Author Topic: Economical option for precision frequency reference?  (Read 155346 times)

0 Members and 1 Guest are viewing this topic.

Offline Shock

  • Super Contributor
  • ***
  • Posts: 4200
  • Country: au
Re: Economical option for precision frequency reference?
« Reply #250 on: September 05, 2015, 08:22:10 am »
FYI the LUCENT/SYMMETRICOM Z3810AS, KS24361 has gone up to $200 now. 9 left.
Soldering/Rework: Pace ADS200, Pace MBT350
Multimeters: Fluke 189, 87V, 117, 112   >>> WANTED STUFF <<<
Oszilloskopen: Lecroy 9314, Phillips PM3065, Tektronix 2215a, 314
 

Offline motocoderTopic starter

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #251 on: September 05, 2015, 05:11:13 pm »
Huawei/BG7TBL 2015-07-08 (Star4+ board)
...
  Cons: Requires minor change (solder two wires) to fix management interface mis-wiring from BG7TBL. Limited availability now it seems. No software to monitor.

Can you give details or link for these changes?

Here are the details for the two simple mods that I have done to the Huawei.

The Huwai/"BG7TBL" 2015-07-08 is actually a Star4+ board from Oscilloquartz that BG7TBL has strapped onto a PCB of his design and jammed into an enclosure. The BG7TBL PCB is a nicely designed board that allows him to connect a variety of 3rd-party GPSDO and provide consistent front panel connections. He includes a MAX3232 chip in there so that internal TTL-level signals can be converted to RS-232 levels before connecting to the DB9, connection points for power, LEDs, and the BNC connectors.

In this post I will show you how to reconfigure the DB9 connector to connect to the actual RS232 management interface on the Star4+ board, as well as how to add a square wave output to the front panel. Since the actual GPSDO is a Star4+ board, the first thing I would recommend is to download the user manual for that which usagi links in this post:

https://www.eevblog.com/forum/testgear/bg7tbl-gpsdo-master-reference/msg740339/#msg740339

If you look near the bottom of this manual, you will see a diagram for a connector on the Star4+ board. THis is where we are going to be tapping into the various signals. Let's discuss the management interface first, as I think that's the mod most people will be interested in.

As the board comes from the eBay seller, the front panel DB9 is a "talk only" interface and is connected to an RS-232 level version of the Tx interface from the uBlox GPS chip. This will sort of let you run the uBlox u-center software. I say "sort of" because this software gets confused by the interface being effectively read-only, and it has some issues sometimes and much of the functionality doesn't work. However, it's enough to see the satellite strength and some basic stats. If having this sort of nice graphical interface is important to you, stop here, because you won't get that after the mod. If you proceed with this mod, you will (until some nicer software is available) just be able to use a serial terminal to connect to the board. However, you will have access to more information, including the state of the actual GPSDO, and the ability to set parameters such as the antenna delay. The board has a set of commands that are documented in the aforementioned manual; take a look at those to see what's possible.

On the Star4+ board is a Xilinx FPGA, which has inside it a soft-core processor. This processor is what actually implements the control logic of the GPSDO. It talks to the uBlox chip via a serial connection  on the uBlox, which is why BG7TBL was not able to route the uBlox Rx pin out to the DB9 connector. This soft-core processor has its own serial connection which we can talk to send commands and view status.

To proceed with the mod, remove the 4 screws from the front panel and slide out the board. The picture below shows a top-down view of the Star4+ board sitting on top of the BG7TBL adapter board. Note the black connector that the red arrow is pointing to. THis is where we are going to make our connections. Also note the arrow pointing to the uBlox chip. This is where the uBlox Tx connection is being made.


The next picture shows a detail view of the connector. The yellow wire is the connection to the management Tx pin, and the green wire is the connection to the management Rx pin. Solder some suitable wire there, and route these wires around to the bottom of the board. If you're interested in also adding the square wave output, you can also see the connection for that here.



Here is a close-up of the bottom side of the BG7TBL adapter board. Note the yellow and green wires, and the dangling red wire. The red wire is coming from the uBlox Tx pin, and used to connect to where the yellow wire connects now. These connections go through the MAX3232 chip, and then onto the front panel.



