Fluke 27, well at least the DC Amps, mA and uA works...

[Robomeds]

So I decided to invest in a project.  For less that $15 I ended up with a Fluke 27 meter that was stated as bad when I purchased it (I knew that going in).  I've been able to successfully repair more than a few meters but this one has me stumped thus far.  So here are some of the symptoms.

In resistance mode it reads 1074 plus the decimal.  I don't mean it has a fixed resistance of 10.74 or 104.7, I mean the digits always read 1074 regardless of manual range.  Voltage AC is off but lacking anything other than 120V I've only verified the AC voltage is off.  DC voltage is also wrong as well as the diode check.  The diode check seems to have a decay almost like a discharing cap.   

OK, I have consulted the service manual for this thing:
http://exodus.poly.edu/~kurt/manuals/manuals/Fluke/FLUKE%2027%20Service.pdf

Section 3-25 suggests a bad main IC.  I failed step 4 suggesting Z1 (the thin film resistor) or U1 may be bad.  I also fail step 5, again suggesting a problem with U1.  However, the DC current measurements are all spot on.  So does that still point at U1?  Anyway, any suggestions?  I've checked the input MOVs and diodes, all look good.  Given the fuses are worth more than I spent on the meter I'm not worried if this project fails but the challenge is still there.  Suggestions?
 

[retiredcaps]

For less that $15 I ended up with a Fluke 27 meter that was stated as bad when I purchased it (I knew that going in).

$15 + free shipping was the going price for this deal.  Looks like a lot of people benefited by getting working units.

http://www.antiqueradios.com/forums/viewtopic.php?f=8&t=310451&sid=5bbd8e0efcd1faf9af07bb3e8b671305

I've checked the input MOVs and diodes, all look good.

Can you post your measurements for the MOVs, fusible resistor and PTC please?

Also, can you post clear focused picture of the pcb, both sides, rotary switch. close up of U1 so we can see what revision board you have?

[Robomeds]

For less that $15 I ended up with a Fluke 27 meter that was stated as bad when I purchased it (I knew that going in).

$15 + free shipping was the going price for this deal.  Looks like a lot of people benefited by getting working units.

http://www.antiqueradios.com/forums/viewtopic.php?f=8&t=310451&sid=5bbd8e0efcd1faf9af07bb3e8b671305

I've checked the input MOVs and diodes, all look good.

Can you post your measurements for the MOVs, fusible resistor and PTC please?

Also, can you post clear focused picture of the pcb, both sides, rotary switch. close up of U1 so we can see what revision board you have?

After I bought the $14 non-working model I got one of the $15 "working" ones (free shipping).  The one I got had a cracked housing and was no where near as nice as the sample unit in the picture.  I ended up swapping the PCB between the one that was stated as non-functional.  So now I have a ~1988 Frankenstein meter that works just fine and one with a bad PCB and rough (but usable) housing.  So I'm now I'm trying to figure out what's wrong with the bad PCB.  U1 seems the most likely failure but I'm still investigating.

I'll try to get those measurements in the next few days (family in town is slowing hobby work).

[ModemHead]

If it can measure current, that means the ADC is functional, so there is hope.  Sometimes the current shunt voltage sense uses a different input to the main chip than the voltage section though.  Each input can become faulty by itself.

Sorry I can't do much research to help right now, I have an elderly father who has fallen and broken his hip.  Thank goodness for hospital wifi, so I can follow the batteriser saga to relieve boredom.

[retiredcaps]

So I'm now I'm trying to figure out what's wrong with the bad PCB.  U1 seems the most likely failure but I'm still investigating.

U1 is a possibility, but I will have to hold off further judgement until I see more results and study the service manual a bit more.

Excavatoree has stated that he has seen bad U1s in some older Fluke 70 series.  I think he said that sometimes reflowing the solder to the pins works despite them looking okay?

[BMack]

For less that $15 I ended up with a Fluke 27 meter that was stated as bad when I purchased it (I knew that going in).

