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

[ZigmundRat]

or item 252304765793, or 231861995746. Got my second unit from pm_shop_uk (item 322024166858). He was slow to ship (like a week), but the unit arrived and worked fine first time. No missing parts or board damage on this one. And only a little oil For $43 a bargain.

[NivagSwerdna]

I also used pm_shop_uk for my replacement unit.  It was complete but still needed minor repair. Their return policy is strict so I think it is pot luck with all these units and you should be prepared to gamble and potentially absorb some cost if it doesn't work.
Also... in retrospect I think the initialisation time of this symmetricom uccm-p unit could be an issue, depends if you are prepared to leave it always running or not.

[Bryan]

For those in the US who are interested in one of the Symmetricom GPSDO's there is a seller on ebay who has a few for 43.99 shipped. To find them you can search for "SYMMETRICOM UCCM", I have no relation to seller, but if they work it is a pretty decent price.

Searching with "SYMMETRICOM UCCM" on EBay doesn't bring anything up other than the seller from China. Do you have a EBAY listing number.

Cheers

Try eBay # 322024166858

Thanks, although looks like only shipping to the US. Was contemplating having a spare

[uncle_bob]

Thanks, although looks like only shipping to the US. Was contemplating having a spare

Hi

Getting a spare (or three) while these things are cheap is a *really* good idea. All of these gizmos come out in a flood at a low price. Once that initial surge of parts is sold, the prices creep up. A replacement might cost you $250 four years from now. You also might be completely unable to find one at all.

Bob

[NivagSwerdna]

Making good progress...  Have got my TFT Touch Screen working with ATMega2560 and added a RS232 converter.

So now I have USB to/from PC as required, RS232 serial to/from UCCM-P and TTL level RX from Fugano GPS for NMEA messages.

The converter was http://www.ebay.co.uk/itm/111927731957 so my total spend now up to £9. 

Still considering whether to route the 6 onboard LEDs straight to equivalents on the front-panel or to Ardunio inputs and drive front-panel LEDs in software...

Incidentally I notice that the device goes to FFOM 0 very quickly, so I don't think survey time is a real issue.

[Nuno_pt]

Well, I just hit the "buy" button on one of this units from this seller here < http://www.ebay.com/itm/331805445560?_trksid=p2055119.m1438.l2649&ssPageName=STRK%3AMEBIDX%3AIT >, for 40€ free shipping, let's see how it plays and what conditions it arrives.

NivagSwerdna, that display is looking good.
The FFOM=0 is very good, another one should be TFOM=3 (I think is TFOM).

[NivagSwerdna]

FWIW The maximum input voltage for the TI DC-DC converter is 5v5.  I would recommend never exceeding 5v1 on the input and not to worry about any high pitched noises during OCXO warm up!  If you go too high on the TI input voltage... it will break.

I found a couple of secret commands tonight..

UCCM-P > :GPS:POS:SURV:STAT?
1
Command Complete

1=Surveying, 0=Position Hold


UCCM-P > :GPS:POS:SURV:PROG?
+2
Command Complete

Survey Progress... 2%

Here's one I like...

UCCM-P > :DIAG:LOG:READ?
"L48:2016-04-05 22:07:56:GPS lock started"
Command Complete

That is the last log entry. Log Entry#, datetime, text


There is this one (mentioned before)...

UCCM-P > :PTIM:TCOD?
T1#H442EFE0020000E2
Command Complete
UCCM-P >

Tells you the time it will be on the next PPS pulse (will block until about 100mS before pulse).  I'm not sure of the format.

also..

UCCM-P > :GPS:POS:SURV ONCE
Command Complete

to force a survey

:SYNC:HOLD:INIT

go to HoldOver

UCCM-P > :SYNC:HOLD:REC:INIT
Command Complete

recover from Hold Over

That's enough for tonight. Time for bed

[NivagSwerdna]

If anyone is interested... here are some TCODE outputs...

