Re: [time-nuts] CPLDs for clock dividers
On 02/03/2010 07:59 PM, Matt Ettus wrote: Does anyone have any experience using CPLDs for very low phase noise dividers? You can get an XC9536XL from Xilinx for around $1, and I thought it would make a good divide by 2 through 10 device. 9500XL (3.3V) is, I believe, similar to 9500 (5V). A downside to the way the 9500 is built is the rather high static power consumption, and that the static current consumption is dependent on the internal state of the signals. If you have a single divider in a part it is not a problem. But if you divide by 2 and 10, the '2' output will be modulated by the divide-by-10 in ways one may not expect since the input threshold of the clock signal changes. An alternative I like better is XCR3032XL, which is entirely CMOS. No current-source-driven AND/OR, with a static current consumption that isn't dependent on the internal state. Thus, the internal state does not move the threshold of the global clock around. Output slew rate is faster than 1ns when driving into 150 ohm. /Kasper Pedersen ___ 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] Rb Oscillator - rather fundamental question
On 02/23/2010 11:10 PM, Matthew Smith wrote: BTW: does anyone know if a 0.55V p-t-p sine wave from an Rb source would be enough to clock an Atmel AVR microcontroller? The crystal/clock input *is* an amplifier, but didn't know if I'd need to do anything to the signal first, to get it closer to the 5V logic level. It will. Set the fuses as you would have for a 10MHz crystal, and capacitively couple the source to XTAL1. Leave XTAL2 open. Do not set the fuses for 'external clock mode'. Do put something like 100pF+1k Ohm in series with the input. While they won't promise anything, I have deliberately run 1A into the protection diodes of an ATMega16 for many seconds and still had a functional part. /Kasper Pedersen ___ 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] simple, cheap clock for the local LAN
On 04/05/2010 03:47 PM, Eugen Leitl wrote: Hi, I'm thinking about putting a local clock standard (nothing too fancy, quartz would probably do) for the local LAN so that I have more or less stable clocks when GPS is down for whatever reason. I have zero clue about time standards for the low end. Can anyone recommend anything affordable? Thanks. You need to give a ballpark figure for what you want. A beyond-awful homecooked OCXO like this one http://n1.taur.dk/timenuts/img_1017_crop.jpg http://n1.taur.dk/timenuts/img_1020_crop.jpg http://n1.taur.dk/clock/rate.png is the trace for something of similar build quality, and as can be seen, it is on the order of 3ppb/day. Noise is due to this being a cheap DCF77 clock driven server. Combined with something like a D945GSEJT board with the crystal removed does give sub-ms 24 hour holdover with plain ntpd. Now, what is the requirement? /Kasper ___ 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] A weekend project
(mostly for entertainment value) As the talk fell on what is a weekend project for timenuts, here is my Easter project. I have previously built a few GPSDOs with my own home-rolled OCXOs. The VC-OCXOs are a problem: I run into linearity problems, noise, the residue of the flux changing with time, and DAC drift. Voltage control is annoying. So, I built a PLL that will take a 14.7456MHz inaccurate non-adjustable OCXO (something that does not have a large common denominator with 20MHz) and generate 20MHz with 1E-12 or arbitrarily better resolution. The method is this: A microcontroller is clocked from a 20MHz VCXO. It samples the 14.7MHz using a flipflop (something I did before), calculates the phase from a number of samples (211ps resolution single shot, 22ps multiple), calculates the predicted phase at the phase acquisition time (it turns rapidly as there is a 5.3MHz beat), and uses the difference to lock the 20MHz VCXO. The whole thing, with the exception of the VCXO, is software. Stepping the synthesizer in 1E-12 steps is strangely satisfying. So is plotting the noise, as the VCXO is deliberately awful; The close-in noise (30ms tau) is the VCXO done with the microcontroller's built-in oscillator, above that it is that of the OCXO. http://n1.taur.dk/timenuts/synth_v2.pdf (2-hour very much draft version - you may want to skip to the very last page) Probably not a new design, but I did re-invent it. Proper measurements will have to wait a few weeks. /Kasper Pedersen ___ 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] A different timenuts interest
On 07/22/2010 02:13 AM, Morris Odell wrote: The pendulum requires a sustaining system to compensate for the inevitable energy loss with each swing. The system is located in the building and therefore rotates relative to the pendulum. It needs to provide an impulse which does not affect the plane of oscillation of the pendulum. I was thinking of an electromagnet located below the centre of the swing which would be pulsed appropriately as the bob passes over it. As a kid, I did a self sustaining pendulum with no moving parts and no magnets: The bob was suspended by two parallel wires, lacquered together, and shorted at the bob end. As the bob passed over the center, a one-shot sent a good-sized current pulse through the wire, heating it, making it slightly longer, and then shrink again as it cooled at the outside of the swing. For a heavy pendulum, and thick wire, the time constant in the cooling phase will likely make this infeasible. /Kasper Pedersen ___ 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] Reference source 1E10^8 only ...Ideas
On 07/24/2010 02:33 PM, Don Collie jnr wrote: I need a frequency reference for my frequency counters, it needs to be accurate to at least +or- 1 part in 10,000,000. For a home-cooked gpsdo in the really low end I have this bit to offer: http://n1.taur.dk/simplexdo/ It was built to be an always calibrated and ready, pocketable, instant-on source. Not to compete with something like a second hand Tbolt. /Kasper Pedersen ___ 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] Homebrew frequency counter, need help
