MOV Failure Mode 2

[buzzp]

I am interested to hear what failure modes people have found failed MOV's in, ie open or short circuit. I am questioning the use of an MOV across a DC relay coil vs. a diode with a series resistor. I am concerned with an application (safety device) which is calling out to put an MOV across the coil. If it were up to me, I would put in a diode and series resistor for the following reasons:
- series resistor will minimize the time to de-energize vs. just a diode or just an MOV
- MOV is one device so it could potentiall fail shorted causing the safety device to not function when called upon to do so
- MOV's lose their energy dissipation abilities with each 'surge' of voltage making it less affective as it absorbs more 'spikes'
- Their is concerns with an 'short' occurring with either configuration but there is two devices that will have to fail shorted and I have not seen a resistor fail short circuited but I have seen a diode fail shorted

Any thoughts on failure modes of MOV's or other additional input will be appreciated.

 

[cbarn24050]

Hi buzz, a MOV is a good choice for this application as there is very little energy in the system to destroy it. When used across a supply, ie. the mains, they can explode and make a real mess. The relay will de-energise fastest with the MOV, because it allows the the back emf to rise to MOV clamp voltage, a diode limits it to its Vf. A series resistor with a diode can work but you have to find the correct resistor value by experiment. I've never seen a component failure in this position but you could allways use a MOV and diode if your still worried.

 

[Skogsgurra]

buzzp,

The main concern is not the kick-back protection but the fact that a safety device uses an activated relay for the safety function. You will never get a fail-safe function there, regardless of what kind of protection you use.

Instead make sure that the relay is activated in normal situations and drops in abnormal (dangerous) situations. Use forced breaking and NO contacts. For even better safety use two relays and two NO contacts series connected.

A diode plus a resistor is probably the best choice in this application, but I would not hesitate using a MOV. I could even think of an RC combination, even if they are mostly used with AC coils.

BTW; if this is a safety device, it should not operate very often so you may be able to operate without protection (if you do not need to protect the driver).

 

[BrianR]

MIL-SPEC relays generally use back to back zeners as this has minimum effect on the relay dropout time. A diode will block the back emf kick but slow the relay deenergising time as well, (although the methods discused earlier are also OK). Use 40V zeners for 24VDC and 20V zeners for 12VDC circuits.

I strongly support skogsgurras comments re the NO relays for fail safe applications.

 

[Paulcet]

Skogsgurra is right on the money here.

To answer Buzzp's question on the failure mode of MOVs: I have only seen them fail short.

 

[buzzp]

The safety aspects etc are being addressed as they arise. Apparently this is how things "have been done", industry wide I am told. In any case, these will be addressed as I become familiar with the different systems and the interaction of each. Lets talk about the safety of MOVs vs other means of kickback suppression.
I guess after the input I got, I am still not convinced I would be using the MOV in this application just because of the possibility it will fail shorted. I only know of cases where they fail shorted or simply burn up or explode.
Cbarn, You bring up a good point about an MOV vs just a diode. This is why I would use a diode with a series resistor. That brings up another item. If we use a diode and series resistor, the resistor will have to handle this high voltage for a short time. The question is how large will this resistor have to be as far as the transient (short time) voltage is concerned? Now I have to go check some resistor data sheets and calculate the amplitude of this voltage.

 

[Skogsgurra]

buzzp,

Let's take an example: The relay coil has 500 ohms resistance and 24 V nominal voltage. You will then have 24/0,5 = 48 mA through the coil. This current will flow through the diode-resistor combination when you switch off and if you choose a 1 kohm resistor the initial voltage will be 48 mA times 1 kohm = 48 V. The current will decline three times faster than the risetime when switched on (the total resistance is now 1,5 kohms) and that is very often enough. So, a diode plus a resistor is a very good choice. The resistor wattage is almost always very low. You seldom need more than 1/4 W if you are talking normal relays.

 

[OperaHouse]

Been a few years since I designed an intrinsically safe circuit but at that time they considered metal oxide film resistors the safest. There is a voltage rating associated with resistors for voltage breakdown. If you are worried about a MOV shorting, I wouldn't use anything less than a 1 watt or higher resistor. I have had .5W carbon film resistors open from surges that physically looked perfect.

 

[60hzhmmm]

paulcet: I have seen them fail "gone" two bare leads standing up off the board...

buzzp: I'm still a little confused... If you are talking only about anti-kick back on the relay coil and it's DC why are you not just putting a reverse diode across the relay and being done with it?! I must be missing something... Why do we care about how fast we damp things on a relay coil unless you're trying to collapse the field extra super hyper weird fast for some pecular situation? You might want to put a mov across the CONTACTS to keep the noise of switching an inductive load down.. resistors and capacitors in "this" application will leak and drive technicians with DMM's insane tho.

 

[Skogsgurra]

60hz...

The resistor is there for more than one reason. The obvious one is that it will speed up the drop out (no pun). The other, not so obvious reason, is that diodes can also break down. If that happens, then the resistor will prevent shorting the power supply, blowing a fuse or tripping an overcurrent protection or whatever. It would actually be possible to use only a resistor parallel to the coil, but you will get a lot more power consumption that way. And the resistor has to be dimensioned accordingly.

OperaHouse mentions that resistors also have voltage ratings (not just wattage) and that is correct, but in this applications the voltage rise will be no more than 48 V and the lowest voltage limit you see in ordinary low wattage resistors is usually 200 V. The "power pulse" will be less than 2 watts during 100 milliseconds. And that will not heat an ordinary 1/4 W resistor more than temporarily and not more than, say, 50 - 80 degrees C.

 

[Paulcet]

Buzzp:
If you use a diode and 1K-ohm resistor as Skogs suggests, and if the diode fails as you suggest could happen, the resistor would have to dissipate over 1/2 watt while the relay coil is energized.

 

[Skogsgurra]

Yes, right. I forgot that.

 

[buzzp]

Quick note on the last comments; of course the resistor would have to dissipate more power if the diode fails shorted. I do not believe this will be the case since the current through the diode is no higher than the current through the coil. This will reduce the chance of it failing shorted. It is more likely to fail open. Although there is a chance it could fail shorted. In any case, there are now two components that need to fail to cause the relay not to energize.
Well let me update everyone. This particular device is a lock-out relay (125VDC, 23 ohm coil). It is normally not energized. When a problem occurs, the relay is momentarily energized. The way this particular relay works, the coil is way underrated or the voltage applied is much higher than the rated voltage. This is done to energize the relay the quickest. Now you can see why the relay is only energized for a short time, the coil would burn out otherwise. It is de-energized with its own NC contacts. As I understand it, this is how most utilities handle their lock out relays. I do not necessarily agree with this but it is what I have to work with for now. You can see, if the MOV fails shorted, then when the relay needs to energize it may not due to too much current being routed around the relay coil.
I have done a lot of research on this and I still feel, for this application, if suppression is required that only a diode with a series resistor be used. This will be the quickest as far as the de-energize time if suppression is required. I have heard claims that if the series resistor is selected such that the value is equal to the coil resistance then this is usually the optimum value. Has anyone have any experience with this statement?
I will run through the calculations to see what kind of power rating I need on the resistor. Thanks to all for your input.

 

[BrianR]

This is an arbitary value. When a resistor equal in value to the coil resistance is used in parallel the back emf spike is limited to 2 x supply volts. The diode merely stops the resistor from drawing current all the time the relay is energised. If your application calls for a brief energising pulse then the combination diode/resistor is not needed. Either by itself will do the job.

A diode by itself offers better suppression than any combination using a resistor however it slows the deenergising time of the relay but that does not sound like a problem in your case.

 

[buzzp]

Well, de-energizing time is somewhat critical due to the coil being underrated. I like the idea of having two parts across the coil rather than one. Just in case (discussed above-two part failure for problems, rather than one).

 

[BrianR]

The slowdown is only ms, eg for a relay it might take 30ms to dropout instead of 15ms.

 

[buzzp]

That is significant when your applying a voltage across a coil that is four times its rated voltage. In addition, 30mS is almost two cycles where as 15mS is one. This will reduce the damage done to equipment.

 

[BrianR]

Yes 4x is quite an overload so you'll be safer with an diode and resistor with R no larger than the relay R.

 

[Skogsgurra]

buzzp,

Is a lock-out relay the same thing as a "shunt release" coil in a MV or HV breaker? (I am not familiar with the word usage in your language) If so, we are having the same problem universally; Coils that are supposed to work when needed, but tested only once a year. If you are lucky.

Anyhow, in these cases, there is not much need for either kick-back protection or EMI reduction, which is the other reason for putting a MOV or diode or RC combination parallel to the coil. I think that you can live happily without any protection at all. But, of course, you shall decide that yourself.

 

[melone]

As R increases, the voltage developed across R during flyback increases. Therefore, the energy from the coil / relay will be disapated by the R and diode. Bigger R means that the energy is disapated quicker. HOWEVER, bigger R also means higher flyback voltage.