What's a good rating for a MOV suppressor?

[J M Fahey]

You got it!

In my experience, every valve amp I've tried has been happy without a load, at idle and moderate signal levels. When you get past what would be the clipping point with a normal load, things start to go wrong. When I tried it with EL34s, the screen grids lit up, and glowed brighter and brighter as the drive level increased, until... :-O

OK, I stopped the experiment to avoid trashing a functional pair of tubes.

On the basis of these results I once tried to make an "open speaker detection circuit" that shut the amp down on excessive screen current. The problem was, it was vulnerable to nuisance tripping, even though it was set greatly in excess of the rated screen current. This made me suspect that EL34s are often abused in practice, it seems to be normal in the guitar amp world to have the screens merrily glowing at twice the rated dissipation.

Maybe one day I'll revisit the idea with a current clamp instead of a trip, or even hook it up to a Vactrol compressor.

I bet you could skip 1/2 the Vactrol and tape an LDR pointing to the screens ... and I'm *barely* kidding here.
Try it the next Glasgow winter when you are bored stuck in your house.

[R.G.]

Wow - I really had no idea the line was THAT noisy. But that's a big honking inductor, and 1/3-1/2A primary current at idle for a 50W tube amp by my figuring (using 4.5A@6.3V + 20w idle bias), and a really really short dt. Divide by almost zero gives almost infinity. I would expect that each and every time you cut it off, you'd get a big spike. I believe you, maybe L or dt arent as extreme as I thought.

There are a couple of subtleties in this. First, the primary inductance isn't carrying the load current. The primary current at idle is probably 95+% going into the rectifiers, heaters and idle loads. The primary inductance is an idealization of the reality that you have to pay with some electrical power to set up the M-field in the iron. Once the field is set up in the core, the core's M-field causes an voltage in the primary windings that keeps the current into just the iron part down to a few percent of the full load current. You can find this current by running the transformer with only the primary connected; the no-load current is one thing that transformer makers sometimes do specify. In really good ones, the no-load current may be down in the 1-2% range.

Stay with me for a minute. The primary current sets up the M-field in the core, which in turn sets up the counter-voltage which keeps the current down in the primary. When you pull current out through a secondary, this removes some of the counter-voltage, and that lets in just enough current to supply the secondary's needs. A secondary load gets power by removing some of the magnetic props keeping it limited in the core. The actual M-field in the core and the primary current causing it stays remarkably constant. The current to power a secondary load flows through the magnetic field, but doesn't change the size of the M-field much.

In effect, the only current the core *ever* sees is the magnetizing current. The rest of the secondary loading passes through without changing the core conditions much.

So your 1/2A at idle isn't really in the primary inductance. It's passing through to the rectifiers, heater, etc. and when you turn off the power switch, the secondary load, which is already sucking out 50 times the stored energy in the primary happily sucks out the remaining primary inductance field.

I can hear several thought-streams of readers of this going "butbutbutbut" and "satuation... saturation... saturation". We'll get to that.

What the switch (and MOV) really do have to cope with is the instantaneous current in the leakage inductance between primary and secondary. That's outside the primary inductance and is an idealization of the M-field that does not couple to the secondary. However, the leakage inductance is in general much smaller than the primary inductance, usually millihenries to fractional-henries in the primary, so there is much less energy there. If the PT has a loaded secondary, that eats the primary inductance field very effectively. It can't eat the energy stored in the leakage inductance, because by definition, that's magnetic field the secondary can't get at. The leakage energy has to be dissipated on the primary side.

In a way, yes, you're right. L isn't as extreme as you think. It's the leakage inductance, not the primary inductance. Even dt is not so bad, because any primary capacitance turns a zero dt into a damped ring by allowing current to flow into the cap, and making dt stretch out by the resonance of the L and C. Even wires gots Cs.

This is not to say that the energy in there is trivial. It's just smaller than you'd expect, and the energy has other places to go.


butbutbut and saturation

It's hard to get to this understanding of transformers, or it was for me. The M-field is kind of like trying to drain water from a tank with an automatic fill valve. You can get water out of the tank by opening a drain valve. The drain valve drops the level of the water a fraction and that lets more water flow in through the fill valve. The level of the water in the tank is what cuts off the flow of water into the tank through the primary fill valve.

