor limit of a good low noise AMP can give about 1 fs of >> phase differences between the two Oscillators OVER the Bandwidth of >> importance. >> If you do not have a favorite low noise op amp to use, one can use the op-27 >> for their calculations, (or many others), >> >> If Charles or anyone would like to do and post the SIMPLE math to show that >> ANY of femtosecond stuff above is not true, >> and their answer turns out to be different than mine, I'd be more than >> willing to show what they did wrong or different than me. >> >> The fact that Charles and others seem to be confusing 10 MHz Phase jitter >> with 100 Hz and below bandwidth limited Phase differences do >> show they have a few major things missing in their understanding about >> what ADEV is and how it is a frequency stability value over a limited time >> and Bandwidth called tau. >> >> Also if anyone still thinks they
Re: [time-nuts] crystal oscillators & TPLL
I imagine that there are people on this list like myself with limited engineering experience but a healthy appetite for knowledge and an interest in time/frequency measurement. I have little interest in complex math - an inherent brain block methinks, but enjoy building hardware and incorporating such 'blocks' into my projects. I am not necessarily unintelligent - being a Consultant Anaesthetist with 3 degrees, 2 diplomas and currently doing a Masters. My interest in this list comes form having a HP106, 2 rubidium standards and 2 z3801a gps units. Why am I posting on this thread? - simple - I wish to build this beastie described by Warren and try it out on my equipment. I need more guidance than a block diagram. Shame on me for being so ignorant but I respectfully request more info (circuit diagram) emailed privately if necessary. I am not here to criticise, but enjoy reading the discussions of people so knowledgeable in this area. I have my own areas of expertise. Thanks for reading. Steve st...@borisone.demon.co.uk ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
Steve, I think using a voltage-to-frequency converter would solve that problem. They are not too expensive and there are several flavors from Amalog devices and some others. Just set it for a 1KHz start point or maybe 10KHz. BillWB6BNQ Steve Rooke wrote: > I would like to track the EFC voltage in hardware using something > cheap and ready to hand. I was thinking of using a sound card as it > has good resolution but it's obviously only AC coupled so it would not > measure the DC of the EFC. I thought about modifying a sound card to > make it DC coupled but most of them seem to reference the 0V point to > some internal reference voltage hence there is a DC shift there. I > next thought about turning the DC into AC by chopping it, IE. > inverting 50% of the voltage via an oscillator. This way I could pass > the square wave directly into an unmodified sound card, take > measurements and then do an RMS calculation on them (really just need > to flip the sign on, say, the negative readings). > > I wonder if anyone has done something like this before and could share > their experiences. I've attached a diagram image (hope it is accepted > by the list) which is my first go with Eagle so I'm not exactly very > familiar with it, sorry. The R's and C's in the astable would be set > to a clock frequency that enables this to work without bias given the > sampling frequency. I'm not sure if this clock should be slower than > the sampling frequency or higher, just haven't got my head around that > yet. The R's around the op-amp would need to be set in a ratio that > transforms the EFC voltage into the range that the sound card can > handle (that is yet to be calculated by measuring the limits). If you > have any suggestions or ways of doing this in a better way, I'd be > very grateful for the advice. > > Thanks, > Steve > -- > Steve Rooke - ZL3TUV ? G8KVD > The only reason for time is so that everything doesn't happen at once. > - Einstein > > > Name: DCchop.png >DCchop.pngType: PNG image (image/png) > Encoding: base64 > > > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] crystal oscillators & TPLL