T1#H4430443220000B0
T1#H4430443420000B2
T1#H4430443620000B4

I cannot get the checksum calculation to work with these.  I thought it was meant to be the last 2 characters were the HEX of a sum of the ASCII values of the preceeding characters... weird I get 7A, 7C, 7E not B0, B2 and B4.

Must be missing something simple.

I can't find a second serial connection so I am giving up on that.  In fact I think I have investigated as much as I can now. Just need to put it in a box, route the LEDs to the front panel and write a bit of software to display FFOM and TFOM as well as time and position to front-panel.

If you explicitly set position the FFOM will go to zero in just a couple of minutes once the OCXO is warm so I plan to add a function to store surveyed position to EEPROM and restore on start-up.

Too much fun, too little time!

(Also added output from 10MHz and PPS connectors for reference) PPS is 50us width.  [10MHz should read 10MHz not 9.99993MHz but I didn't win a scope in scope month so that's as good as it gets!]

[Nuno_pt]

NivagSwerdna,

Will you publish the schematic and code?

I think that the display should have position, time, date, TFOM & FFOM, Satellites, for now it's what I can think.

 

[NivagSwerdna]

Will you publish the schematic and code?

Sure.  There isn't much in the way of schematic... it is just Mega2560 clone eBay auction: #272041591038, Display eBay auction: #221831832114 and RS232 Level converter eBay auction: #400461786687.
I will upload the code to github very soon but it's very much a work in progress as I have a day job (actually several!)

[Nuno_pt]

I know what you mean, to much to do and little time to do it.

I like the infos on the VK4GHZ display on is Trimble, that I suppose that have much more info then the Symmetricom.

BTW my seller add an tracking number to the board for free.

[NivagSwerdna]

I like the infos on the VK4GHZ display on is Trimble, that I suppose that have much more info then the Symmetricom.

I wasn't aware of that display but looking at the pictures the Symmetricom UCCM-P has all those details and more (I assume the locator is calculated).
I have a couple of design decisions with the clock display...

• use NMEA messages and accept the display will be 0.5s late and will only operate when the GPS is receiving
• use PPS from GPS but keep a local RTC running (fixed to the NMEA messages), gets the clock right but only when GPS is receiving
• use PPS from the PPS output and keep RTC running and sync to TCODE and/or NMEA... allows the clock to run during HOLDOVER
• Use flashing LED from board as 1PPS source

Might stay simple with Option 1 for now.

[Nuno_pt]

Is Trimble LCD is showing several parameters but they are on menus, you can go trough the menus.

This the PDF from the LCD < http://vk4ghz.com/wordpress/wp-content/uploads/Thunderbolt-Display-3-User-Guide.pdf >.

The locator (Gridsquare) are calculate base on the Lat. and Long, there was some code for arduino to do this, I'll post the link here when I find it.

[NivagSwerdna]

The algorithm for Maidenhead Grid is quite simple if you are converting from WGS-84 so I have added that.... I seem to be in IO91UK

I don't know what the meaning of "10Mhz offst" in ppb on the Thunderbolt display.  For the UCCM-P

The DIAG:LOOP gives...

DIAG:LOOP?
-------------------------------------------------------------------------------
4/10/2016 21:42:57
-------------------------------------------------------------------------------

   LOS      MEAS       NCO STATUS WEIGHT PBUC FBUC DBUC LBUC IBUC G  M TC
-------------------------------------------------------------------------------
 0:  1  0.000e+00  0.000e+00 0x07AC   0   700  700    0    0    2  2  4 29
 1:  1  0.000e+00  0.000e+00 0x068C   0   700  700    0    0    0  2 14 226
GPS: 0 -2.082e-09 -6.762e-09 0x0000   1  ---- 1000 1000 ---- ----  3  6 31
-------------------------------------------------------------------------------
freq cor  = -6.761820e-09
phase cor = 2.000000e-11
gps phase = 3.333333e-09
temp cor  = -2.301801e-11
Command Complete

Not sure which of those many numbers might be equivalent.  Maybe it doesn't matter... FFOM and TFOM enough?