On 11/27/2014 03:08 PM, lllaaa wrote: > Hi guys, > I've just get my homebrew counter working. And the resolution seems 10x > better than my RACAL DANA 1992. > This counter is heavily inspired by the idea from Kasper Pedersen. > http://n1.taur.dk/permanent/frequencymeasurement.pdf > STM32F051RB & EMP240T100C5 do the control and counting job. TDC-GP22 as > the interpolator. Linear regression is done by CPU. > There are no fancy analog front for both signal path and refclk path. > I'm using two SN75ALS176 and the schmitt input of CPLD to do the job. > I've noticed that the 10s gate does not get more meaningful > digits(looks worse than 1s gate). So here are the questions: > 1) I'm wondering if I could say this is an 11 digits/s counter? > 2) How can I improve that? Is it limited by the 485 transceiver? I can > switch to a faster MCU, that gets more measures per second, but I think > that only gets no more than 2 stable bits. > A few things to try, and learned: Try measuring the reference against itself, triggering on the same edge you clock the cpld on. If your VCC is wandering, your threshold will wander, and you get wandering phase out of the schmitt trigger in the CPLD. When I built my counter, I had much fun with my 'front end' (AC04s) having variable heating, and thus variable delay, depending on slew rate. I ended up giving each input channel its own low noise regulator to keep crosstalk from going through VCC. I think I calculated that, for a 10MHz 10dBm signal, 6mV threshold error is 100ps. I assume you can pick which edge to trigger on. Measure the reference against itself, and read out interpolator (phase) data on either edge. When I did my counter, I had ground current flowing through the coax between the counter reference input, and the house standard. I had been silly and chosen a low cutoff frequency for the dc-block capacitor in the reference input, which meant that the resulting voltage over the coax shield resistance got through the dc-block, and caused phase modulation. On the rising edge, the noise was low. On the falling edge, it was nasty and wandering, since when you add LF to 10MHz, and then slice it, the pulsewidth varies. >From bad experience, try dumping out adjusted timestamps of almost-10MHz, and plot actual timestamp vs predicted timestamp. It will show you if you have 10MHz crosstalk, or, if as I did, you added the interpolator value instead of subtracting it. In my case the counter appeared to work most of the time, while giving wrong readings all of the time. And congratulations on getting it working. /Kasper Pedersen (When getting 10MHz out of FEI5680As, mine had ferrite blocks around the dsub connectors, and while I could get a cleaner signal by shorting GND to shield on the connector, it was better again when I bypassed the connector entirely and ran coax.) ___ 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] Leap Second results: cheap GPS/1PPS receivers
On 07/01/2015 05:23 PM, Tom Van Baak wrote: > I logged NMEA from three cheap ($15-$50) GPS/1PPS receivers, the kind popular > with hobbyists: parallax(good), reyax(good), adafruit(bad). > > 3) Adafruit Ultimate GPS Breakout > http://www.adafruit.com/products/746 > > If someone else has data from this model receiver, please let me know. > I have: - MC-1010 (MTK3339, same chipset and code as the adafruit one) - GPS-1513R (Venus 624) <-- FUN! - LTE-Lite - old GPS18x The MTK3339 with the same (standard) firmware is also used in the MC-1010. I have output from the 28th and ahead. Same behaviour. $GPGGA,235958.000,5610.4160,N,00929.5793,E,1,10,0.80,84.8,M,43.3,M,,*5E $GPRMC,235958.000,A,5610.4160,N,00929.5793,E,0.01,171.62,300615,,,A*66 $GPGGA,235959.000,5610.4160,N,00929.5793,E,1,10,0.80,84.8,M,43.3,M,,*5F $GPRMC,235959.000,A,5610.4160,N,00929.5793,E,0.01,171.62,300615,,,A*67 $GPGGA,235959.000,5610.4160,N,00929.5793,E,1,10,0.80,84.8,M,43.3,M,,*5F $GPRMC,235959.000,A,5610.4160,N,00929.5793,E,0.01,171.62,300615,,,A*67 $GPGGA,00.000,5610.4160,N,00929.5793,E,1,10,0.80,84.8,M,43.3,M,,*5E $GPRMC,00.000,A,5610.4160,N,00929.5793,E,0.02,171.62,010715,,,A*66 $GPGGA,01.000,5610.4160,N,00929.5793,E,1,10,0.80,84.8,M,43.3,M,,*5F $GPRMC,01.000,A,5610.4160,N,00929.5793,E,0.01,171.62,010715,,,A*64 But, SkyTraq venus 624 (RF Solutions GPS-1513R) is much more fun. It backsteps one minute(!), in addition to the other 'features' it has. $GPGGA,235958.187,5610.4163,N,00929.5757,E,1,11,0.8,84.6,M,41.0,M,,*65 $GPRMC,235958.187,A,5610.4163,N,00929.5757,E,000.0,308.5,300615,,,A*6F $GPGGA,235959.187,5610.4163,N,00929.5757,E,1,11,0.8,84.6,M,41.0,M,,*64 $GPRMC,235959.187,A,5610.4163,N,00929.5757,E,000.0,308.5,300615,,,A*6E $GPGGA,235900.000,5610.4163,N,00929.5757,E,1,11,0.8,84.6,M,41.0,M,,*66 $GPRMC,235900.000,A,5610.4163,N,00929.5757,E,000.0,308.5,300615,,,A*6C $GPGGA,00.187,5610.4163,N,00929.5757,E,1,11,0.8,84.5,M,41.0,M,,*66 $GPRMC,00.187,A,5610.4163,N,00929.5757,E,000.0,308.5,010715,,,A*6C $GPGGA,01.187,5610.4163,N,00929.5757,E,1,11,0.8,84.5,M,41.0,M,,*67 $GPRMC,01.187,A,5610.4163,N,00929.5757,E,000.0,308.5,010715,,,A*6D The Venus on the LTE-Lite, with the updated timing firmware, is sane. I also have the 20MHz output timestamped every 5ms, nothing odd happened. $GPGGA,235959.000,5610.4155,N,00929.5739,E,2,09,0.9,85.4,M,41.0,M,,*6F $GPRMC,235959.000,A,5610.4155,N,00929.5739,E,000.0,000.0,300615,,,D*66 $GPGGA,235960.000,5610.4155,N,00929.5739,E,2,09,0.9,85.4,M,41.0,M,,*65 $GPRMC,235960.000,A,5610.4155,N,00929.5739,E,000.0,000.0,300615,,,D*6C $GPGGA,00.000,5610.4155,N,00929.5739,E,2,09,0.9,85.4,M,41.0,M,,*6E $GPRMC,00.000,A,5610.4155,N,00929.5739,E,000.0,000.0,010715,,,D*64 Old Garmin GPS18x did the same as the MTK3339: $GPRMC,235958,A,5611.0119,N,00932.1092,E,000.0,136.1,300615,002.0,E,D*10 $GPGGA,235958,5611.0119,N,00932.1092,E,2,10,0.8,41.5,M,41.7,M,,*72 $GPRMC,235959,A,5611.0119,N,00932.1093,E,000.0,136.1,300615,002.0,E,D*10 $GPGGA,235959,5611.0119,N,00932.1093,E,2,10,0.8,41.4,M,41.7,M,,*73 $GPRMC,235959,A,5611.0119,N,00932.1094,E,000.0,136.1,300615,002.0,E,D*17 $GPGGA,235959,5611.0119,N,00932.1094,E,2,10,0.8,41.3,M,41.7,M,,*73 $GPRMC,00,A,5611.0120,N,00932.1094,E,000.0,136.1,010715,002.0,E,D*1F $GPGGA,00,5611.0120,N,00932.1094,E,2,10,0.8,41.2,M,41.7,M,,*79 $GPRMC,01,A,5611.0120,N,00932.1094,E,000.0,136.1,010715,002.0,E,D*1E $GPGGA,01,5611.0120,N,00932.1094,E,2,10,0.8,41.2,M,41.7,M,,*78 /Kasper Pedersen ___ 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] UPS for my time rack
On 10/11/2015 12:07 AM, Dr. David Kirkby (Kirkby Microwave Ltd) wrote: > Essentially the charging circuits are not designed to run as long as needed > to charge big batteries. Even on ones designed for external batteries, > there's a recommended limit on the size of them. So if you think you might > want to increase runtime by adding some batteries, buy one designed for > that service. I have gone down that route, so I have some real data to share: My (soon to be replaced) backup is an old back-ups CS 500, with a rewired battery pack out of an RT3000 UPS. So instead of 7Ah, the UPS has 40Ah. With plenty of fuses. When charging the standard 7Ah battery, the UPS delivers about 0.7A (from memory) for many hours, and sits at about 14C above ambient. When charging the 40Ah, the current is the same, the temperature is the same, just for longer, as it should be, since the thermal time constant is much shorter than the time it takes to charge the 7Ah. Where this has problems is during discharge: I have about 55W load on it, which in turn is at least 5A on the battery. After 2 hours a timer in the UPS shuts it off, regardless of battery voltage. Also, if you run the UPS at high load where the standard battery lasts shorter than the thermal time constant, then there might well be trouble. The replacement, a back-ups pro 1500 behaves differently. It has support for external battery packs, and will happily run for at least 5 hours, even when the external-pack-present signal is not connected. The external-pack-present signal does make a difference when charging; Without it, it charges at 0.7A. with it, it charges at 1.5A, and the fan is on continuously. /Kasper Pedersen ___ 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] adafruit $40 GPS