Gerard, On 26 June 2010 09:46, Gerard PG5G wrote: > Warren, > > I couldn't care less whether your or any method works or not. I have no > vested interest or opinion whatsoever. Well, if your not interested in this, why are you bothering us by your opinions. Perhaps *you* should "leave us alone" unless you have something useful to say. > I can say however that in the short time I have been on this list I have > grown very tired of the way you hijack any thread that comes along. I wasn't aware that Warren was hijacking "any thread", he has only been commenting on the TPLL thread that he started a long time back now. > Most people who think they have something that is better than ANYTHING done > before have one of two motives: fame or fortune. Or wishing to gift something to the community. > If you are after fame, build one of your TPPL thingies, test it (or better > yet, have it tested) and submit results to a peer reviewed magazine. Glory > will be yours. He has done that here. > If you are after fortune, build one of your TPPL thingies, test it (or > better yet, have it tested), manufacture it, sell it. Money will be yours. He is not interested in fortune, well not per say. > So I guess what I am trying to say is: build one of your TPPL thingies and > have it tested. He already has had this done by a third pay on this list and the results have been published. Please would you like to review the archives of this group before you wade in like this. > I wish you either or both, fame and/or money. I honestly do. However, until > you have decided what you want out of this and how to go about it, please > leave us alone. Well that is very polite coming from someone who has only been here for a self proclaimed "short time". Perhaps you would like to sit this one out and just hit Del whenever you see any of the postings on this subject. Thank you, Steve - ZL3TUV & G8KVD > Thank you. > > SK PG5G > > > WarrenS wrote: >> >> Charles Posted a bunch of stuff (below), >> >> Most think I should just ignore him, but I can not help myself, >> he has after all made this one just too easy and silly not to respond to. >> >> I hope Charles did not consider this to be just another "good example" of >> all the 'constructive helpful criticism' I've received. >> >>> His childish tantrums, insults, and outlandish claims are his and his >>> alone. >> >> Funny, I have to wonder if maybe Charles was just reading and referring to >> his own posting. >> Charles's past and latest posting does show that he has several problems, >> both technical as well as emotionally. >> Don't we all?, It is just that most have the good sense and taste not to >> make them so public. >> Sounds to me that he is someone that does need a lot of help, but >> certainly not the kind of help I can give. >> >> Some of the more ironic things, I find in the latest 'attack facts' is his >> statement: [paraphrased] >> >>> [Warren does not understand that all the name calling and insults and >>> attacks have been fair attempts by professional engineers to understand >>> Warren's TPLL implementation] >>> so that they can TRY to ascertain to what degree the TPLL is likely to >>> provide useful results over a broader range of conditions than those that >>> have been publicly demonstrated. >> >> 1) why would a professional engineer have to resort to attacks in a fair >> attempt to understand something so simple or so old and basic? >> >> 2) why would the professional engineers need to have more information, >> when everything that is needed is already on John's site? >> especially if AS CLAMED over and over, they are asking to get this >> information from someone that does not even know what he is doing. >> >> 3) Just how much broader range do they want or need than has already been >> publicly demonstrated, that it works good enough from near DC to 100Hz for >> every device and noise type it has been tested WITH NO exceptions (limited >> only by the controlled OCXO). >> >> There are some things that I do not Understand, Such as: >> >> I do not understand, Nor do I really care, what part of this Charles does >> not understand. >> I will not let his or others shortcoming and non-understandings be MY >> problem >> >> Also posted: >>> >>> I know [some] have said more than once that we should just ignore "the >>> femtosecond thing," but why? >>> (Not that anything turns on this one claim anyway) >> >> Just because of Charles's and others own non-understandings and >> limitations, >> why one would then feel it is MY reasonability to try and educate someone >> like that is way beyond my understanding. >> but >> >> I'll try again to comment on the femto second thing, since some seem to be >> hung up on that part most of all. >> In order to work good (which no one seems to be denying any more), the >> TPLL method has to hold the two Oscillator's phase differences real, real >> close OVER the Bandwidth of interest. >> Anyone that can understand what limits a noise
Re: [time-nuts] EFC tracking
