Prototype HP/Agilent GPS based refererence clocks generator

[TheSteve]

More late night ebay trolling lead to this purchase. This is a "Ericsson GE Precision Reference Clocks" - HP part # 58XXXA. The unit produces 10 MHz, 19.2 MHz, 9600 kHz and 100/300/600 Hz. It was most likely made for cellular base station testing based on its outputs. It looks to have been made 95/96 just as the Ericsson/GE partnership came to an end. The part number on the front makes it pretty clear it was a prototype. HP has several GPS based time/frequency references with part numbers with the format of 58xxxA. The serial number is "EGE001P", but the most convincing part are the two hand built boards inside. The front and rear labels are also simple printed labels. I believe the true part number is 58501A.

Inside is has a main board with part # 58501-60001, a 10 MHz OCXO part # 05071-60219, a Motorola GPS part # 84D43215M02, a hand built board which generates (likely phase locked) 19.2 MHz from the 10 MHz, a hand built board to hold the three front mounted LED's and a third party power supply.

At power up the front panel power LED comes on and a second later the GPS and ALARM LED's toggle on/off in sequence. The main board LED's also all toggle and then one of them flashes once per second. I can communicate with the unit via SCPI using 9600/8/N/1 RS232 data.  I was able to verify the GPS does receive a signal as the time is correct and the date updates but is 1024 weeks behind(GPS rollover bug). However the "GPS LOCK" and "ALARM" LED's never do anything. The 10 MHz out the back does travel between 9,999,997 and 10,000,003 Hz which follows the EFC test point voltage that goes between -5 and +5 volts. I have found very few SCPI commands that work. At this point they are:

*IDN? which gives: HEWLETT-PACKARD,58501A,0,Fiji_EGE
*CLS
*TST? which gives 1 (possibly indicating a rom error)
PTIME:DATE?
PTIME:TIME?
SYSTEM:ERROR?

I have pulled the four firmware roms and reseated them as well as the Xilinx chip, it made no difference. So the question is, did it ever work or has something failed?

I find it to be a fascinating piece of equipment and think the hand built boards are super cool to see, but lets face it, it would be more fun if it worked.
So I welcome all comments/thoughts on things to test or try.

[TheSteve]

More pics

[TheSteve]

Here is the firmware if anyone feels like taking a look.

[uncle_bob]

Hi

Given the construction techniques, who knows if you'll ever get it working.

Bob

[Vgkid]

That certainly looks like a prototype. Lets see how it performs.

[TheSteve]

If nothing else it was cheap. Will need to take a good look at the firmware to see if I can find more commands.

[uncle_bob]

If nothing else it was cheap. Will need to take a good look at the firmware to see if I can find more commands.

Hi

I'd try the normal SCPI HP commands. There are a number of lists running around to provide the "usual suspects".

Bob

[Vgkid]

Atleast the ocxo is of the 10811B variety, so it is worth something.

[FFFF00]

The main board looks very similar to the 58503a just with many of the components shuffled around, try checking the programming manual for commands.

Try SYSTEM:STATUS?, it might print out a formatted status page.

[TheSteve]

I have tried all of the commands I can find in the various manuals for the HP GPSDO products with pretty much no success. I have a 58540A here so I am pretty familiar with the normal SCPI commands. I may even try the firmware from one of the other units if I can find it posted. I should also look for a program that will combine the MSB and LSB firmware files so it is easier to look through for ASCII readable commands.

[FFFF00]

Ah, that is unfortunate.

I can help you with the last bit, I wrote a program to accomplish the same thing for the 58503a. I've attached the combined firmware files you uploaded previously.

[DimitriP]

attached the combined firmware files

 
Go o
odJb

[FFFF00]

attached the combined firmware files

 
Go o
odJb

I may be missing something (it is late), but the contents looked reasonable... Inspecting the bin files with a hex editor shows properly aligned data where I could see anything human readable.

[DimitriP]

I saw "ntued Saestig" in one file and "Isrmn tt etns"  in onother
and when I searched for "Instrument State Settings" in your combined file, it was there in all it's glory !




 

[TheSteve]

Ah, that is unfortunate.

I can help you with the last bit, I wrote a program to accomplish the same thing for the 58503a. I've attached the combined firmware files you uploaded previously.

Very nice, thank you.

I'll have a tour through, maybe myself or someone else will get lucky and find something useful.

edit - I see lots of good stuff, perhaps even a status screen, but how to get any of it displayed remains a mystery.

[DimitriP]

All there were bunched up together:

WAI  TST  STB  SRE  RST  OPC  IDN  ESR  ESE  CLS  DOUBLE FLOAT INTEGER BOOLEAN SAMPLE LANGUAGE ERROR SYSTEM INITIATE RECOVERY
DURATION AUTO  HOLD OVER ROSCILLATOR TIME  SURVEY
 STATE
LOCATION HOLD  POSITION EMAN GLE ADEL AY GPS YSTEM DATE  PTIME DELAY FIXED FREQUENCY SOURCE
W  E  S  N 
MINIMUM MAXIMUM

STATE looks like a good one to try

[TheSteve]

All there were bunched up together:

WAI  TST  STB  SRE  RST  OPC  IDN  ESR  ESE  CLS  DOUBLE FLOAT INTEGER BOOLEAN SAMPLE LANGUAGE ERROR SYSTEM INITIATE RECOVERY
DURATION AUTO  HOLD OVER ROSCILLATOR TIME  SURVEY
 STATE
LOCATION HOLD  POSITION EMAN GLE ADEL AY GPS YSTEM DATE  PTIME DELAY FIXED FREQUENCY SOURCE
W  E  S  N 
MINIMUM MAXIMUM

STATE looks like a good one to try

I agree, but trying it only gives me the standard error 113.

But it does look like it recognizes WAI  TST  STB  SRE  RST  OPC  IDN  ESR  ESE  CLS as you can see here:

scpi > *WAI
scpi > *TST?
1
scpi > *STB?
+0
scpi > *SRE?
+128
scpi > *RST
scpi > *OPC?
1
scpi > *IDN?
HEWLETT-PACKARD,58501A,0,Fiji_EGE
scpi > *ESR?
+0
scpi > *ESE?
+0
scpi > *CLS


While the unit boots I think the root of the unit not functioning is related to the *TST giving a result of 1. On pretty much all HP gear it should return a 0 if everything has passed.

[uncle_bob]

Hi

The Forth in the ROM's looks pretty standard for HP. I suspect that when operating properly it *will* respond to all the usual SCPI commands. The question now becomes: What's broke and how do you fix it?

Bob

[scopeman]

Steve,

Any chance another FPGA (or something else) should be in that socket that is unpopulated or can you see if perhaps someone pulled it for proprietary reasons.

 Just a thought.

Sam

Worst case you have a great case OCXO and a power supply and a basket of chips!

[uncle_bob]

Hi

I suppose there is another possibility here:

What we are looking at is an attempt to turn the unit into something else. It's a bit odd that they have a nicely done stencil printed back panel and point to point boards inside. A normal design process usually does it the other way around. The connector locations are "firm" once the board is debugged. You can get the print for the panel in near zero time, you do it at the last minute.

Indeed the giant empty FPGA sized socket does look a bit alarming.

Bob

[FFFF00]

I would guess the empty socket is fine. Going back to my previous post that the main board is very similar to the 58503a, the extra socket is used for the 59551a "Measurements Synchronization Module" configuration. You can also see unpopulated (or just direct solder) attachment points to the left. Picture attached for reference from a working unit.

Can you verify there is a PPS output from the GPS module?

[TheSteve]

I agree, I don't think the empty socket is an issue. The board also does generate all of the requires signals out the back.
I can see the serial packets and the 1 PPS out from the GPS receiver.

[TheSteve]

Just a little update. I am communicating with the GPS directly now and it looks to be in position hold mode and the last location stored in it is from Maryland, USA. I will attempt to have it perform a self survey.

[uncle_bob]

Just a little update. I am communicating with the GPS directly now and it looks to be in position hold mode and the last location stored in it is from Maryland, USA. I will attempt to have it perform a self survey.

Hi

Record the location before you erase it. You never know when you might want to call up the previous owner to have a chat.

Until you blow that location away, you will have very little luck with the gizmo. It *is* odd that it would prevent the status screen from coming up... :SYST:STAT? has always worked for me on the HP boxes.

Bob

[TheSteve]

OK, we have lock!

I manually set a reference location and the date to get around the 1024 week bug this receiver has. I can see the EFS voltage is nice and stable at -56 millivolts and it is putting out 10,000,000,00X,X MHz. Obviously how much phase noise there is remains to be seen. The power supply packed it in earlier this afternoon so it is running on three external supplies.
I expect the unit will not work properly if power cycled as it not store the week correctly and who knows if the GPS has the proper location stored or if the main unit has a stored location in its eeprom. It is nice to see it really does work though so it worth further investigation and likely the purchase of an updated Motorola GPS. This one is so old it communicates via RS232 to the unit, not even TTL, the sensitivity is also beyond terrible.

I have still not been able to find any other SCPI commands. I figure there must be more, and there should be a command to change the 600 Hz output to 300 or 100 Hz(base on the back panel labeling).

EDIT:

Removed power for a few minutes and it seems to still have the proper date some how, yet the location is back to 37.3586547 -79.1746364 degrees. I do see packets from the 58501A to the GPS receiver so I assume it is sending configuration data. Now to determine if it sending an initial location or is it saved inside the GPS.

EDIT 2:

The GPS receiver is storing the correct location but the 58501A is programming a fixed location 10 seconds after powering on into the GPS and messing it up, it also programs the GPS to position hold mode so it will never get a proper fix.