On 05/14/2013 07:57 PM, George Lu wrote: > According to the documentation at > http://www.adafruit.com/datasheets/GlobalTop-FGPMMOPA6H-Datasheet-V0A.pdf: > > High accuracy 1-PPS timing support for Timing Appli > cations (10ns jitter) > > George Here, have some data. I am testing a few MT3339 modules at the moment; They all seem to have the same firmware. The (live, updates every minute) plots of one are at http://cm.kasperkp.dk/mc1010p.png (phase) http://cm.kasperkp.dk/mc1010a.png (adev of this and others) http://cm.kasperkp.dk/mc1010p.txt (pps phase in seconds) It is far better than others cheap navigation receivers I have tested, but it does peculiar things sometimes, such as the 60ns excursion at 2500 seconds. And I am looking for a navigation receiver at the moment, not a timing one, so this is just because I can. /Kasper Pedersen Note: The receiver is fed off of a splitter from an external antenna. The counter is one of my homebrew ones with ~250ps-ish resolution, and the FEI-5680A in it was turned on 15 mins before starting the run, so it's still settling. ___ 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] Advice on good reception for radio clocks
On 06/27/2012 04:04 PM, Tony Finch wrote: > Are there any basic steps I should take to improve the reception quality > of a radio clock? I have a cheap and cheerful DCF77 receiver for > connecting to some GPIO pins, but its PPS output is basically noise with > maybe a one-second period. Perhaps it's just cheap and nasty. > Be careful what you connect the ground of the receiver to. When I did my DCF77 receiver, my first source of interference was the common noise on the output of the supply I was powering it off of. I went to a linear power supply, and things were good for a few years. Then they installed remote-reading power meters in the neighbourhood, and DCF77 was completely jammed. The meters talk back on 75kHz with ~6kHz bandwidth. Halfway by accident I found out that if I earth the receiver well enough, thereby shunting off some of the 75kHz common mode signal, I get mostly reliable reception all day. I would suggest, at least for development, a battery and an optocoupler to isolate the receiver section from conducted interference. Hmm, I do have a Pi. And when you have trouble decoding the signal at around 04 in the morning, you too will have rediscovered sferics, and the need for a filter that handles that. /Kasper Pedersen ___ 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] have 10MHz need 19.5Mhz
On 06/02/2013 10:21 PM, Attila Kinali wrote: > > Another stupid idea: You have a 19.5MHz crystal, probably some > odd PLL and a varicap diode lying around: Build a VCXO out of the > crystal and lock it to the 10MHz using the PLL. > It has been done with an xor gate, a couple of passives, and the varicap diode. There processor performs as the divider: This bit of code will configure the gpclk0 block to output 10MHz on pin 7 of the connector: http://n1.taur.dk/pi10MHz.tgz The signal on the pin is the 19.2MHz crystal multiplied to 500MHz (the GPU PLL), and then divided back down to 10MHz. Feed that signal, and the external reference signal into an xor gate or other phase detector of choice. Put an RC filter of around 50us on the output, and connect that to the varicap. Picture of the thing as built: http://n1.taur.dk/piclock.jpg The only thing on the other side of the board is the connector and the 74HC86 in DIP. I don't know what my varicap is, it came out of an old gps receiver. /Kasper Pedersen ___ 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] Computing GPS Distance Error in Time
On 11/04/2013 06:30 PM, Bill Dailey wrote: > The pps on that unit is good. Compares favorably to the ublox6T. I can't say > how favorably. > > I am working on quantifying its pps as compared to my 2 fury units and the > ublox 6T. They are pretty good, being dirt cheap nav receivers. In this plot, it is identical to the MC-1010 (same reference design, same firmware). http://cm.kasperkp.dk/mc1010ga.png (adev of various) http://cm.kasperkp.dk/mc1010p.png (phase, fei5680) The MC-1010 is using the 3339 chipset, and is GPS-only. The MC-1010-G has the chipset, and is GPS+GLONASS. When testing, try fast temperature shifts; I found that it would shift the PPS much more than I expected, as if there is a rather slow filter driving the PPS output. > > Have another trick up my sleeve but need a Linux kernel hacker. I don't want > to go down another rabbit hole if I can avoid it. > I have stuck my fingers in the timekeeping code in the kernel, and managed to get away unscathed. If it is not too complicated, and interesting, shoot me a mail directly. I reserve the right to decline, and the right to give up:-) /Kasper Pedersen ___ 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] Happy Holidays
Bruce Griffiths wrote: > Yet another option is to use the simple TAC (figure 2) in: > http://n1.taur.dk/permanent/frequencymeasurement.pdf > > The only drawback with this circuit is that the capacitor voltage at > reset isn't zero. > The capacitor reset voltage is also temperature dependent. > Since the change in the capacitor voltage is proportional to the > synchroniser delay, the solution to this is obvious: > Measure the capacitor reset voltage with the ADC and subtract it from > the capacitor voltage held on the capacitor after charging is terminated > by the STOP input. > Dedicated ACMOS (or equivalent low output impedance) flipflops can be > used to drive the diode switches directly. > The drawback turned out to be a useful feature. For one, the TLC072 opamp I used does not swing below V-+0.5V, and I had no negative supply. Also, the 0.8V reset keeps me out of the worst varactor-behaviour of the opamp's clamp diodes, and ditto the 'i' diode, which would have caused some nonlinearity for short pulses. Finally, the leakage of 'i' in hold has less variation over 0.8V-2.5V (1nA) than 0.0V-2.0V (5nA), http://n1.taur.dk/permanent/1n4148.png , which in turn means that the delay from the stop to the adc s/h closes doesn't introduce as much nonlinearity, as long as it is constant. The way I do the subtraction is to measure the reset voltage after the start-stop-convert-reset sequence (one could possibly misread your description as suggesting converting the reset voltage first) and before rearming - that way drift doesn't creep in while waiting for the edge. The charge redistribution adc would have been good, true, but there wasn't one in my junkbox. /Kasper Pedersen ___ 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] DSP WWVB Receiver Idea