I wonder if anyone has done something like this before and could share their experiences. The general principle should work. However, as you're interested in slow changes, there are some error sources that might be unacceptable, including the drift of (differential) channel resistances for the 4066 over temperature, voltage and time. As shown the scheme is also sensitive to impedance mismatch/drift on the two inputs. Charge injection is a bit on the high side on a 4066; a more expensive (A)DG4xx-series chip may improve on that. I don't know if it qualifies as simple/cheap, but Analog Devices and others have single chip low-rate sigma/delta converters with good to excellent properties; these were meant for strain gauges but should be able to track slow-moving control voltages just fine. Interfacing them to a parallel port (or USB PP adapter) should be close to trivial. Do have a close look at the data sheet: some parts have unbuffered inputs, and present a fluctuating input impedance which might couple onto EFC lines. A simple isolation amp with one or two precision op-amps should fix that. JDB. [had just been looking into this for a transistor matcher/noise test rig I'm working on] -- LART. 250 MIPS under one Watt. Free hardware design files. http://www.lartmaker.nl/ ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] EFC tracking
On 6/26/2010 7:12 AM, Steve Rooke wrote: > > I next thought about turning the DC into AC by chopping it, IE. > inverting 50% of the voltage via an oscillator. This way I could pass > the square wave directly into an unmodified sound card, take > measurements and then do an RMS calculation on them (really just need > to flip the sign on, say, the negative readings). > I've done similar stuff in work projects, but never written code. I've thought about this some as well. I'd consider a few things; 1. Use the sound card output as the chopper control signal instead of the discrete unit. You'll have more control and phase sync will be easier. * I'd be temped to take the sound card output and run it through a comparator to square it up, but I'm almost certain this isn't needed. 2. Buffer the input so that your waveform is not so dependent on source impedance. 3. Make the input buffer differential so that you can get some small amount of ground isolation and CMRR 4. look at the 4053 mux, it might make your interconnect life easier. 5. The probelm with chopping is that signal levels around zero don't have much amplitude and are a challenge to extract from noise. 6. If you mix (in the RF receiver sense, not sum in the audio studio sense) rather than chop the DC offset becomes a phase shift, generally pretty easy to calibrate for and decode from the output samples of a sound card. See http://en.wikipedia.org/wiki/Frequency_mixer > I wonder if anyone has done something like this before and could share > their experiences. I've attached a diagram image (hope it is accepted > by the list) which is my first go with Eagle so I'm not exactly very > familiar with it, sorry. The R's and C's in the astable would be set > to a clock frequency that enables this to work without bias given the > sampling frequency. I'm not sure if this clock should be slower than > the sampling frequency or higher, just haven't got my head around that > yet. The clock needs to be much higher than the highest frequency of the input waveform to keep Nyquist happy and things simple. You can do this inband, but you don't want to. If you chop very close to half the soundcard sample rate I suspect you'll get no output because you'll be in the roofing filter cutoff and your waveform will integrate to zero. I suspect you want to be 5 - 10X below that to make waveform recovery easier, and even lower is better. So, if you use a 44.1 ksps default rate, Nyquist is 22.05. I'd run the chopper at less than 1 kHz. The good news is that your input waveform period is hours (maybe ~100 microhertz) and chopping at 1 Khz will make 100 Hz response easy and 500 Hz possible with great care and some effort. > The R's around the op-amp would need to be set in a ratio that > transforms the EFC voltage into the range that the sound card can > handle (that is yet to be calculated by measuring the limits). Most sound cards I've seen are ~ 1V pk to peak, though some are MUCH higher. > If you > have any suggestions or ways of doing this in a better way, I'd be > very grateful for the advice. > It's worth exactly what you've paid for it... > Thanks, > Steve > > Oz -- mailto:o...@ozindfw.net Oz POB 93167 Southlake, TX 76092 (Near DFW Airport) ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] crystal oscillators & TPLL
