Agilent E3648A eeprom content lost

[chronos42]

Hello everyone,
I am currently repairing an Agilent E3648A power supply. The device's eeprom (93C86) is defective and has lost parts of its content. The eeprom contains the adjustment data and the setups. After replacing the eeprom I now have the problem that the device cannot be calibrated, since the eeprom apparently also contains the information about the exactly device type of the E364X series. In the current state the device doesn´t "know" that it should be an E3648A, it is in a undefined state. Therefore it cannot be adjusted. Voltage is possible, but current calibration fails because of wrong calibration requirements like 1.500A instead of 2.500A. I have no idea how to set the device to an E3648a via GPIB, there are no hints in the service manual.
Can someone please readout the eeprom content from a 3648A? That seems the only way to bring back the device to life. Or is there another way to configure the device type?

Thanks

[ddcc]

Are you referring to U12 on the digital board? Is it labeled E3646-60023 REV 1.4 or REV 1.7? I think I previously uploaded both to KO4BB:

http://www.ko4bb.com/getsimple/index.php?id=download&file=HP_Agilent/Agilent_E3646A/E3646A_Power_Supply_EPROM_v1.4_E3646-60023.zip
http://www.ko4bb.com/getsimple/index.php?id=download&file=HP_Agilent/Agilent_E3648A/Power_Supply_EPROM_v1.7_E3646-60023_17.zip

[chronos42]

Hi ddcc, thanks, but the link shows to the firmware rom content, but I need the content of the eeprom or at least a hint how to configure the eeprom in a way that the device "remebers" that it should be an E3648A.
Firmware of my device is rev. 1.7, I assume the firmware is the same for all E364X devices, only an entry in the eeprom determines the type. I have also an E3649A and copied the eeprom content of the E3649A to the replaced eeprom of the E3648A with the result that the device now behaves like a E3649A and therfore can't be adjusted. There must be an initial command in the production to set the device to the correct type bevore calibrating, but I have no idea how to do that. So the only way seems a copy of the eeprom (not the rom). This is really annoying, i am so close but now I am stucking because of this silly problem.

[ddcc]

What EEPROM are you referring to (reference designator)? The schematic shows U12 as the only EPROM.

[chronos42]

Hi, U12 is an OTP ROM for the firmware, I am referring to U13, a serial EEPROM  type 93C86 for the setups and calibration data.
(There is an error in the schematic, it is actually a 93C86, not a 93C66 as designated in the schematic. Older power supply series (363X) had a 93C66, Agilent forgot to update the eeprom type in the E364X schematics.)

[ddcc]

Oops, missed that earlier. Here you go.

Mine is a CAT93C86VI, configured as 1 x 16Kb.

[chronos42]

Thank you very much, ddcc!
This brought back my E3648A to life.

[feedback.loop]

Thank you very much, ddcc!
This brought back my E3648A to life.

Did you desolder the chip to program it or is it possible through GPIB somehow?

[ddcc]

I had to desolder it to read it. I initially tried reading it with a Pomona SOIC-8 clip and a TL866, but kept getting bad data and wasn't sure if it was an issue with the programmer or the connection, as there's a bunch of J/K/S/U-suffix variants for the chip in MiniPro that don't match the datasheet. In the end, after I desoldered it, I was able to read it using the non-suffixed variant in the software.

Oh, and as for desoldering, there's too much plastic nearby for hot air, so I had to use low-melting point solder since I don't have the right tweezer tips. Ugh.

[chronos42]

Thank you very much, ddcc!
This brought back my E3648A to life.

Did you desolder the chip to program it or is it possible through GPIB somehow?

Hi feedback.loop,

I did not desolder the chip for programming (beside the fact that I had to replace the faulty eeprom of course)
I am sure the eeprom can be read/write via GPIB, but I have no idea how to do that.

I did it this way:

1: Use a cheap chinese SO8 clip adapter and disconnect pin 8 (5V eeprom supply) from that adapter.
2: Connect the processor /reset input to gnd with a botch wire, so that all processor I/Os are in tristate condition.
3: Switch on the device to supply the eeprom. Because of the interrupted Pin 8 there is no danger of revese-voltage into the programmer.
4: read or write the eeprom.

That worked for me.

BTW: I used a ST 93C86WP as a replacement, that works perfectly.