Author Topic: Stumped trying to repair Fluke 45  (Read 7116 times)

0 Members and 1 Guest are viewing this topic.

Offline c4757p

  • Super Contributor
  • ***
  • Posts: 7805
  • Country: us
  • adieu
Stumped trying to repair Fluke 45
« on: January 31, 2013, 03:01:47 pm »
I recently acquired a broken Fluke 45 multimeter in an attempt to repair it. The meter had a bad power supply - four times out of five, it would fail to achieve regulation and the rails would only hit half of their specified voltage. I traced the problem to the switching transistors, replaced them, and now the power supply works perfectly. Unfortunately I seem to have caused more problems (before I did anything, on the one time of five that the power supply started up properly, the whole meter worked just fine). I kind of suspect some digital chip is fried, which does not make me happy, but I can't really imagine how it happened.

When the meter powers on, the system is inactive. The display does not initialize, no relay clicks, and the keypad scan lines are silent. All voltages are fine. The microprocessor clock oscillates briefly (about 400 ms) at the correct frequency. During this time, there is activity on the address and data bus. However, the "E" pin on the microprocessor, which the service manual says is the first thing to check, is dead. It should have a clock signal on it, and it simply pulls high as soon as the meter is powered on and stays there. It doesn't appear to be shorted to anything, and since the only thing it connects to is the GPIB option card, I guess lack of this signal is just a symptom of whatever the problem is, not the cause. I really don't know anything about traditional microprocessor architectures, so I have absolutely no clue what the pin does. The service manual just says to check it (and what to expect on it), it says nothing about what it means if it is dead.

Clearly something is preventing the processor from starting. The power-on reset works fine, as does the watchdog (it fires in about four seconds and the process repeats). As mentioned above, the oscillator works, as does whatever circuit shuts it down when the processor goes to "sleep". The behavior is identical if I unplug the socketed ROM, so I'm pretty sure the processor isn't running any code at all - not sure what the bus activity is. (I'd guess it's trying to read code out of the ROM, but I was under the impression that it would just jump to the reset vector and start trying to run, not screw around with the bus for 400ms.)

Am I correct in guessing that the processor has died? Is there anything else I may have missed? (And would anybody be willing to dump the ROM out of the same meter so I can compare to my own and check for corruption?)

Service manual is here.
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 7405
  • Country: us
    • SiliconValleyGarage
Re: Stumped trying to repair Fluke 45
« Reply #1 on: January 31, 2013, 03:31:35 pm »
Do you see signals on the optocouplers ? At i it the cpu tries to talk to the measurement processer. If he doesnt answer ....

Doublecheck all your voltages.
Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).
 

Offline c4757p

  • Super Contributor
  • ***
  • Posts: 7805
  • Country: us
  • adieu
Re: Stumped trying to repair Fluke 45
« Reply #2 on: January 31, 2013, 04:15:33 pm »
There are no signals at all on either side of the optocouplers. Anyway, the manual says that it should be able to display an error message if it can't communicate with the measurement processor, so I imagine that comes at a later stage.

One voltage is actually slightly out of spec, which I didn't notice (thank you for insisting I check again), but because of which voltage it is and how close it is, I doubt it is the problem:
Out-guard VCC: 5.217V (spec: 4.85 to 5.35)
Out-guard VEE: -4.923V (spec: -5.0 to -6.0)
Out-guard VLOAD: -29.73V (spec: -28.5 to -32.0)
Out-guard filament supply: 10.0Vp-p square wave, 29.3 kHz, offset -24.4VDC

In-guard VDD: 5.093V (spec: 4.95 to 5.45)
In-guard VSS: -5.173V (spec: -5.0 to -6.0)
In-guard +VAC: 4.891V (spec: 4.7 to 5.35)
In-guard -VAC: -4.987V (spec: -4.7 to -5.35)

I have checked power supply connections to: U6 (microprocessor) all power pins, U5 (EEPROM), U8 (EPROM), U10 (RAM), U9 (address decode), and all are fine, and I have checked for shorts to both rails and adjacent lines on the address and data bus. Could VEE really be the problem? The only place it goes is to the display assembly, where it supplies a bias to the VFD driver, and I doubt the VFD driver minds being 80 mV out of spec...
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 7405
  • Country: us
    • SiliconValleyGarage
Re: Stumped trying to repair Fluke 45
« Reply #3 on: January 31, 2013, 04:34:35 pm »
Nah. Thats not the problem.

Do the address lines of the eprom toggle ?
Ill see if i can make a romdump. I think i have a 45 somewhere (wrecked)
Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).
 

Offline c4757p

  • Super Contributor
  • ***
  • Posts: 7805
  • Country: us
  • adieu
Re: Stumped trying to repair Fluke 45
« Reply #4 on: January 31, 2013, 04:48:28 pm »
Yes, they do.