And you can't overflow the tank by sucking water out of it on the drain valve as long as the inlet valve is set up right. Same with transformers. You can't saturate a transformer from the secondary, because the secondary current demands are instantly balanced by primary current flows, and the net difference in current that keeps the iron magnetized is maintained the same.

That doesn't say you can't overload a transformer from the secondary. You can happily overheat the wires and burn it up by secondary loading. That heats both the secondary and primary wires because of this energy-balance thing in the M-field.

Saturation is a forced failing of the iron's ability to support field changes to link energy from primary to secondary. You put in more volt-time on the primary, but this does not result in a matching increase in magnetic flux, so the secondary winding cannot even see that you've done something on the primary side. It's not coupled to the change through the field. And the failure of the iron to increase magnetic flux means that the opposing voltage field can't build up, so primary current pours through the primary winding, limited only by the resistance of the wires. That's saturation. It's a decoupling of the secondary and primary windings because the iron can't support bigger M-field changes.

This is in fact how constant-voltage transformers work. The primary winding never saturates. But there is a BIG capacitor on the secondary side that pumps part of the secondary iron into saturation. So the voltage that is created on that saturated part is limited by the amount of flux the iron can sustain there. The primary voltage is "clipped" in the secondary by the iron not supporting any more voltage and you get regulation.

Sorry, I'm ramgling. It's late.

[bob p]

I'd like to resurrect this thread with a question about selecting a specific MOV for OT protection.

My application has B+ = 530 VDC. I really need to protect the OT because it's for a 6x6L6 amp and there are no suitable replacement OT.

I'm thinking about MOV placement either from plate to ground on both sides of the OT, or plate to plate across both sides of the OT. Needless to say, it's kind of hard to find devices that have a suitable DC rating that's high enough to stand the B+ while staying low enough that the max voltage that the windings could see doesn't scare me. The other thing that bothers me is that the capacitance of these things looks really high.

Any suggestions on a suitable part?

Thanks!

[trobbins]

I suggest you move your focus to a MOV-R that is connected to each OT plate back to B+. As such, the MOV minimum DCV at 1mA rating needs to be at least about 560-600V, and a series resistor with R value that is equivalent to Plate impedance loading. Unfortuntely that means finding an available MOV with a data sheet and looking up the DCV rating. Physically smaller MOVs, ie. 7mm disk, give suitably low capacitance. I have a batch of 7mm disk MOV's with 330VDC rating and 90pF capacitance - so I would use two of those MOVs in series if I had your OT.

[bob p]

the MOV minimum DCV at 1mA rating needs to be at least about 560-600V, and a series resistor with R value that is equivalent to Plate impedance loading.... Physically smaller MOVs, ie. 7mm disk, give suitably low capacitance. I have a batch of 7mm disk MOV's with 330VDC rating and 90pF capacitance - so I would use two of those MOVs in series if I had your OT.

Finding data sheets isn't that hard. Littelfuse makes them easy to find:

LA Series Datasheet.pdf

UltraMOV Series Datasheet.pdf

It's not to hard to find 7mm disks that have low capacitance, but the voltages are lower and it looks to me like you need to trade away the joule rating to get a small size. You end up needing two in series, which isn't much (any?) better than selecting a larger diameter disk.

I'm finding 460 VAC MOVs that are rated for 615 VDC; the 1 mA test current voltage ranges are 640-790. But the max clamping voltages are kind of scary: 1190 V. To get a decent joule rating requires that I accept a fairly high capacitance and a larger size. Example: V460LA20AP; V10E460P.

Am I missing something?

[trobbins]

I suggest that joule rating of MOV has no consequence at all, and so can be as small a level as you can find.

The MOV-R circuit is there to restrict any over-voltage transient, which may typically arise from leakage inductances and the impact of high di/dt (eg. disconnected speaker when playing). There is little 'energy' in the clipped part of the transient (the voltage in excess of the MOV voltage rating). The MOV is not normally 'working' - so the circuit operation is different from say a simple RC circuit that was/is sometimes used (often from plate-plate), although the MOV's capacitance and the series R do add a subtle RC load on each winding (which may be an advantage to tame any very high frequency self-resonance within the primary winding).

