Agilent 34970A, Battery replacement and repair

[Jwalling]

I have two of these Agilent 34970A units.
How do I test what FW version is installed?
Can I do it over the UI?

Version of FW is in response to *IDN? command, it is first number ( XX-X-X).
Do you have PLCC 44 package programmer?

Keysight Connection Expert also shows the firmware version from the ident string. My unit returns 12-1-2. Does anyone know what the differences are?

Max

Take a look at service notes - there's a lot of info on the firmware (and other) changes:
http://servicenotes.literature.keysight.com/litapp/SearchSN.do?method=openExternalSNSearch&prodNum=34970A

Would be nice if you could download and burn your own PROM.
Jay

[plesa]

Found following thread about FRAM exchange
https://www.eevblog.com/forum/testgear/agilent-34970a-errors-and-battery-replacement/
And buch of others with usefult information for 34970A restoration
https://www.eevblog.com/forum/testgear/hp-34970a-error-741-fix/
https://www.eevblog.com/forum/testgear/hpagilent-34970a-teardown/
https://www.eevblog.com/forum/testgear/hp-34970a-errors-740-741-747-748/
Thanks Free_Electron
So ordered two pieces  of  Cypress FM24C04B-G and hopefully I will replace also 27C4096 OTP EPROM with firmware and K102 relay.
Latest still seems to be v13.

[plesa]

Scored on 34901A (  20 Channel Multiplexer (2/4-wire) Module)  this is the most universal module for 34970A.
Started with check according to service manual and found that five relays does not work (K107-K111).
According to odometer all relays has about 500k cycles, which is minimal relay life ( electrical ).
Contact resistance on relays is according to spec in manual ( < 2Ohm).
In the service manual for 34970A is schematic for module http://cp.literature.agilent.com/litweb/pdf/34970-90012.pdf on page 245.
Suspected IC was immediately 74HC574 ( U104  and U103)  and U104 it is dead according to IC tester. Logo on IC is Phillips, manufactured in 2003, so Agilent needs to have plenty of them, module has been manufactured in 2010 ( firmware 2.3 )
uC soldered directly to board without socket. Another strange things is that the 20 channels relays are made by Panasonic TX2 and not by Omron G6SU like on previous version of module.

[douardda]

Hi,

I just  bought a sick 34970A (with a 901 and a 907 for 200€ with shipment).

It has a sick display (at least one of the 2 flv drivers is dead, and the display itself is a quite dim). I'm waiting for the replacement parts, meanwhile I'm playing with it via the GPIB interface (I'm trying to implement a sigrok driver for it).
 
And it gives enough 700 errors (including several DMM cal data lost) each time I power it up that I can tell its 2 FRAMs are dead.

It also have a very old firmware revision (8-1-2), so since I am in the process of trying to give it a chance for a second life, I'd like to upgrade its firmware.

Has anyone dumped a recent firmware?  plesa have you found a v13 dump somewhere?

Thanks,

David

   

[HighVoltage]

Keep us updates with your repair and show some pictures.

I have an older version and a newer version.
May be I can get the FW versions and read them out for you.

[douardda]

Hi

so my attempt of repairing my 34970A is not going so well so far.
I've replaced the 2 VFD drivers (putting them in PLCC sockets instead of soldering them directly).


But the VFD is not reacting well to my rejuvenation assaults.


With the VFD being dead, I was thinking about hacking the device by replacing the VFD display with a blue OLED display managed by an arduino or a nucleo that would take place of the 2 shift registers (in the VFD drivers) to decode what to display.

But I'm afraid I killed (don't know exactly how, I did nothing specially wrong when it started to fail) the U1 microcontroller of the display board yesterday evening since it does not give me any more signal on SEGDATA (U1 pin 5).

From there, I see 2 options:

- I find a replacement front-panel unit (they seem to be still available from Keysight for a bit more than 200€)
- I inspect the serial communication to rewrite a implement a new front-panel management controller.
  I've looked a bit at what happens there, and it should not be too difficult.