Warren, advice from a friend, stop doing this please. Everyone else, please stop feeding the troll. Steve On 26 June 2010 08:32, WarrenS wrote: > > Charles Posted a bunch of stuff (below), > > Most think I should just ignore him, but I can not help myself, > he has after all made this one just too easy and silly not to respond to. > > I hope Charles did not consider this to be just another "good example" of > all the 'constructive helpful criticism' I've received. > >> His childish tantrums, insults, and outlandish claims are his and his >> alone. > > Funny, I have to wonder if maybe Charles was just reading and referring to > his own posting. > Charles's past and latest posting does show that he has several problems, > both technical as well as emotionally. > Don't we all?, It is just that most have the good sense and taste not to > make them so public. > Sounds to me that he is someone that does need a lot of help, but certainly > not the kind of help I can give. > > Some of the more ironic things, I find in the latest 'attack facts' is his > statement: [paraphrased] > >> [Warren does not understand that all the name calling and insults and >> attacks have been fair attempts by professional engineers to understand >> Warren's TPLL implementation] >> so that they can TRY to ascertain to what degree the TPLL is likely to >> provide useful results over a broader range of conditions than those that >> have been publicly demonstrated. > > 1) why would a professional engineer have to resort to attacks in a fair > attempt to understand something so simple or so old and basic? > > 2) why would the professional engineers need to have more information, when > everything that is needed is already on John's site? > especially if AS CLAMED over and over, they are asking to get this > information from someone that does not even know what he is doing. > > 3) Just how much broader range do they want or need than has already been > publicly demonstrated, that it works good enough from near DC to 100Hz for > every device and noise type it has been tested WITH NO exceptions (limited > only by the controlled OCXO). > > There are some things that I do not Understand, Such as: > > I do not understand, Nor do I really care, what part of this Charles does > not understand. > I will not let his or others shortcoming and non-understandings be MY > problem > > Also posted: >> >> I know [some] have said more than once that we should just ignore "the >> femtosecond thing," but why? >> (Not that anything turns on this one claim anyway) > > Just because of Charles's and others own non-understandings and limitations, > why one would then feel it is MY reasonability to try and educate someone > like that is way beyond my understanding. > but > > I'll try again to comment on the femto second thing, since some seem to be > hung up on that part most of all. > In order to work good (which no one seems to be denying any more), the TPLL > method has to hold the two Oscillator's phase differences real, real close > OVER the Bandwidth of interest. > Anyone that can understand what limits a noise floor plot, can see that the > phase differences are being held to about 10 fs at 100 Hz, from the data > posted on John's site. > > Anyone that can do simple math and has a vary basic understanding of the > TPLLs could calculate for their self with the BW information given in John's > site, > that the TPLL is "trying" to hold the phase difference over the Bandwidth of > interest from DC to 1 KHz down to single digit Femtoseconds varation for > low noise oscillators. > If you do not have a favorite Phase detector to use, can use the > mini-Circuits SYPD-1 for you calculations, (or any other), > > A little less obvious but still very easy to calculate with simple math (OK, > just a little harder than 2+2, but not by too much), > is that the noise floor limit of a good low noise AMP can give about 1 fs of > phase differences between the two Oscillators OVER the Bandwidth of > importance. > If you do not have a favorite low noise op amp to use, one can use the op-27 > for their calculations, (or many others), > > If Charles or anyone would like to do and post the SIMPLE math to show that > ANY of femtosecond stuff above is not true, > and their answer turns out to be different than mine, I'd be more than > willing to show what they did wrong or different than me. > > The fact that Charles and others seem to be confusing 10 MHz Phase jitter > with 100 Hz and below bandwidth limited Phase differences do > show they have a few major things missing in their understanding about > what ADEV is and how it is a frequency stability value over a limited time > and Bandwidth called tau. > > Also if anyone still thinks they can make a reasonable data set file that > shows where the TPLL will mess up, Go for it. > I'm still willing to try and prove to all that will NOT EVER be the case. > OR is it still OK for some expert to make an unsubstantiated and false c
[time-nuts] EFC tracking
I would like to track the EFC voltage in hardware using something cheap and ready to hand. I was thinking of using a sound card as it has good resolution but it's obviously only AC coupled so it would not measure the DC of the EFC. I thought about modifying a sound card to make it DC coupled but most of them seem to reference the 0V point to some internal reference voltage hence there is a DC shift there. I next thought about turning the DC into AC by chopping it, IE. inverting 50% of the voltage via an oscillator. This way I could pass the square wave directly into an unmodified sound card, take measurements and then do an RMS calculation on them (really just need to flip the sign on, say, the negative readings). I wonder if anyone has done something like this before and could share their experiences. I've attached a diagram image (hope it is accepted by the list) which is my first go with Eagle so I'm not exactly very familiar with it, sorry. The R's and C's in the astable would be set to a clock frequency that enables this to work without bias given the sampling frequency. I'm not sure if this clock should be slower than the sampling frequency or higher, just haven't got my head around that yet. The R's around the op-amp would need to be set in a ratio that transforms the EFC voltage into the range that the sound card can handle (that is yet to be calculated by measuring the limits). If you have any suggestions or ways of doing this in a better way, I'd be very grateful for the advice. Thanks, Steve -- Steve Rooke - ZL3TUV & G8KVD The only reason for time is so that everything doesn't happen at once. - Einstein <>___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] PICTIC II
Ah! Sounds like you were subjected to the "Bruce factor". It's a case of publish a design and then expect to be insulted about it. I'm glad you weathered the storm and wish to thank you for your present. Steve On 26 June 2010 19:11, Richard H McCorkle wrote: > Fellow Time-Nuts, > When I first uploaded the Simple PICTIC interpolating time interval > counter to the K04BB site in 12/08 and presented it to the group as > a Christmas present my goal was to get amateurs building their own > interpolating time interval counters for GPS monitoring and making > improvements to my design. The interpolator in the PICTIC was > “borrowed” from an early HP counter with minor modifications so I > didn’t design it and had no personal attachment to it. My forte is > writing PIC assembly code for time and frequency applications and > I make no claims of being a hardware design engineer. I am just > an amateur interested in writing PIC code and was building my own > PICTIC. I wanted to share my design, as I couldn’t afford multiple > commercial high-resolution time interval counters to monitor my > standards and thought there might be others with similar needs that > could benefit from a low cost TIC design. > Testing the HP interpolator with my code showed it had sufficient > capability to be useful in a low cost TIC for GPS monitoring where > a modest resolution increase was desired. In response to my posting > I was subjected to a barrage on the shortcomings of the interpolator > with few comments on the rest of the design. I responded to the > comments based on test results showing it was sufficient for purpose > and suggested trying the design before saying it wasn’t usable. The > response returned was “I wouldn't waste my time with a circuit that > is inherently as non-linear as yours.” so the rest of the design was > discounted by others as not being worth the time to evaluate. I began > a discussion on line about the interpolator, which was destined to > turn into a long and annoying thread similar to the recent TPLL > discussion. Rather than getting into it on line I elected to put a > stop to it by saying publicly I was at fault and would begin further > study, as I am not as thick skinned as Warren. Subsequent independent > testing of the PICTIC by William Riley showed the interpolators were > linear to the 10-bit ADC resolution over the measurement range with > suitable accuracy for the intended application of GPS monitoring > as originally stated. > Over the last 18 months I have developed a new diode switched > interpolator based on the comments made on line and have thoroughly > tested it. Some suggestions made improvements in the performance and > some resulted in poorer performance. I incorporated those suggestions > that made improvements, eliminated those that made things worse, and > once I was satisfied I sent the revised interpolator design directly > to Bruce off line. Based on comments and suggestions he returned > during a long series of emails I incorporated additional changes in > the code, front-end, and interpolator designs and tested those > changes until I was satisfied with the performance and he had no > further comments. I am finally satisfied with the new design and > admit that by incorporating the majority of the suggestions made > the new interpolator has significantly better linearity than the > HP interpolator used in the PICTIC and it is now suitable for > higher resolutions. > I was reluctant to post the PICTIC II in this forum, as I don’t > want to get in another public discussion of its