HP 58503A EFC error [FIXED]

[Samogon]

Hi,
I have GPS standard HP 58503A which in past had broken power supply i have fixed it and unit started to work, but fails to lock, it can track birds and even syncs time, but reports health error EFC err (see print of system:status?)
Also diagnostic shows no errors.
It had problems with +15V being out of tolerance ~19V.
I wonder if it could affect some componets that now it has this EFC error.
Can someone point me where to dig with this error?
Thanks
------------------------------- Receiver Status -------------------------------
SYNCHRONIZATION ........................................... [ Outputs Invalid ]
SmartClock Mode ___________________________   Reference Outputs _______________
   Locked                                     TFOM     9             FFOM     3
   Recovery                                   1PPS TI  --
   Holdover                                   HOLD THR 1.000 us
>> Power-up: coarse freq adj [TI +162.0 ns]   Holdover Uncertainty ____________
                                              Predict  --

ACQUISITION ................................................ [ GPS 1PPS Valid ]
Tracking: 7 ____   Not Tracking: 3 ________   Time ____________________________
PRN  El  Az   SS   PRN  El  Az                LOCAL    08:17:16 [?] 18 Sep 1997
 10  26 158   30     1  12 273                GPS 1PPS Synchronized to UTC
 14  56  39   99     3  17 319                ANT DLY  0 ns
 16  26 191   20   *25  20  41                Position ________________________
 22  42 304  114                              MODE     Hold
 26  55 176   38
 31  59 358  129                              LAT      N  bla-bla-bla
 32  46  76  135                              LON      W  bla-bla-bla
                                              HGT               +20.50 m  (MSL)
ELEV MASK 10 deg   *attempting to track
HEALTH MONITOR ...................................................... [ ERROR ]
Self Test: OK    Int Pwr: OK   Oven Pwr: OK   OCXO: OK  EFC: Err  GPS Rcv: OK
E-113>  DIAG:LOG:READ:ALL?
Log status: 3   entries

Log 001:20001016.16:16:27:  Log cleared
Log 002:20001016.16:22:00:  GPS reference valid at 19970918.03:27:55
Log 003:20001016.20:37:02:  Position hold mode started

[texaspyro]

You might want to try running Lady Heather (download from ke5fx.com).  But before Lady Heather will work with it, you need to send it the command to turn off echo on the serial port (the command that Lady Heather sends to turn off echo does not work with the 58503A... fixed in the next release).

You also need to start Heather with the /rx5 command line option to specify that the receiver is a HP5xxxx.  The other command line options that may be needed is to specify the serial port and maybe the baud rate parameters.

You want to look at the DAC plot and see if it is pegged around +100% or -100% or if any other errors show up.

I just got in a 58503A... power supply is bad.  The power supply was good when I tested the unit at an equipment auction preview inspection.  I'm replacing it with a $30 Meanwell RQ-15C.   Just last month the same series PowerOne supply (+/- 24V version) died in my UCT-2008 rubidium oscillator.  You can buy new ones from Mouser/Digikey... only $350+  Luckily Ebay had one for $25... no such luck with the 58503A supply... cheapest one was around $200 on Ebay.

[Samogon]

Thanks, will try today.
If you want i can try to fix your power supply.

[texaspyro]

Thanks, will try today.
If you want i can try to fix your power supply.

Thank you for the offer,  but these supplies are around 20 years old and all of them are dying.   Best to replace them with something new.

The bad supply had a toasted etch and at least one shorted transistor.  I replaced the transistor, but no joy.   No obvious failed caps were found.

[Samogon]

Replace all(!) caps even if they look OK. Once you touch it with iron tip you will smell it .

[texaspyro]

Replace all(!) caps even if they look OK. Once you touch it with iron tip you will smell it .

I went over them with an in-circuit cap checker and nothing obvious popped up.  Way too much work to desolder and check/replace them...  the Meanwell supplies can be had for $20-$30.

[texaspyro]

I got the Meanwell supply in today.  It was trivial to install.  The screw terminals on the Meanwell are in the same order as the PowerOne (with the AC input on the same terminal strip instead of a separate set of terminals).

The screw terminals are spaced .312", so I took two .156" spaced pin strips and soldered them at right angles.  Then I loosened the screws on the Meanwell terminal strip and slid the .156" pins into the terminal strip with each pair of pins straddling the screw and tightened the screws.  The result was a row of .156" spaced pins sticking up just like with the PowerOne supply.

To mount the supply I will need to extend the green wire ground from the chassis to give more freedom in choosing a mounting point and do a little surgery on the PowerOne mounting brackets.

I connected the 58503A power cables to the pins, plugged it in, and it powered up just fine.   Status is all good.  It's trying to acquire sats right now...  these things can take several hours to start tracking sats (even with time and position hints).

It has around 85,000 hours on it.  The VFD display is a little dim and blotchy.

[Samogon]

Good job!
Mine catches birds almost instantly. Then foes few hours survey.
Looks like i have cold oven, it should be warm on touch and it should draw ~500mA at start, it does not. Will
check it on weekend

[texaspyro]

Looks like i have cold oven, it should be warm on touch and it should draw ~500mA at start, it does not. Will
check it on weekend

The oven is powered off of 24V derived from the +/- 15V supply.  On the motherboard near the right side of the front panel board are two 12V regulators (?) that I think create the oven power.   On my motherboard by those regulators there was a ceramic (not tantalum) cap that was glowing orange hot.  Maybe that killed the original supply.

You can unplug the oscillator cable from the motherboard and check for the voltages.

There is a Ebay auction for one of the oscillators ($100).  The listing shows the cable pinouts:

http://www.ebay.com/itm/Functional-HP10811-60159-Crystal-OSCILLATOR-10MHz-/332194400295?hash=item4d58544827:g:q1IAAOSwkShY~rfx

Mine started tracking 1-2 sats and got an almanac and is showing sat positions, but is not yet tracking all the sats.   It could be that my power supply is pretty much sitting on top of the GPS with a piece of cardboard insulating it.   My Z3801A took around 6 hours to lock in when I first got it.   It looks like the 58503B (it's a B, not A version) was last used in 2013.

[Samogon]

Got it fixed. +12V regulator was dead on the main board, as you predicted

[texaspyro]

I replaced the GPS receiver in my unit and it is now locked and tracking sats.  Everything now seems to be working properly.  The Motorola receivers in these units are known to have problems with the TCXO when they get old. 

A lot of them also have GPS week rollover issues and they report a date 1024 weeks in the past, but there is a way to compensate for that by manually setting the date.  You have to power down the unit, disconnect the antenna, then issue the command to set the date.  The unit won't accept a date set once the GPS receiver starts tracking sats.   The date adjustment is stored in EEPROM, so you should only need to do this every 1024 weeks.

The DAC EFC control voltage in mine is at around +50% of full scale (the range is -100% to +100%).  In a perfect world you would like it to around 0%,  but after 90,000 hours of aging on the oscillator it drifted a little bit.   Unfortunately the HP-10811 in these units is a hermetically sealed version and there is no frequency adjustment screw to re-center the control voltage.

[Samogon]

1024 weeks means once in lifetime.

[deepskyridge]

A lot of them also have GPS week rollover issues and they report a date 1024 weeks in the past, but there is a way to compensate for that by manually setting the date.  You have to power down the unit, disconnect the antenna, then issue the command to set the date.  The unit won't accept a date set once the GPS receiver starts tracking sats.   The date adjustment is stored in EEPROM, so you should only need to do this every 1024 weeks.

I have a HP 55300A that has the GPS week rollover issue and reports a date 19.7 years in the past. Which command is used to correct this ?
Any info would be helpful.

Thanks
Gary

[Samogon]

Disconnect antenna, reboot and use ptime:date yyyy, mm, dd
Then connect antenna and reboot again. Should be good

[deepskyridge]

Thanks for the info:

I tried both the Date and InitDate command after powerup but I see no change, it still shows 30 Sept 1997 for date.

I did the power cycles and unplugged antenna.

Gary

[Samogon]

oops
 gps:initial:date yyyy, mm, dd

And make sure that you have no errors. on system:status?

[texaspyro]

oops
 gps:initial:date yyyy, mm, dd

And make sure that you have no errors. on system:status?

Before sending the command send a "*CLS" command to clear any error (E-xxxx prompt), then send the date set command.  You should get back a "scpi >" prompt.  If you get back a "E-xxx" prompt the command was not accepted...  usually either a bad command that was not recognized or the receiver was already tracking sats and rejected the command.  You can look up the error code and see why.

[deepskyridge]

OK, got the date set. Unit works like it should. Command docs shows a 2 digit year but will only take a 4 digit year.

Thanks
Gary

[Ordinaryman1971]

I'm having problems with changing the date on my GPSDO. It's Z3805A, firmware if I'm correct is 3543B-A.
gps:initial:date yyyy, mm, dd goes up to 2007... and reports error with any year further.
Any way around it..... ?