Brooke Clarke wrote: On the PICLIST there has been a discussion about the CMAX WWVB front ends and noise. Olin mentioned that you could use a dsPIC to look at the I and Q signals resulting from mixing the WWVB signal with a carrier at 60 kHz. His example case was to use a cheap crystal (+ or - 3 Hz) and so use a 10 Hz low pass filter on the I and Q signals prior to squaring and adding them. I've built such a thing ( http://n1.taur.dk/dcf/ ). The zero-if I/Q approach has a few things that make it less ideal than it sounds. There's the 1/f noise, discovering and compensating for DC offset on each of the channels requires that you remove the input, and it might not be a nice divider from 10MHz. If you choose a small arbitrary offset you can solve these problems in software, only the filters in hardware need to be wider. Having the first filters wide, I found, was a good thing: In the very early morning I get a lot of sferics, and my steep filter rang like a bell with every crackle. A low-Q front end allowed throwing those samples away. Since that was done I have added a narrow bandwidth phase integrator (2mHz) in software, and it will happily pull out ~10ns rms phase with a +60dB carrier 1Hz from center. It even stayed locked when the antenna amplifier broke and output 5Vp-p instead. The real advantage of the I/Q method is that the bandpass filter becomes two lowpass, and two lowpass is easier than a similar width bandpass with enough precision and phase stability to be centered around 60kHz (and if you use crystal resonators in the front end you can't track anything else, and you get a problem with suppressing sferics). You might not be able to get continuous reception no matter how hard you try; I've seen inversions where the carrier just slowly fades and comes back inverted with no apparent phase jumps (it looks like extremely slow bpsk). If I did it today I'd try phk's approach first. Preferably with a somewhat tuned antenna to keep harmonics from PAL horizontal retrace from clipping the converter. The one above was built with what was available in the junkbox at the time. /Kasper Pedersen ___ 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] Heated crystal? & Rb tube corrosion (FE-5680A)
On 12/01/2011 11:18 AM, Attila Kinali wrote: > > Other than that one spot, i have to say that the whole device looks > like new. No dirt, no corrosion where you'd expect it. Even the solder > joints look like new. > And I did not attempt to clean anything before I took the pictures, either. The only dirt was a grey tint above and below the oven, where the convection had deposited all non-gaseous dirt. I have a few of these, two of which have appear to have actual corrosion (rust red spots) on the outside, but are just as clean inside. I wonder what that does to the magnetic properties to the case. Posting to nuts is dangerous:-) So far the page has had 243 non-bot visitors, 20 people downloaded all the full resolution images, and it cost me 10GB this week. An impressive figure considering the topic. /Kasper Pedersen ___ 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] Ebay FE-5680A Rb: Are they that good?
On 12/08/2011 07:58 PM, Chuck Forsberg WA7KGX N2469R wrote: > The price seems reasonable, esp. with free shipping. > Does anyone have experience with these? I ran one of this type against a tbolt. I was testing a super simple 6-channel phase comparator at the time, and it was not quite as good as hoped, so the plot is limited by the 25ps noise of the comparator. time offset plot: http://n1.taur.dk/timenuts/p444.png adev plot: http://n1.taur.dk/timenuts/p445.png I had (rather coarsely) adjusted the unit the day before, and a 2*10^-12 offset has been removed before plotting p444. It has been quite useful, moreso since I put it in a rugged box. /Kasper Pedersen ___ 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] Thunderbolt? (re simple gpsdo.)
On 01/01/2012 12:23 PM, Tom Harris wrote: > I too have been mulling over a minimal GPSDO. > > Has anyone mentioned a PWM DAC? Even a "toy" microcontroller like the > Atmel AVRs can generate 16 bit resolution PWM, which sounds like it .. > So a minimal GPSDO looks like a simple microcontroller (I would use > Arduino since we use these at work whenever we need to build something > in an afternoon) with a filter for a PWM DAC, so one opamp and a few > discretes. Of course as I write software for such beasts as a trade I > am giving the software a zero cost. > A 'bike light' sized microcontroller can do the job. Today I would have used something larger, and saved development time. On the other hand, having only room for 512 instructions prevented spending time on non-essential features. http://n1.taur.dk/simplexdo/ I used 8 bit PWM into a two-stage RC filter, making it easier to filter. To get 16 bit resolution I added delta-sigma modulation on top of the PWM. The ATTiny13 is not the right part for the job; It has no input capture. So space- and time consuming tricks are necessary to get cycle-accurate capture. Recently I discovered that some of the microcontrollers in the drawer would do nanosecond capture all on their own, but with a working tbolt, need has been lacking. /Kasper Pedersen ___ 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] GPS Splitter
On 04/13/2012 12:19 AM, iov...@inwind.it wrote: > Just found on the auction site: > Item number: 220957196441 > Specs available at the manufacturer's website. > I have no idea if it is worth the price. It feels mechanically indestructible, and works. I have a tbolt powering it, and a 3.3V (skytraq venus cheap nav) receiver on another port to allow me to make plots like http://wap.taur.dk/gc8.png which is the view from a puck on the balcony - it has workable view, east-northeast is through roof. And the tbolt reports slightly better SNR (the antenna alone has less gain than the tbolt likes) with it in series. The N-covers it comes with are just covers, not terminators, so I suspect it to be the high-iso (+10dB) version. I have not yet actually measured the gain, work's generator stops at 1GHz so I need to build a doubler first. /Kasper Pedersen ___ 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] GPS Week 1536 causing problems?
Magnus Danielson wrote: Unfortunatly I have heard a few reports of failures, one which I can't disclose the details of right now, but replacing GPS equipment solve the issue. I didn't think this would hit me, and I don't know if it's related, but: I have (had) 2 Garmin GPS-18x fw 3.00 on the windowsill, one driving a homecooked GPSDO, the other just a separate pps. This morning both of them were quiet; There's no NMEA data coming out of them, no pps, nothing. Garmin's tool won't talk to them. They were on separate supplies, and one of them has RXD tied hard to ground (nothing speaks to it). They did share windowsill, and it hasn't rained. Anyone else lose an 18x? /Kasper Pedersen ___ 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] GPS 18x death and resurrection - was: GPS Week 1536 causing problems?
Hal Murray wrote: I don't think it has a battery inside. That seems like a poor design. Too many reasonable use cases would include sitting in a drawer for extended periods of time. There is a battery inside. A tiny little Panasonic rechargeable lithium cell. Mine were at 2.7V. Mine was less than a year old. It had been plugged in and working fine for several months. Then it just died. Six months. The most recent component is dated 0751, so they've been in storage for a while before I got them. As I was poking around, double checking things to make sure I didn't inject too much noise into this discussion, it started working again. I discharged the batteries on my pair. 0.6V was not low enough, then I tried again with a dead short for a few minutes, and they both came alive again. So just discharging in the drawer for a month or two might bring it back to life. So I support the 'yet another buggy firmware' theory. MediaTek.tw does not give me a warm fuzzy feeling. In case it dies again... Does anybody know how to take one of them apart? The case is two cups. The top has two 'rings' 2mm apart, the outer ring being the outside of the top. The bottom also has two rings, 1mm smaller than the top. The cavity between the two rings on the bottom is filled with silicone before the top is pushed on. It is held in place with 3 snap locks, one of which is right next to the thickest bulge on the strain relief. If you insert a flat screwdriver and pry the top up exactly opposite that bulge, you can get the silicone to release and it's then relatively easy to see where the snaps are. Apply force. If you have compressed air, pull off the bottom label, pull the green seal off the breather hole, and apply compressed air. The magnets are glued on from the inside, there's nothing on the outside. (and on a side note, the tiny little TCXOs in them are both ~1.4 ppm low) /Kasper Pedersen ___ 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] Reference oscillator accuracy
On 11/13/2009 07:15 PM, Don Latham wrote: The ground wave path of WWVB varies due to a very small changes in the index of refraction (temperature and absolute humidity) over the path. It is not much, but is measurable. Don My own nuttiness started with that, and the innocent question "How much? What does it take to measure it? What do I need to build?" http://n1.taur.dk/dcf/ (raw data and plots, time is UTC+1) I am no more than 6-700 km north of DCF77, and get ~5us p-p. Now that is quite large, when I started out that number seemed a lot smaller. /Kasper Pedersen ___ 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] power line common view - a silly experiment
I have servers at two sites: At home (northern part of Silkeborg, .dk), and in the center of Silkeborg. Population count about 4. The location matters, since these two sites are on different 60kV radials. I have clocks at both locations, so I decided to see how well I could do time transfer between the two sites, only using the power line. A poor man's common view experiment. http://mx15.kasperkp.dk/plcv.png I expected it to be a lot worse. 100us is 1.8 degrees phase error, or 10V slicer offset on 230V. Since the sense circuits are really cheap opto couplers, and the temperature variations on the 'city server' are rather extreme (parking basement), they can be responsible for an error at least this large. I have a set of slightly linux specific sources that generate a shm refclock from the local power line and a remote observing server with a real refclock, if anyone wants to play with this. /Kasper Pedersen ___ 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] 4 KV Power Supply Recommendations
On 01/18/2010 11:10 PM, Didier Juges wrote: > I could not resist, so I checked my relatively expensive Sears/Kenmore > microwave oven. > The results are there: > http://www.ko4bb.com/Test_Equipment/Microwave_oven_leakage/ > > In one word, dismal. Almost 1W peak power leakage at 1 foot, and almost > 100MHz occupied bandwidth. 1W into the analyzer with a sub-optimal antenna. If we are really optimistic and assume 3dBi, the effective area is (12cm)2/(4*pi)*2(for 3dBi) = 23cm2, or 12 cm2 for 0dBi. 1W becomes 43..83mW/cm2. Ow. The ballpark figure I learned when I worked with ISM was 50mW/cm2 - the point at which I have already left, and the point where the heating of the lens in your eye becomes a problem. Of the permanent kind. /Kasper Pedersen ___ 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] GPS ceramic patch in what plastic housing?
On 09/05/2010 06:12 PM, Peter Krengel wrote: > Hello, > > I just did some experiments using a ceramic patch antenna inside a > small plastic (80 x 30 x 20mm) screw box and experienced much bader signals > at lower elevations. > The plastic is marked as PS (I guess polystyrole ?). Further experiments > covering a GPS with the same kind of box seemed to effect the signals > too. On the other hand covering the patch with a flat pcs of the same > material didnt effect the antenna. > Finding a suitable top for that gps module/patch/.. that is both the right material and the right shape is hard. Just the right shape is much easier. So: I have had good success with vacuum forming such a beast. One particular glass cup I have in the kitchen is flat top conical, of just the right size. I put it on top of the ground plane, a 1mm sheet of clear polycarbonate above that, and with an air pump and heat gun I pulled the PC down around it. A bit of white paint on the inside, and it looks really good. /Kasper Pedersen ___ 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] A real-world precision timing need....
