HP 34970A Ser.Nr. US37007297 repair help needed -FIXED -

[Shiv]

Hi,
replaced the new 24C04-S in the DMM but the errors are always the same.  The 24C04-S is supplied. I need to initialize the unit or something is wrong anyway ?

https://community.keysight.com/servlet/JiveServlet/download/63858-1-5176/Error%20740%20Series%20for%20fix%2034970A.doc

Here are the commands you should send to 34970A

You can do it with vee, labview or another software like NI-MAX or MetCal

[Spectrum]

Thank guys, I have an Agilent 82357B USB and I installed the software from owner CD. It is the first time I do that, so I need help if I get some issue.
The VEE I read from the link is the same sw I can find in the Agilent's CD ?
This afternoon or next I try to connect my PC to the unit by 82357B and tell you news

[jogri]

Thank guys, I have an Agilent 82357B USB and I installed the software from owner CD.

Let me guess: Brandnew one for 80 bucks from ebay? It's a fake, i recommend that you download the software directly from Agilent/Keysight (the software from the CD is probably an older version, but that shouldn't really matter). (Those fake GPIB adapters work 95% of the time, so don't worry there, just keep in mind that it might have some weird quirks)

But yeah, install a software of your choice (Agilent VEE, NI-MAX, TekVISA, etc) , connect the 34970 to the PC and start it (look at the display, it will show you its GPIB adress when it starts) and then search for instruments. The 34970 should be there, you can make sure that it communicates by querying the string "*IDN?" (without the "") (there should be a Query, Write and Read button, use the "Query"). That command works with all GPIB devices and gives you the name/ID of it back. In your case, it should return "Agilent, 34970A".

If that's the case, it's time to send the two commands: First, write "DIAG:POKE 15,0,0" and hit "Write", then write "DIAG:POKE 0,0,0" and hit "write". After that, restart the 34970 and follow the rest of the procedure from the document.

Protip: Remove the cards of the 34970 before you do this.

[Spectrum]

Hi,
I don't know where my 82357B come from. After read your message, I got dismantle it and inside the pcb is Agilent marked (serigrafy) also the IC are marked Agilent. If it is a fake, is really well done.

Is this a fake ? https://www.ebay.it/itm/New-In-Box-HP-Agilent-82357B-USB-GPIB-Interface-High-Speed-USB-2-0/264258218741?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2060353.m1438.l2649
Mine is not Keysight marked, it is older...USB 2.0 but Agilent marked. It stood still for a long time

Before dinner, I quickly tried to connect 34970. No result, no instrument connected, but it was set for RS-232....when I realized it was late and I went to dinner

[coromonadalix]

For an original usb gpib,  you're supposed to have copper paint on the shells and have some can be been seen near the two screws when they enter the shell,  thats a good start to know if its genuine

And i think you can do a serial number search on/at Keysight to check/validate it ??

[Spectrum]

No copper paint

[Spectrum]

I opened Keysight VEE Pro, click Instrument Manager, then Find and it appears in the window GPIB0 and bottom newInstrument(hp34970a@GPIB0::9::INSTR).
Now I can do anything else, I do not have the windows I can see in the link "Error 740 series for fix 34970A".
If I click something else, it is appears a blue message in the bottom "An object cannot be placed in the workspace (press Esc to cancel)"

In the unit nothing change, you can see Insert Card and red light Error.

[Shiv]

You can also do it from connection expert:

Select board, scan for instruments, select yours and interactive I/O.

[Spectrum]

Thanks Shiv and yes I did it by Keysight Connection Expert 2020 and Keysight Interactive IO. For first I unsecured the unit by front panel, then send and read this commands:

-> *IDN?
<- HEWLETT-PACKARD,34970A,0,6-1-2
-> DIAG:POKE 15,0,0
 ! VI_ERROR_TMO: A timeout occurred
Visa ErrorCode: 0xBFFF0015 (-1073807339)
<-
-> DIAG:POKE 0,0,0
 ! VI_ERROR_TMO: A timeout occurred
Visa ErrorCode: 0xBFFF0015 (-1073807339)
<-
-> SYST:ERR?
<- +745,"Cal data lost: FRES corrections"

In the display I can see the ADRS and RMT leds light on and of course INSERT CARD and red light ERROR, nothing else...

24C04-S or interface 82357B do not working ?

[Shiv]

Thanks Shiv and yes I did it by Keysight Connection Expert 2020 and Keysight Interactive IO. For first I unsecured the unit by front panel, then send and read this commands:

-> *IDN?
<- HEWLETT-PACKARD,34970A,0,6-1-2
-> DIAG:POKE 15,0,0
 ! VI_ERROR_TMO: A timeout occurred
Visa ErrorCode: 0xBFFF0015 (-1073807339)
<-
-> DIAG:POKE 0,0,0
 ! VI_ERROR_TMO: A timeout occurred
Visa ErrorCode: 0xBFFF0015 (-1073807339)
<-
-> SYST:ERR?
<- +745,"Cal data lost: FRES corrections"

In the display I can see the ADRS and RMT leds light on and of course INSERT CARD and red light ERROR, nothing else...

24C04-S or interface 82357B do not working ?

Did you send the POKE commands as "write" or as "query" ? you must send it as write
They shouldn´t reply, so the timeouts are because the PC keeps waiting for an answer. (Query)

The 82357B is working, it writes and reads (*IDN? and SYST:ERR? are writing and reading ok)


Follow the entire procedure step by step:
https://community.keysight.com/servlet/JiveServlet/download/63858-1-5176/Error%20740%20Series%20for%20fix%2034970A.doc

After sending PUKE commands, the memory is initialized, so the CAL DATA LOST should be normal.


You need to check if calibration procedurement is working and the memory can be written/readed by the 34970A.
To do this, after puke, go to the service manual and make the first steps of adjustment procedure.

It´s just Zero adjustment, you only need to short terminals.

Once it´s done, the cal data lost should dissappear and is time to send the 34970 for adjustment with a BIG NOTE saying all cal factors are lost, so tecnician will go directly to adjustments and don´t think is a broken unit.

[Spectrum]

Hi, I have no Write or Query but Send Command, Read Response, Send and Read.
Anyway after some attempts the errors codes 700 disappeared, well done. Thank you  I also tested this old 82357B

Now, during the extended test, error 621 remained, plus error 620 and 622  I think some OP-AMP has gone...

[Shiv]

Query = Send and Read.
Write = Send

620 = AC rms zero failed This test configures to the 100 mV ac range
with the ac input grounded through A4K103. The internal residual
noise of the ac section is measured and checked against a limit of
–10 mV to 70 mV at the output of the rms-to-dc converter.

621 = AC rms full scale failed This test configures for the 100 mV ac
range. The 1 mA ohms current source is switched on the charge the
ac input capacitor A4C301. This produces a pulse on the output of the
rms-to-dc converter which is sampled 100 ms after the current is
applied. A 20 ms A/D measurement is performed and checked against
a limit of 10 V ± 8.5 V into the ADC

622  = Frequency counter failed This test configures for the 100 mV ac
range. This test immediately follows test 621. With A4C301 holding
charge from test 621 the ac input is now switched to ground through
A4K103. This produces a positive pulse on the input to the frequency
comparator A4U310. While C301 discharges, the ENAB FREQ bit is
toggled four times to produce a frequency input to the counter logic in
A1U205. A failure occurs if the counter can not measure the
frequency input.

Did you do the zero adjustment?
If not, try to do it.

If yes, I would try to check A4 k103 relay for bad contacts. (fisical switch is more likely to fail by age.)

After that, follow the signal through U301,U304,U305, U306, U312,U307 and U308

[Spectrum]

Hi,
Zero Adj done, no results.

All three relays (K101,102,103) seem to me to be faulty with alternate functioning.
I have exactly the same but with 12V coil, not 3V. I have nothing at 3V

However they seem to me of poor quality (even if OMRON) considering the cost of the instrument.

Now I am going to looking for.. OMRON G6SU-2 3VDC. Probably is better to purchase a supply for future repairs.

[coromonadalix]

i dont think youre doing good repairs at some point,  Omron Does make good relays,  do proper analysis before thinking this and that,  follow the service manual properly and follow signal paths

The meter got busted in some ways, you have to check many things as you know, dont assume things and do the opposite by changing too many parts for nothing, stay focused

If they where this bad (relays) it would not have been put in it

[Shiv]

Hi,
Zero Adj done, no results.

