-
While continuously monitoring a 900Vdc supply for almost an hour, using a Fluke 87v and a Keysight 1282a as backup, the programmed output rose to 1300Vdc instead of dropping to 500Vdc by operator error.
The 87V registered OL above 1100Vdc while the Keysight continued to read the actual output. The 87V output then blanked out, and when the supply was shut off, began to register correctly, voltage <=600Vdc. The Vdc range was off by 1% and the ohms range was increasingly inaccurate >= 1Meg ohm. As the measured output exceeded 700Vdc, the 87V readings began to drifted down and sped up the higher the input voltage was.
Postmortem revealed blown input varistors.
I could not locate an exact replacement part and nearest values were obtained as discussed in an old eevblog thread.
https://www.eevblog.com/forum/testgear/fixing-a-fluke-87v/
There is a size difference but there was enough space around the PCB to force fit the larger varistors in the space of the original parts, without making alterations to the blast shield or the case.
In the final repair not pictured, the expose varistor leads are covered with tape to further reduce arc risk.
Function check of the 87V now shows it operating as new, but surprisingly better: it seems faster in response and accuracy is now spot on to the 1282a Keysight.
This Fluke 87V was not purchased new, but operationally like new and refurbished by me 10 years ago. It may have been exposed to high transients in the past to allow its varistors to fail early or the Fluke varistors are overly sensitive. The Keysight meter was unharmed.
-
Nice repair!
-
I don't remember what the meter looked like on the front end. Looks like the V input goes through the PTC/resistor before it hits the MOV. Looks like on MOV shared for the common return. What limits the current for the second leg, third MOV? Were the two black ones damaged and the blue fine? Had the back ones been replaced before? Maybe they were the wrong value? The last one I had ran, all the MOVs were blue.
-
Hi Joe
Yes, its roughly similar to the old Dave video on the Fluke 27 input protection:
I do not know if the desoldered blue MOV is still fine, but given all the MOV are across the inputs, they were exposed to the same insult, and will be bound later to fail. However, I will destructively test them later and report back, Imcurious to know what state they are in; I'll also sacrifice one of the replacement MOVs to see what will trigger a failure.
When I first 'refurbished' my 87V 10 years ago, all the MOV appeared "blue" virgin from the factory.
Of interest, at this link, from another user back in 2013, my failure is the same two MOV.
https://www.eevblog.com/forum/testgear/fixing-a-fluke-87v/
From another user, the native undamaged MOV area:
I could find no equivalent 575V varistor that was ~ 5mm in diameter, this a fairly uncommon part. At most it was 420V. Given the 87V failed at ~ 1300VDC with the 575 "blue" rated MOV, using a reduced working voltage 420V MOV could make it fail sooner, unless, there is something about the part Fluke uses.
Overall, I think Fluke made a safe choice, because 5mm MOV at ~14J will blow less violent: burning, heat and smoke. But the bad side, is it could be a bit sensitive as it has less reserve if sees a high voltage transient, which maybe my case. Regardless, Fluke's lifetime warranty won't cover usage damage as its for product build defects only.
In this video starting at 16:00, the 87V is shown repeated hit with high voltage transients, but while it appears to function, whether its remains factory accurate is unknown but suggested by Fluke literature. At most one can say it can take a transient and not fall apart. In my case, the meter was functional and appeared undamaged, but not factory accurate, which is useless from a DMM point of view.
To diagnose the issue, I desoldered one foot of one MOV and it returned to factory accurate, so the in series resistors/PTC were unaffected. It stands to reason that if a transient were to partially damage any input MOV, the Fluke would start to act 'out of factory' since MOV across the input would affect the input impedance. I suspect this was the case as since repaired it seems to be faster or solid reading that in the past, and I use it fairly often.
The replacement I used are 3-4x bigger and rated 65J, if they blow it can burn with greater energy and likely damage nearby non affected components compared to the more isolated placed MOV from the factory.I don't remember what the meter looked like on the front end. Looks like the V input goes through the PTC/resistor before it hits the MOV. Looks like on MOV shared for the common return. What limits the current for the second leg, third MOV? Were the two black ones damaged and the blue fine? Had the back ones been replaced before? Maybe they were the wrong value? The last one I had ran, all the MOVs were blue.
-
Quote
The replacement I used are 3-4x bigger and rated 65J, if they blow it can burn with greater energy and likely damage nearby non affected components compared to the more isolated placed MOV from the factory.
