MOV Failures

[SG-1]

I need some advice.  I am having trouble with this product. We have been using it for years with no issues.  The vendor says no changes have been made in the last 7 years. Look at the CTD-2 model.

http://www.electromagnetic.biz/Portals/0/Docs/CDT/CTD1%20&%202%20UL.pdf

The MOV is failing in a few units, maybe 1 out of  50 or a 100.  This is 20 or 30 units so far.  We are experiencing failures on our test floor. Some fail when first energized, others function a few times, then the output goes to 20VDC or 30 VDC some down to 3VDC. The units were removed & bench tested. The output remained at 20 or 30, or 3 VDC.  The unit is replaced & the second one works just fine. No external wiring defects.  This device is a critical component of our product.  No failures can be tolerated.

I have some specs for the MOV & capacitor.

MOV is a EPCOS / Type (untaped) SIOV- S14K150 / 200 VDC /4500 Imax (8/20 micro S)  / 40.0 Wmax (2ms)  / Pmax W .60

Capacitor is electrolytic 1500 micro-F  250VDC

Thermister reads 20 ohms at room temperature. 

The vendor is offering to replace the MOV with a 175 volt model. Or any thing else we desire, for a cost, of course.

What think ye ??

Thanks in advance for your help.

[migsantiago]

Hello

We've had problems with big capacitors. They were sent from Israel and when they arrived to Mexico City, they were leaking oil. I thought that the problem was that the pressure in Mexico City is different to the air pressure in Israel (or China, the original manufacturing place). The guys who sent them had no problems with them.

Just my 2 cents, check if the pressure rating is good enough for your capacitors.

[Feanor]

I am also guessing capacitors leaking current. Could be just about any element in this circuit though.

MOV, capacitor, freewheel diode : - If anyone of those is allowing current to pass then your output voltage may drop.

You really need to open one up, get a probe in there and see what all the voltage levels are. Then cut out the electrolytic cap and replace it with one you know is good.

Test again.

Cut out the MOV and test again. Usual debugging of problems just remove one element at a time till you find the problem.

Obviously this will be connected to the mains so take due precautions. Good luck!

[SG-1]

The problem child has been identified as the MOV. They are shorting out, cause unknown. cut one lead loose from the circuit board and the Cap Trip Device is back in business.  The vendor is replacing them & then sending them back to us.

[SG-1]

I am also guessing capacitors leaking current. Could be just about any element in this circuit though.

MOV, capacitor, freewheel diode : - If anyone of those is allowing current to pass then your output voltage may drop.

You really need to open one up, get a probe in there and see what all the voltage levels are. Then cut out the electrolytic cap and replace it with one you know is good.

Test again.

Cut out the MOV and test again. Usual debugging of problems just remove one element at a time till you find the problem.



Obviously this will be connected to the mains so take due precautions. Good luck!

I do get your point, but I do not understand how a leaking cap will damage the MOV.  We think the vendor got some bad MOVs. The vendor says we are doing something wrong.  He has repaired about 20 or 30 of these so far by replacing the MOV.  He wants to up the voltage on the MOV to 175VAC RMS to stop the unknown problem, that he believes is voltage spikes or transients.

I suppose my question really is what is the highest value for the MOV so it can still protect the downstream componets ??

[Feanor]

If the capacitor is "leaking" (drawing current to ground) the MOV will not be damaged you are right. It will mean a voltage drop across the thermistor and that will reduce your output voltage.

As for is a 175V MOV enough protection or not? This is not my field so hopefully someone else also answers this.

Your capacitor is rated at 250V so I am guessing that will be ok. Guessing again but the diodes are probably rated for well over 175V so they should be fine. That just leaves your solenoid, the device you are switching on and off.

And of course all of this depends on the type of transient being dealt with. If it gets hit by lightning then probably the whole lot will go up in smoke anyway.

[Simon]

what mains voltage are you supplying this thing ? even with 120VAC RMS supply you are looking at 170V peaks, usually things like MOV's (I'm not wholly familiar with MOV's) will start conducting something more or less small at a little under the rated voltage, the rated voltage being the clamping voltage = the maximum voltage the mov will hold at. I think the MOV voltage is too low

[Zero999]

Check the datasheet, MOVs are normally rated way below their clamping voltage. Some also have their rating given in RMS not DC.

