Fluke 8845a gets surprisingly warm.

[sigmatechnica]

So I scored a fluke 8845a on ebay for a good price.  Not done any thorougher tests yet but it seems to work fine.  Only thing is it gets what i would call 'surprisingly warm' after a few hours, even switched 'off' (with soft power button)  Had the cover off and it seems to be mostly the transformer/voltage regulator/capacitor area.  No signs of any discoloring form heat, but things are definitely hotter than i'd like if i was designing such an instrument. 

Was just wondering if anyone else's also got warm?  I'd rather not have to take the board out on this 6.5 digit meter that's probably the most accurate thing i own now unless it's really necessary, then again  i'd quite like it not to go bang or catch fire either    (like my Solartron did  )

Thanks!

[Specmaster]

I don't about the 8845a but I do have a 8840a and a 8842a which don't get warm at all. As for the Solartron, which model was it? I have 2 x Solartron 7150 plus's and they have Schaffner Filters built into the mains' inlet socket which will explode and catch fire from time to time as they age unless you check them from time to time and replace when they start getting around 20 years old,

[bdunham7]

My 8846A has been in standby all day (8 hours) and is just barely noticeably warmish near the transformer, pretty much stone cold elsewhere.  It measured 88F near the transformer with a temp gun in an 81F ambient temp.  I haven't noticed it getting any warmer in use.  What voltage are you running at and do you have a way of measuring how much power it is using?

[sigmatechnica]

As for the Solartron, which model was it?

yes it was the 7150.  i found out about the faulty mains filter when it let all the smoke and flames out!  fixed it of course but all that sticky black tar all over the insides, followed by scrubbing all the PCBs with soapy water probably didn't help the calibration much!

What voltage are you running at and do you have a way of measuring how much power it is using?

running at 240v.  measuring the power is a good idea,  power meter ordered!  this is definitely getting warmer than that though.

[Specmaster]

I would not recommend using soapy water for cleaning PCB's, use instead 70% IPA and then make sure that it is totally dry before powering up again as IPA is flammable.

[Specmaster]

You may find this link useful to you. I would advise getting an infrared temperature probe to take spot readings from various components to see where the excessive heat is coming from.

https://xdevs.com/fix/f8846a/#manual

I would also ensure that there is at least 10mm airspace above and below the meter if it is in a stack and at least 20mm on the sides to allow some airflow around and through the meter otherwise you are trapping any heat inside.

It seems that there are no service manuals available for this meter and therefore no schematics either, but the area that will generate the most heat is going to the power supply, so that is the transformer and maybe the rectifiers / diodes and or the smoothing caps if they are beginning to fail.

[bdunham7]

running at 240v.  measuring the power is a good idea,  power meter ordered!  this is definitely getting warmer than that though.

Fluke says 12 watts.  I would measure mine, but there will be enough variance at 50Hz that we probably won't match.  I'm assuming you are at 50Hz.

Is the line voltage block set correctly? 

[drussell]

If it is getting significantly warm, even in standby, something is probably wrong.

A shorted turn or two in the transformer could do that, for example.  It will still work, but the transformer is always seeing an extra, parasitic load.

Measuring the total power draw, especially in standby where it should be drawing virtually nothing, is indeed probably the best course of action for a first step.  (Well, after double-checking the voltage settings, that is... Like is it set for 220V instead of 240V?)

[sigmatechnica]

Thanks for all the replies everyone, you have all been super helpful!  An update:

It's just stopped doing it.  Had the cover off this morning trying to locate the hot bit without much luck. checked the voltages in case something was amiss, which it wasn't.  it was only getting slightly warm.  put the cover back on and still only slightly warm.  Power meter arrived so i measured it, 15W on and 10W at idle.  Checked the voltage selector and actually it was on 220v, so changed that to 240 and predictably now it draws 9w at idle.  not enough of a difference to be the cause but can't have helped. 

The only thing i can think is maybe the caps needed to reform and everything got a bit warm in the process.  Apparently it had been in storage for some time,  still surprising on a relatively modern instrument with quality caps though.  In any case i'm content to let it be at this stage, old adage of not fixing that which is not broken being at the forefront of my mind with all those fragile looking fancy custom resistors and guard traces all over the place. 

Now some fun, assuming this is the most accurate thing in the lab now i did some comparisons on the voltage range
the Brymen 867s is surprisingly good, almost to all it's digits in standard mode
the Solartron that caught fire isn't too bad either considering what i subjected the PCBs to getting all the sticky black tarry residue off it, good to 3 decimal places anyway

[bdunham7]

Actually I think you fixed it by setting the voltage correctly.  I haven't tried this on the 8846A yet, but I know on the 8840/42A models they tolerate overvoltage up to a point, then the current starts to increase a bit more steeply.  I presume this is overvoltage protection in the regulator somewhere and not the transformer saturating, but I don't know.  Eventually the fuse blows, but until that point, they can just draw 2-3X normal current indefinitely.  I suspect your line voltage was just a bit higher before and this morning perhaps it was just under that threshold.  Now with the correct setting that threshold will be 20 volts higher.  I doubt it has anything to do with capacitors.

[Specmaster]

It is difficult trying to get say a Brymen 867 to agree exactly with a Solartron 7150 beyond 3 digits as they employ different timings etc, from memory the 867 takes far more reading per second than the 7150 so there in itself is just one reason for the slight variation. I personally would say that 3 digits is for 99,9999% of the time good enough anyway.

[Fretec]

As for the Solartron, which model was it?

yes it was the 7150.  i found out about the faulty mains filter when it let all the smoke and flames out!  fixed it of course but all that sticky black tar all over the insides, followed by scrubbing all the PCBs with soapy water probably didn't help the calibration much!

This happens with all of them. If you have an old 7150 (or any of that series) that didn't blow up the mains filter yet, replace it asap!
It will blow up and make a huge mess, and the smell is really not pleasent.

You can still get replacement ones of the same size that fits perfectly.