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

[rastro]

Location of L20.

[PA0PBZ]

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.

The continuing story of waikeionline_us (not _uk but the same guy obviously)

Today I received his replacement shipment (to my surprise I must say).

Guess what, not 2 but 4 pcb's! But there is where the fun ends: 2 are missing the TXCO's, they are all damaged and cannibalized (push buttons, crystals, connectors, coils and caps).
On top of that he never picked up my return shipment so that also landed on my desk today. I wonder if I can get 2 working ones from the pile of rubbish 
I think I should ask PayPal for a refund of the rest of the payment (he already refunded $20 for the return shipment...) for the trouble I have to get at least something working.
And yes they where listed as working.

Moral of the story: don't buy anything from waikeionline_*



edit: Picture attached.

[Bryan]

I think the seller has probably reached the bottom of the barrel (literally) of these units and selling off what he can. But you are right, best to stay away.

[Bryan]

Does anyone know what these NMEA like commands are outputting

$PFEC,GPtps,160530053705,3,1,2,150701000000,00,17,160530015837,1899,106642
$PFEC,GPrrm,0,0,00,00,00,00,+000,+02

No doubt some custom Furuno/Symmetricom commands, but I can't seem to find anything out there that describes what they mean. Thinking the GPTPS may just by the term"tps" may be describing the 1pps output??

[NivagSwerdna]

The NMEA commands are described in the Furuno protocol spec.... it's linked somewhere earlier in this thread.

(I use $PFEC,GPtps for my NTPd interface see https://gitlab.com/NivagSwerdna/SymmetricomUCCMDisplay/blob/master/sketches/firsttft/nmea.cpp I create a synthetic $GPRMC from a $PFEC,GPtps)

http://www.ko4bb.com/getsimple/index.php?id=download&file=02_GPS_Timing/Furuno/Furuno_GT8031_Protocol_Specification_Rev1.pdf

[Bryan]

The NMEA commands are described in the Furuno protocol spec.... it's linked somewhere earlier in this thread.

(I use $PFEC,GPtps for my NTPd interface see https://gitlab.com/NivagSwerdna/SymmetricomUCCMDisplay/blob/master/sketches/firsttft/nmea.cpp. I create a synthetic $GPRMC from a $PFEC,GPtps)

http://www.ko4bb.com/getsimple/index.php?id=download&file=02_GPS_Timing/Furuno/Furuno_GT8031_Protocol_Specification_Rev1.pdf

Excellent, thank you very much.

p.s. The period at the end of first link needs to be removed, otherwise link is broken.

[kutte]

Location of L20.

yes, it looks like a transformer, but it is not. I just put in a replacement inductor of 2.2uH onto the tabs next to R105 and this worked. Actual value does not seem to be critical.
Kutte

[Confucian]

Hi All
Great thread, info overload. lol
My unit arrived today with oil and damage.
See below .  With values and actions suggested in this thread.
I have stolen shempe's picture to show location of damage (hope he dosnt mind)
GPS Ant Socket missing ------------ I will solder coax pigtail to Board.
RST & MACT switches both Broken ----------- Double Pole Momentary action -------- move to front panel of enclosure
C23 & 35 missing -------------- 100nf
C4 missing and pads ripped off ---------- 220uf 25v --------- need to scrape away green layer to discover tracks
U24 Missing   --------- Has anyone any Idea what U24 is and can you spot any deliberate mistakes above :-)
Thanks

[PA0PBZ]

U24 Missing   --------- Has anyone any Idea what U24 is

Z04, must be 1/6^th of a ls04, inverter.

[Macbeth]

Location of L20.

yes, it looks like a transformer, but it is not. I just put in a replacement inductor of 2.2uH onto the tabs next to R105 and this worked. Actual value does not seem to be critical.
Kutte

I think its a common mode choke with the winding with the red dot going across unpopulated R105 measuring 5.4uH and the lower winding isolated from the above. The lower winding on mine appears to be a dead short. In circuit measurements using a DER EE-5000. I suspect the short on the lower winding is deliberate and elsewhere on the PCB.

rastro, check your unpopulated L20 pads and if the lower measures a short then surely you could get away with merely turning the bad L20 upside down and soldering it back on

ETA: Just checked your previous post with the 3 turn winding. That is so tiny it would show up short. Hmm...

[Bryan]

(I use $PFEC,GPtps for my NTPd interface see https://gitlab.com/NivagSwerdna/SymmetricomUCCMDisplay/blob/master/sketches/firsttft/nmea.cpp I create a synthetic $GPRMC from a $PFEC,GPtps)

MacBeth
Had a look at your project and it looks quite impressive, are you able to share any screenshots of the TFT display with the project running.

[rastro]

Kutte & Macbeth,

Good information I'll probably use a 2-5uH inductor.  I don't think the original L20 could stand a round of soldering - plus the 2 pads are broken.

Thanks for the feedback.

-rastro

[NivagSwerdna]

are you able to share any screenshots of the TFT display with the project running.

There are some pictures earlier in the thread.. The display is really sweet... touch screen menus...
https://www.eevblog.com/forum/projects/a-look-at-my-symmetricom-gpsdo-(ocxo-furuno-receiver)/msg911644/#msg911644
https://www.eevblog.com/forum/projects/a-look-at-my-symmetricom-gpsdo-(ocxo-furuno-receiver)/msg934318/#msg934318
I'm just finishing off my 1:4 splitter (took me ages to get the LT6551 bias correct).
https://www.eevblog.com/forum/projects/a-look-at-my-symmetricom-gpsdo-(ocxo-furuno-receiver)/msg947866/#msg947866
and may use the pulse stretcher that is also on that board to run the NTPd PPS rather than currently taking it off the Furuno
and then a bit of work on the software to add the ability to also interact via the USB at the same time (would be really good if we could find some pins for a 2nd comm port on the UCCM-P)...

[Confucian]

Has anyone any Idea what C4 is connected to ? I have pads missing, have scraped away at the green layer but can only find a broken track to one end of C6.

[NivagSwerdna]

See photo... one end attaches to a small nearby via and disappears downwards, the other end heads off to C6

[Confucian]

See photo... one end attaches to a small nearby via and disappears downwards, the other end heads off to C6

Thank you for the Pics, looks like I may have to create my own "via" and connect to the other side of that one as not a lot of it left.

[NivagSwerdna]

Thank you for the Pics, looks like I may have to create my own "via" and connect to the other side of that one as not a lot of it left.

Once you find the other end you might find its just GND and that is available close by on the top side.  The daughter board is just a boost converter from 5V to 12V so C4 probably is just a smoothing capacitor on the output?

[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

I finally received my Oscilloquartz 8663-XS yesterday. I measured the Vref and found it to be about 7.6v. So as an experiment, I applied 7.6v via a 3.3k resistor to the Vref pad on the Symmetricom board to see if it might get clamped to around 5v. It was about 7.6v so I figured there was no problem with the higher voltage input to the Vref pad. Next, I hooked up the OCXO via wires to the pads. I powered it up and noted that the Vtune voltage was about 4.7v and that it would go past 6.5v when hunting. It was about 2.7v in holdover mode with the original 5.1v Vref. The Fc tuning voltage for this OCXO is about 4.78v. I left it in circuit for a while and it appears to work perfectly. It locked up just fine. My thought is, the Vref output from the OCXO feeds a DAC and the DAC just scales the Vref value to tune the OCXO. Therefore even though the Vref voltage and tuning range are higher and broader than the original 0-5v, it doesn't matter since it is scaled by the DAC. the Kvco is about the same as the original. So it looks like this OCXO will work as a replacement, at least for my board. I got the OCXO from "flyingbest" on Ebay. I previously bought two Morion mv89a's from him and both of them worked good as well. I think he does a pretty good job of testing things he sells.

[rastro]

Location of L20.

yes, it looks like a transformer, but it is not. I just put in a replacement inductor of 2.2uH onto the tabs next to R105 and this worked. Actual value does not seem to be critical.
Kutte

Fixed problem with damaged/lifted "L20".  I ended up soldering a 3.3uH surface mount inductor in the un-populated  R105 pads.  It is electrically the same as the L20 pads closest to the OSC/oven.  This is because the pads that go to L20 where lifted/marginal and the R105 is a better gap/distance for an SMD ferrite inductor that I used.  I ended up leaving the other 2 pads of L20 (closes to the edge of the PCB) un-populated since it essentially reads a short.

Picture attached prior to rework.

Thanks again for the help/advice.

-rastro

Anyway it's doing it's survey 20% complete.  So preliminary results are good.

[NivagSwerdna]

would be really good if we could find some pins for a 2nd comm port on the UCCM-P...

The more I think about this the more I think a 2nd output port doesn't exist.  The H8S/2317 has two serial interfaces (USART/UART/SCI whatever you want to call them).  It makes sense for one to talk to the GPS module and the other to talk to the outside world.  Happy for someone to prove me wrong.
The SCI are on Pin 8 TxD0, Pin 9 TxD1, Pin 10 RxD0 and Pin 11 RxD1 (counting pins anti-clockwise from the corner with the notch).

[Macbeth]

There is a second serial port on the Trimble version of these UCCMs, which I gather spits out a TOD (time of day) message in binary format every 2 seconds. This same feature is available in ASCII hex format on the normal serial channel by issuing the TOD ON command and TOD OFF to disable it.

I actually think this may be a simple way of getting NTP without tapping the 1PPS and the GPS module GPRS.

It bursts forward with its ASCII-HEX stream no matter what you are doing, like in the middle of a SYST:STAT? output, if the 2PS pulse hits it will just corrupt what you would normally expect with a sudden burst of data.

Thankfully the data packet appears consistent with a start byte, end byte plus white space/carriage return and fixed length within. It appears to have a CRC that I haven't worked out yet. However the GPS time and what I think is a UTC seconds offset byte (17) is clearly available in the data.

I've only been playing with RegExp and serial port interrupts to catch these quick and filter them out from normal serial stream.

[ekyle]

Here are a couple images of the box I came up with. I may add a distribution amp eventually.

[Bryan]

Here are a couple images of the box I came up with. I may add a distribution amp eventually.

Think I would keep the switching power supply as far away as possible from the GPSDO.

[ekyle]

Here are a couple images of the box I came up with. I may add a distribution amp eventually.

Think I would keep the switching power supply as far away as possible from the GPSDO.

I didn't notice any noise pick up from the switcher and no degradation in satellite reception. I did add a common mode choke and caps on the 5v output which greatly reduced the switching ripple, etc. The power supply frame and chassis ground are single point grounded to the box. The switcher is a high quality Lambda supply and not some cheap made in China model and is fully enclosed. I looked at the 10MHz output with my HP8568A spectrum analyzer and there are no switching spurs or power supply harmonics visible, all the way down to a frequency span of 100Hz. So I don't think the power supply is a problem. On another note, I have a Racal Dana 1992 frequency counter. I measured the reference output and it has some hideous power supply side bands. It uses a 5MHz oven and has an x2 frequency multiplier board attached to the outside of the oven can. The x2 doubler uses two tunable inductors and these are coupling 60Hz harmonics from the power transformer and modulating the reference signal. It all depends...

[Bryan]

Well, looks like you have certainly done your homework. Makes a big difference on the quality of the switching PS, yup have been there with the cheap Chinese ones.