All three relays (K101,102,103) seem to me to be faulty with alternate functioning.
I have exactly the same but with 12V coil, not 3V. I have nothing at 3V

However they seem to me of poor quality (even if OMRON) considering the cost of the instrument.

Now I am going to looking for.. OMRON G6SU-2 3VDC. Probably is better to purchase a supply for future repairs.

I allways try to fix relays before replacing.

Little more tension than rated on square signal at coil.

1 kHz and let it go for a while.

[Spectrum]

I will try, thanks

[Madmanguruman]

Can confirm that the 3V lithium battery doesn't affect the calibration data of the 34970A - it's only used to back up scanned data / channel configs. I've replaced at least 5 of these so far.

I've also noticed that on a few DAQs, self-test errors related to analog parameters self-corrected as a function of this disassembly/reassembly - perhaps the contacts between the boards would become a bit corroded (some of these DAQs were 20+ years old) and just removing/reinserting them would clean the pins enough to make them 'work'.

(It's a little sad that these units are being discontinued in June. The new DAQ970A is decent enough, but hey. Nostalgia.)

[Spectrum]

Ja, battery doesn't affect, you must replace the FRAM 24C04-S (U450). Unfortunately no oxide in the contacts.

Come back to my unit, relays regenerated and replaced for safety with those of card 34901. The three errors are still there.
The DMM doesn't measure AC Volts.

I followed the signal through C301, R301, R302, U301, U304, U305 ... but I am not able to evaluate if the signals I see with the oscilloscope are the correct ones.

At the input there are 15Vac 50 Hz (DMM in autorange). In U301A pin2 there is a half wave, which becomes smaller at pin6. Through U304C, I find this signal on U305A pin3, in pin 2 I find a square wave with some spike during the flat, after that I didn't go ...

Could the fact that there are 3 errors, all the possible ones for the AC, help to identify the fault?

thanks for your patience

[Spectrum]

in U305 (perhaps also in U312) the -15V is not present (-15B). Probably the zener CR304 and the C314 are knocked out. I have no a BZX84C3V3 (SOT23) thus for a test I will mount a traditional 3.3 zener.

[Spectrum]

Yeesss, FIXED.
Temporarily mounted a BZX55C3V3 (old style, not smd). Extended self test results PASS ... The AC reading works in all ranges but is seriously out of calibration (14.5 Vac instead of 15.2). As soon as I have the BZX84 available, I will replace it and proceed with calibration. The power supply voltage of the U305 / U312 is + 15.2 pin7 and -14.60 pin4 (a little low). We will think about it when I will have the final components, a couple of weeks.

Thanks everyone for the help

[Shiv]

Yeesss, FIXED.
Temporarily mounted a BZX55C3V3 (old style, not smd). Extended self test results PASS ... The AC reading works in all ranges but is seriously out of calibration (14.5 Vac instead of 15.2). As soon as I have the BZX84 available, I will replace it and proceed with calibration. The power supply voltage of the U305 / U312 is + 15.2 pin7 and -14.60 pin4 (a little low). We will think about it when I will have the final components, a couple of weeks.

Thanks everyone for the help

Good to hear that!

I feel like it was mine!

[Spectrum]

Ja thanks Shiv.

I saw earlier that a 34401A also had a problem with these Z-diodes and related C. Obviously it is one of their weakness.

In summary, this unit had 4 issues:

- tired back up battery 3V
- inefficient relays
- FRAM 24C04-S (U450) of calibration data KO
- power supply OP-AMP (U305, U312 etc.) due to Z-diode CR304 (BZX84C3V3) KO

Not bad, all in one shot

[coromonadalix]

glaid it work now 

[Madmanguruman]

I would also consider replacing the fan for such an old DAQ. They're easily available from Digi-key or other vendors.

https://www.digikey.com/en/products/detail/delta-electronics/AFB0405LA-A/5799904

You'll need to crimp on two Molex header inserts (or cut/solder onto the existing fan header).

[Spectrum]

Ja why not, the fan that's there now is noisy