From the Thunderbolt documentation:

10 MHz Offset: This field carries the frequency offset of the 10 MHz output relative to
UTC or GPS as reported by the GPS receiver in ppb (parts-per-billion.) Positive values
indicate that the ThunderBolt’s 10MHz clock is running slow relative to GPS or UTC.

Maybe that is freq cor?  If so the performance of the UCCM -6.761820e-09 is much larger than the very small numbers shown in the Thunderbolt display examples.  Maybe the Thunderbolt is a much better device than the UCCM-P? i.e. at least 2 orders of magnitude better?

[NivagSwerdna]

Making progress

Source: https://gitlab.com/NivagSwerdna/SymmetricomUCCMDisplay
Schematic: Sketch attached.

Bill of Materials: Arduino ATMega2560 Clone (£5.50 EBAY 100fys), 2.4" TFT Touch Screen Shield (£3 EBAY), Max3232 RS232 level converter module (99p EBAY various sellers).

Only a very initial version but shows what you see below.  Contributions welcomed once it settles down.  I haven't explored the touch screen and menus yet.


(The display is very sharp focused... the blur is my own photographic skill in action)

[Nuno_pt]

Looking very good, still waiting for mine to arrive here.

[ghulands]

I received mine today. A small amount of damage. C36 is hanging on by a trace and C24 is missing. It looks like all those capacitors are the same in that area. Does anyone know the value of them? I'm still waiting for my LCR meter to arrive.

[gm8bjf]

Hi ghulands,

I had same problem with one of mine. These caps are associated with the RS232 level shifter IC on the other side of the board. Looking at the data sheetfor the chip I figured that 100nF would not go too far wrong and that worked fine for me.

Brian.

[ghulands]

Hi ghulands,

I had same problem with one of mine. These caps are associated with the RS232 level shifter IC on the other side of the board. Looking at the data sheetfor the chip I figured that 100nF would not go too far wrong and that worked fine for me.

Brian.

Thanks Brian. 0402 sizing?

[NivagSwerdna]

C23,C24,C25,C35 & C36 look similar

[gm8bjf]

Hi ghulands,

I had same problem with one of mine. These caps are associated with the RS232 level shifter IC on the other side of the board. Looking at the data sheetfor the chip I figured that 100nF would not go too far wrong and that worked fine for me.

Brian.

Thanks Brian. 0402 sizing?

Hi,

I think that they should be, but I used 0805s as they were what I had to hand.

Brian.

[NivagSwerdna]

Can someone who understands these things please explain....

Does the accuracy of the frequency increase with time?  Is that observable from the measurements in DIAG:LOOP?

I was thinking that if I am only going to use this thing to calibrate my other instruments (frequency counter and function generator) I do not need to leave it on for long periods of time however if it benefits from staying on for long periods then I should also make it serve NTP to increase its usefulness.

Advice please.

[uncle_bob]

Can someone who understands these things please explain....

Does the accuracy of the frequency increase with time?  Is that observable from the measurements in DIAG:LOOP?

I was thinking that if I am only going to use this thing to calibrate my other instruments (frequency counter and function generator) I do not need to leave it on for long periods of time however if it benefits from staying on for long periods then I should also make it serve NTP to increase its usefulness.

Advice please.

Hi

Yes, the frequency accuracy increases with time. For best results, you need a well surveyed location (which takes time). You also need to allow the OCXO to fully stabilize (which might take a week or two). You also need to get the disciplining loop into it's lowest state. That may only occur after the OCXO has stabilized. It might occur in a couple of days. Digging into the specific firmware on your unit would be the only way to really know.

How much of a difference does a couple days make? In the ideal case, you likely will go from down about 2X every time you double the time on past the first day out to a few weeks. The ultimate floor will depend on your antenna location, the quality of the survey, and the temperature stability of your lab. With a crummy antenna and a drafty lab, you might not see any improvement past the second day ...