The MOV is there to alleviate over-voltage stress on insulation within the OT windings, and doesn't protect the OT from over-current (eg. as normally protected by using PT secondary winding fuse, and individual cathode fuses especially if using parallel output tubes).

[bob p]

I suggest that joule rating of MOV has no consequence at all, and so can be as small a level as you can find.

The MOV is there to alleviate over-voltage stress on insulation within the OT windings,

Thanks for posting. But in focusing on the joule rating and the capacitance, we're ignoring the thing that's most bothersome to me -- I think I mentioned this a couple of times -- the max clamping voltages are pretty high (1200V). That's not bothering you?

[trobbins]

Good question. Not easy to clarify - here's how I see it:

That clamping voltage is for the specified applied pulse - the datasheet may identify the max current level attained for the pulse - I suggest that is a surreal current level in this application, as the test circuit to generate that standardized pulse is quite different from the OT type of circuit, and the series R won't allow it anyway.

My view is that the energy in an OT transient is relatively low. A quick ballpark calc for a leakage inductance of 1H (1% of a Williamson style 100H OT spec), and a peak current of 1A (that's a big big OT scenario) gives a power level of 0.5W. If that power is dumped in a very short time, then the joule level is very low.

Insulation rating for a PP OT would likely be fairly high, given that the cut-off end of the winding swings to at least twice B+, and waveforms from loudthud indicate how much higher it can typically swing. To my mind, the clamping of the MOV-R is aimed to be somewhat 'soft', and the MOV-R protection is likely to be subtly loading the OT primary half windings even during normal cranked use.

As I assess the 'failed OT' situation, there are certainly examples of failed OT's - no question about that - but similarly, there aren't bucket loads of failed OT's in every amplifier service shop.

[trobbins]

My last post was in error on a few points. The energy in the winding in a transient is 0.5 x L x current squared. So my example of 1H and 1A gives 0.7Joule (not W).

But that just related to energy in the leakage inductance of an example OT. In a fault situation (where protection is needed) the instantaneous energy in the transformer is the current flowing in the working sections of windings, and that current wants to continue flowing in those windings, or transfer in to other windings, in whatever escape paths can be taken.

If the current in the secondary winding stops abruptly, then that energy could escape in a number of paths. If one of the PP valves is conducting, then that winding loop is a nice low impedance path to transfer energy in to. But if both valves are not conducting, then the energy starts raising winding voltages, in which case stray capacitance will soak up some energy, and the MOV-R's in the two half windings of the primary will start conducting when the particular winding voltage reaches the MOV conduction level. Another fault scenario could be where a conducting valve stops conducting.

The energy that could be around in the OT could be ballparked as the inductance of a full primary half-winding, and its conducting current. For an example of 100H for a half-winding, and 500mA current level, then energy level is 12J. I typically end up using 2x 20J varistors (7mm disk) across each half-winding - so effectively there is 80J of MOV's available for soaking up the spike energy. Although the series resistance used with a MOV would need to be not too large as to allow sufficient current to bypass through the MOV.

[J M Fahey]

^^^^^^^^^^^^^ Excellent analysis.
Puts hard numbers instead of soft opinions.
Thanks for the contribution.

By the way, it explains why MOVS sometimes seem to do their job for a long time, while others degrade and lose effectiveness.

[bob p]

Finding data sheets isn't that hard. Littelfuse makes them easy to find:

LA Series Datasheet.pdf

UltraMOV Series Datasheet.pdf

Aargh. Website has moved the files. Here are updated links:

Littelfuse_Varistors_UltraMOV_Datasheet.pdf

Littelfuse_Varistors_LA_Datasheet.pdf

[bob p]

I'm moving at a snail's pace on this project, so I'm doing a little bit of necro-threading.

here are some parts that I'm thinking about for my application. I'd appreciate feedback if any of my choices need to be reconsidered.

