Tektronix DM5120 - 6 1/2 digit multimeter (need repair)

[sorenkir]

Hi,

My DM5520/DM5120 has arrived yesterday.
Long and heavy beast!

As advertised, it has Error 453 on startup (bad EEPROM checksum).
It seems that the EEPROM is dead. Program 30, to save current configuration, has no effect.
The EEPROM is referenced 2816A on the schematic, but it is actually a XICOR X2816BP. I don't know the difference...

Visual inspection, fuses, and supplies are OK.
The multimeter kind of works. It can perform measurements, with specified accuracy, but only on some ranges/polarities:
Measurement of negative DC voltages is OK on all ranges, but positive DC voltages read 0 on lower ranges !
AC volts are OK, Ohms work on the lower ranges (below 300k), Current measurements are stuck at 0 (but the fuse is good).

I wonder if this only due to some calibration constants corrupted in the EEPROM ?

Next steps:
- dump/check program UVPROM,
- check if EEPROM control signals look OK and, if yes, order a replacement,
- desolder the EEPROM and try to read it,
- replace EEPROM (on a socket),
- calibrate

To be continued...

Michel.

[coromonadalix]

Check this wiki, i added some stuff on it, some eeprom dump too


http://w140.com/tekwiki/wiki/DM5120

I had this thread too

https://www.eevblog.com/forum/testgear/tektronix-dm5120-a-keitley-196-brother/


Warning  be very careful when removing ic's on the main board, the pcb traces are very thin,  busted 2 traces while removing the eeprom, and i was very careful with an desoldering pump.

I hacked mine with an different main xformer taken from a Keithley 196 and hacked a new regulator board, i dont need the mainframe anymore, have sold it.

You can calibrate it   but needs tons of parts to do so.

[texaspyro]

I have seen several DM5120's with no working current ranges.  Problem has always been a bad (2 ohm?) power resistor.  It looks to be around 2 watts,  is brown,  with prominent end caps.  Sorry, I don't know its reference designator.

[coromonadalix]

Was it tempered / repaired before ?    take note of the inputs wirings before removing them, i did not  and was saved by the service manual  loll

Check the input 3 amp fuse

On the http://w140.com/tek_5120_5520.pdf     service manual

On the  page 11   analog 2  section     you see all the input current dividers


As usual  you need to check all the psu voltages, references voltages  ....

A main 5v regulator,  heat a lot

+15 and - 15v regulators
+5v  for the analog section
+2.6v   in the analog section
-10vdc in the analog section
+4.3v  in the analog section


To do servicing outside the main frame,  i builded an extended  flat cable IDC14 pins extension pcb slot in the page 16 to supply the meter, i had 4x power supply to feed the meter,  since the 2x 5vdc regulator (main and secondary) need to be separated / isolated for the digital and analog section.


EDIT   just powered my meter,  damn  i may have another broken trace in the eeprom section        

[sorenkir]

Hi,
Thanks for the feedbacks and the links!
Fuses and supplies are OK.
I have checked the UVPROM: content is identical to the dump available in the other thread.
Next step is to check if the EEPROM control signals are OK (address decoding: U1, access logic: U17).
I think I will use my HP 1650A logic analyzer, 1987 vintage, two years before this DM5120.
And it is a project in itself!
Maybe I will be able to dump EEPROM content, just by decoding accesses. Lot of fun to come!
Michel.

[sorenkir]

I have made good progress on this repair project!

I have managed to dump the EEPROM content without desoldering it, by recording accesses done with my HP 1650A logic analyzer:




and indeed, the content is corrupted, there are a few lines full of 00 or FF.

Differences with a "good" EEPROM dump:



According to the accesses, done at various moments (eg. power-up or when changing ranges):
Calibration constants are stored from 005 to 0DC (4 bytes per constant, 54 constants = calibration points).
Calibration checksum is at 004 (FF in my EEPROM is obviously wrong, hence the error 453).
Settings (and probably other data) are stored from 0DD to 148, which is the last address used.
10 00 is regularly read/written to addresses 001 & 002.
Addresses 002 & 003 are never accessed.

I have dumped the write and read accesses done when trying to save front panel settings to the EEPROM.
The data is stored and tt is read back correctly on next power-up, but for some reason (probably the checksum error), it is immediately overwritten with default factory settings.
So, in fact, the EEPROM is OK, but some calibration constants have been corrupted.

Then I have tried to calibrate (approximately) the 3V DC range that was only OK for negative voltages, and it worked!
The checksum has been recalculated (no more error) and the 3V DC range is now OK.



I have no appropriate calibration equipment, but I will try to "calibrate" as best as I can with the help of my HP34401A.
By chance most of the AC (V & A) ranges are OK, which is the most difficult to calibrate with the equipment I have.

It will still acceptable for a sub-50€ score!

Michel.

[coromonadalix]

The thing(s) that intrigued me is :  how the analog measurement data is transmitted to the digital section ... i had the stupid idea to convert the front panel display into a big graphical display lol

Not  enough experiece to decypher how it work.