If you're also interested in adding a square wave output, this is what I did there. Remove the front panel, which is a simple matter of removing the nuts from the two BNC connectors and the one SMA connector. Drill a suitable size whole for your BNC. I purchased a BNC connector with a pre-attached pigtail off of eBay. It had an SMA connector on the other end, which I just snipped off. Re-attach the panel, add in your new BNC, and solder it to the location shown in the previous photo. There is a convenient spot to solder the ground/shield right there, so it's a very simple connection other than the fact that the connector pads are very small.



Now before you do this mod, you might want to have a look at the scope capture of the square wave output. It's not the nicest square wave in the world, and in particular note that the duty cycle is not remotely close to 50%. At some point I will probably come back and add a little daughterboard to convert the sine wave to a clean, 50% duty cycle square wave.

« Last Edit: September 06, 2015, 06:50:13 am by motocoder »
 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Economical option for precision frequency reference?
« Reply #252 on: September 06, 2015, 02:12:13 am »
I was just wondering, was the OP question actually answered? ;D

What is the best economical option for a precision frequency reference?
With my experience so far, the FE5680B seems to be it.

What is the easiest to use and have running and have accurate without worrying about being off because of age drift?
IMHO the bg7tbl GPSDO on ebay, any flavor, is the best bet.

So far I've purchased a Nortel NTBW50AA for 80 USD shipped and a second Trimble unit which is a PCB with no case for 48 USD shipped. I also purchased cables to convert the various connectors on the two units to BNC for another 12 USD shipped. And lastly two Trimble GPS antennas which were 15.99 USD shipped each. I already had power supplies on hand. I have the Nortel unit fired up and running, the Trimble PCB has not yet arrived from china. The Nortel is a little large but is known to work extremely well, it has an even second output instead of 1 PPS but 1 PPS is on the board, a cut trace and jumper and you can have the 1 PPS if needed. In my case I really only care about the 10 MHz. The Nortel also has a DB9 factory installed and directly supports the Trimble diag software as well as Lady Heather. The other unit I will be receiving has a text based user interface, a DB9 just needs to be wired in. I will install it in a box when I receive it and after some aging time I plan to compare the two side by side. I already have both antennas installed on my tower outside with exact length runs of coax in to the shop.
I'd still like a rubidium based reference and when I can find one at a reasonable price I will pick it up. Clearly the rubidium is easier for most people to manage as it doesn't need an antenna however if you receive one off ebay you have no idea if it really is outputting 10 MHz or 1 PPS - they still technically need calibration. Also many of them are 10 to 20 years old. So as an only source I prefer a GPS based solution over rubidium. With a little effort and patience it should be no problem to purchase and setup a nice stable reference in your shop for under 100 USD. The Trimble based GPS units also seem to require less power then rubidium, my Nortel is 5 to 6 watts when warm.
VE7FM
 

Offline motocoderTopic starter

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #253 on: September 06, 2015, 04:09:06 am »
I was just wondering, was the OP question actually answered? ;D

What is the best economical option for a precision frequency reference?
With my experience so far, the FE5680B seems to be it.

What is the easiest to use and have running and have accurate without worrying about being off because of age drift?
IMHO the bg7tbl GPSDO on ebay, any flavor, is the best bet.

So far I've purchased a Nortel NTBW50AA for 80 USD shipped and a second Trimble unit which is a PCB with no case for 48 USD shipped. I also purchased cables to convert the various connectors on the two units to BNC for another 12 USD shipped. And lastly two Trimble GPS antennas which were 15.99 USD shipped each. I already had power supplies on hand. I have the Nortel unit fired up and running, the Trimble PCB has not yet arrived from china. The Nortel is a little large but is known to work extremely well, it has an even second output instead of 1 PPS but 1 PPS is on the board, a cut trace and jumper and you can have the 1 PPS if needed. In my case I really only care about the 10 MHz. The Nortel also has a DB9 factory installed and directly supports the Trimble diag software as well as Lady Heather. The other unit I will be receiving has a text based user interface, a DB9 just needs to be wired in. I will install it in a box when I receive it and after some aging time I plan to compare the two side by side. I already have both antennas installed on my tower outside with exact length runs of coax in to the shop.
I'd still like a rubidium based reference and when I can find one at a reasonable price I will pick it up. Clearly the rubidium is easier for most people to manage as it doesn't need an antenna however if you receive one off ebay you have no idea if it really is outputting 10 MHz or 1 PPS - they still technically need calibration. Also many of them are 10 to 20 years old. So as an only source I prefer a GPS based solution over rubidium. With a little effort and patience it should be no problem to purchase and setup a nice stable reference in your shop for under 100 USD. The Trimble based GPS units also seem to require less power then rubidium, my Nortel is 5 to 6 watts when warm.

Nortel NTBW50AA goes on eBay, with antenna but not power supply, for exactly the same price as the options I listed above: $159. I do see a few standalone units for $80 in the sold items search, but they are not generally available at that price and that price does not include the nice antenna that the $159 package does.

So I don't think you're comparing apples to apples here. Let's say  you didn't have power supply on hand, but that one could be had for $15. So after looking and waiting on eBay for some time, you finally snag a deal on the $80 Nortel unit. Then you buy an antenna for $16, and the BNC converter for $6. Your at $115. Ok, you saved $54 with some smart shopping and by being willing to wait to find the right deal. Good on you. However, you also got an 8 channel GPS receiver with relatively poor GPS sensitivity. ALso, have you actually verified the Nortel unit supports the software you mention? A brief scan of the time-nuts mailing list shows a few posts saying there is some model of that (it's actually listed as "Symmetricon Nortel NTBW50AA" that does not.
 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Economical option for precision frequency reference?
« Reply #254 on: September 06, 2015, 04:24:41 am »
I have Lady Heather connected to my Nortel unit as I type this. Started the program and it connected. To build a system for under 100 USD it would take time I do admit. I was more referring to the 48 dollar unit I picked up off ebay. The price has since gone up but they seem to go up/down quite often. They have been as low as 40 USD shipped. The cheaper one also has a newer 12 channel receiver based on posts over on time-nuts, hopefully it arrives soon.

I selected a price of 100.00 USD as to me that seemed to meet the "economical" request. Considering the price of quality test equipment maybe anything under $100 should be considered cheap.

btw, my Nortel/Trimble unit does not mention Symmetricon anywhere on it and is the version with a single PCB.

Maybe we could come up with a list of known working options that includes pictures for people so they can try to acquire the exact same hardware if they really want to get something going as cheap as possible instead of buying one of the ready to go solutions. I am buying various pieces myself as I find it more fun, others may just want to get something that plugs in and runs when it arrives.
« Last Edit: September 06, 2015, 04:39:07 am by TheSteve »
VE7FM
 

Offline EV

  • Frequent Contributor
  • **
  • Posts: 525
  • Country: fi
  • Aficionado
Re: Economical option for precision frequency reference?
« Reply #255 on: September 06, 2015, 05:47:19 am »
Here are the details for the two simple mods that I have done to the Huawei.
...
...

Thanks motocoder!
 

Offline motocoderTopic starter

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #256 on: September 06, 2015, 06:46:14 am »
I have Lady Heather connected to my Nortel unit as I type this. Started the program and it connected. To build a system for under 100 USD it would take time I do admit. I was more referring to the 48 dollar unit I picked up off ebay. The price has since gone up but they seem to go up/down quite often. They have been as low as 40 USD shipped. The cheaper one also has a newer 12 channel receiver based on posts over on time-nuts, hopefully it arrives soon.

I selected a price of 100.00 USD as to me that seemed to meet the "economical" request. Considering the price of quality test equipment maybe anything under $100 should be considered cheap.

btw, my Nortel/Trimble unit does not mention Symmetricon anywhere on it and is the version with a single PCB.

Maybe we could come up with a list of known working options that includes pictures for people so they can try to acquire the exact same hardware if they really want to get something going as cheap as possible instead of buying one of the ready to go solutions. I am buying various pieces myself as I find it more fun, others may just want to get something that plugs in and runs when it arrives.

I'll be interested to see what the Trimble is when you get it. I suspects it's teh surplus board that is in the "Trimble" BG7TBL unit I tested. Hopefully not, because that one was at the bottom of my list in terms of features and bang for the buck.

I agree it's fun to assemble stuff. I had a lot of fun playing around with the Lucent unit, but now I really just want the space back on my bench.
 

Offline motocoderTopic starter

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #257 on: September 06, 2015, 06:51:48 am »
Here are the details for the two simple mods that I have done to the Huawei.
...
...

Thanks motocoder!

You're very welcome. We should both thank usagi for getting the Star4+ manual from OscilloQuartz, and edpalmer42 for numerous helpful posts on this topic.

BTW - you may want to refresh my post; I forgot to embed the picture of the bottom of the board.
 