Bob

[edpalmer42]

Does the accuracy of the frequency increase with time?  Is that observable from the measurements in DIAG:LOOP?

I was thinking that if I am only going to use this thing to calibrate my other instruments (frequency counter and function generator) I do not need to leave it on for long periods of time however if it benefits from staying on for long periods then I should also make it serve NTP to increase its usefulness.

Accuracy is a bit of a slippery concept when applied to GPSDOs.  Assuming that it's working correctly, the average output frequency of a GPSDO doesn't change.  However, it will wander around a little since GPS has to deal with path-length variations due to the ionosphere.  The main purpose of the OCXO is to act as a flywheel to filter out the wander.

If you look here:  http://www.leapsecond.com/pages/gpsdo/ , particularly the last graph, you'll see performance graphs for 4 GPSDOs.  Notice the 'hump' in the graphs that have the dots.  These humps show the crossover between the performance of the OCXO and the GPS system.  It's a balancing act between short-term GPS noise and long-term OCXO aging.  Most commercial GPSDOs have to take a middle-ground, 'good enough' path and the hump is the typical result.  Some commercial GPSDOs actually do some autotunig to adapt to the oscillator's performance.  If you look here:  http://www.leapsecond.com/pages/tbolt-tc/ , you'll see a discussion about tuning a Trimble Thunderbolt to match the performance of the OCXO in that particular unit.  The results are impressive.

The performance of a good quality OCXO will improve if it's allowed to run continuously.  Aging in particular will drop.  This would allow for a longer time constant in your control loop and therefore, a lower, longer flatter portion of the AlDev curve before it hits the sloped GPS performance line.  Proper tuning will minimize or eliminate any hump.

But does this matter?  If you look back to the performance graph for the 4 GPSDOs, they're all around 1e-12 @ 1 sec.  which is a common measurement interval.  You'll have to decide whether improvements at 10, 100, or 1000 seconds are useful to you.  I just measured the frequency of my Z3801A GPSDO.  With a one second gate time, I'm seeing a Standard Deviation of about 200 - 300 uHz.  A 10 second gate time drops that to less than 50 uHz.  My counter is a Fluke PM6681.  My Efratom FRT Rb standard is the external reference.  These numbers are not as good as the ones shown in the graph.  Each OCXO has its own unique performance level.  The OCXO in my Z3801A might not be as good as the one shown in the graphs.

As for the DIAG:LOOP parameters, there isn't really enough info to say much about them.  Watch and record them.  See how they change as the unit runs for days or weeks.  Do they wander up and down?  Do they slowly stabilize?  I would expect frequency offset to wander because the OCXO will continuously flywheel to buffer the wander from GPS.  There will always be an offset that varies up and down.  But if that 'TC' parameter stands for 'Time Constant', you might see it slowly increase as the OCXO settles down.  That would show you that your GPSDO is doing some autotuning.

Ed

[uncle_bob]

Hi

If *frequency* accuracy is of interest, then ADEV (especially at 1 second) is probably not a good measure on a GPSDO. At 1 second, you are looking at the free running performance of the OCXO. You get great numbers, but the accuracy might only be as good as the OCXO's. You need to wait long enough that the GPS section has had time to get the OCXO settled down in order for the accuracy of the GPS system to transfer in a meaningful way to the OCXO.

A simple experiment:

Yank the antenna on the GPSDO and look at the ADEV. The numbers inside 10 seconds look no different than the numbers with the antenna connected. The OCXO *could* be off anywhere.

Leave the antenna off and watch the ADEV plot of the OCXO. As the unit stabilizes (over days / weeks), the curve past 10 seconds flattens out. The longer it is on, (in a stable environment) the flatter it is likely to get. That may continue out past 1,000 seconds or it may start to turn up a bit. The flatter that curve is, the better the short term frequency of the OCXO will be compared to the long term.

Yes, it would be nice if the ADEV curve dropped off as 1/sqrt(tau). That's not what happens in an OCXO. The physics don't work that way.

Bob