$15 + free shipping was the going price for this deal.  Looks like a lot of people benefited by getting working units.

http://www.antiqueradios.com/forums/viewtopic.php?f=8&t=310451&sid=5bbd8e0efcd1faf9af07bb3e8b671305

I've checked the input MOVs and diodes, all look good.

Can you post your measurements for the MOVs, fusible resistor and PTC please?

Also, can you post clear focused picture of the pcb, both sides, rotary switch. close up of U1 so we can see what revision board you have?

After I bought the $14 non-working model I got one of the $15 "working" ones (free shipping).  The one I got had a cracked housing and was no where near as nice as the sample unit in the picture.  I ended up swapping the PCB between the one that was stated as non-functional.  So now I have a ~1988 Frankenstein meter that works just fine and one with a bad PCB and rough (but usable) housing.  So I'm now I'm trying to figure out what's wrong with the bad PCB.  U1 seems the most likely failure but I'm still investigating.

I'll try to get those measurements in the next few days (family in town is slowing hobby work).

You had some terrible luck man, mine arrived near perfect. It has some sticky residue from a sticker on the side. I shared the auction when I found it on another forum and one guy bought two, his were perfect. If they relist and you want a good one, I'd try your luck again $15 is a steal and they did relist once after the initial listing. Heck at $30 for one working and a Fluke doorstop, you still did really well.

Good luck with the repair, I wish I could offer some real help here but I've never repaired a DMM component level.

[Robomeds]

OK, finally got some pictures and a few measurements. 

The MOVs are all fine (resistance shows OL).  The diodes are good and the input resistors on the good vs bad PCB are all the same within tolerance.  Currently I'm leaning more to a U1 (the main IC) failure.  It is nice to have a known good PCB as well as the "almost" correct schematics.

[Robomeds]

One more picture

[retiredcaps]

1. Let's check your fusible resistor circled in red.  It should read around 1k ohm. 

2. Check your PTC circled in blue.  It should be around 1.1k ohm.

3. What is the input impedance of the Fluke 27 when it is turned on and set to DCV?  It should be around 11.11M ohm as measured with another multimeter.

4. If you measure an AA cell, do you get roughly 1.5V DC?

[Robomeds]

So where's what we have...
1.  Checks out at 1K
2.  Checks at 1.1K
3.  Checks at 11.13Mohm  Interestingly it takes a bit to settle, almost like there is a bad capacitor impacting things. 
4.  Not even close   Absolutely static at 0.074.

[retiredcaps]

So where's what we have...
1.  Checks out at 1K
2.  Checks at 1.1K
3.  Checks at 11.13Mohm  Interestingly it takes a bit to settle, almost like there is a bad capacitor impacting things. 

The above look correct.

4.  Not even close   Absolutely static at 0.074.

Hmm the 74 digits showed up in your resistance readings as well.  I'll need to study the manual and schematics a bit before I can offer another suggestion at this point.

[retiredcaps]

Have you tried the DC signal tracing?  If yes, which steps work and which fail?

If U1 is the likely culprit, do any of the solder joints look suspicious or dodgy under a 10x jewelers loupe or microscope?

[Ordinaryman1971]

Reflow the U1 or if you first want to check if that's the problem just press it down to the pcb with your finger and turn the meter on, see if you get the same display values with the U1 being pressed down. It helped me few times to determine if the chipset is being the one if question. Sometimes the cracks are so minuscule you can't really see them. You have to either reflow it or just press it down when checking the display.
To me it actually looks like a short between legs of one of the ics, wash it really well with alcohol one more time and look if you can see the signs of battery spillage.

[Robomeds]

I did go through the service manual's voltage tracing steps.  As I recall I failed at step 4.  All the U1 solder joints look good and none moved when poked and prodded.  Still, just in case I reflowed all of them.  No change.  I'm as much puzzled by how this thing died as anything.  I just don't see any obvious damage or issues.  Now that I think about it, one thing I haven't checked is current draw.  Last time I had a bad Fluke IC it was clear when I looked at the load on the battery.  I haven't checked that here.

[retiredcaps]

Last time I had a bad Fluke IC it was clear when I looked at the load on the battery.  I haven't checked that here.

The Fluke 27 has 1000 hour battery life.  Assuming a 500mAh 9V battery, that suggests a 0.5 mA or 500uA current draw for normal operation.

[Robomeds]

Well now I'm less inclined to think it's U1.  I've found a few new things.

1. The current consumption seems normal.  I measured .52mA in DCV range.  Other ranges were slightly higher at .6 or so.  The only one that was much higher was diode at .9 but that included the continuity tone (no leads connected so any continuity was inside the meter. 

2.  The DC voles isn't consistent as I thought.  It's almost like I'm charging a capacitor.  When I start by selecting diode I get a high enough reading as to not trip the continuity buzzer.  Depending on what ranges I select first I can trip the buzzer.  It also changes what I see in the DCV range.  Note that Resistance seems very consistent and stable regardless of previous range tested.  It's almost like I have some type of bad capacitor in the system.  However, if ti were a bad cap I would expect to see a change in R depending on open vs shorted leads. 

As an aside, I just picked up a second "project" meter, a made in China Fluke 73-3 "Lincoln Tech" edition.  The seller said it just stopped working.  It turns out the problem was nothing more than the included 9V battery's negative terminal was too loose to ensure a good connection.  Closing it up slightly with a pair of pliers fixed the issue.  Now I have a 73-3 to join the recent working 27 in the garage.  Talk about extreme ends of the diagnostic range!

[ModemHead]

Don't be misled into thinking the fault is a capacitor.  With the high input impedance of the ADC and just parasitic capacitance, "open" faults can lead to that sort of charging/discharging behavior.

[retiredcaps]

Well now I'm less inclined to think it's U1.  I've found a few new things.

For completeness sake, let's test all the components in step 6, namely R13, R16, C18 and C19.

[Robomeds]

I didn't state it but I did check items 5 and 6 with no success.  That was part of why I thought it might be U1 even though the current readings look good. 

[retiredcaps]

If you short the probes on DCV, do you consistently get 0V? I have a feeling you may have to cut a trace on the pcb to isolate U1? I know modemhead has done this in the past to see if the main IC is properly functioning with the analog input section out of the picture.

[Robomeds]

OK, shorting the inputs makes no difference what so ever.  However, when I turned the meter on (from Off to VDC) the reading was ~0.010V and climbing slowly.  When I turned it to diode it read 0.060.  When I turned it back to VDC it reads 0.290 and is dropping slowly.  Turning it back to diode and my reading is 0.03.  Back to DCV and the reading is now 0.015 and climbing again!  It does seem like there might be an odd blown trace somewhere.

[Robomeds]

Just doing a bit more testing... When connected in series with my power supply and two other meter to test DCmA I get 320.48 (Fluke 187), 320.5 (27/FM - Frankenstein) and 321.5 (troubled meter).  I'm guessing this is nothing more than a bit of adjustment error (and in spec).  The DCA range in a similar test reads 2.2583A, 2.24 (franken-Fluke), 2.26 (troubled meter).  Those readings all seem reasonable to me. 
I looked at the current drain into the voltage terminal with a 10V input (10.05V).  My 187 sees 0.98uA of current and the bad 27 just counts up (currently .118 but it was under .1 earlier). 
 

[Robomeds]

OK, it's getting more interesting.  I'm not sure why I decided to try the DCmV range but I did.  It was spot on.  So it seems unlikely that U1 is bad given the proper current draw and the accurate measurements in DC mV, mA, uA and A.  The VAC range is way off (based on testing a wall outlet).  The diode, continuity and resistance are also way off.  It's really seeming like I should have a single bad part somewhere in the path between the positive input and the IC.  I'm not seeing any board damage even on close inspection but I might still be missing something (well clearly missing something since it doesn't work). 

Incidentally, I found the Fluke 8025B service manual seems like it's probably closer to the correct schematic for this meter.  The Fluke 27 service manual I found seems to be from the final version of the meter, not this earlier iteration.

Incidentally, I find it interesting that the 8025B had a third fuse.  In addition to the two buss fuses, it also has a cheaper glass fuse in series with the low current fuse.  I guess the idea was the cheap one goes in most cases but the pricey one is still there if the poop hits the fan.