MOV Power calculation

[yashrk]

Hey everyone!   

I am writing a blog post on MOV selection and sizing for 8/20 uS Surge test.
The method I normally use

• Calculate the impedance or the series resistance added to the circuit
• Then will calculate the current that is going to flow via the TVS
• Using that current rating and considering the clamping voltage I will select the MOV

But I have also found that energy has also been listed in the MOV's datasheet. So I though may be I should calculate the energy dumped in the MOV to confirm my selection.
So I started to by calculating E = V x I x t
But then I found out due the current waveform for surge test we have to consider a constant K due to the shape of the current waveform
So the surge wave is divided into first ramp signal with t = 8 uS and exponentially decaying signal for t = 20-8 uS
Reference => https://m.littelfuse.com/~/media/electronics_technical/application_notes/varistors/littelfuse_selecting_a_littelfuse_varistor_application_note.pdf - Page number 121
Hence, I did the calculation according to the constants mentioned in the above selection guide.
During same time I found this calculator from Vishay (https://www.vishay.com/resistors/pulse-energy-calculator/) which when configured for exponentially decaying signal and given the same values will give different answer.

I am not sure which method is right. I also read some where that power calculation shouldn't be used for MOV selection.

I also want to know which method do you use the current rating & clamping voltage method or energy calculation method?

[trobbins]

What have you gleaned about one-off versus repeated application of some type of standard transient, and service life of a MOV ?  You need to clarify whether there is degradation occurring for any one transient, and if so then how many such transients can be accommodated before end of life.  There is also a transient level where no degradation occurs for a given transient repetition, which relates to internal temp rise and hence the MOV package size.  Those aspects are typically presented in statistical performance charts and based on some long term testing of batches subject to a small number of given transient/rate, with the rest of the charts filled in by interpolation, with some manufacturers then making a statement about no degradation at some level of transient.

[yashrk]

What have you gleaned about one-off versus repeated application of some type of standard transient, and service life of a MOV ?  You need to clarify whether there is degradation occurring for any one transient, and if so then how many such transients can be accommodated before end of life.  There is also a transient level where no degradation occurs for a given transient repetition, which relates to internal temp rise and hence the MOV package size.  Those aspects are typically presented in statistical performance charts and based on some long term testing of batches subject to a small number of given transient/rate, with the rest of the charts filled in by interpolation, with some manufacturers then making a statement about no degradation at some level of transient.

These are truly valid points that I will include in the blog post.
Using this calculation I am attempting to find peak power that can be dissipated in one surge event.

[trobbins]

Are you saying for a life of one event only.

[T3sl4co1l]

FYI, if you're in need of a more accurate model, this is probably representative of most generators:



Tim

[jonpaul]

Yashak, Bonjour, : We used MOVs since 1970s and had some blowup and start fires.

See the IEC standard, /IEC_61000-4-5
https://en.wikipedia.org/wiki/IEC_61000-4-5

MOVs data and physics is documented app notes  from GE and other vendors, back to 1970s..1980s.

The degradation is a log function on # hits and surge energy.

They are NOT rated for continuous power.

Finally the very high capacitance and fire hazards and degradation of protection of MOVs, cause most modern transient protection design to use other methods.

 Bon Chance,

Jon

[Terry Bites]

Most MOV manufacturers have selection tables app notes on thier products.

[coppercone2]

t3sla, what are the power level of those components?

I just noticed I have a bunch of old 10uH ferrite inductors that are rated at 15A and it would not be much trouble to unwind them a little to get the magnetics at the correct value.

Normally I would not have the budget for a random tool like this but it could free up my attic and be a free project save the box.

[TimFox]

The last time I blew up an MOV (packaged like a large disc capacitor), the case fractured with two circular plastic discs flying off in opposite directions (like the gammas from positron annihilation), and one hit a software guy.  That helped my reputation at work.

[floobydust]

[...] I just noticed I have a bunch of old 10uH ferrite inductors that are rated at 15A and it would not be much trouble to unwind them a little to get the magnetics at the correct value. [...]

I would think those must be air-core inductors capable of high voltage end-to-end, so a bit larger turns spacings. But the resistors are inductive too at that power level.
joeqsmith won't divulge what parts or the circuit in his. I was interested in the switch.

I've seen Leader uses MOV's as high voltage zener diodes, but at very low current.

[coppercone2]

ah, figures, turns out their 1mH inductors too, not 10uH

4 turns brought it to 4uH, if you space em its less.

not that air core is hard to make either but I wanted to use that ferrite for something

the switch.. maybe oil filled relay?

still might try it because I got 4 fully unwound inductors that I stripped the wire from for my LISN stuff, pity that is unfinished because the storm shutter steel salavage operation to weld a box up requires a insane amount of metal beating work to make the chassis somewhat rectangular, I am stuck with what looks like a king arthurs war tent at the moment and have little will to proceed

[T3sl4co1l]

Uh, well if you charge the 20uF to 2.5kV or so, and the output impedance is around 2 ohms... fill in the rest.

Thyristors are more than adequate for switching such (µs+ surge) networks.  I have a unit in storage that uses exactly this (a stack of SCRs that is).  (It's not for this waveform; some oddball one, I suppose hence why I got it for free.)

EFT (IEC 61000-4-4) needs to be done with something faster, I suspect hydrogen thyratron being the most common.  Superjunction and SiC MOSFETs are also fast enough to do that.

Tim

[T3sl4co1l]

I've seen Leader uses MOV's as high voltage zener diodes, but at very low current.

I have a possible application for this, at intermediate power levels.  It should work; duty would be down in the "infinite" endurance range as given on the datasheet.  The real question is, is that reliable, what is the actual peak voltage at that current, and does the leakage or clamping voltage "age" at all in the process (which will be switched by semiconductors, so needs to be well known).  It's a low priority so I still haven't built anything of it...

Tim

[trobbins]

MOV's are an excellent choice for some secondary-side applications where a soft clamp doesn't impose on operation up until an inadvertent load is applied or part failure occurs, and the MOV is asked to instantly apply a soft load to the peak of over-voltage transients (single or continuous) and hence reduce peak voltage stress to other parts.  I use them for protection of output transformer primary windings in valve amplifiers.  In that application the MOV operates in the 'infinite' endurance range for many types of known fault scenarios, and so I disagree with Jon in an earlier post about NOT being rated for continuous power dissipation (perhaps Jon just has mains side or high energy blinkers on for all MOV applications he has come across).

[coppercone2]

what the hell do you use for high voltage inductor winding? 2kv seems kinda hard

searching class 3 magnet wire, is that the best?

https://www.eevblog.com/forum/beginners/copper-magnet-wire-breakdown-voltage/

[trobbins]

coppercone - perhaps learn how to wind on to a core to alleviate wire insulation stress, then it won't primarily matter what grade enamel insulation is used.

[floobydust]

[...]  I use them for protection of output transformer primary windings in valve amplifiers. [...]

Protecting valve amplifier O/P transformer's primary I would not use MOV's because they have such high capacitance and the last thing you want is resonance with the leakage inductance.
Guitar amps i.e. modern Marshall are just putting a 1.6kV reverse-diode across the pentodes' plate-cathode, I think it clamps the transformer's primary legs to 2x B+ maximum.
edit: added schematic pic of the approach, old is the 1N4007's and newer use SF1600

[trobbins]

floobydust, perhaps if you look up a MOV datasheet and choose something like a 660Vdc 1mA part with 45pF spec, and then look at a variety of output transformers and measure their primary shunt capacitance and ball-park the change in resonance frequency, and then look at the most sensitive application of a hi-fi amp with global negative feedback and its stability, you may well find that adding the MOV has negligible influence. 

Yes a few manufacturers have used the flyback/catch diode scheme for decades, some even going back to using valve diodes, but that has an inherent disadvantage that it doesn't circulate the flyback current just around the half-primary winding that is forcing the inductive voltage spike, but rather it clamps the other half-primary winding and directs that current through the B+ power supply.

[jonpaul]

Rebonjour:



History: MOV is a ZnO frit ceramic, first invented by Charles Proteus STEINMETZ of  General Electric, Schenectady ~ 1920s.

Application was  lightning protection  for power distribution and transmission  ...still in use today!



Fire: In 1986, place a transient protective  MOC across the mains ~ 25 mm dia GE,

 A  fault on a 12 kW movie light in the next room caused line transient,   MOV to exploded, and startd a fire on the carpet.

My engineer shouted...."Jon..there is fire!"

That experience ended our use of MOV.

Snubber:

The practice for repetitive  transients in SMPS has been a snubber  R-C,  diode, perhaps a TAZ.

I have not seen MOV on an SMPS as a snubber.

MOVs high capacitance and protection deterioration as well  as chance of  fire or explosion preclude use in production designs.



Just the ramblings of an old retired EE!

Enjoy,
Jon

[coppercone2]

I highly recommend wrapping a MOV with kapton tape and a thermal fuse, but a grizzly transient will still probobly make it explode, I think the MOV is just under rated. The one that works in the worst case probobly is not something we are used to seeing


usually you see more then 1 25mm mov together, but when they blow its not 100% equal damage sharing, one usually takes the brunt of the damage.

[floobydust]

floobydust, perhaps if you look up a MOV datasheet and choose something like a 660Vdc 1mA part with 45pF spec, and then look at a variety of output transformers and measure their primary shunt capacitance and ball-park the change in resonance frequency, and then look at the most sensitive application of a hi-fi amp with global negative feedback and its stability, you may well find that adding the MOV has negligible influence.  [...]

If I wanted to make an RF transmitter, I would put a tank circuit on the plate    At least a series resistor, something to lower the Q.
With proper loop feedback (including the output transformer) valve amplifiers don't Tesla coil when operated with no load. The Marshall amps I imagine are frequently operated in overload and have no NFB to the loudspeaker, so they need coverage which the diodes must be for.
So I'm not sure what your situation is, what you are protecting against using the MOVs. I can't find any MOVs with appreciable power handling and that low capacitance 45pF. They seem to be 2x to 3x that value even for a small 7mm part. I have no idea how MOV's are at high frequencies for dielectric heating, if that is a concern beyond mains frequency use.


There are many papers on varistor degradation and aging, must be good for a PhD thesis. The utility industry does a lot of research on them.
Degradation of ZnO Varistors as Estimated by Aging Tests Kansai Electric Japan, did some 10uA leakage current tests on aging so no transients. It seems oxygen and ZnO ions evaporate, in addition to the crystal and grain size changes that occur due to transients and hot spots. The physics can get very complex.

Application Note EC635: Designing with Thermally Protected Varistors in SPD and AC Line Applications

[trobbins]

For hi-fi amps where the higher frequency resonances of the output transformer are a significant issue for stability margins, adding a MOV with series R may provide an RC benefit.

I have a batch of 330Vdc 1mA MOV's (Littlefuse V250LA2) that I regularly use, and series up for higher voltage applications.

I've tried to go through a few failure/fault scenarios for valve amps in the linked doc, where inductive energy can force an over-voltage transient.  It's an interesting topic as it goes into many aspects of operation.
https://dalmura.com.au/static/Output%20transformer%20protection.pdf

I've used a few different MOV based products over the decades for industrial grid applications, but secondary side applications where fault energy can be quite limited often requires a different mindset as to how MOVs can be used to benefit - not that that is on-topic wrt lightning induced protection!