Offline voltsandjolts

  • Supporter
  • ****
  • Posts: 2282
  • Country: gb
Re: Economical option for precision frequency reference?
« Reply #258 on: September 13, 2015, 08:04:18 pm »

I'd just like to point out a very cheap and stable reference for anyone like me who has fairly basic accuracy requirements. I only need an accurate reference intermittantly to calibrate my old but reliable OCXO based frequency counter. As has been discussed in this thread, the UBLOX LEA-6T has improved timing specs over some other UBLOX modules, can output a MHz signal locked to GPS atomic clocks and is available $30 on ebay complete with antenna.

http://www.ebay.com/sch/i.html?_from=R40&_sacat=0&_nkw=lea-6t&_sop=15

Unfortunately, neither of the timepulse outputs are wired outside the unit - so need to add some RG178 and BNC ($3)

OK, this unit has some nasty jitter but is it good enough to calibrate my frequency counter using a 10 second gatetime? I tested it against a (borrowed) Agilent 53230A with UOCXO option (nice!). I set the UBLOX output to 8MHz as recommened in datasheet for minimum jitter.

Answer: Over one hour the frequency spread was mostly within 4ppb and showed the Agilent to be 8ppb off frequency (well, that would need verified I guess).

So, I think LEA-6T is good enough for me.


 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Economical option for precision frequency reference?
« Reply #259 on: September 14, 2015, 03:26:59 am »
Messed around with my Trimble/Nortel NTBW50AA for a bit today. I converted the Chip Clock/8X output from 9.8304 MHz to 19.6608 MHz and the even second PP2S output to a more normal 1 PPS output that has a positive going pulse.

The 19.6608 MHz clock is useful to me as it is one of the standard reference in options on my Agilent N1996A spectrum analyzer.

The 9.8304 MHz to 19.6608 MHz mod is very simple. The 19.6608 MHz signal is normally divided by two by U18 (74HC74). The divide is bypassed by lifting U18 pin 6 and running a jumper(white wire) from TP30 to TP33.

The PP2S signal is supplied to R82 by an output from the Xilinx chip, it has an input(from the Trimble chip) which is fed a more useful 1 PPS (TP13). I ran a jumper (yellow wire) from the 1 PPS signal going into the Xilinx chip to R82 which was rotated 90 degrees to allow connection to the jumper and disconnect the PP2S input.
The last mod was to lift pin 9 of U18 and run a jumper (blue wire) from pin 8 of U18 to TP29. This inverts the signal so that the 1 PPS signal pulse on the output is positive going.

All of the signals are phase locked to each other and that should be preserved just fine with this mod. The actual delay of the PPS signal would need to be measured but should be a consistent fixed value as I have read the PPS signal going into the Xilinx(from the Trimble chip) is fully corrected.

Hopefully someone else may find this mod useful and sorry for the crap quality pic, no daylight here right now to get a really nice shot.

edit - there seems to be more then version of this GPSTM around, my board is marked as a "NTBW50AA 10" on the front, on the back it says it is PN 45321-00. It is a single board design that I believe came from a Sprint site. It talks to Lady Heather etc with no problems.
« Last Edit: September 14, 2015, 04:24:38 am by TheSteve »
VE7FM
 
The following users thanked this post: Dwaine

Offline Vgkid

  • Super Contributor
  • ***
  • Posts: 2710
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #260 on: September 14, 2015, 05:27:11 am »
What is the ocxo on your board?
If you own any North Hills Electronics gear, message me. L&N Fan
 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Economical option for precision frequency reference?
« Reply #261 on: September 14, 2015, 05:28:15 am »
What is the ocxo on your board?

34310-T
VE7FM
 

Offline SoundTech-LG

  • Frequent Contributor
  • **
  • Posts: 788
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #262 on: September 23, 2015, 03:48:07 pm »
Got what I think was a very ECONOMICAL Spectracom SecureSync on the bay. Can't wait to set it up...

Anyone have experience with these???



http://spectracom.com/Desktopmodules/Bring2Mind/DMX/Download.aspx?EntryId=419&PortalId=0


 

Offline SoundTech-LG

  • Frequent Contributor
  • **
  • Posts: 788
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #263 on: September 24, 2015, 01:38:28 pm »
 

Offline motocoderTopic starter

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #264 on: September 24, 2015, 02:16:51 pm »
New Lucent KS-24361 RFTG-u info...

https://www.mail-archive.com/time-nuts@febo.com/msg74968.html

So to summarize this for people that don't want to wade through the whole document. Someone has figured out how to modify the KS-24361 REF 0 unit to make it functionally equivalent to a REF 1 unit. You just need to come up with a surplus Motorola OnCore GPS unit from somewhere, and do a little SMD rework, and now you have the equivalent of a standalone REF-1 unit with the 10 MHz test-point output
 

Offline Vgkid

  • Super Contributor
  • ***
  • Posts: 2710
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #265 on: September 24, 2015, 03:01:28 pm »
Got what I think was a very ECONOMICAL Spectracom SecureSync on the bay. Can't wait to set it up...
very cool, expecting a teardown. Do you know what options it has?
If you own any North Hills Electronics gear, message me. L&N Fan
 

Offline SoundTech-LG

  • Frequent Contributor
  • **
  • Posts: 788
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #266 on: September 24, 2015, 07:11:15 pm »
yes... cool, probably not worthy of such a fine instrument, but I'll be trying my best... :-+

No option cards installed, and it has the standard OCXO, so nothing extremely exciting, but at least it doesn't have the downgrade TCXO  :phew:. I was thinking of doing a teardown with photos, or video. Spectracom claims it is impossible to upgrade the reference oscillator after manufacturing, but maybe they are not counting on an EEVblogger investigating that possibility. I noticed the unit weight increases by roughly a half pound with the high stability Rb Osc. installed. That does not sound like a surface mount device! Should be in my hands early next week some time.
 

Offline Vgkid

  • Super Contributor
  • ***
  • Posts: 2710
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #267 on: September 24, 2015, 07:30:47 pm »
When i first was getting into this...
I saw a spectracom 9183(i think right model) gps freq std. Reading the manual i determined to have the rb option. I bid,  put a max bid of 205. And lost. Another one appeared bid first, max bid of 343... lost again. The only person bidding against me was the same person. I was upset.
If you own any North Hills Electronics gear, message me. L&N Fan
 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Economical option for precision frequency reference?
« Reply #268 on: October 10, 2015, 11:34:06 pm »
Finally received my Trimble unit a few days ago. It is the same or a very similar unit used in the BG7TBL trimble units. I have it mounted in a box and it has been running for a few days now. So far I have no complaints. While I have no primary source to compare it to the jitter between it and my Nortel has been under 30ns over a 24 hour period. My unit has the 63090 OCXO in it. The part # on the board itself is 57964-15. The board appears to use an Xemics 3330R GPS RF front end IC which feeds I/Q outputs to the processor for decoding.

The output from the command "*IDN?" is:

TRIMBLE,57964-15,86972377,V3.0.0.11-01

If you have one of these units or the equivalent BG7TBL unit you can force a site survey with the command "SYSTem:PRESet"
The board stores the last survey location and maintains it without power. However as motocoder mentioned earlier in this thread it does not store the elevation mask setting. At power on it always defaults to 5 degrees. If you're performing a site survey I'd set it to 10 to 20 degrees.
The default serial comm is 57600,8,N,1

You can enable once per second time packets to be output via the serial port if needed. It has two dual color LEDs on the bottom of the board that I moved to the front of my enclosure. I didn't want to install 4 LEDs and didn't have any dual color common anode LEDs so I cheated and cut the tips off some 5 MM opaque LEDs and then used CA(glue) to attach the original SMD LEDs to the tips of the 5MM LEDs - it is working great and they look like regular 5MM dual color LEDs in use.

The board requires 6 volts @ 1.8 amps during warmup and 0.85 amps after it is warm. It is a possible cheap option for people if they already have an antenna/power/box etc and like building stuff. I got mine for 48 USD shipped off ebay but the price has since gone up quite a bit. I do see there is a Symmetricom version of the same board for for 48 USD shipped though which is likely a potential good option.
If you want a complete solution it would be tough to beat the regular BG7TBL packages offered on ebay.

I also came across a working Cesium reference for 1K today - sadly it is out of my budget for a fun toy.

Attached a few pics of the unit. I've since ordered a black enclosure so I may swap the top and bottom halves so its totally black when it comes in next week.
« Last Edit: October 11, 2015, 04:30:22 am by TheSteve »
VE7FM
 

Offline Macbeth

  • Super Contributor
  • ***
  • Posts: 2571
  • Country: gb
Re: Economical option for precision frequency reference?
« Reply #269 on: October 11, 2015, 08:13:49 am »
I do see there is a Symmetricom version of the same board for for 48 USD shipped though which is likely a potential good option.
If you want a complete solution it would be tough to beat the regular BG7TBL packages offered on ebay.

I believe this is what you refer to? I've just bought one anyway. Time to dip my toe in the time-nuts water. Perhaps I should buy some RF connectors while I wait for it to arrive. What are they? SMA?
 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Economical option for precision frequency reference?
« Reply #270 on: October 11, 2015, 08:20:10 am »
I do see there is a Symmetricom version of the same board for for 48 USD shipped though which is likely a potential good option.
If you want a complete solution it would be tough to beat the regular BG7TBL packages offered on ebay.

I believe this is what you refer to? I've just bought one anyway. Time to dip my toe in the time-nuts water. Perhaps I should buy some RF connectors while I wait for it to arrive. What are they? SMA?

Yep, that is the one. The mating connectors for the jacks on the board are MCX - this is the cable I used with the BNC connectors - http://www.ebay.com/itm/281313869844
VE7FM
 

Offline motocoderTopic starter

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #271 on: October 11, 2015, 08:03:56 pm »
I have decided to sell the Lucent, BG7TBL Trimble, and BG7TBL units that I bought to do the test/evaluation.

 The two BG7TBL units come with the antenna and power supply that the eBay seller ships with them.

The Trimble unit has some ventilation holes in the case. I added these because its oven runs quite hot, and I was concerned about longevity of the other components in the case.

The Lucent comes with a high quality medical grade power supply to which I have afixed the DB9 power connectors with the appropriate latches. It also comes with a high quality, FTDI-based RS-422 to USB adapter and a DB9 adapter cable. I'll throw in a TNC to N patch cable as well.

PM me if intetested.
 

Offline SoundTech-LG

  • Frequent Contributor
  • **
  • Posts: 788
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #272 on: October 12, 2015, 12:58:10 am »
I know there are curious types around here, so I wanted to do this minor tear-down, actually, more a description of the interior of the Spectracom SecureSync unit. There is not really much to disassemble.

 I also wanted to test it right away for 100% functionality, and no error conditions. So it has been powered on for a little more than a week now,
 and has begun to settle down a bit. It does seem quite sensitive to room temperature change. I live in a canyon, so reception is poor here.
 Removing no less than 20 screws just to pop the cover while powered up to try to prevent any fluctuations in performance,
 but taking the cover off lowered the temperature severely. I started wondering about the thermal environment it had been designed for.
I think in a very warm rack space. The unit has 6 option slots on the rear panel, but each are perforated with plenty of holes.
So if all 6 option slots are filled with modules, the holes go away, and now added heat from the modules, all keeping it quite toasty inside.
 It does have a fan, but according to the instructions you may never hear it cycle on, except briefly at power on. I decided covering the option
 blanks with tape to isolate it from the outside air was a good idea. Fan still does not cycle on. Temp slightly higher now, and less fluctuation.

The GPS radio is a Trimble resolution T. The standard 10 mHz OCXO is a Russian made Morion MV-197. It has a Western Digital SSD of 1 GB.
 A small computer on board, an SOM-4455 by Advantek, featuring an AMD Geode LX 800 CPU, AMD CS5536 chipset, and 1GB of DDR-400 RAM.
The backup battery is a huge coin cell BR3032.  You can see the area of the board for the optional DC power, no fuse installed, and J33 not connected,
2 rather large inductors sitting there as well.

The Rubidium option jacks are there, no info about what that is. Even if you could ID it, Firmware for it would need to be installed.
According to Spectracom there are no upgrades after production. It appears to be a late 2012 production unit. The SSD is dated 9/5/12,
 the Morion MV-197 dated 12/28, which is the 28th week of 2012. All in all a pretty interesting, and quite sophisticated unit.

Here's a few photos of the interior, front, and rear. and some screen shots of interest.
« Last Edit: October 12, 2015, 01:15:32 am by SoundTech-LG »
 

Offline SoundTech-LG

  • Frequent Contributor
  • **
  • Posts: 788
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #273 on: October 12, 2015, 12:59:36 am »
more photos
 

Offline SoundTech-LG

  • Frequent Contributor
  • **
  • Posts: 788
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #274 on: October 12, 2015, 01:01:01 am »
more...
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf