Fluke 8505A Repair - Current Shunts and AC RMS Converter optional modules

[cat87]

Hi all,

   This will be a short post where I sum up what I did to get two modules of my Fluke 8505A up and running again. The full article can be read on xDevs site here
I got this meter a while back and I have to admit, for the years it has, it's aged quite gracefully. I was surpsised to see the meter fully kitted out with nifty optional modules like AC RMS Converter, Ohms Converter, Current Shunts module and GPIB module.
After initial turn-on, the DC voltage measurement was working nice, though not the same could be said about two of the modules, the Current Shunts and the AC RMS Converter.



Let's take these one at a time.

The Current Shunts Module:
Initial power-up showed that Option 03 - Current Shunts Module was not being identified by the DMM Controller, hence no  "3" displayed. When trying to access the module, the display would
show "Error 9"



According to the Service Manual, this may be due to a number of faults, like U6, CR14 and CR15 may be leaky, U5 and Q18.
But before I can get started, I had to make some sort of adapter for the module so I could have access to all the components on it. Because this isn't high speed, I can get away with just having some wire extensions. Apparently the hardest part was to actually figure out the weird 3.96 mm pin-pitch connector Fluke used and actually find it for sale somewhere. Also I had to reverse-engineer the board pin layout. Only took me two board designs to get it right. I'm getting better at this . When this was all done, I had this monstruosity:



   I initially started with the digital section, U5 in particular. Whenever a module is addressed by the Controller, it answers back with an initial ACK signal. Then the module goes into some kind of self-test and of this passes, then a second ACK signal is sent, letting the Controller know that it can use that  module.
  After some brain storming, probing U5 with my L.A. during the power-up sequence showed that the addressing was indeed being  decoded and the first ACK signal sent out. But  no luck on the second ACK. This meant that the self-test was failing.



  After measuring CR15 and 14, my attention turned to U6.
After following the first rule of troubleshooting, I had some pretty weird readings. I had 27 V on C10, however the inputs to U6 were at about the same voltage. Looking elsewhere in the circuit didn't pan out much, so I was beginning to get more and more suspicious of U6 and thinking that maybe something inside it released the magic smoke.

 Going with my gut, I took out the old U6 and put in a socket and a "new" LM358.
And to my surprise, it really was that simple. But the surprise was short lived. After leaving the meter on overnight, the module failed again. So back to square one.
However, this time my measurements were different. Something was indeed triggering U6 but it didn't seem like it was coming from the measurement bus.
Continuity tester in hand, I went at it to see what information I could gather.  C12 seemed suspicious so I took it out and measured it with the LCR meter. Sure enough, old age was taking its toll and poor C12 croaked.
Changing this together with C13 for good measure fixed the issue.



AC RMS Converter:

   Unlike the previous module, the RMS Converter can be accessed and used, it's just that  for some reason, it displays "HHHHHH" on all ranges. So first thought might be one of the stages in the module is misbehaving.



Care to wager it's some unobtanium part?
But for the moment, let's look over the schematic and see how Fluke designed this.



  There's three stages, that in the end output an RMS value for whatever amplitude signal you input. The first stage is the Range Amplifier, which brings the input signal down (or up) in amplitude to about 1V (composed of Q1 and U3).
  The signal is then passed to the second stage, made up of the Balance Amplifier (U4 and Q10), which converts the signal to its absolute value. The remaining Squaring Amplifier, Integrator and
Square Root Amplifier that make up the last stage is where all the maths happen and these will output a signal which is  the square root of the integral of the square of the current in the emitters of Q8A and Q12A.  Obviously the first thing to look at is the Range Amplifier stage output, which is Pin 4 of U3. Keep in mind that this should be somewhere around 1V. Let's see what we actually have...
 
 Hmm, 13.3 V. There's your problem. Ok, now I wanted to know it it's U3 itself that's faulty or something upstream is driving it hard to the rails. Checked Q1, nothing there. Also took out C40 and C7 to check them, seeing how caps of that type also failed in the Current Shunts.
 As everything checked out, the conclusion is U3 is indeed at fault.



 I didn't have a direct replacement for it, but I did find  a 741 op-amp in a TO-99 package that was pin-compatible. Ok, that'll do. So took out U3, put in a socket and put the LM741 in there. Powered up the
 DMM and wouldn't you know it, it worked first go.



Yes, I know the reading's way off, but once I'll get the proper replacement part in there and do some internal  alignment, things should be OK and in-spec.
There's also lots of stuff to do, to see what tempco this thing has and how much it went out of spec, but that's work in progress right now.