[SG-1]

If the capacitor is "leaking" (drawing current to ground) the MOV will not be damaged you are right. It will mean a voltage drop across the thermistor and that will reduce your output voltage.

As for is a 175V MOV enough protection or not? This is not my field so hopefully someone else also answers this.

Your capacitor is rated at 250V so I am guessing that will be ok. Guessing again but the diodes are probably rated for well over 175V so they should be fine. That just leaves your solenoid, the device you are switching on and off.

And of course all of this depends on the type of transient being dealt with. If it gets hit by lightning then probably the whole lot will go up in smoke anyway.

You are correct, the thermister is choking off the output voltage.  The MOV mode of failure is a short or very low resistance.  The solonoid ( actually a clapper ) is very robust.  It is only activated to open a large circuit breaker.  It is energized for 34 milli-seconds & then de-energized again.  Of course there can be some very nasty things happening during that time on the medium voltage side.

[SG-1]

what mains voltage are you supplying this thing ? even with 120VAC RMS supply you are looking at 170V peaks, usually things like MOV's (I'm not wholly familiar with MOV's) will start conducting something more or less small at a little under the rated voltage, the rated voltage being the clamping voltage = the maximum voltage the mov will hold at. I think the MOV voltage is too low

The normal supply voltage is 120VAC RMS.

The MOV voltage rating of 150V is given in RMS.  I see I missed that in my OP. 

[Simon]

maybe you need to remove a mov and test it separately, if it's faulty/incorrect you need to ask your supplier some serious questions

[SG-1]

maybe you need to remove a mov and test it separately, if it's faulty/incorrect you need to ask your supplier some serious questions

The MOV has all the right numbers printed on them.  I like your suggestion, then send it back to him for replacement, either way.  We have been beating on him pretty good. My boss just keeps telling him that he is not the only supplier.  I wanted him to test the unit at 150VAC RMS for a few seconds.  The inital reply was that any more than 130 would damage something ( unspecified something ).  He then asked if our spec had changed.  I then clobbered him with his own spec from his own website. Clearly states the device can withstand 140VAC RMS continously.  They will now raise the test voltage to 145 VAC RMS.  Will they have any in-house failures ?  Would they tell me if they did ??  They seem willing to work with us.  I think they may have been testing at 120 volts...

I found a 14 volt MOV here at the house & ran an experiment of my own to see where the clipping starts. 

The circuit consists of the MOV in series with a 10K resistor.

V _INPUT                     V_MOV

14.0                         14.0
15.0                         15.1
18.0                         17.4
25.0                         19.9
92.5                         24.6  resistor smoking, smells bad in here
155                          25.4  need new resistor

155 volts is the top of that power supply.  I think I can get more out of the Rotek.  Plenty of resistors in-house.  Actually for that last shot I wrapped the resistor in some C130 rubber splicing tape, nice thick tape. Then added a wrap of Permecil, 600 volt tape that is super tough.  The resistor still smelled bad after the pop.

[Time]

The obvious answer is to test the MOVs by themselves.  I do not know why this thread made it this far without actually doing so!

[Simon]

you needed a proper QC inspector to point you on you way 

[Zero999]

Your tests reveal that the MOV starts clipping around 17V which is approximately the rated voltage plus 20%; this confirms what I initially thought so if the MOV is rated to 130VAC, it shouldn't clip until 156VAC or 221VDC.

Did the MOV get hot?

What power rating was the resistor? Let me guess 0.25W?

No wonder 130V (1.7W) killed it, next time use 9 resistors, three paralleled stings of three in series. Resistors may be cheap enough to be disposable but operating them above their maximum ratings can change their value causing the current to be unpredictable.

[SG-1]

The obvious answer is to test the MOVs by themselves.  I do not know why this thread made it this far without actually doing so!

The vendor says he can do this, but it will result in a substantial price increase.  I think if he uses the correct test voltage for the finished product that would suffice. 

[SG-1]

Your tests reveal that the MOV starts clipping around 17V which is approximately the rated voltage plus 20%; this confirms what I initially thought so if the MOV is rated to 130VAC, it shouldn't clip until 156VAC or 221VDC.

Did the MOV get hot?

What power rating was the resistor? Let me guess 0.25W?