One trick I have seen in some dmms is to put heat shrink around the body of the MOV. Leave the top of the MOV uncovered to allow venting at the top. -
Quote
The replacement I used are 3-4x bigger and rated 65J, if they blow it can burn with greater energy and likely damage nearby non affected components compared to the more isolated placed MOV from the factory.
One trick I have seen in some dmms is to put heat shrink around the body of the MOV. Leave the top of the MOV uncovered to allow venting at the top.
I've damage a fair number of PTCs during my testing. Some companies use piss ant PTCs that don't seem to do well with the higher voltages. I have seen companies use heat shrink over the PTCs but I don't think I have ever seen it used on MOVs. What meter/s have you seen this? -
Rather than heat shrink I would use some fibreglass reinforced silicone sleeving, to provide extra protection from blast fragments flying around and safely constraining the explosion of the PTC or MOV during failure. Just a small bit that is a push fit over the unit possibly held in position with a small blob of silastic so the sleeve and the MOV/PTC leads do not fracture when dropped.
-
Hi Joe
Yes, its roughly similar to the old Dave video on the Fluke 27 input protection:
...
I do not know if the desoldered blue MOV is still fine, but given all the MOV are across the inputs, they were exposed to the same insult, and will be bound later to fail. However, I will destructively test them later and report back, Imcurious to know what state they are in; I'll also sacrifice one of the replacement MOVs to see what will trigger a failure.
When I first 'refurbished' my 87V 10 years ago, all the MOV appeared "blue" virgin from the factory.
..
I could find no equivalent 575V varistor that was ~ 5mm in diameter, this a fairly uncommon part. At most it was 420V. Given the 87V failed at ~ 1300VDC with the 575 "blue" rated MOV, using a reduced working voltage 420V MOV could make it fail sooner, unless, there is something about the part Fluke uses.
Overall, I think Fluke made a safe choice, because 5mm MOV at ~14J will blow less violent: burning, heat and smoke. But the bad side, is it could be a bit sensitive as it has less reserve if sees a high voltage transient, which maybe my case. Regardless, Fluke's lifetime warranty won't cover usage damage as its for product build defects only.
...
To diagnose the issue, I desoldered one foot of one MOV and it returned to factory accurate, so the in series resistors/PTC were unaffected. It stands to reason that if a transient were to partially damage any input MOV, the Fluke would start to act 'out of factory' since MOV across the input would affect the input impedance. I suspect this was the case as since repaired it seems to be faster or solid reading that in the past, and I use it fairly often.
The replacement I used are 3-4x bigger and rated 65J, if they blow it can burn with greater energy and likely damage nearby non affected components compared to the more isolated placed MOV from the factory.
When I was modifying the UT61E, I ran into the same problem trying to find smaller body MOVs with low enough capacitance and high enough breakdown. When you replaced the MOVs the first time, any idea what caused the damage and what where the black ones you replaced them with? Why did you switch from the black to the red ones you now have instead of just getting a new set of black ones?
I've yet to see a MOV get damaged to the point where it throws the meter out of calibration. But I test at such low levels and the MOV is not normally directly absorbing the pulse (except once the PTCs arc).
-
Quote
I don't think I have ever seen it used on MOVs. What meter/s have you seen this?
Off the top of my head, cannot remember.
Maybe, I thought the covered device was MOV instead of PTC.
In any case is it bad to do this to MOVs?
If a MOV fails catestrophically, is it bad to try and direct the resulting spray if debris? -
Quote
I don't think I have ever seen it used on MOVs. What meter/s have you seen this?
Off the top of my head, cannot remember.
Maybe, I thought the covered device was MOV instead of PTC.
In any case is it bad to do this to MOVs?
If a MOV fails catestrophically, is it bad to try and direct the resulting spray if debris?
I am not sure. You would really need to look into it. My guess is you will find it is related to how the MOV is used and the MOV itself.
In the case of the handheld meter, I would like to think that the cases are designed to contain any sort of blast (assuming the meter is being used in it's rated environment) but I doubt very much this is the case. But I also doubt a MOV coming apart inside a meter would cause as much pressure as copper going from a solid to a vapor.
In all of the meters I have looked at except one, if they use MOV/s they will have one ore more limiting devices in front of them. I saw one meter that had a placeholder for a MOV that was directly across the inputs but it was not populated. With enough voltage, you may breakdown or arc over the limiting parts and the MOV/s could see the full force of the line.... A 20J device no longer seems like it would do much.
IMO, the MOVs, PTCs, fast clamps and such are there to lower the meters failure rate more than anything. Conductor spacing, material selection, case design and such are there for safety. Again, just my take on it.
-
He he, here we are