A look at my Symmetricom GPSDO / 10MHz reference (OCXO + Furuno receiver)

[Leo Bodnar]

Looks like a typical PIC16 layout.
Leo

[MKatana]

Some info I came by trying to get NMEA out of my UCCM modules;
On the Samsung EP92-01764A module I ran magnet wire form pins 3 (TxD1) and 28 (Timepulse) of the LEA-6T module to pins 18 and 19 of the 50-pin flat flex with a MAX3232 on the other end sending the data to the PC.
Unfortunately instead of NMEA all I got was UBX messages  .
U-Center decoded them but I'm doing all of this to update my local clock or as a reference for a NTP server.
On Trimble 57963-B there's a XEMICS 3330R/59213 GPS front-end and all the smarts are done in either the FPGA or the TMS320 so, again, no luck.

Anyways, here are some pictures of my GPSDO project so far.
Beside the UCCM module lies a Si5351A PLL and a Arduino nano thus making this project a bit more than a simple GPSDO.
The 10MHz square wave output is routed to the Xtal_in of the PLL and using the awesome etherkit library https://github.com/etherkit/Si5351Arduino it can be easily configured to output three different frequencies form tens of kHz up to 200MHz.
My initial plan was to drive a PC motherboard's main 25MHz and maybe even the RTC clk with this so the system clock doesn't drift once synchronised to a NTP server or the GPSDO's own reciever.

[PA0KTF]

The Arduino Nano is 5Volt, the SI5351 is 3V3. So you should lower the voltage for the I2C interface by means of two resistors.

[MKatana]

The Arduino Nano is 5Volt, the SI5351 is 3V3. So you should lower the voltage for the I2C interface by means of two resistors.

It is very convinient that the SI5351a breakout has a LDO and a MOSFET level shifter and it costs under $4USD  .
https://cdn-learn.adafruit.com/assets/assets/000/018/765/original/adafruit_products_schem.png?1407873628

[ve7xen]

I have a UCCM 'packaged' GPSDO I've had sitting around on the bench for a year or more, and with the increased amount of time I have sitting around at home, I'm trying to get it running again.

In large part the device seems to work. It boots up, achieves GPS lock (even with my terrible sky view it can complete a survey, though it does seem quite deaf), and outputs 10MHz that is fairly stable. However it seems to get stuck in a 'settling' mode twiddling the DAC but not locking. There aren't any alarms or other obvious issues, and at least from LH's graphs, the twiddling does seem to be affecting the phase measurement. I'm kind of assuming there's some kind of hardware fault here, but I don't understand what 'settling' means, so don't really know where to start looking. At least from its own internal measurements, the EFC seems to work as does the phase measurements, so it should have everything it needs to run the PLL, no? Nothing interesting in the log.

First, how long after a successful survey does a 'normal' unit go into a PLL mode other than "OFFSET OBSERVATION MODE"? I've left it for a couple of hours now today, and I'm fairly certain left it for several days when I first got it, but it doesn't transition out of this mode.

Code: [Select]

UCCM-P >SYST:STAT?

-------------------------------------------------------------------------------
57964-80     serial number  30798886     firmware ver  2.0.1.6-01 LINK    mode
-------------------------------------------------------------------------------
Reference Status __________________________   Reference Outputs _______________
XX Ref 8KHz 0: [LOS]                                                           
XX Ref 8KHz 1: [LOS]                          TFOM     2            FFOM      3
XX Ref 8KHz 2: [LOS]                          UCCM A Status[ACTIVE]           
XX Ref 8KHz 3: [LOS]                                                           
>> GPS: [phase:-1.1E-07, settling]                                             
ACQUISITION ...................................................[GPS 1PPS Valid]
Tracking: 4 ____   Not Tracking: 4 ________   Time ____________________________
PRN  El  Az  C/N   PRN  El  Az                GPS      02:52:24     28 Mar 2020
  2  44  60   42     6  11  56                                                 
 21  13 237   34     5  46 120                ANT DLY  0 ns                   
 25  66 213   42    12  42 164                Position ________________________
 29  57 294   32    31  21 300                MODE     Hold                   
                                                                               
                                              LAT      N  49:16:3.485         
                                              LON      W 123:04:21.643         
                                              HGT               +12.98 m (MSL)
                                                                               
                                                                               
                                                                               
                                                                               
ELEV MASK  5 deg                                                               
-------------------------------------------------------------------------------

Command complete
UCCM-P >STATus

               - UCCM Slot STATE-

1-1. #Now ACTIVE STATUS ---------------- [Master Manual]
1-2. #Before ACTIVE STATUS ------------- [OCXO Warm]
2-1. #Reference Clock Operation -------- [Not Used]
2-2. #Current Reference Type ----------- [GPS]
2-3. #Current Select Reference --------- [GPS 1PPS]
2-4. #Current Reference Status --------- [Good Accuracy & Stable]
     #GPS STATUS ----------------------- [Available]
     #Priority Level ------------------- [LINK > GPS]
     #ALARM STATUS
     #H/W FAIL [ LINK ]
     #OPERATION ALARM ------------------ [LINK ]
3-1. #PLL STATUS ----------------------- [ENABLE]
3-2. #Current PLL MODE ----------------- [OFFSET OBSERVATION MODE]

Command complete

Edit: Experimented some more with the board out of the enclosure. I can confirm that the EFC voltage is set properly by the DAC (I can swing it between ~0V and ~10V (!)) with the DIAGnostic:ROSCillator:EFControl:DATA command, and the device does this itself early on to calibrate itself (I guess). Also confirmed that the crystal is pulled in frequency at what seems like an appropriate amount (full tuning range approx 10Hz). For long periods of time it locks itself at 0.4Hz off my 'reference' frequency (an HP 8657A with OCXO), and the LH chart shows phase error of < 5ns, so it seems to be locking in a coarse mode at least...and then randomly decides to go wild and slam the EFC to the rails.

[PA0KTF]

Did you check the board for physical impairments, components missing ?

[uncle_bob]

If your sky view is still terrible, that may be the source of the problem.

[thinkfat]

Try providing an accurate antenna position manually. If you have bad visibility, the self-survey might not be able to compute a good position by itself.

[ve7xen]

Did you check the board for physical impairments, components missing ?

Nothing obviously damaged. There are quite a lot of unpopulated positions so it may be easy to miss something carefully removed.

If your sky view is still terrible, that may be the source of the problem.

Not much I can do about it unfortunately . I have a timing antenna on order that may help with multipath...

Try providing an accurate antenna position manually. If you have bad visibility, the self-survey might not be able to compute a good position by itself.

This was the first thing I tried. I may try surveying position with a more sensitive u-blox receiver and inputting that position instead of just copying the self-survey parameters.

[uncle_bob]

It's not so much the antenna type as the location of the antenna that matters. You need a pretty good sky view.

[thinkfat]

Try providing an accurate antenna position manually. If you have bad visibility, the self-survey might not be able to compute a good position by itself.

This was the first thing I tried. I may try surveying position with a more sensitive u-blox receiver and inputting that position instead of just copying the self-survey parameters.

That was not what I meant you to do. You need to input an accurate position. Just entering the position the receiver calculated by itself is useless. Are there no maps of your area? Even using Google Maps you can possibly pinpoint the antenna position with a sufficient precision.

[PA0KTF]

The number of satellites received is low. But in your LH screenshot I can see that you sometimes receive 3 or 4 sats. And more sats are detected.
The board is detected as Trimble etc. You detect your position, so your receiver is ok.
I have had several Trimble boards. I repaired a few because some components were ripped off. 
The green and purple graphs are irregular. I think it must be a hardware problem.

Earlier in this blog you can find pictures of top and bottom of this board. Perhaps you can compare.
Is there a 10 MHz output ? If you use the SMA output, then press the MACT button first.

You might check the ocxo pins. There should be a reference voltage on one pin (the middle of the 3) and a pin next to it on which there should be a voltage between 0 and Vref.
Success.

[uncle_bob]

You need a sky view that typically provides > 8 stats and *never* goes below 4 sats in use ( that's different than sats in view). If you can't hit those numbers with your antenna location, it's time to find a new antenna location. Will this or that minimal number of sats cause exactly what you are seeing in *every* case? Certainly not. Multpath propagation effects will never be the same in one location as another.

[SHF]

I bought 3 Symmetricom units on eBay from 4 years ago, all of them were not complete and had to be repaired, I paid 35-45 € per unit.

[ve7xen]

Try providing an accurate antenna position manually. If you have bad visibility, the self-survey might not be able to compute a good position by itself.

This was the first thing I tried. I may try surveying position with a more sensitive u-blox receiver and inputting that position instead of just copying the self-survey parameters.

That was not what I meant you to do. You need to input an accurate position. Just entering the position the receiver calculated by itself is useless. Are there no maps of your area? Even using Google Maps you can possibly pinpoint the antenna position with a sufficient precision.

I am quite sure that the surveyed location with a u-blox NEO-M8T will be more accurate than what I can get with guesswork and Google Maps. I don't have access to survey-grade equipment to be certain about the antenna position better than that, and particularly the elevation isn't practical to obtain any other way.

You might check the ocxo pins. There should be a reference voltage on one pin (the middle of the 3) and a pin next to it on which there should be a voltage between 0 and Vref.

Yeah I monitored this for a few hours the other day. It seems to be working as expected, the control voltage tracks the EFC value as expected, the 10MHz is stable to my ability to measure it.

I guess there could be a hardware failure, but it seems odd, since the key elements of the control loop are all there - the phase measurement, GPS receiver PPS, and EFC. Without a block diagram anyway, maybe there is some other oscillator it is trying to lock which is broken. I will review the boards again more carefully against the photos.

Does anyone with a working unit ever see this 'settling' in 'SYSTem:STATus?'? I've seen very little mention of it or outputs including this, and the couple that have mentioned it have all been having problems.

It was nice out today and with nothing better to do I took the GPSDO to the park with a LiPo battery and my laptop. I let it run for 2 hours with an excellent sky view. Survey completed in maybe 20-30 minutes. After that it behaves very similarly to from the normal location. It seems to lock somewhat better, but stays in 'Settling' mode, despite clearly being in some sort of control loop, as it does pull and keep the phase offset much lower than at startup.

From the park it does this:

Code: [Select]

-------------------------------------------------------------------------------
57964-80     serial number  30798886     firmware ver  2.0.1.6-01 LINK    mode
-------------------------------------------------------------------------------
Reference Status __________________________   Reference Outputs _______________
XX Ref 8KHz 0: [LOS]                                                           
XX Ref 8KHz 1: [LOS]                          TFOM     2            FFOM      3
XX Ref 8KHz 2: [LOS]                          UCCM A Status[ACTIVE]               
XX Ref 8KHz 3: [LOS]                                                             
>> GPS: [phase:-1.8E-09, settling]                                               
ACQUISITION ...................................................[GPS 1PPS Valid]   
Tracking: 6 ____   Not Tracking: 3 ________   Time ____________________________   
PRN  El  Az  C/N   PRN  El  Az                GPS      22:48:03     29 Mar 2020
 24  53 284   36    15   7 233                                                 
 17  58  58   41     2  16 179                ANT DLY  0 ns                   
  6  35 153   37     1   5  28                Position ________________________
 12  25 285   34                              MODE     Hold                   
 19  82  89   44                                                               
 28  31  94   40                              LAT      N  49:16:8.355         
                                              LON      W 123:04:25.279         
                                              HGT               +10.70 m (MSL)
                                                                               
                                                                               
                                                                               
                                                                               
ELEV MASK  5 deg                                                               
-------------------------------------------------------------------------------

Command complete
UCCM-P >STATUS?
Command error
UCCM-P >STATUS

               - UCCM Slot STATE-

1-1. #Now ACTIVE STATUS ---------------- [Master Manual]
1-2. #Before ACTIVE STATUS ------------- [OCXO Warm]
2-1. #Reference Clock Operation -------- [Not Used]
2-2. #Current Reference Type ----------- [GPS]
2-3. #Current Select Reference --------- [GPS 1PPS]
2-4. #Current Reference Status --------- [Good Accuracy & Stable]
     #GPS STATUS ----------------------- [Available]
     #Priority Level ------------------- [LINK > GPS]
     #ALARM STATUS
     #H/W FAIL [ LINK ]
     #OPERATION ALARM ------------------ [LINK ]
3-1. #PLL STATUS ----------------------- [ENABLE]
3-2. #Current PLL MODE ----------------- [OFFSET OBSERVATION MODE]


[PA0KTF]

Totally different screenshot. It seems to work ok. It takes hours to settle so be patient.
If you use a magnetic GPS antenna: put a small metal plate underneath ~10 x 10 cm.

[uncle_bob]

Settling = you are still in the early part of the loop filter convergence process. The typical reason for this is poor GPS antenna location. Any other reason is "non-repairable".

[ve7xen]

Settling = you are still in the early part of the loop filter convergence process. The typical reason for this is poor GPS antenna location. Any other reason is "non-repairable".

Kind of what I figured. How long does it 'normally' take?

[uncle_bob]

It very much is a "that depends" sort of thing. An hour or so once the survey has completed gets most GPSDO's going. For full performance, it is rare to see one get the job done in under 24 hours. There are many very good designs out there that are still "settling" a week or three later. These parts all had a  target application of being turned on and left on "forever".

[texaspyro]

The typical reason for this is poor GPS antenna location. Any other reason is "non-repairable".

Or a wonky antenna / cable.

[Nuno_pt]

Hi guys, I´ve some spare time now due to the facts happening around the world now, so I was looking at something to pass the time and decide to start putting together the GPSDO project that I´ve put aside about some 3 years ago.

I´ve a Trimble 57963-B board, i´ve it plug in at 6.5v, for now I´ve no antenna since the one that was in use is broken by a fall.

At first power on the 2 led´s flash red, and one of them change to green, I left it plug for the night.

When I went back this morning both led´s were red fix, if I connect it to my HP53131, it read 10.000.084.~~~.~, so the OCXO is working from what I can see.

The 53131 is with Trigger:
Auto ON, Level 50%, Slope POS, Sense Hi, 50 DC coupled with 100kHz filter on, the Gate time is 5.000s

On the J5 I´ve put some pinheads to be easy to connect, and plug it to the PC RS-232, facing the board the 1PPS MCX is at the left, so pin 5 (RS-232 GND) at the left pin, pin2 (RS-232 RxD) at the 4 (square hole), pin3 (RS-232 TxD) at the pin3, and at 57600 8N1N on PuTTy but it doesen´t put out any prompt line like UCCM

At the board on the J5 pin 3 I´ve -5.6v, on the MCX for the antenna i´ve the +5v for the antenna, from what I can see on all the test points i´ve 3.27v, teh MACT and Reset buttons don´t do nothing.

Can any of you guys give any hint, or should I trash the board

[thinkfat]

Both LEDs solid red is likely undervoltage. How do you power the module?

[Nuno_pt]

I'm powering it from a PSU, I can adjust it, I´m measuring it at the fuse, it's around 6.5v, I can put more if need.

[thinkfat]

6.5V at the fuse is plenty, maybe even too much. These modules work with 5.5V nicely. How do you supply the power, directly to the fuse? Or through the flex cable? If you supply directly to the fuse, try lowering the voltage. Also, no antenna might be the reason for the LEDs solid red as well.

[Nuno_pt]

I try to lower the voltage still the same, the OCXO is ok since is outputting the 10Mhz and it´s warm, has the 5v in and voltages on the other two pins.

Still can´t get nothing out of the RS232, can´t write on PuTTy, maybe sipex3232is dead , this is the strange thing if I could make it talk, I could see the reason of the alarms.

Where can I pick some signal to connect to one MAX232 that I´ve here, that way I can eliminate the sipex3232 problem.

The two led´s are still red no matter what voltage is set from 5v to 7v.

Maybe it´s better start to thinking about buy some new board