Thank you! I'm going to rig up a breadboard to dump my ROM (sadly, my dog decided recently that my programmer looked an awful lot like a chew toy...) to compare. Seems to be a pretty much jelly bean ROM, anyway, so even if mine is junked I can always flash your image onto a new one.
No longer active here - try the IRC channel if you just can't be without me :)
 

Online PA0PBZ

  • Super Contributor
  • ***
  • Posts: 4343
  • Country: nl
Re: Stumped trying to repair Fluke 45
« Reply #5 on: January 31, 2013, 08:54:48 pm »
According to the datasheet the "E" pin should output a system clock at 1/4th of the cpu clock frequency. So either your CPU does not get a clock or it is fried....
Keyboard error: Press F1 to continue.
 

Offline c4757p

  • Super Contributor
  • ***
  • Posts: 7805
  • Country: us
  • adieu
Re: Stumped trying to repair Fluke 45
« Reply #6 on: January 31, 2013, 10:45:53 pm »
Aw, damn. That's kind of what I was expecting. Guess I'll just have to scour the corners of the Internet and see if I can find an HD63A03Y - and be fifty times more careful next time I'm fixing something next to an 80s CMOS chip... Sadly the PLCC version of the chip seems rather hard to find.
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline c4757p

  • Super Contributor
  • ***
  • Posts: 7805
  • Country: us
  • adieu
Re: Stumped trying to repair Fluke 45
« Reply #7 on: February 01, 2013, 02:09:35 am »
Hmm... I'm not sure where I pulled "at the correct frequency" out of, considering I don't actually remember measuring it now. Not only that, but I was too dense to notice that it's off by a couple orders of magnitude! Instead of the nominal 3.6864 MHz, the "brief oscillation" I see on startup is actually around 14 kHz (yes, kilohertz). I wonder if I managed to damage the crystal? Maybe bashed it with something when I was trying to get to the hard-to-reach switching transistors. I think I'll poke around for a temporary replacement at a similar frequency, to test that. Sure would be nice if that were the problem.  :)

(Logical Me speaking now. Shhhh... don't tell Overly Optimistic Me, but it probably is the unobtanium built-in oscillator that's broken, not the crystal. It can't be that easy.)

Edit: I told you so, Overly Optimistic Me.
« Last Edit: February 01, 2013, 02:21:01 am by c4757p »
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline c4757p

  • Super Contributor
  • ***
  • Posts: 7805
  • Country: us
  • adieu
Re: Stumped trying to repair Fluke 45
« Reply #8 on: February 01, 2013, 02:52:22 am »
God, this is bizarre. I tried injecting an external clock (the datasheet says that the clock pin can be driven directly), and while that didn't make the thing work, now the proper frequency is back! It still doesn't work, and the E pin is still dead, but I have an oscillator now.

TBH, I don't even know why I'm still messing around with it, I'm sure the chip is gone for good... Somehow I suspect "magical, unexplainable behavior" is a good sign of fried silicon.
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline c4757p

  • Super Contributor
  • ***
  • Posts: 7805
  • Country: us
  • adieu
Re: Stumped trying to repair Fluke 45
« Reply #9 on: February 22, 2013, 10:48:00 pm »
Update: Replaced the processor and it works again.  :-+ But the EEPROM where it stores the cal data is dead (not just erased - dead - as in "always a diode drop between SDO and GND"). I think I'll recalibrate it myself with my 3468A as a transfer standard - it has better accuracy specs across the board, and while its own calibration is out of date, it appears to be spot-on.
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline Raff

  • Regular Contributor
  • *
  • Posts: 83
  • Country: au
Re: Stumped trying to repair Fluke 45
« Reply #10 on: February 23, 2013, 02:32:00 am »
Well done! Thats a nice classic example of trouble shooting. Great story and good to hear its up and running once again.
 

Offline c4757p

  • Super Contributor
  • ***
  • Posts: 7805
  • Country: us
  • adieu
Re: Stumped trying to repair Fluke 45
« Reply #11 on: February 23, 2013, 02:44:23 am »
Trouble shooting or trouble causing? If I hadn't been such a klutz in the first place and zapped the chip, I'd have fixed it for the cost of two PN2222 and two PN2907...  ::)
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline c4757p

  • Super Contributor
  • ***
  • Posts: 7805
  • Country: us
  • adieu
Re: Stumped trying to repair Fluke 45
« Reply #12 on: February 23, 2013, 02:48:18 am »
Gah!! Now the power supply is dying again!! Guess it's time to probe around that yet again - maybe something failed and it's generating high voltage spikes that are blowing out transistors. Thankfully the 5V rail is clean, so whatever is happening, it's not going to fry the new processor...
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline Raff

  • Regular Contributor
  • *
  • Posts: 83
  • Country: au
Re: Stumped trying to repair Fluke 45
« Reply #13 on: February 24, 2013, 02:26:38 am »
Sometimes troubleshooting can cause troublecausing as a side effect, just chalk it up to a learning experience.  8)
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf