Fluke 8050 Battery Power Supply Repair Advice

[tango17]

This is the power supply for a Fluke 8050 with the battery - 01 option.  The batteries are new,  and it seems to work,  except the readings are all about 40% too high.  One of the battery leads broke off,  the manual says don't run it without batteries,  and that's about when the problem started. Tracking it down,  the VR1 reference is too low,  causing the deintegrate to take longer,  causing the high readings. Further checking finds that all the power supply voltages are low.  I've marked the voltages in red that I'm seeing now. 
Advice?

The zener VR2 has about 3.6v across it,  but out of circuit it's ok at 5.6v.
Am I correct assuming Q5 and Q6 must be ok,  otherwise I'd get no output voltages?

My problem is I'm busted and can only sit or stand for,, 5 minutes at a time.  I run down,  make a measurement,  and then consider it while lying on my back,  so I need to plan carefully for each test.

Thanks for suggestions. 

[ModemHead]

The +6V and -5V is definitely a problem.  I have several of these meters and I've seen quite a wide range on the +13V and -10V supplies on working meters, basically anything 10V or more (absolute value) seems to be OK. Your -10V is high because it is relative to the ailing -5V, the difference being the battery voltage, 4.8V in this case. The pack must be flat?

Keep in mind the -5V powers all the CMOS logic, relative to ground.

I tend to agree with you about Q5 nd Q6, but the amplitude may still be down for some reason, maybe something on the rail is loading the +6 or -5 down. Hard to say.

Away from home right now, but in 2 or 3 days I can get you some scope traces on the primary if it will help.

[tango17]

The pack has some gairly new NiCd in it,  although they are a bit lightweight in mAh.  The battery warning isn't on,  and it reads pretty much the same with and without AC power on.  Just to be sure,  I did run a power supply across the batteries at 5.4v, but it didn't help. I was trying to decide if it was a loading problem or something in the supply itself.  I haven't finished checking all the caps with the ESR meter.  The waveforms would help, I've got a scope.  Any suggestions on what to pull to see if it's a loading problem downstream of the supply?

[ModemHead]

The only reason I questioned the battery status is that a healthy pack will usually rise to 1.4V per cell under trickle charge.

My memory isn't good enough to advise on break points, I need to look at one of my units first. I'm pretty gonzo about those things, I'd just start cutting tracks.    But do continue to check the caps and also the diodes in the area, esp. CR4.

[tango17]

Actually,  CR4 was next on my list.  I'm glad we're on the same wwavelength. Then I'll  finish the cap testing.  Under load, the NiCds should be about 1.2, per cell so the 4.8 differential looks ok to me.  Then I'll cut some traces to see if it's a loading problem. However,  my  first thought was that a loading issue wouldn't affect both the positive and negative supplies together.  That's why I posted the supply circuit...  suspecting it was probably in there somewhere.  Plus,  it might have been damaged when it ran without the batteries connected. 
Thanks for the help!

[ModemHead]

Ah, you had the batteries uner load, I was assuming we were on mains, with current being pumped into the pack by the LM317.

This power supply is an interesting animal.  It's a basic inverter circuit for the +13 and +6 rails from the secondary. But it also uses the primary for a boost to get 5V above the battery, thru CR4, VR2, and C22. Since its all inter-related, I figure a problem on one rail would affect the others.  Given  your measurements, my intuition tells me to be most concerned about the -5.  Could easily be wrong...

[tango17]

Testing in circuit with the esr meter,  C20 seems to be open.  That could be it.  CR4 tests good.  When I can stand/sit again,  I'll swap it out and report back.

Nope,  it wasn't C20.  On closer inspection, I'd measured the wrong terminals.  C20 is ok.

[tango17]

So where would you go next?  Cut the -5 line after C22 to see if it's a loading problem there?

[eugenenine]

I have one of these and even when working properly its a simple truckle charge circuit with no shutoff so it kills the batteries every few years.  I finally removed the batteries and mounted a regulator so it works like the non battery option.

[tango17]

Yes, I know about the battery removal conversion and the constant charge current design of this model.  I thought about converting it,  but I was buying some NiCds for another project,  and it only cost $6 to get a set for the 8050, too.  It's occasionally useful to be able to run it on batteries. If the display goes,  I'll do the led conversion and that will probably take more power,  making the battery life shorter,  so I'll remove the batteries then. That's assuming I figure out what is wrong with it now.

CR4 reads about 3.8v reverse biased on the meter.  I think its time to connect the scope and get more info.

[ModemHead]

So where would you go next?  Cut the -5 line after C22 to see if it's a loading problem there?

Well it is a quick way to answer the loading question, at least on that rail.  Being that I'm both gonzo and also cautious, I'd probably cut the whole supply loose, to avoid any possible damage from energizing inputs on unpowered devices.

Just saw your latest... I should be able to get you some traces Sunday afternoon.

[ModemHead]

I have one of these and even when working properly its a simple truckle charge circuit with no shutoff so it kills the batteries every few years.  I finally removed the batteries and mounted a regulator so it works like the non battery option.

I have one unit that I converted to AC-only by modifying the LM317 circuit from a current regulator to a voltage regulator.  Works OK.

[eugenenine]

My problems was it only runs from batteries.  If they had designed it to run while charging I might have kept the battery option.

[tango17]

CR4 is switching at about 5-6 kHz.  For 75 uS it's forward biased.  It then goes about 8v reverse biased and drops to about 6v over the next 100 uS.  Looks to me like it's a loading issue.  I'll start removing loads. 

[JFJ]

... If they had designed it to run while charging I might have kept the battery option.

It was designed to be run while charging*:



When the power button is pressed, half of the switch connects R25 in parallel with R26 (re: diagram in opening post). Thereby, increasing the regulator (U26) output current by 100mA, which should be sufficient to power the meter. Initially, my secondhand Fluke 8050A wouldn't operate while charging, but replacing R25 (which had gone high) cured the problem.

*The charge current while running the meter is sufficient to maintain the battery pack charge level, but not to increase it.

[tango17]

The+13v rail has steady dc, if a bit low.  The others all have the same 60/40 duty cycle at 6 kHz as CR4.
Following photos are all 200mv/div AC coupled 50uS/div. CR4 is forward biased for the 40% period of about75 uS.
Photos are +6v, then - 5v,then - 10v.

[ModemHead]

The ripple looks excessive, esp. on the -10V.  Have you considered just replacing the caps?  They're old enough and cheap enough to do that as a matter of course.

[tango17]

Well,  I've checked esr on most of the pwr supply caps and they seem fine.  In theory,  the -10 is 4.8 offset from the -5 rail,  so why the big ripple difference? Also,  that ripple is on a DC average that's way off the nominal +6 and -5.  Something's going on. 

It makes no difference if iit's running on batteries or batteries plus AC.  As expected,  there's lots of ripple on the battery pack.

Notice how  the -10v levels out.  TThat's happening at about 8v plus.

[JFJ]

Following photos are all 200mv/div AC coupled 50uS/div.

The attached photographs are a repeat of your test point checks, on my Fluke 8050A, using the same V/div and Time/div settings.

NB My meter's power converter operates at a frequency of 25 kHz (measurement shown in last image).

[tango17]

Thanks for the photos.  Removing loads has been tricky, but so far I've seen no improvement.  Your frequency is 4 times mine. I know C20 is good. Do you think I could have a problem with Q5/Q6?

[ModemHead]

Had some difficulty getting non-blurry shots off of an analog scope.  The frequency is modulating a tiny bit at about 2Hz.  The trace looks like it has a heartbeat.  I assume this is the variation in loading as the ADC is doing it's thing.

Running from freshly charged battery pack:

TP2: 14.4V
TP3:  6.1V
TP4: -5.3V
TP5:-10.7V

Oscillator frequency here is about 33kHz. 

For the following shots, note that I was using 100mV per division instead of 200mV.  The ripple on the +6V was too low to be interesting, so no shot of that.  The last shot is the collector of Q6 (the tab), which is essentially the voltage across CR4.

[tango17]

That last shot of CR4 looks a lot like mine,  except the frequency of yours is higher and mine sags down the voltage more.  I've removed the display board,  with no effect,  and I'm  cutting traces, but I'm having trouble finding the right spot for -5. I guess its a loading issue,  but I admit I don't fully understand how that  affects frequency.

Thanks to both of you for the help. It's odd that my only symptom I saw was high readings.  I later saw the display start to fade out after a few minutes of operation, but I figured that was a typical bad LCD display issue. While testing,  I ran down the batteries, so I connected 5.3v to replace them,  and saw 500 mA draw,  which seems high to run the meter,  so I'm back to a loading problem.

I'm only able to sit/stand for short periods, so it will be slow going before I can report back, but I will. Again,  thanks!

[ModemHead]

While testing,  I ran down the batteries, so I connected 5.3v to replace them,  and saw 500 mA draw

500mA?   

This one is averaging 143mA off the battery pack, which is at 5.07V at the moment.

Sorry to hear of your sitting/standing difficulties, I know that must be frustrating.

[ModemHead]

Quick experiment, I added 100mA extra load on the -5V and the frequency dropped to 16.7kHz.

[tango17]

I'd previously heard that 150mA was about the draw to be expected,  so,  given the 500mA I saw,  it's looking more like a loading problem. The frequency drop you saw under heavier load also points that way.

We had a power outage all day,  so nothing yet,  but I'll get the power supply isolated eventually,  and verify it's working,  then track down the load.