A. For the primary side, which is a 120 VAC supply:

The amp will have DPDT switching of both hot and ground so that the MOV aren't subjected to line noise when the amp is powered-off. I'm planning to use 3 of these, putting them across hot, neutral and ground:

Littelfuse UltraMOV series, Part # V20E150P
Ratings: 150 VAC, 200 VDC, 120J
V20E150P - LITTELFUSE - METAL OXIDE VARISTOR, 200V 395V | Newark

B. For the OT:

Amp is a 6x6L6 Fender Super Twin Reverb, with a B+ of 530 VDC. I haven't taken actual measurements of the OT primary Z, but for a sextet of 6L6 I would imagine it's in the range of 1800-2200R. Call it 2k.

So I'm thinking about putting two of these in series, and adding a 2k/2W resistor in series with the 2 MOV, and putting the string of 3 series devices between B+ and each plate winding:

Littelfuse UltraMov series, Part # V07E275P
Ratings: 275VAC, 350VDC, 28J, Itm= 1750A, 80pF
V07E275P - LITTELFUSE - METAL OXIDE VARISTOR, 350V 710V | Newark

thanks!

[trobbins]

I'm not sure what you are trying to achieve on the PT primary side. A single MOV across PT primary, with hot line switched, would manage primary winding induced voltage spikes when the switch is opened. Are you worried about externally generated spikes damaging the amp parts - if so then I'd be me more worried about other equipment in your house first and perhaps putting something at the entrance to your house first if spike were from the mains.

The OT primary MOV (or series MOVs) shouldn't have much series resistance - at most that resistance would be something like the half-winding impedance (eg. 500 ohm for a 2k PP).

[bob p]

Thanks for posting.

I'm not sure what you are trying to achieve on the PT primary side.

When I posted I wasn't thinking about primary winding induced voltage spikes with switching, but I was thinking about externally generated spikes damaging the amp. That, and for the possibility of a sustained over-voltage condition damaging the equipment. My plan was to add embedded supplementary "point-of-use" protection against line surges and over-voltage conditions.

As I understand it, electrical equipment can be damaged by any one of five anomalies in the AC power supply: 1) open neutral events (caused by external service supply equipment failures), 2) catastrophic over-voltages (caused by fallen trees, traffic accidents, storms), 3) sustained AC over-voltages (malfunctioning utility equipment), 4) AC under-voltages (brownouts), and 5) utility switching transients.

From what I've read, the service-entrance type of whole-house surge protection devices will protect against lightning, utility switching transients, and transient over-voltages that last less than a second, but they offer no protection whatsoever against sustained over-voltage conditions (like you'd get with a utility regulator failure, an open-neutral condition when a transformer fails, or high voltage power crosses. For these types of protection, a downstream "point-of-use" device is required to provide over-voltage protection. My understanding of the situation is that two separate devices that appear in series on the line are required to provide full protection, one at the service entrance, and then secondary devices in "point-of-use" locations.

My idea for adding the MOV on the primary side of the transformer is to provide embedded "point-of-use" protection in the amp, so that I don't have to worry about dragging a surge-protection outlet when the amp travels from location to location. I had decided to put the MOV after the line & neutral switch in order to limit the lifetime exposure of the MOV to AC, just for concerns about longevity of the device.

In case this topic interests anyone, here are a couple of interesting reads:

IEEE Guide for Surge Protection of Equipment Connected to AC Power and Communication Circuits

National Fire Protection Assoc Pub 780: Standard for the Installation of Lightning Protection Systems 2004 Edition

The OT primary MOV (or series MOVs) shouldn't have much series resistance - at most that resistance would be something like the half-winding impedance (eg. 500 ohm for a 2k PP).

Half of the primary Z. That's the answer I was looking for. Thanks.

Insofar as you've remained mute on the question of the specific devices that I asked about, should I interpret that as ignoring the question or approving the choice?

Thanks!

[J M Fahey]

FWIW Electrical Safety rating Agencies around the world agree that up to 3KV pulses and transients are *often* present in home power lines, so much so that they test any and all home equipment for it, and HIPOT tests are run at 3000VAC or 5000VDC before issuing approval.

I think it's not *mandatory* in USA (UL) but definitely is in the rest of the World.