I don't event want to try to dump the ROM from U1, since being a 87C51, it's most probably "encrypted"
(XORed with a 32bit key). And on my unit, if P1.3 pin is dead, I guess I won't be able to dump the
ROM (unless it still work as an input, sice P1 is used to set the address when dumping the ROM). I may give a try. I don't know yet.

The good news is that I've changed the FRAM chips (both on the main board and the DMM)
and it does not complain any more about these.

I have an older version and a newer version.
May be I can get the FW versions and read them out for you.

That would be awesome! Yes please!


PS: I've written a blog post on this work on my modest blog (https://whatever.sdfa3.org).

[douardda]

For the record, I've asked keysight if they would accept to make the ROM content of the display panel microcontroller available, but they are reluctant to respond favorably to my request:

I’m sorry but there is nothing I can do for you. The only Display part that is available is P/N 34970-66502 which price is $222.00. Keysight will not make available parts that are not set up for Trade Sale. Drivers, Computer Software and Programming Examples can be found under the “Drivers, Firmware and Software Tab at Keysight.com

I don't really understand why they can't make these kind of "firmware" available (besides the time required to find the correct file and
upload it on the support web site). Any Keysight guy is hanging around? ;-)

So now I guess I'm good for a reverse ingeeniring of the serial protocol between the 2 boards...

David

[HighVoltage]

This usually is the problems with larger companies.

As you know, I have a 34970A around and could do a readout for you.
But I am not sure if it is the same version.

[douardda]

This usually is the problems with larger companies.

yup, I did not expect too much, so I was not disappointed :-)

As you know, I have a 34970A around and could do a readout for you.
But I am not sure if it is the same version.

Yes, please! that would be very nice. I'm looking for both ROMs in fact: the display panel's 80c51
one (if it's not encrypted), and a reasonably recent version of main CPU one.

David

 

[julian1]

Could someone tell me if it's normal behavior for the 34970a to perform a full startup when power is applied to the unit (eg. by plugging in the power cord).

Or does it need the soft-start button on the front panel to be pressed.

[Jwalling]

Could someone tell me if it's normal behavior for the 34970a to perform a full startup when power is applied to the unit (eg. by plugging in the power cord).

Or does it need the soft-start button on the front panel to be pressed.

I've never seen one do that. Maybe there is a persistent setting that enables this? If you want to stop it from doing the full test, there's probably a way to reset the unit to factory defaults.

Jay

[julian1]

I've never seen one do that. Maybe there is a persistent setting that enables this? If you want to stop it from doing the full test, there's probably a way to reset the unit to factory defaults.

Jay

Thanks. It's wasn't doing a full-test, but plugging in the power cord, made it start up automatically. Without needing to also press the front-panel button.

I suspect the power-management u104 chip may have been incorrectly issuing a reset due to sensed a voltage drop when power was first applied.

[Jwalling]

I've never seen one do that. Maybe there is a persistent setting that enables this? If you want to stop it from doing the full test, there's probably a way to reset the unit to factory defaults.

Jay

Thanks. It's wasn't doing a full-test, but plugging in the power cord, made it start up automatically. Without needing to also press the front-panel button.

I suspect the power-management u104 chip may have been incorrectly issuing a reset due to sensed a voltage drop when power was first applied.

OK, I misunderstood what you were saying.
They will do that when the battery is replaced. Maybe yours is getting low? They will also do that when AC power is removed then restored if it was powered on.

Jay

[julian1]

They will also do that when AC power is removed then restored if it was powered on.

Good. thanks for clarifying.

[douardda]

Could someone tell me if it's normal behavior for the 34970a to perform a full startup when power is applied to the unit (eg. by plugging in the power cord).

Or does it need the soft-start button on the front panel to be pressed.

Actually, there is an option to configure this behavior in the "Utility" menu (Shift + Advanced).

David

[gamalot]

I have just upgraded my 34970A, using CR2302 (in holder) to replace the original CR14250 battery 

[carl_lab]

Which lifetime do you expect using your CR2032?

Capacity of CR2032 (230mAh) is only 1/4 of CR14250SE (~900mAh)...

http://www.varta-microbattery.com/applications/mb_data/documents/data_sheets/DS6032.pdf
http://www.batteryjunction.com/varta-cr12aa.html

[Jwalling]

Which lifetime do you expect using your CR2032?

Capacity of CR2032 (230mAh) is only 1/4 of CR14250SE (~900mAh)...

http://www.varta-microbattery.com/applications/mb_data/documents/data_sheets/DS6032.pdf
http://www.batteryjunction.com/varta-cr12aa.html

I did some measurements a few years ago, and one could expect at least a year or so.

Then figure:
1. Ease of repair - no de-soldering required and no stressing the same solder pads over and over. Just pop a new one in. If you're careful, you can do it without losing your setups and logging memory with power applied.
2. Cost - about 30 cents (US) versus $5 or more.

Jay

[gamalot]

Which lifetime do you expect using your CR2032?

Capacity of CR2032 (230mAh) is only 1/4 of CR14250SE (~900mAh)...

http://www.varta-microbattery.com/applications/mb_data/documents/data_sheets/DS6032.pdf
http://www.batteryjunction.com/varta-cr12aa.html

I did some measurements a few years ago, and one could expect at least a year or so.

Then figure:
1. Ease of repair - no de-soldering required and no stressing the same solder pads over and over. Just pop a new one in. If you're careful, you can do it without losing your setups and logging memory with power applied.
2. Cost - about 30 cents (US) versus $5 or more.

Jay

 

[HighVoltage]

I was lucky to get a 34972A very cheap, because it was "broken" and I expected another repair.
This is the successor of the 34970A with LXI and LAN and USB connection.

Turns out to be that the display did not turn on, because it was turned OFF
The instrument works and looks like new, even the fan had no dust on it.
This unit was built in 2012 and even the calibration certificate was online at the Keysight website.
So, in other words, NO repair needed.

Now I took the opportunity to look inside and take some pictures.
This new model looks very different in the inside, compared to the older models.
Mostly noticeable, a new battery holder at the bottom of the unit.

[HighVoltage]

Pictures of the inside

[HighVoltage]

More pictures of the bottom

[nfmax]

Oh wow! I was only half serious when I suggested checking if the display was turned off, and that turned out to be the problem after all.

I gather the 34970A is basically a 34401A type DMM + scanner, while the 34972A is a similar scanner but based on a later DVM - the 34410A, I think (never had either of these in my hands to look)

[HighVoltage]

Oh wow! I was only half serious when I suggested checking if the display was turned off, and that turned out to be the problem after all.

I gather the 34970A is basically a 34401A type DMM + scanner, while the 34972A is a similar scanner but based on a later DVM - the 34410A, I think (never had either of these in my hands to look)

Yes, you were right with the display, that was the first I checked.

That is an interesting suggestion you make there, that this 34972A model is based on the 34410A.
I will have a closer look.
It should be easy to find out, since the 34410A was much better in low voltage AC measurements, than the 34401A
That will be an interesting test.

Interestingly, the bootup time takes a little longer than the older models.

[HighVoltage]

Unfortunately this is not confirmed...

It seems the 34972A is also using the circuitry of the 34401A, at least when it comes to low AC voltage measurements.
It is a known fact that the 34401A has problems to accurately measure low AC voltages

Here I did a test with 250 mV AC, 400 Hz
This is the lowest setting of my calibrator.

and these are the measurement results:

- 34401A:  54 uV, jumping around +/- 8 uV
- 34410A:  250.8 uV, stable
- 34411A:  253.7 uV, stable
- 34970A:  53.4 uV jumping around +/- 8 uV
- 34972A:  52.2 uV jumping around +/- 8 uV