HP/Agilent 34970A Teardown

[AF6LJ]

Thanks for the teardown, good stuff.

[free_electron]

Hi free_electron,

I've just purchased a Keysight 34970a. Was listed as parts on ebay with error 625. Should be arriving in the next day or two. I have seen one thread where low battery voltage was being linked to this error? (I should be so lucky) Just wondered if you have had any dealings with this particular error?  Service manual tells me lack of communication between main processor and outguard processor. I'm hoping for faulty 5V earth reference supply, reset or opto isolator problems. If it is the outguard processor will I be able to obtain a replacement? or have I spent £140 on a door stop?

Thanks in advance
Jim

most likely : battery leaked and acid ate a few traces around the glue logic. i have repaired one in that state...

[DoricLoon]

most likely : battery leaked and acid ate a few traces around the glue logic. i have repaired one in that state...

Thank you free-electron, I'd settle for that type of repair.
Will let you know what I find when it arrives.

[HighVoltage]

I have fixed already two of these leaky batteries and it worked both times and took care of the error as well.

free_electron:
your great post here, got me started on the 34470A
Thank you for that

[DoricLoon]

Well the unit arrived and sure enough error 625 is present. Outer condition is excellent, opened it up and removed the logic board.

My heart sank when I saw four 8 pin devices with pretty grotty looking soldering. Turned out they were the opto isolators for the data between processors. Gave the soldering a touch up and cleaned the mess of old flux off the board. Nothing else visually but someone obviously had a go. Put the card back in and powered it up. Supplies are all present, battery voltage at 2.82V Will fit a replacement once I can lay my hands on a Panasonic BR2477A/VAN in the UK. Digi-key lists them but can't transport them! Outgoing processor 5V_ER supply present, both the 3.8MHz and 12MHz clocks are running the 5V_ER power on reset goes high, low then back to high. The main processor sends some data via the opto and buffer to the outgoing processor but no reply back from the outgoing processor. While in the chassis it is difficult to probe the outgoing processor for activity on all the pins. Think I have to get the card out and under power to test further. I'm hoping it's not the outgoing processor itself U305-PROG PART LOWER LEVEL 1821-1876 Manufacturers part number 34970-88803. Don't know how easy it will be to obtain one? Are Keysight likely to provide components rather than complete assemblies? Get another go at it next week sometime hopefully something else holding the outgoing processor in this state?

5V_BP is switching and the IO_RST to the GPIB controller is happening so U305 can't be completely dead as it switches both of those, processor must be running. Just not sending data back to the main processor?

Was thinking perhaps an interupt from either of the io devices may be holding things up?
With GPIB interface selected
UART interupt is low and remains low, GPIB interupt goes from low to high on power up. Which inturn sends the common interupt line low to high on power up?
When the RS232 is selected UART interupt goes low to high GPIB interupt goes low to high and Common interupt goes high to low.
Does that sound about right?

[DoricLoon]

Much to the wifes frustration I've spent a while fault finding this error 625.

There is indeed two way communication between main processor and IO processor. When the power on button is pressed the main processor sends out a 400mS pulse on the data line after 15mS of that pulse the IO processor reset is taken high and remains high for the rest of the pulse. 62mS after the reset returns low the IO processor sends a 53uS pulse to main processor. Further 15mS and main processor sends 2 11 bit frames of data to the IO processor.

The fact that I can change interface type from RS232 to GPIB and see the interupts change means the IO processor has responded to data from the main processor. I don't have a card for this thing yet. Waiting for a couple to arrive. Going to try to connect RS232 to see if I can actually talk to the unit?

Any ideas?

[DoricLoon]

Have connected to RS232 and Connection manager finds an instrument but doesn't identify iteslf and doesn't respond to *IDN just gives a time out. I have set COM port 1 with same settings as those of the instrument.

[DoricLoon]

Had another go at the 34970a today and think I may have found something?

I can change the baud rate of the RS232 interface and see the baud clock on the TL16C550CFN alter so the IO processor is definately responding to main processor command.

I checked the data lines between the IO processor and the two bus management IC's and could see data switching. I then selected GPIB and each of the data lines shows the same signal. As per attached PNG. I'm an electrician trying to learn electronics so my knowledge is pretty rudimentary. Looking at the signal it appears to me that the rise time of the signal is way off. With GPIB selected the signal is constant. RS232 selected and data is changing but after taking a closer look that horrible rise time still exists.

Should I desolder the GPIB controller and see if the data lines improve? 

[voltsandjolts]

FYI this is a long dead thread, just waking it up to add some info.

This thread pops up when searching for a part number for the 34970A front panel encoder.
The Keysight info for the encoder and knob are:
0960-0961      MECHANICAL ENCODER 2 BIT 24 POSITION
34970-87401  34970A KNOB

As mentioned above, it's a Bourns ECW and I suspect the exact PN is

ECW0J-C24-BE0024L        no detents, smooth rotation (as supplied by Keysight)
Currently available at Mouser for £3.40

...or if you want detents...
ECW1J-C24-BC0024L        24 detents
Currently available at DigiKey for £4.80