No wonder 130V (1.7W) killed it, next time use 9 resistors, three paralleled stings of three in series. Resistors may be cheap enough to be disposable but operating them above their maximum ratings can change their value causing the current to be unpredictable.

The  MOV stayed at room temperature.  I was checking it with an IR therometer.  The MOV is good for 1 watt.

You guess good, 1/4 watt.   It survived the 130V.  The 155V finished it off.  It reads 9.7 Mega-ohms now.

I have a 130V MOV to play with also.  I need to calculate the watts this time for sure.  First I am going to see if I can generate some switching transients and try to see the difference with & without the 14V MOV.  Maybe a light switch & capacitor ? At the the 14V level.

It will be Tuesday before I can do this.  Going to the mountains tomorrrow for the day.  Looking for fall folage.

Thanks !

[Simon]

to make spikes turn a large fan on and off or other inductive load, i got 400V spikes out of a 24V system once. The vendor should not be charging you extra for testing, if his products have suddenly gone bad that is his problem not yours

[TechGuy]

I would recommend checking the devices with failed MOVs for high voltage spikes (without the MOVs connected). If your device has a coil in it (inductor, motor, smps transformer) it could be causing excessive voltage spikes. The MOV will suppress the spikes for a period, but MOVs are perishable devices. Every time a MOV clamps a voltage spike it will degrade, to the point it completely fails. MOVs are nothing more than a Zinc Oxide semiconductor that transistions from a high impediance state to a low impledance state when encounters a high voltage pulse. Over time the zinc oxide will break down, usually in a state of low impediance.  If this is indeed the case, then you will need to alter the device to including snubbers to attenuate the voltage spikes so that they don't trip the MOVs. You may wish to to include a small snubber in your design to clamp small line spikes in your device to extend the life of your MOVs. a small RC snubber will cull all of the lower power spikes so the MOVs don't have to trigger everytime something with a motor turns on (ie refig, HVAC, etc).

I have a 130V MOV to play with also

Remember that 120 VAC is RMS voltage not peak voltage. 120VAC real peak voltage is 170V! (1.414*120V = 170V) A 130V MOV will burn out if you connecting to AC mains. A proper voltage rating for an AC Mains MOV is 180V.

The MOV is good for 1 watt

Not a lot of room to handle a moderate spike. You need at some kind of current choke such as an air core inductor to slew the spike, otherwise your MOV is a Wiley E Coyote with a lit stick of dynamite. I would recommend using a more robust MOV something in the 100J to 200 J range.

FWIW, I recommend avoid using MOVs in high reliable devices since they are perishable devices. TVS diodes in the correct places provide a reliable way to clamp TVS events since the do not degrade.A  TVS diode and a PTC fuse makes a good combo. The TVS diode clamps the spike the current inrush forces the PTC fuse into a high-impediance state attentuating the current flow to protect the TVS diode. For high power application I use a set of Air core Inductors, TVS diodes coupled with a Cap. A secondary Crowbar circuit with a higher TVS clamping voltage provides secondary defense. The first line of defense is the TVS diode Cap combo, which clamps small to moderate spikes.
The Air core inductor serves as a choke to slow the change of current flow to avoid blowing the TVS diode. An air core inductor is more desirable because it will not saturate and you don't need a very large inductance. If the cap becomes fully charged, the input voltage will rise and trip the secondary crowbar circuit to clamp any all voltage. However when the crowbar trips, it usually blows the fuse as an SCR used in a crowbar circuit will not shut off until the voltage drops to zero. So its undesirable to make the crowbar circuit as your first line of defense.

[SG-1]

I would recommend checking the devices with failed MOVs for high voltage spikes (without the MOVs connected). If your device has a coil in it (inductor, motor, smps transformer) it could be causing excessive voltage spikes. The MOV will suppress the spikes for a period, but MOVs are perishable devices. Every time a MOV clamps a voltage spike it will degrade, to the point it completely fails. MOVs are nothing more than a Zinc Oxide semiconductor that transistions from a high impediance state to a low impledance state when encounters a high voltage pulse. Over time the zinc oxide will break down, usually in a state of low impediance.  If this is indeed the case, then you will need to alter the device to including snubbers to attenuate the voltage spikes so that they don't trip the MOVs. You may wish to to include a small snubber in your design to clamp small line spikes in your device to extend the life of your MOVs. a small RC snubber will cull all of the lower power spikes so the MOVs don't have to trigger everytime something with a motor turns on (ie refig, HVAC, etc).

I have a 130V MOV to play with also

Remember that 120 VAC is RMS voltage not peak voltage. 120VAC real peak voltage is 170V! (1.414*120V = 170V) A 130V MOV will burn out if you connecting to AC mains. A proper voltage rating for an AC Mains MOV is 180V.

The MOV is good for 1 watt

Not a lot of room to handle a moderate spike. You need at some kind of current choke such as an air core inductor to slew the spike, otherwise your MOV is a Wiley E Coyote with a lit stick of dynamite. I would recommend using a more robust MOV something in the 100J to 200 J range.

FWIW, I recommend avoid using MOVs in high reliable devices since they are perishable devices. TVS diodes in the correct places provide a reliable way to clamp TVS events since the do not degrade.A  TVS diode and a PTC fuse makes a good combo. The TVS diode clamps the spike the current inrush forces the PTC fuse into a high-impediance state attentuating the current flow to protect the TVS diode. For high power application I use a set of Air core Inductors, TVS diodes coupled with a Cap. A secondary Crowbar circuit with a higher TVS clamping voltage provides secondary defense. The first line of defense is the TVS diode Cap combo, which clamps small to moderate spikes.
The Air core inductor serves as a choke to slow the change of current flow to avoid blowing the TVS diode. An air core inductor is more desirable because it will not saturate and you don't need a very large inductance. If the cap becomes fully charged, the input voltage will rise and trip the secondary crowbar circuit to clamp any all voltage. However when the crowbar trips, it usually blows the fuse as an SCR used in a crowbar circuit will not shut off until the voltage drops to zero. So its undesirable to make the crowbar circuit as your first line of defense.

Here is the device.  The MOV is supposed to protect two diodes & a capacitor.
http://www.electromagnetic.biz/Portals/0/Docs/CDT/CTD1%20&%202%20UL.pdf

Here is the data sheet for the 130V MOV.  The maxium continous voltage is given in RMS.  Or am I reading this incorrectly ??
http://www.alliedelec.com/Images/Products/Datasheets/BM/LITTELFUSE/846-0244.PDF

The manufacturer is using a .6 watt (continuous )

I have sent an e-mail to the manufacturer asking, why not a transil ??
I will pass along your recomendation for not using the MOV in a high reliability devices, that they degrade every time they clamp.

Thanks

[SG-1]

to make spikes turn a large fan on and off or other inductive load, i got 400V spikes out of a 24V system once. The vendor should not be charging you extra for testing, if his products have suddenly gone bad that is his problem not yours

I agree with your premise, but if his device goes bad in our product, the customer will blame us.  We will pay the warrenty cost to replace each unit.  This is like the switch in your car that controls the window going up & down.  The car manufacturer did not make the switch, but if it dies under warrenty he replaces it at his expence.

[Simon]

you should be passing onto your supplier any costs you incur in fixing the issue, if they sell you faulty gear and it causes £££ of damage that is theirs to pay for

[FreeThinker]

Just my 2 cents but have you checked the mains supply?.High voltage harmonics may be a possible cause.

[PetrosA]

I'm with FreeThinker on this. I think some power quality analysis is in order at point of use on one or more of the customer sites who have had your product fail. At minimum, I would want to see some data logs for voltage, even better, a fuller analysis with harmonics shown. At the least, this will show whether the component is being exposed to conditions outside its rating or not and if not, will go far in proving that it's a problem with the component itself.

[FreeThinker]

<snip from wikipeadia>While a MOV is designed to conduct significant power for very short durations (? 8/20 microseconds), such as caused by lightning strikes, it typically does not have the capacity to conduct sustained energy. Under normal utility voltage conditions, this is not a problem. However, certain types of faults on the utility power grid can result in sustained over-voltage conditions. Examples include a loss of a neutral conductor or shorted lines on the high voltage system.<snip>
The neutral conductor in modern times cannot be relied on to remain at 0 volts WRT earth.This can cause significant voltage fluctuations for short periods of time causing havoc with voltage sensitive devices, and it's getting worse! see thread in this forum https://www.eevblog.com/forum/index.php?topic=1561.0