Electronics > Metrology

I bought the very Last Symmetricom SA.22C in the shop!

(1/8) > >>

Johnny B Good:
 @wa4whl

 Did you have any luck getting hold of information on your Symmetricom SA.22C ? Im hoping you clicked the notification button when you made this post just over nine years ago!  ::)

 I recently bought a later (2006) 15MHz square wave output version after downloading data sheets and user guides which suggested that it might be possible to program it to produce a 10MHz sine wave via the customer interface using the serial port connectivity feature. At the very least, I could always phase lock one of my many 10MHz sine output OXCO spares to the 15MHz square wave output if the worst came to the worst (excluding of course, the even worse case of receiving a non functioning unit (the very last one available via any on line shopping sites).

 As it happens, it does appear to be a fully functional 15 MHz unit as advertised and, after making up an 18 pin dual row 2 to 2.54 mm pitch adaptor and a test break out board plus connecting up an FTD-232 usb to serial adaptor, was finally able to access the serial interface and experiment with the customer run commands two days after taking delivery.

I was able to change the frequency to 10MHz using the o command) (but still square wave) and fine tune the initial +5.5mHz offset out using the f command. Unfortunately, these settings revert to defaults when next powered up and the 't' save tuning settings command is not clearly explained by the example given in the user guide. I'm hoping you managed to at least figure out the various user commands and how to interpret the 'health' data (w command). Basically, I need an "Idiot's Guide to control and configuration of an SA22C".

 I'm planning on disciplining to the 1PPS of an M8T gps receiver module so, apart from the need to reset it back to 10MHz using a nano3 mcu on every power up, I might not need to worry any further about reprogramming it. I'm going to attach one of my spare M8Ts and experiment with this built in GPSDRO feature.

[EDIT] Finally figured out why my attempt to reply was being aborted by eevblog's cackamaimee posting system - it was the now smaller 4000kB limit error with its carefully hidden error message which was the cause of my grief  :wtf:


Johnny B Good:
 I see my attached images raised some interest so it seems the subject of DIY rubidium atomic frequency standards isn't totally dead (just rather moribund  :( ).

 I did finally get round to sending my reply to our newbie friend, wa4whl but, with only 2 posts to his credit in over nine years, I doubt I'll ever get a reply. :palm:

 I only jumped on that post out of desperation since I couldn't track down other discussions on the SA22C anywhere on the 'net, just that one reference. I ticked the notification button just in case before closing that tab to post more on my 'new toy' here in this now renamed topic thread.

 I've been busy testing it out against my home brewed gpsdo (basically a revamp of the G3RUH design with an M8T replacing the Jupiter T, phase locking the ocxo at 100KHz with a 1200 seconds TC filter) and was finally persuaded to reorganise the bench to accommodate my LPR101 based RFS and give my SDS2502X+ something useful to do for its 46 second boot up time and its 54W consumption (versus the 16 seconds/22W of the SDS1202X-E that usually suffices as a graphic TIC).

 The reason for the reorganisation to show both rubidium oscillators against the GPS reference simultaneously being the reappearance of a curious symptom of instability I'd seen with the LPRO a few weeks ago (the impetus for blowing another 290 quid in total on acquiring the SA22C).

 Initially I'd assumed the rapid 10 to 20 ns transient backward phase shifts to be a very peculiar intermittent fault being developed by my now 25+ yo lpro but I am now starting to suspect the GPS system may be suffering fall out from the war in Ukraine due to electronic countermeasures by either the Russians or possibly even by the Americans as part of a strategy to wrong foot the Russians who may be using GPS as a sanity check/backup to their GLONAS systems  :-\

 Yesterday I started seeing similar symptoms with the SA22C which had me dialling in offset corrections via its serial interface to try and keep it syntonised to the GPS reference which raised my suspicions once more about the stability of the GPS system as a high(ish) precision reference source. That was the moment I decided to make room for my LPRO RFS in order to confirm or lay my suspicions to rest. Of course, this 'weirdness' only occurs once every few days or so and the setup now seems to back to its normal state of precision (ionospheric disturbances aside).

 Unfortunately, the SA22C is only at the bench testing stage and subject to the variations of ambient temperature (just like the LPRO did when I was first checking it out almost 2 years ago now) which makes the task of pinning down 'the guilty party' just a little more tricky than I like. Still, I only need to witness another such 'event' to determine which, between the LPRO RFS and the GPS reference. is the true cause of these phase jumps.

 In the meantime, I'm trying to make sense of the information in the Symmetricom / Microsemi user guides and data sheet specifications. Of more immediate interest to me, is how to interpret the health data which is simply listed in hexadecimal form as per the w command output shown below:-

r>w

AData:
  SCont: 6012
  SerNum: A12
  PwrHrs: 5F18
  PwrTicks: 1282841
  LHHrs: 5ECF
  LHTicks: 10F5542
  RHHrs: 5ECF
  RHTicks: 1063C65
  dMP17: 4159F035.
  dMP5: 3D4E9CE6.
  dHtrVolt: 4172E210.
  PLmp: 3F93465F.
  PRes: 3FC3C440.
  dLVthermC: B8520000.
  dRVthermC: B99CD000.
  dLVolt: 3F8F52C5.
  dMVoutC: C9D122BB.
  dTempLo: 41000000.
  dTempHi: 42B10000.
  dVoltLo: 4156B1A8.
  dVoltHi: 4173473E.
  iFpgaCtl: 204C
  dCurTemp: 425A0000.
  dLVoutC: 3E0508D5.
  dRVoutC: 3DF08B1B.
  dMV2demAvg: 3F0C7C1A.