While in the shower: You have the advantage that most of the equipment will be more or less in line, and you will have line of sight. What if, at each station, you let your microcontroller generate a 10kHz carrier modulated by a 5kHz PN code, through an IR led, and through a plastic lens (children's toy magnifier? cheap plastic reading glasses?), pointed back at the shooting stand. Invert at 50bps, and you have your data backhaul. At the receiver end you have an IR receiver, go into a sound card, and in software recover the code from each of the all simultaneously transmitting stations. In the process you get the clock offsets of each of your stations. Subtract this out before post processing. 25 usec should be achievable, 25us*3200f/s=0.08f Very similar to something called 'GPS'. Only using NIR light. /Kasper Pedersen On 10/31/2010 01:56 PM, Michael Baker wrote: >Hello, Time-Nutters-- >A real-world precision timing need: >As a dedicated long-range rifle shooter and >ballistics enthusiast, I am in the early stages >of a project I am getting started on... >The object is to measure the velocity of a >rifle bullet both at the muzzle and downrange at >various distances up to 800 yards/meters or so. >Conventional optical sky-screens will will be >used for measuring the velocity at both ends. >However, I also need time-of-flight and this >requires knowing the timing relationship between >the time the bullet crosses the muzzle sky-screen >and the downrange sky-screen. Bullet muzzle velocities >will be between 1900 to 3200 feet-per-second. >Additionally, I will be using the output from an >array of 4 ultrasonic sensors located on the >corners of a 4-foot PVC pipe square to determine >the size of the shot group at the far end and >telemeter this info back to a laptop at the >shooting bench. >I can use a 10-MHz crystal for the sky-screen clocks >and the for the 4 ultrasonic bullet shot location >sensors. However, determining the time-of-flight is >a more difficult task as this requires syncing clocks >together at both ends to a moderate degree of accuracy. >Out to 100 yards I can send the time-of-flight >far-end pulse back by wire and compare it to the >muzzle-end sky-screen pulse but this is not practical >to do by wire out at 800 yards. >This project is on a tight budget-- namely, MY >wallet, so cost is a major concern. Suggestions >will be most welcome!! >Thanks!! >Mike Baker >Gainesville, Florida, USA >- > ___ > 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] GPS USB dongle for time server
On 11/09/2010 03:07 AM, Hal Murray wrote: > The Garmin USB 18 was much better. Unfortunately, it wasn't as sensitive as > competing units and it's been replaced by the 18x which has the typical > horrible jitter problems. I don't have a graph of the 18x handy, but here is > data from a USB 18. > http://www.megapathdsl.net/~hmurray/ntp/GPS18USB-off.gif > (The Garmin USB 18/18x units don't speak NMEA, but that's just SMOP.) > The 18x-lvc (rs232): http://n1.taur.dk/permanent/hist3.png http://n1.taur.dk/permanent/adev3.png http://n1.taur.dk/permanent/gpst3.png (in the last plot the y axis is seconds, not nanoseconds) It is no timing receiver, but it suffices to drive ntpd. Two notes: the 18x here is configured for n...@115200, as that is what the homecooked GPSDO speaks. Otherwise the mesages pile up, moving the time message around excessively. This unit is decapsulated and sitting indoors. /Kasper Pedersen ___ 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] Linux timekeeping / jiffy source
On 12/03/2010 04:17 AM, Mike S wrote: > Anyone familiar with Linux kernel timekeeping? > On boot, the kernel picks a clocksource. If the cpu is recent, and it is not forced, the tsc (cpu) counter gets chosen. The tsc increment rate is initially unknown, and is measured against pit (8253/8254), and the result may be off if SMI interrupts arrive (these are BIOS interrupts and can not be masked by the OS). The newer the kernel, the more attempts to work around SMI in that part of the code. If this code works, it does not matter what frequency the cpu (and thus tsc) is running, so I suspect this code is failing. Or pit is on drugs. http://n1.taur.dk/permanent/testpmt.c this bit of code will compare the timekeeping clock to the PMTimer, which runs off of plain 14.31818MHz/4. It gives you the frequency offset (in ppm) in a few seconds, and provided ntpd is not started on boot, gives you the scaling error. If you force the kernel to use another clocksource, either clocksource=hpet clocksource=acpi_pm then you may see a rather large error (12 or 127ppm) between PMTimer, and the clock running off of PMTimer. This is due to a rounding error in timekeeping.c, and for that there is this patch: http://n1.taur.dk/timefix2.patch With a good* 14.31818MHz, running the clock off of acpi_pm is now good. With hpet there's another bug, causing a 60e-9 offset. My patch above is not perfect. When it corrects for, say, a 127ppm rounding error, it also inadvertently changes the gain of adjustments. Thus, when you (or ntpd) ask for 1ppm trim, you get 1.000127 ppm trim. Also, my comment about this starting with 2.6.32 is false, it affects much older kernels too. /Kasper Pedersen *as defined by time-nuts. ___ 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] My Garmin 18x, Ver 3.50, currently 1 second slow to UTC
On 12/31/2010 10:20 AM, Joseph Gray wrote: If I am using the NMEA data (as in the Garmin is the only time source), what firmware is recommended by all those using a GPS18x-LVC? Here are measurements on the bad versions: http://n1.taur.dk/permanent/gps18x-330-pps.jpg http://n1.taur.dk/permanent/gps18x-350-pps.jpg Note I am running 115200, not 4800 3.00 kills itself. I had 2 die simultaneously. 3.10 starts transmitting 350ms .. 480ms after PPS 3.30 starts transmitting 920ms .. 1220ms (!) after PPS 3.50 starts transmitting 700ms .. 1220ms (!) after PPS /Kasper Pedersen ___ 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] advice: frequency calibration to 1 ppm possible without GPSDO?
On 02/17/2011 01:58 AM, beale wrote: Hello time enthusiasts! I'm hoping for your advice on my (perhaps modest, by this list's standards) project. I would like to make a frequency calibration of a 10 MHz oscillator to 1 ppm (1E-6) or better, using some basic equipment. I do not have a GPSDO or any serious lab equipment, or budget for same as this is just a personal project. What I do have access to: I did a piece of software to provide calibration to equipment-deficient microcontroller-hobbyist frequency counter builders (think 100ppm crystal from the bin) over ntp: http://n1.taur.dk/nft/nft.pdf (usage guide with relevant screenshots, required reading) http://n1.taur.dk/nft/nft.exe http://n1.taur.dk/nft/nftsrc.zip (delphi source) Depending on how you are connected, 0.1ppm is quite doable. Measure, adjust, measure 3 times to verify. When you decide that this is too cumbersome and buy a GPS with pps output, you could make it self-adjusting: http://n1.taur.dk/simplexdo/ /Kasper Pedersen ___ 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] 50/60 Hz clocks
On 03/10/2011 11:41 PM, Robert LaJeunesse wrote:
Poor man's solution: Use an Arduino to read the Thunderbolt 1PPS and lock a 50Hz
(or 60Hz) square wave to the 1PPS. Any resulting jitter can likely be kept in
Here is an even poorer man's solution (and plug):
A DDS using both compare outputs of an 8 pin part to get
a phase-centered PWM with half the usual ripple.
With PPS input as a bonus so the zero crossing occurs where
you want it to.
http://n1.taur.dk/gen60a.jpg
Output 5Vpp@60Hz +1.1mVpp@39kHz. Very pretty sinewave.
//
TinyAWG.c
// 60Hz generator - 2011 Kasper Pedersen - Beerware license
//
// This is an arbirtrary waveform generator set up to produce 60Hz sine
// Compile with GCC -Os
//
//
// 2-5V ---+-+
// | |
// |__ |
// | __|* |__|
// | | |__| VCC |__|---+--||--+
// 3k| | | | 1u |
// | __| |_ _|_
// 10MHz ---||--+|__| CLK LOCK|__| GND
// 1n | | |
// | |__| |_____ ___
// 3k| | |__| PPS PWM1|__|---___--+---___---+
// | | | 2k2 | 3k3 |
// | __| |_____ | |
// GND ---+|__| GND PWM0|__|---___--+--||--+--||--+-
// | |__| 2k2100n | 100n
// | ATTINY13V |60Hz out
// | |
// +-+
//
//
// Rising edge on PPS input (optional) will steer the output
// so that, after 128 edges, the positive zero crossing
// of the output will coincide with PPS.
// When this happens, LOCK will go high.
//
// PWM frequency is 39kHz
// first filter stage attenuates 27x
// second filter stage attenuates 81x and pulls phase 1 deg.
#include
#include
#include
#define DCBIAS 127
PROGMEM unsigned char table[256]={
127,130,133,136,139,142,145,149,152,155,158,161,164,167,169,172,
175,178,181,184,186,189,192,194,197,200,202,205,207,209,212,214,
216,218,220,222,224,226,228,230,232,233,235,237,238,240,241,242,
243,245,246,247,248,248,249,250,251,251,252,252,252,253,253,253,
253,253,253,253,252,252,252,251,251,250,249,248,248,247,246,245,
243,242,241,240,238,237,235,233,232,230,228,226,224,222,220,218,
216,214,212,209,207,205,202,200,197,194,192,189,186,184,181,178,
175,172,169,167,164,161,158,155,152,149,145,142,139,136,133,130,
127,124,121,118,115,112,109,105,102,99,96,93,90,87,85,82,
79,76,73,70,68,65,62,60,57,54,52,49,47,45,42,40,
38,36,34,32,30,28,26,24,22,21,19,17,16,14,13,12,
11,9,8,7,6,6,5,4,3,3,2,2,2,1,1,1,
1,1,1,1,2,2,2,3,3,4,5,6,6,7,8,9,
11,12,13,14,16,17,19,21,22,24,26,28,30,32,34,36,
38,40,42,45,47,49,52,54,57,60,62,65,68,70,73,76,
79,82,85,87,90,93,96,99,102,105,109,112,115,118,121,124};
unsigned char phase;
signed acc;
unsigned char lastp=1;
ISR(SIG_OVERFLOW0)
{
signed s;
unsigned char v;
//60Hz*256=15360Hz increment rate.
//irq rate is 10MHz/256=39062.5Hz.
//we need to increment at: 60*256*256 / 10M
//split into primes and eliminate common factors:
//10MHz = 2^7 * 5* 5^6
//60*256 *256 = 2^7 *2 * 5 *2*2*3 * 2^8
//scaler = 2*2*2*3*256 / 5*5*5*5*5*5
// = 6144 / 15625
v=__LPM(&table[phase]); //generate output
OCR0A=v;
OCR0B=(2*DCBIAS)-v;
s=acc; //generate phase
s-=6144;
if (s<0) {
acc= s+15625;
++phase;
} else {
acc= s;
}
if (PINB&16) { //on rising edge: adjust phase so this conincides with
the positive zero crossing.
if (!lastp) { //we need 128 pulses to become adjusted
lastp=1;
if (!phase) {
//phase is 0. At 15kHz we are within 65us
PORTB|=4;
} else
if (phase&0x80) {
++phase; // 65us adjustments
PORTB&=~4;
} else {
--phase;
PORTB&=~4;
}
}
} else {
lastp=0;
}
}
void main(void)
{
TCCR0A=0xB3; //A is clear on match, positive output when bigger
TCCR0B=0x01;
TIMSK0=0x02;
DDRB|=1; //output
DDRB|=2;
DDRB|=4; //"locked" output
PORTB|=16; //~50uA pu
Re: [time-nuts] 60hz disciplined watch
On 04/20/2011 05:26 PM, Poul-Henning Kamp wrote: > Are you sure ? > > Here in europe that was lost in the "privatization" of the grid: Nobody > was charged with paying for the extra power needed to capture lost > cycles, so now they just try to keep it close to 50.0Hz and don't > care about the integral. The mainland Europe grid (which you are not on, I know..) looks like this: http://n1.taur.dk/grid/plt.png http://n1.taur.dk/grid/pltw.png (very large with grid) y axis is phase in seconds, x axis is time The plot is 130 days long, 60sec/130d= 5ppm but in that plot is also 5 days in a row where the frequency is 230ppm high. /Kasper Pedersen ___ 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] 60 Hz measurement party
On 06/26/2011 01:07 AM, Tom Van Baak wrote: > Extrapolating further, I wonder if anyone has done common view > time transfer based on synchronized power grids? Although not > as precise as LF or TV or GPS methods it would make a nice > demo of the concept. > > /tvb http://n1.taur.dk/plcv/ (with pretty plots. Posted to 'nuts 1.5 years ago) The two sites were on different 60kV radials, but within the same city. I wonder what fun things one might deduce by watching the difference in phase across the continent? /Kasper Pedersen The files on the site are pretty old. I still have this in operation to a time server where I can get the space, power, and bandwidth, but not the antenna space. ___ 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] Three Phase Power
On 06/27/2011 10:38 PM, Jason Rabel wrote: > Centerpoint Energy has reached the epitome of laziness and is converting > everyone over to "smart meters" so people can better > monitor their usage... In reality it does nothing more than offer real-time > monitoring for the power company, and they no longer > have a use for meter-readers... > If the communications to/from them is in the CENELEC A band (or similar), one may have to go to great lengths in order to have 60..77.5kHz clocks/receivers working. http://n1.taur.dk/dcf/tm.png (scope) http://n1.taur.dk/dcf/idleannot.png (quiet) http://n1.taur.dk/dcf/active.png (active on the high channel in A band) I am only 500km from DCF77 (Denmark), and since they replaced the meter with a 'smart' one, none of the radio clocks have been able to sync. Only the phase tracking receiver, having a noise blanker, still works. So I do not like 'smart' meters. /Kasper Pedersen ___ 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] 60 Hz measurement party
On 06/25/2011 09:34 PM, Tom Van Baak wrote: > We are more a group of experimenters than lamenters, so > here's an open invitation to all of you in the US to join me > on a 60 Hz measurement party, starting as soon as you > can and lasting as many weeks or months that it takes to > get interesting plots. > Here is what I have used with far better success than transformers. For some reason the transformers I get hold of have quite large phase error (ie. they are cheap and get warm). http://wap.taur.dk/zcd/sch.jpg http://wap.taur.dk/zcd/plug.jpg Cp has the dual role of killing RF, and compensating for the H->L threshold of the HC14. It sits on the hot side, and charges up through the protection diodes for the next 200us pulse. It will drive 10m of POF. Substitute an opto coupler of choice if you need less isolation. /Kasper Pedersen ___ 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] SLIP vs Ethernet for NTP
