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

[NivagSwerdna]

I'm pretty close and I want it to fit in my box... PCB on the way.... just need to tweak a few values to bias it right....

[Nuno_pt]

Still waiting for the rest of my stuff to arrive, antenna, cables, raspberry 3, and other small stuff.
Then put it all together to see the size of the box, with the 8 way distribution amplifier.

I'm thinking to power the UCCM separate from the raspberry and arduino.

One entrance for +12V 8A, then divided in +5V 3A for UCCM, +5V 1A to arduino, +5V 2A for raspberry.

[Macbeth]

You don't think a close to 100W supply is a little bit overkill?

[Vgkid]

I figured 5v at 5A would easily be sufficient for everything. Thinking about it a 15W 5V supply would be the least power needed. Use a precision comparator or timer to turn on a relay powering the digital stuff. After either about 10 minutes, or when the shunt voltage drops.

[NivagSwerdna]

Very interested to hear about power supply solutions... I bought a 6V 2A cheapo adapter from EBAY... gives me the same brown out behaviour I get with my bench supply until the OCXO is warm but after that everything stabilises and I get around 800mA constant from my bits and pieces.

The early start brown-out isn't very nice... especially to the Rpi.

I guess the sensible solution is to provide 12V and regulate down for the controllers?

[Nuno_pt]

It can be overkill, I've to see what I've around here, but I was talking about taking out the +5V from the +12V 8A supply.

My thinking is provide one Input for 12V and then divided by the UCCM, Arduino and Raspberry.

UCCM - 5V 2.5A for start on OCXO warmup
Arduino - 5V 500mA with TFT
Raspberry 3 - 5V 1A

This should be around 12V 4A or maybe 3A

+12V 4A to one LT1085 feeding the UCCM, picking up the same +12V feeding one LM7805 to the Arduino, and another to the Raspberry.

Maybe one comparator, that compares the dropout of the OCXO warmup from 3A to 800mA, and then close one reed relay that will feed the Arduino and Raspberry.
In this approach there will be no display when power up, but since the GPSDO is to be run in contínuos mode, and even if it be unplug from time to time, no problem in staying 5 minutes without display.
 
 

[ekyle]

It can be overkill, I've to see what I've around here, but I was talking about taking out the +5V from the +12V 8A supply.

My thinking is provide one Input for 12V and then divided by the UCCM, Arduino and Raspberry.

UCCM - 5V 2.5A for start on OCXO warmup
Arduino - 5V 500mA with TFT
Raspberry 3 - 5V 1A

This should be around 12V 4A or maybe 3A

+12V 4A to one LT1085 feeding the UCCM, picking up the same +12V feeding one LM7805 to the Arduino, and another to the Raspberry.

Keep in mind, if you are using a 12v supply and are using linear regulators to get 5v, you will be dissipating more power in the regulators than in the devices themselves. 7v * the current draw for each device.

[ekyle]

Wow. This is a great thread! I've learned a lot of info about the Symmetricom GPSDO here. Mine came with the oil on the board and a few damaged caps. All of which were repairable. One thing I noticed is that it really is sensitive to having a LOS view to the satellites. I had my GPS antenna (A lucent antenna with 26db of gain) on a pole in some trees. It had a terrible time even getting a fix. It took hours. I moved the antenna to the unused chimney on the roof of the house and it made a world of difference. Before it would only track about 4 or 5 satellites and they were never stable. Now it tracks about 8-10 on average and they are stable. So, I had mine running for a day or so and I got an error that the frequency control was at the upper limit. The Fc to the OCXO was pegged at 5.1v. I also was getting low OCXO current warnings. I removed the OCXO and poped the cover off and reflowed all of the connections on the lower board. It's a single oven if anyone wants to know. I put it back together and it ran for a few days and then I got the same error. So, I happened to see an exact replacement OCXO on Ebay: 222113598831. I ordered one because is was at a reasonable price and it was not overseas. Well, the "new" oscillator arrived a few days ago and it was defective. I sent it back and got a refund. I wasn't going to take another chance on a replacement fearing that maybe they all came from a bad lot. I'm wondering if anyone else has had trouble with the Symmetricom OCXOs? I bought a used Oscilloquartz OCXO and hopefully it will arrive next week and will be a drop in replacement.

[uncle_bob]

Wow. This is a great thread! I've learned a lot of info about the Symmetricom GPSDO here. Mine came with the oil on the board and a few damaged caps. All of which were repairable. One thing I noticed is that it really is sensitive to having a LOS view to the satellites. I had my GPS antenna (A lucent antenna with 26db of gain) on a pole in some trees. It had a terrible time even getting a fix. It took hours. I moved the antenna to the unused chimney on the roof of the house and it made a world of difference. Before it would only track about 4 or 5 satellites and they were never stable. Now it tracks about 8-10 on average and they are stable. So, I had mine running for a day or so and I got an error that the frequency control was at the upper limit. The Fc to the OCXO was pegged at 5.1v. I also was getting low OCXO current warnings. I removed the OCXO and poped the cover off and reflowed all of the connections on the lower board. It's a single oven if anyone wants to know. I put it back together and it ran for a few days and then I got the same error. So, I happened to see an exact replacement OCXO on Ebay: 222113598831. I ordered one because is was at a reasonable price and it was not overseas. Well, the "new" oscillator arrived a few days ago and it was defective. I sent it back and got a refund. I wasn't going to take another chance on a replacement fearing that maybe they all came from a bad lot. I'm wondering if anyone else has had trouble with the Symmetricom OCXOs? I bought a used Oscilloquartz OCXO and hopefully it will arrive next week and will be a drop in replacement.

Hi

All of the OCXO's you see for the most part are salvaged off boards in China. Every once in a while there is a "brand new" one that pops up. The other 99 batches are all salvage parts. They all go through a pretty brutal process getting them off of the boards. If you run a full test of the parts when you get them (including temperature, aging, and phase noise) it is rare to find even one in 40 that meet spec. With most batches it does not take a lot of testing to spot the failures, 10 out of 10 have an output < 0 dbm when the spec is +7 or some such thing. It's not the OCXO that is the issue as much as using a weed burning torch to pull them off the board. Tossing them into a 55 gallon drum from 10 feet away likely is not the best next step after they come off the board. Having the drum sit out in the rain until it fills with OCXO's seems to be the next phase.

Good Luck!

Bob

[NivagSwerdna]

And the most damaged ones at the bottom of the barrel are taken by flyxy2015

DO NOT PURCHASE FROM flyxy2015.  I agree with others... he is a THIEF.  I ended up with nothing and paid the full amount.

(My working board came from waikeionline_uk)

[PA0PBZ]

(My working board came from waikeionline_uk)

He sent me 2 damaged ones I returned, now he supposedly sent me "working" ones but I have to see it to believe it. Also, he refunded me $20 from my original payment to pay for the return. The tracking on the return shipment shows that he still did not pick them up (but that could be an error of course). Oh well, you get what you pay for I guess...

[NivagSwerdna]

Also, he refunded me $20 from my original payment to pay for the return. The tracking on the return shipment shows that he still did not pick them up (but that could be an error of course). Oh well, you get what you pay for I guess...

That's where he got me... the tracking never showed completed so EBAY sided with him.  Returns don't work with some people.  He also timed is so it was too late to leave feedback.

[Nuno_pt]

I got mine from waikeionline_uk too, and the board work after 2 caps, and the GPS connector.
I'll probably get another board from him, after I finish this one.
About feedback, I only leave feedbacks after I receive the order and try it, only then I go there and leave the feedback.

Received Friday some more bits, still missing the antenna, cables and some connectors.

[kutte]

I bought a used Oscilloquartz OCXO and hopefully it will arrive next week and will be a drop in replacement.

my Oscilloquartz OCXO 8663-XS did not work as an exchange although beeing compatible in its dimensions it is not with its Vref and Vefc.
Kutte

[kutte]

made some Adev measurements of my Symmetricom GPSDO.  Adev of GPS-locked GPSDO is very similar to ZigmundRat's (https://www.eevblog.com/forum/projects/a-look-at-my-symmetricom-gpsdo-%28ocxo-furuno-receiver%29/msg890574/#msg890574) measurement, see my blue curve. For reference I measured my Lucent KS-24361, see red curve under same conditions. And finally I measured the Symmetricom OCXO in "holdover-mode" (i.e. antenna input left open), see green curve. As you can see, the Symmetricom GPSDO is not optimal disciplined but the OCXO alone is one of the best I own. Please note: Adev-reference is always my Trimble Tunderbolt since I do not have any better.
Kutte

[NivagSwerdna]

My first ever PCB!  Isn't she a beauty!  10MHz in and 4x 10MHz outputs, Pulse stretch and multiple PPS outputs and a TTL/RS232 level converter on the flip side.

10MHz in... via LT6551 gives 4x 10MHz out...

[Nuno_pt]

Very nice board.

[Vgkid]

Nice job, looks tiny.

[ekyle]

I bought a used Oscilloquartz OCXO and hopefully it will arrive next week and will be a drop in replacement.

my Oscilloquartz OCXO 8663-XS did not work as an exchange although beeing compatible in its dimensions it is not with its Vref and Vefc.
Kutte

Rats! I saw the Vref and Vefc were around 8v and 0-10v respectively. Did you use a different OCXO? Do you know if the Vref output is actually used?
I don't think it would be too difficult to scale them with a couple op-amps using the 12v as the supply. Or maybe a single shunt resistor for the Vref output since it has about a 1k impedance and then just a x2 op-amp for the Vefc.

[Vgkid]

I bought a used Oscilloquartz OCXO and hopefully it will arrive next week and will be a drop in replacement.

my Oscilloquartz OCXO 8663-XS did not work as an exchange although beeing compatible in its dimensions it is not with its Vref and Vefc.
Kutte

Rats! I saw the Vref and Vefc were around 8v and 0-10v respectively. Did you use a different OCXO? Do you know if the Vref output is actually used?
I don't think it would be too difficult to scale them with a couple op-amps using the 12v as the supply. Or maybe a single shunt resistor for the Vref output since it has about a 1k impedance and then just a x2 op-amp for the Vefc.

Besides tuning voltages, you will also need to look into the tuning slope of the ocxo.  Positive vs negative ones.

[kutte]

Rats! I saw the Vref and Vefc were around 8v and 0-10v respectively. Did you use a different OCXO? Do you know if the Vref output is actually used?
I don't think it would be too difficult to scale them with a couple op-amps using the 12v as the supply. Or maybe a single shunt resistor for the Vref output since it has about a 1k impedance and then just a x2 op-amp for the Vefc.

a data sheet for the OSA 8663-XS you may find here:
http://www.2917.com/EBAY-images/DOCXO/8663.pdf
I did not try any further adaptations.
Kutte

[NivagSwerdna]

The dissipation of heat by the GPSDO is a bit of a problem... it is making my room a bit hot.

What material might be suitable to put around the OCXO to thermally isolate it?  (I have seen some odd foamy stuff in Rb standards before)

[Nuno_pt]

Some guys put some kind of foamy on all 4 sides of the OCXO, but from what I've read on the Time nuts list, this will make the OCXO incapable of heat regulation, because the heat would have no place to go, making the OCXO temp rise so the OCXO will lower the EFC to compensate, then it decompensating the OCXO.

Some guys put the foamy but leave about 3 to 4cm around the OCXO, but it will lead to the same result.

Try to read some of this on Time nuts list.

Still waiting for my antenna to show up here, it-s been in the customs for about 10 days, but the rest is here.

[uncle_bob]

Some guys put some kind of foamy on all 4 sides of the OCXO, but from what I've read on the Time nuts list, this will make the OCXO incapable of heat regulation, because the heat would have no place to go, making the OCXO temp rise so the OCXO will lower the EFC to compensate, then it decompensating the OCXO.

Some guys put the foamy but leave about 3 to 4cm around the OCXO, but it will lead to the same result.

Try to read some of this on Time nuts list.

Still waiting for my antenna to show up here, it-s been in the customs for about 10 days, but the rest is here.

Hi

The thermal gain of the heater in the OCXO is determined by the power in the heater times the thermal resistance from the heater to ambient. If you add insulation, you increase the thermal gain. A modern OCXO is "tweaked" for optimum gain at the factory. If you increase or decrease the gain, the unit will no longer be at the optimum setting.

The second issue is that there is only so much energy controlled by the heater. Stuff like the oscillator draw power regardless of the temperature. That means that the oven must always be some fixed amount above the upper ambient temperature. The classic term for this is "dead power" and thus "dead heat". In a typical design you may see 10C for this.

Next up is the fact that an economical crystal has a tolerance on it. You don't get every single one at exactly the same turn temperature. The tighter you spec them, the more you throw away. Net result is a typical tolerance of about 15 to 20C on the crystal temperature.

So, lots of variables. Lots of things to go wrong. Lots of "worked for you" / "failed for me". Unless you have the ability to do a full temperature run on the OCXO (say -20 to +70C), there is no real way to even know how much you degraded the device.

Bob

[rastro]

I received a unit with L20 pulled up and damaged.   I need to find a replacement but its not clear what part is really needed.

This has was previously identified as a 5uH Inductor.  However after looking at the damaged part it seems like part has 4 separate pads and comprises what appears to be 2 separate coil winding.  If I were to see this part without knowing it's designation "L" I would think it was a transformer.

hi to all,
received my Symmetricom board yesterday. As to be exspected some damage could be seen. On the top side xtal y1 is missing, inductor L20 was ripped off. On the botton side capacitor C26 and resistor R83 were missing.
R83 could be easily identified as 103 (10kOhm), C26 is in neighbourhood of C27 which I measured as 1.2 uF. I put in 0.56 uF as a provisory measure.
Inductor L20 I measured as 5uH and replaced with a new one. ...
regards Goetz

I've attached some pictures showing the bottom of the damaged L20.  There appears to be a 3 turn winding between 2-pads sitting next to a larger double layer winding connecting to 2 pads on the other end.  So I can either find a replacement (ideal) or try and cobble together a repair (difficult).  Any further insight would be greatly appreciated.