  Does anyone know how to interpret this data. Unless the  PwrHrs: (5F18) counter has wrapped around a 16 bit limit, it looks like it's clocked up just a tad over 1014  24344 hours, a mere 2.7771 years worth of run time, which if true, is just 27% of its design life.

 The i command displays this:

r>i


SA22C by Symmetricom, Inc., Copyright 2006
   SA22 Version 6.02C of 7/2006; Loader Version 3
   Mode CN41  Flag 0004

Unit serial code is 1211EC3234-h, current tuning state is 6
Crystal: 3938700hz, ACMOS: E4E1C0.00000000hz, Sine: 989680.00000000hz
Ctl Reg: 004C, Res temp off: BFC53F7D., Lamp temp off: BFF9418E.
FC: enabled, Srvc: low

 The final line just indicates the sense of the service output pin when it asserts a service required warning. The service pin reads 4.9 volts so no warning of imminent failure within the next few months at least. Worthy of note is the firmware version which predates the later version 6.05C which offers customer access to the 1PPS configuration, along with being able to enable sine wave output.

 The user guide suggested that the early version would normally default to 10MHz sine wave and enable the PPS feature  by default, the implication being that the end user wouldn't be able to tune the PPS control algorithm with the earlier firmware version. To test this, I added an SMA-F socket to my breakout board and connected an M8T to it. Sad to say, there was absolutely no indication that the PPS input option was enabled so it seems I'm out of luck regarding a quick solution to creating a GPSDRO by simply adding nothing more than a suitable GPS receiver module into the mix. :( I wasn't too surprised at this after reading this note in the user guide:
================================================================================
Setting the 1PPS Synchronization
The following are the steps for setting 1PPS signal synchronization (SA.22c with 1PPS enabled customer
version of firmware at revision 6.05c or higher installed.)
================================================================================

 I did a search to try and track down a firmware update but came up completely dry. I guess such an update has now long since become yet another item of "Unobtainium", so it's looking like I'll just have to create a 10MHz sine wave output GPSDRO the old fashioned way. :( I rather think I'm on my own in this endeavour but if anyone knows anything useful about these SA22C units, I'd really appreciate the help.

 I've attached four photos of the current bench setup.


Johnny B Good:
Notes regarding the photos attached to the previous post (plus a set of ten screenshots taken earlier this evening)

The SDS2000X+ series conveniently display the date and time in the bottom right hand corner of their displays. The time is UTC and running about 1 1/2 minutes behind.

 The first image shows the SA22C with the heatsink plate temporarily removed in the left hand side of the foreground with the LPRO based RFS to its right (that big 12 by 10 by 8 inch WxDxH aluminium enclosure sized to accommodate a 25mm polystyrene foam lining and an LPRO wrapped in another 25mm thick polystyrene foam overcoat covering all sides other than the fan cooled heat spreader bolted to its base plate).

 The oled display is showing (top to bottom respectively) ambient temperature, base plate temperature and barometric pressure. The lockable ten turn counter drives a ten turn 5K helipot used to trim the frequency. The LED to the LHS of the oled display is a bi-colour red/green LED which shows red for powered up awaiting atomic lock and green to indicate lock. Power is courtesy of a 19v laptop charging brick. Its 10MHz sine wave output exits via an SMA-F connector mounted on the rear panel which I have connected to CH3 on the SDS2504X+ in the background. CH1 is the GPSDO reference from which the timebase is triggered whilst CH2 is the SA22C's square wave output. I have infinite persistence enabled to record the relative shift between the GPS reference (the source of the 5ns pk-pk wobbles on the LPRO trace). The larger drift of the SA22C obliges me to clear the persistence every half hour or so to keep track of its behaviour otherwise I'd only need to clear the persistence once every 6 to 24 hours or so for the LPRO on its own

 The second image includes the SDM3065X displaying the LPRO's lamp voltage. The second image is a more encompassing view to include the cheap 'n' cheerful Owon XDM1041 displaying the GPSDO's EFC voltage. The final image is just a slightly different angle to reduce the glare on the SDS's display.

 I've attached ten screen shots taken over a 2 hour and 20 minute time period to demonstrate the difference between a temperature stabilised rubidium oscillator (the LPRO) and one attached to a heatsink simply to prevent it overheating (typical hobbyist usage case going by the various youtube videos I've watched on the subject of setting up a home lab with an atomic frequency standard).

 Enjoy!

P.S. I almost forgot: I ran some temperature excursion tests on the SA22C using a small USB powered desk fan to chill it down and I was able to ascertain that it has a negative frequency tempco for what that's worth. My particular LPRO for example, has a positive frequency tempco but a sample of 19, according to the user guide showed a scattering of both positive and negative tempcos so that information is not a hard fact regarding the tempco of other SA22Cs, more a point of mild interest.

Johnny B Good:
 Since there's some interest being shown in the SA22C, I thought I'd pop the lid off and take a few teardown photos for those either too shy to ask or daring enough to break the warranty sticker on theirs. :)

 It's not as if it's still in warranty and it does appear to be functioning ok after clocking up some 24,344 hours of service life out of a design life of better than 88,000 hours. It's obviously survived the 'infant mortality' phase of the MTBF bathtub curve and a long way off hitting the old age mortality end of said curve. IOW, it's just nicely matured :) That's assuming of course, that I've correctly interpreted the PwrHrs: data -fingers crossed that I have.

 I've attached the 1st 5 images of the teardown sequence and there'll be several more posts to follow with the later high detail images.

Johnny B Good:
More images

Navigation

[0] Message Index

[#] Next page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod