Fluke 89 dim display

[dexar]

I have a Fluke 89 where the display is faded. All the digits are equally faded. Turning on the backlight max on makes the digits almost unreadable. The strange thing is that it looks good in some modes (capacitance) but very dim in other (VDC). I have replaced the elastomer and display. I also cleaned the mode switch contacts and display contacts but nothing helps. Anybody have any tips?

[alm]

You're not providing many details, but I assume you're talking about the Fluke 89-IV, the predecessor to the Fluke 189? The symptoms sounds like something is dragging down the power rail in certain modes. The first thing I would look at if there is a leaking supercap as has been reported for the Fluke 189, and otherwise I would see if any component is heating up in DCV mode compared to capacitance mode. A thermal camera would be useful for this, since it may not get scorching hot.

[Martin72]

Anybody have any tips?

As ever, pics are telling more than 1000 words could...

[dexar]

Thanks a lot for the tips!

Yes, it is a Fluke 89-IV.

The supercap looks good without any leaking (and measured voltage is 3.1V). Current consumption is 40mA in DCV and 19mA in capacitance mode (when powering using 6V). When looking with a thermal camera the only warm (+15C ambient) component is what I think is a LT1307 DCDC located at the upper right corner in the picture.

I have attached a picture showing the PCB, display with max backlight with a fluke 83-v for comparison and pictures showing the difference between difference modes. This is the only 89-iv I have seen live but I assume that the display should be similar to the 83-v.

[shakalnokturn]

I haven't measured on my 189 for comparison, I think there are a couple of "supercap" related topics that mention current draw for reference, anyway I think that even 19mA is too much.

First try measuring the supply voltages on the tantalum capacitors (IIRC +3.3V, +5V, -5V) I'm expecting the +5V to be low.

Next if supply voltages make sense, I don't see any reason why the CPU would double current draw changing functions, more likely the analog ASIC under the shielding would so I'd try to thermal image that.

[Axtman]

Have you cleaned the zebra strips?

[shakalnokturn]

Have you cleaned the zebra strips?

The zebra strips don't really fit the symptoms here, it doesn't take much to try though.

[Black Phoenix]

Have you cleaned the zebra strips?

The zebra strips don't really fit the symptoms here, it doesn't take much to try though.

Exactly, he could try but I doubt that it is the fault in question.

[dexar]

I have cleaned the zebra strips. Even replaced them and the display.

There is a 3.3v LDO and one DCDC driving a transformer that generates three voltages that I think is +3.3V, +5V and -5V.
The feedback of the DCDC goes to the +3.3V rail.

The LDO 3.3V is 3.3V so that rail seems to be ok.
When I measured the tantals the DCDC rails are +4.8v/-4.8v/+2.7v. When in capacitor mode all the voltage goes up
and I can see that the DCDC is going in and out of burst mode (in contrast to being in constant pwm mode).
It therefore looks like something is taking to much current at +3.3V
when set to some multimeter modes. The ASIC below the shield and the supercap  is just a couple of degrees above ambient.

I also tried to remove the LCD all together but the current consumption is still 40mA@6V.

[shakalnokturn]

Well... A couple of °C above ambient for something that is meant to be low power is significant, to me the PWM and analog ASIC being the warmest components is already saying something.

I've had good results reflowing analog ASICs on fussy Fluke DMMs but none of them were dragging their supply low...

Next thing I'd do in your situation is to try to confirm by resistance or current measurements which supply is overloaded.

It's looking as if the ASIC has been damaged somehow.

[Hamelec]

for me it looks like a foggy acrylic bezel..

[dexar]

Well... A couple of °C above ambient for something that is meant to be low power is significant, to me the PWM and analog ASIC being the warmest components is already saying something.

Next thing I'd do in your situation is to try to confirm by resistance or current measurements which supply is overloaded.

It's looking as if the ASIC has been damaged somehow.

I have now removed the supercap but it still draws 40mA. I agree with you that it looks more and more like it is the ASIC.
Another theory I have is that is a specific IO pin that is shortcircuit and therefore draws to much current.

What however surprises me is that the fluke seems to work despite it drawing to much current. The VDC mode that has a very dim display still measures voltages correctly?!   

[robert.rozee]

what about a filter capacitor having simply gone leaky? this seems like a far simpler scenario.

cheers,
rob   :-)

[shakalnokturn]

Another theory I have is that is a specific IO pin that is shortcircuit and therefore draws to much current.

I've run into that on a DMM before, analog ASIC had been partly destroyed by someone abusing the autoranging Amp input. Mostly worked except for a few functions where battery draw shot right up...

Yes it is quite odd that everything otherwise works on your Fluke.
It could be a specific in or out pin that is shorted but this could be external or internal and when it's internal it's much harder to solve.
As this meter does AC+DC measurements the RMS converter is likely powered as soon as Volts measurements are selected, is the converter external? Does it get warm?

I assume you have checked the switch contacts? I doubt they'd get that dirty that they'd be a heavy leakage across them though.

If you have another Fluke with the same ASIC you could try comparative measurements to try to find the bad pin. My favourite method for this is the Huntron or Hameg or Fluke 867 curve tracer.

A more brute-force approach could be a carefully set lab power supply to up the loaded power rail, carefully watching with the thermal camera. (Just be aware that you could end up with it working less than it is... This approach can either just make things hotter, finish shorting them or even sometimes clear the short.)

I don't really believe in the leaky capacitor, that should have shown on the thermal camera. I could be wrong of course.

[dexar]

I found the cause of the dim display!

It was the DCDC converter. After replacing it the display is as bright and nice as on my Fluke 83-v. I really do not know how I could miss it previously.

For anybody that have the same problem. The current consumption should be about 23mA (differs a little between different modes) with 6V at the battery port.
The DCDC is a Linear LT1307. I accidentally ordered a LT1307B (which does not shift into burst mode) which seem to also work.