Products > Other Equipment & Products

Vintage Racal-Dana 5004 Digital Multimeter Fault

(1/2) > >>

davec:
I wonder if anyone on here can help. I have a large rack mount 6 digit voltmeter, a Racal-Dana 5004 from around 1979, but it has a problem. When there isn't a voltage across it's input, it counts down from the last voltage applied.  By that I mean, if I measure an AA batter, it says 1.54356 volts (the same cell on a 4 digit calibrated meter reads the same without the extra digits). As long as the battery is connected, it stays the same. But when it is removed, it then goes to about 1.4 volts, then 1.3, and so on. You could probably get the same effect by adding a capacitor across the input, but it doesn't stop at 0v, it keeps going -0.1v, -0.2v etc. It gets to -30 then goes back to 0v and starts counting down again.

A bit of background, this thing was going to be skipped many years ago, so I thought I might be able to do something with in. 15 years later, I'm here watching the eevblog and I see the one about the precision voltage reference, and that reminds me that at the back of a cupboard somewhere is a huge faulty multimeter (and when I say huge, it's 2U 19" rackmount, about 20" deep). When I dug it out, it said saying 'Error 7' on power up. Once that cleared, it started counting backwards. I managed to get hold of a service manual, and that says it was faulty RAM. I replaced that and the 'Error 7' message has gone, but the countdown still continues.

It's a nice sold old bit of test kit, and it would be nice to have it working properly. Any suggestions on where to start with this kind of problem.

Thanks,

Dave

tekfan:
This isn't a problem at all. Most high percision voltmeters have many 100 giga ohm input impedance. This causes the voltmeter to pick up all sorts of electromagnetic radiation from the air. The lower ranges have very high input impedance so the charge builds up. The higher ranges (100V, 1000V ranges) have a impedance of 10M ohm.

When the charge is high enough that the multimeter wants to switch to a higher range with 10M ohm impedance the charge bleeds off and then the voltmeter goes back down to one of the lower ranges. Then the process repeats its self.

davec:
Oh right, I had never thought that it was actually meant to do that. I was just expecting to see zero, but I suppose that makes sense.

So the actual problem leading to it being disposed of was the memory problem. It has a 4MHz 6802 microprocessor with 1K of memory, the service manual I got of the internet showed this as 2 x 2114 SRAM. It seems this must have been a slightly later revision of the board as everything else on was the same, but it was laid out to take a single 4118 instead. Shame as I had some 2114's available and the 4118 seems to be a rare beast. I managed to find some SRAM that had a similar pinout, although larger at 32K and 28 pins. I made a bodge socket to fit it into the 24 pin socket, tie the spare address lines to ground and reroute power and WE to the correct pins. That seemed to work, and the error message is no longer appearing.

So in that case, now the memory fault is fixed, and the main problem of user education has also been addressed, it's back to being a working meter again.

Thanks for the help (and not laughing too much at my ignorance)

Dave

tekfan:
You should check the voltmeter for accuracy. The RAM chips may have had calibration constants on them. The 3 volt battery powers the RAM even if you unplug the multimeter from the wall.

You may want to replace the battery as it is probably more than 15 years old now. If you don't want to loose the calibration constants you must remove the battery and solder in a new one while the multimeter is turned ON. You should heat up the soldering iron then unplug it from the wall and then quickly remove the old and solder in the new battery. Once the new battery is in you can use the multimeter normally again.

davec:
According to the schematic, there is a separate 1K of sequential CMOS RAM for the calibration data (a 6508). The error I was getting was specific to the main SRAM, so I presume that the non-volatile memory is OK. The battery looked fine, no signs of leakage and still holding a charge. Now that this is back as being a potentially useful bit of kit I suppose I should replace the battery as suggested.

As for the actual calibration, I don't have a transfer standard with sufficient accuracy to check it fully, but the first 4 digits are showing the same as the best calibrated meter I have to compare it to.

Thanks,

Dave

Navigation

[0] Message Index

[#] Next page