So to verify the initial results we saw we tested another 10MHz DIP-14 unit on a different ADEV test-set, and the results are similar. See the attached ADEV plot of a random unit. This was actually the synthesized output from the board set to 20MHz from the 10MHz TCXO, then fed through a divide-by-2 CMOS circuit utilized as a 50 Ohms buffer then anaylzed by our TSC5125A. It shows that the synthesizer output nicely follows the TCXO, and it shows the ADEV performance of the DIP-14 TCXO. bye, Said In a message dated 11/22/2014 18:23:35 Pacific Standard Time, [email protected] writes:
Nigel, I would not worry too much about the RTV, more about damaging the nice TCXO. That part is by far the most expensive and valuable component on the boards. It's not your good old $10 style TCXO.. We were not aware and did not know of the beat frequency causing some spurs until the day I posted my findings and suggestions, but there is an easy way around the issue as I had discussed at that time as well. The easiest way to avoid the beat frequency spurs when using an external oscillator is to use a splitter with high isolation on the RF output of the oscillator, and to put an LC filter on the EFC pin of that oscillator. One (CMOS compatible) output of the splitter goes to the LTE-Lite MMCX connector, the other output can be used for your circuitry or instruments etc. Depending on the isolation capability of the splitter I bet there won't be anything coming through. Instead of a passive splitter a simple CMOS gate as a buffer also provides massive amounts of isolation from the 20MHz RF floating around the LTE Lite board. In either way the spurs are quite low if I remember correctly, so even a passive splitter with 40dB or so isolation will likely push the spurs below the noise floor of the oscillator I would think. The secondary path into the OCXO could be through the EFC output, and that can be easily filtered out by using a small RC or LC filter, say 22uH into a 100nF cap mounted onto the OCXO EFC pin. That should squelch any RF on the EFC line. Since most OCXOs have an RC filter built into the EFC pin anyways (usually a 30K Ohms or higher series resistor into the capacitance of the varactor diode at the least) I don't think much of the 20MHz would come through the EFC line, most of it probably through the RF pin. There is a 220 Ohms series resistor on the EFC output of the LTE-Lite anyways, so just putting a 10nF to ground on the OCXO EFC pin would already create a low pass filter with a 70KHz cut off frequency, so that single cap by itself may be sufficient already to get rid of any 20MHz RF floating around the EFC line. Definitely use a shielded coax cable for the EFC line though, not just a simple wire. On the 10MHz boards the DIP-14 TCXO is soldered onto the board for best performance and mechanical stability, but since it only has four pins in a DIP-14 configuration it is very easily removed from the PCB with a proper solder-iron if that is desired. Bye, Said Begin forwarded message: Cc: [email protected]_ (mailto:[email protected]) Subject: Re: [time-nuts] update on LTE-Lite Reply-To: [email protected]_ (mailto:[email protected]) , Discussion of precise time and frequency measurement <[email protected]_ (mailto:[email protected]) > Hi Said, Many thanks for your reply. I'd realised that R2 and R3 were in the paths of alternative power feeds but had managed to convince myself that both resistors were shown to be fitted in your auction photos, although I can now see quite clearly that whilst R2 is fitted, R3 isn't, whoops:-) I was hoping that removing the TXCO might be reasonably straightforward, it was after all your suggestion:-), but am now concerned that the RTV might have flowed under any components perhaps not fully seated and didn't want to take the risk of damaging anything. I'd also hoped it might prove possible to remove the onboard TCXO power by breaking one of the tracks on the LTE-Lite module itself but it sounds like that might have been a bit too optimistic. Whilst I'm pretty sure I would have stuck with the 20MHz version anyway, the initial enthusiastic rush did mean nearly 40 of us had already bought them before your comments regarding the harmonic mixing and the suggestion to use the 19.2MHz version reached the list, so to some extent at least that might help explain the numbers:-) Anyway, thanks very much again for making these available, I certainly don't see this as a significant issue so it's time to dive in and worry about the finer details later. Just out of interest, on the 10MHz version without the onboard TCXO, did the DIL oscillator end up socketed or soldered? Regards Nigel GM8PZR In a message dated 22/11/2014 18:01:18 GMT Standard Time, [email protected]_ (mailto:[email protected]) writes: Nigel, CC'ing time nuts.. R2 and R3 are stuffing options, see the schematics in the user manual. Typically you don't have to solder anything. The default is set for the low-noise 3.0V to be fed to the DIP-14 tcxo for best performance. On removing the SMT Tcxo, this is not easy, but possible with a heat gun. Its easy to melt the adjacent switches though when doing that. The RTV over it should just peel off, but we have not tried that yet. There is no way to just remove power to the internal Tcxo unfortunately. This is why I had suggested the 19.2MHz version for people who want to use the external oscillator option, because that won't beat with your oscillator close-in. I was a bit surprised that so few of those 19.2MHz units sold compared to the significant numbers for 10/20MHz units. In either case I think the resulting beat spurs are typically lower than many other GPSDOs and oscillators have in their spurs, the CSAC with its spurs comes to mind.. Hope that helps, Said Sent from my iPad On Nov 22, 2014, at 6:08, [email protected]_ (mailto:[email protected]) _ (mailto:[email protected]) wrote: Hi Said, Sorry to bother you again, I'm not sure is this is an oversight or deliberate but I've just noticed that R2 has not been fitted to my evaluation board. I'm a bit concerned as this will affect power to the external oscillator and R2 would be difficult to fit now that SW2 is installed. I may have missed this if it was mentioned on the Time-Nuts list but there's a lot there to look through so thought it best to ask you direct. Regards Nigel GM8PZR In a message dated 30/10/2014 23:30:40 GMT Standard Time, [email protected]_ (mailto:[email protected]) _ (mailto:[email protected]) writes: Hello everyone, We wanted to give you an interim update on the LTE-Lite Evaluation kit builds. We are making good progress on the assembly. We are expecting to ship within the announced time frame, and are guiding the factory on a daily basis. We are using the time to improve the firmware and have made some upgrades with the help and efforts of a very well-known Time Nut (thanks much!) who I am certain will identify himself in due time. This development activity has revealed a sensitivity in the GPS receiver: the unit should not be changed from mobile mode to Position Hold (Timing) mode ROM/Flash firmware or vice versa while power is enabled, this could lead to Flash memory corruption. We are thus shipping the unit with a solder jumper that will require soldering-in two pins to get out of Timing mode and into mobile mode so as to avoid the possibility of a user accidentally inserting a jumper with power supplied to the unit, which could then corrupt the GPS Flash memory. We have also been able to upgrade the GPS Flash firmware in close cooperation with the vendor of the GPS receiver to make it more optimal for operation in a GPSDO, and these efforts have given us a GPS Flash firmware release that works well with our GPSDO algorithms and TCXOs. We will upgrade all of the GPS receivers prior to shipment with this specialized firmware release. We also have some good news on the 10MHz versions of the board: the 10MHz DIP-14 TCXOs are scheduled to arrive at our factory in about 5 weeks, so that will pull-in the delivery schedule significantly. We will continue to keep track of the schedule and update you with any changes. Lastly there has been some fairly intense discussion recently about counter-fit FTDI serial chips getting permanently disabled by the latest FTDI software drivers and thus rendering affected equipment unusable. Many links have been published on Time-Nuts so please consult the archives for more information on this issue. We use FTDI chips on many of our products, such as the CSAC GPSDO, Mini-JLT, LC_XO Plus kit, and also this LTE-Lite Eval kit. We are pleased to announce that we have verified with our suppliers that all of our shipments in the past and in the future have and will continue to use genuine FTDI chips. Close inventory control and counter-fit policies are one advantage of manufacturing in the USA versus in Asia. More later, Bye, Said Said Jackson Jackson Labs Technologies, Inc. 1635 Village Center Circle, Suite 150 Las Vegas, NV 89134 Tel. (408) 596-0600 Fax. (702) 233-1073 [email protected]_ (mailto:[email protected]) _ (mailto:[email protected]) __www.jackson-labs.com_ (http://www.jackson-labs.com/) _ (http://www.jackson-labs.com/) = _______________________________________________ time-nuts mailing list -- [email protected]_ (mailto:[email protected]) To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there. = _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
LTE-Lite_DIP14_ADEV_long.pdf
Description: Adobe PDF document
_______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