On 10/23/2011 11:28 AM, Iain Young wrote: > Hi Guys, > > I have often heard it said that since RS-232 is more "deterministic", > and suffers from less jitter, and uncertainties, than ethernet, that > it makes a better medium for time distribution (no CDMA for a start). Old HDX ethernet uses cSma-cd. On what you are likely to use (junkbox 100Mbit switch and up) this is not a problem. I once built a 'frame generator' to send precisely&accurately timed ethernet frames, and did a bit of testing on this: http://n1.taur.dk/etherpps/ In short, when I disabled interrupt coalescing on a Pro/1000, the jitter was better than the jitter from the same signal through the DCD line of a serial port (1us vs 4us). Also the packet timestamp on the nic was ~5us faster than the DCD one, likely because the RS232 level translator is pretty slow. Beware of ARP on ethernet. Either poll quickly enough that the entry stays cached, or set them static on both ends. And beware of windows boxes. I did the same experiment across work's network: http://n1.taur.dk/ethertest/ The right answer is obviously to connect them using both, and compare. /Kasper Pedersen ___ 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] FE 5680A Warning
On 11/13/2011 10:36 AM, Peter Bell wrote: > I can let you know what I've found out about these units - I got given > a couple of dead ones and did a bit of poking around trying to get > them working. I didn't get that far, since one of the units was > easily fixed (the MAX882 3.3V regulator had failed) and the other > seems to have a faulty Rb absorption cell (it works if you swap the > cell from the other unit into it). Here is a pile of photos I did while fixing one of mine. In this one the oscillator had drifted high, and wouldn't lock when warm. A tiny nudge to C217 and it is happy, the output should sweep to about 4ppm high. I suspect the PTC element on the xtal is dodgy too; It seems slower to settle than the other units. http://n1.taur.dk/fe5680a-2/ I figured they might be useful when trying to point out components and test points. /Kasper Pedersen ___ 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] quick and very dirty phase comparator
This is to poke my head in, and to share a simple multichannel phase comparator/monitor that turned out to be useful. I have a few homebuilt boxes that will do ~100ps timestamps, I have other uses for them, and tying them up with a single long term experiment is unacceptable. So I needed a phase comparator that I could rebuild in an hour if I wanted to, something in the spirit of TVB's PIC16 divider. Something simple. The idea is this: Beat a number of 10MHz inputs against another clock frequency using the input flipflops in a microcontroller as samplers, sample a number of times to cover the phase circle, and calculate the phase. Beating against 11.0592MHz gives a period of 3456 clocks, sampling every 8 clocks gives 432 samples, that should give 230ps resolution. Add clock jitter and aperture jitter, and precision should be around 500ps. How much performance can one get for $10? schematic: http://n1.taur.dk/timenuts/phasecomp8.pdf samples: http://n1.taur.dk/timenuts/phasesamples.png When comparing a 10MHz source with itself through a cable delay, the peak noise is 400ps. That's for a 300-microsecond acquisition time, which means the update rate is limited by how fast I can compute the vector and shovel data over to a PC. When monitoring beating 10MHz sources with an adjusted (-3ppm) microcontroller clock, it's about 600ps peak. When misadjusted to be -30 ppm off, it's 1.1ns peak. At the moment there is only a crude windows program for turning the output into decimal plottable data, so some programming skill or beer drinking friend with programming skill is required, as well as knowledge of what one wants to measure (that's usually the hard part). source, hex: http://n1.taur.dk/timenuts/phasecomp.zip converter: http://n1.taur.dk/timenuts/phasehex2dec_win.zip I'm aware that 600ps is a very large figure in a lot of cases but for my oscillators it's good enough. /Kasper Pedersen. ___ 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] Near-perfect chip for Loran-C frequency receiver
Bruce Griffiths wrote: > Your original loop antenna should have a lower phase tempco > (particularly if a lower TCE coil former than a plastic lid is used) > than any ferrite antenna, particularly so if the antenna is resonated. > Quartzlock have abandoned the use of ferrite antennas and substituted a > balanced shielded loop antenna for this reason for their VLF phase > tracking receivers. > When I built a DCF77-DO (selective I/Q front end, the rest a software radio in an AVR), I ran into this problem of phase tempco. A loop was better, but it was unhandy, and I still had phase drift in the amplifiers, filters and mixer propagation delay. My fix was to have a tiny transmitter loop near the antenna (just the feeder cable termination resistor). Once in a while the microcontroller generates a 77.519kHz carrier, gets a phase measurement using the same frontend and I/Q mixer, and uses this for compensation. DCF77 does wiggle quite a bit at dawn, plus there is the occasional 1us phase adjustments at the transmitter end. http://n1.taur.dk/dcf/(~100km NNW of PHK) I recall having seen better than this with Loran? /Kasper Pedersen ___ 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] bounce - apologies
I must apologize for screwing up the mail server configuration. I do not believe in generating replies to usually falsified addresses. And I'm also rather confused as to why it happened now, and not a lot earlier. To make it short, it is supposedly fixed. /Kasper Pedersen ___ 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.