faults without any > discussion of the merits of a $50 serial output interpolating TIC > on a 3.8” x 2.5” thru-hole board designed for amateur construction. > The K04BB WIKI PICTIC page was recently updated to include the > PICTIC II code and ExpressPCB board layout and schematic files for > those that might be interested. > > http://www.ko4bb.com/dokuwiki/doku.php?id=precision_timing:pictic > > The PICTIC II incorporates the new interpolator, requires a delay > between the inputs, and uses a low stability XO timebase with > software peak detection for calibration with provision for an > external timebase like the PICTIC to minimize size and cost. > > Richard > > > > > ___ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > -- Steve Rooke - ZL3TUV & G8KVD The only reason for time is so that everything doesn't happen at once. - Einstein ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
[time-nuts] PICTIC II
Fellow Time-Nuts, When I first uploaded the Simple PICTIC interpolating time interval counter to the K04BB site in 12/08 and presented it to the group as a Christmas present my goal was to get amateurs building their own interpolating time interval counters for GPS monitoring and making improvements to my design. The interpolator in the PICTIC was borrowed from an early HP counter with minor modifications so I didnt design it and had no personal attachment to it. My forte is writing PIC assembly code for time and frequency applications and I make no claims of being a hardware design engineer. I am just an amateur interested in writing PIC code and was building my own PICTIC. I wanted to share my design, as I couldnt afford multiple commercial high-resolution time interval counters to monitor my standards and thought there might be others with similar needs that could benefit from a low cost TIC design. Testing the HP interpolator with my code showed it had sufficient capability to be useful in a low cost TIC for GPS monitoring where a modest resolution increase was desired. In response to my posting I was subjected to a barrage on the shortcomings of the interpolator with few comments on the rest of the design. I responded to the comments based on test results showing it was sufficient for purpose and suggested trying the design before saying it wasnt usable. The response returned was I wouldn't waste my time with a circuit that is inherently as non-linear as yours. so the rest of the design was discounted by others as not being worth the time to evaluate. I began a discussion on line about the interpolator, which was destined to turn into a long and annoying thread similar to the recent TPLL discussion. Rather than getting into it on line I elected to put a stop to it by saying publicly I was at fault and would begin further study, as I am not as thick skinned as Warren. Subsequent independent testing of the PICTIC by William Riley showed the interpolators were linear to the 10-bit ADC resolution over the measurement range with suitable accuracy for the intended application of GPS monitoring as originally stated. Over the last 18 months I have developed a new diode switched interpolator based on the comments made on line and have thoroughly tested it. Some suggestions made improvements in the performance and some resulted in poorer performance. I incorporated those suggestions that made improvements, eliminated those that made things worse, and once I was satisfied I sent the revised interpolator design directly to Bruce off line. Based on comments and suggestions he returned during a long series of emails I incorporated additional changes in the code, front-end, and interpolator designs and tested those changes until I was satisfied with the performance and he had no further comments. I am finally satisfied with the new design and admit that by incorporating the majority of the suggestions made the new interpolator has significantly better linearity than the HP interpolator used in the PICTIC and it is now suitable for higher resolutions. I was reluctant to post the PICTIC II in this forum, as I dont want to get in another public discussion of its faults without any discussion of the merits of a $50 serial output interpolating TIC on a 3.8 x 2.5 thru-hole board designed for amateur construction. The K04BB WIKI PICTIC page was recently updated to include the PICTIC II code and ExpressPCB board layout and schematic files for those that might be interested. http://www.ko4bb.com/dokuwiki/doku.php?id=precision_timing:pictic The PICTIC II incorporates the new interpolator, requires a delay between the inputs, and uses a low stability XO timebase with software peak detection for calibration with provision for an external timebase like the PICTIC to minimize size and cost. Richard ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
