Output transformer protection, flyback diodes, MOVs, 2KV capacitors,

[nsubulysses]

Hi I have a question of preference or good ideas regarding TH1 and TH2 on Traynor YGL-3A Mk III.

I have one here with those parts shorted. They're rated at 1KV and I always see similar parts short in Musicmans, but in that case they're 2KV capacitors from plate lead to ground. Some amps have a string of flyback diodes 1N4007 from plate lead to ground like Ampeg V4. I have never seeen those shorted but I have seen a handful of cooked V4 transformers, typically power transformer though. And then most amps have nothing.

Removing TH1 and TH2 from circuit makes the amp work again but what is best to do in this scenario? Should I just leave it or.....I usually replace plate lead to ground 2KV capacitors in Musicmans with 3KV and have had no problems so far.




Wild Bill does say to just leave it but then again he does seem pretty wild. Loudthud is the jury in yet???

Thanks guys

What is it? - AMPAGE Archive

[R.G.]

Protection is always a tricky business. First you have to know what's going to fail, and what the likelihood is of that happening, and then trade that off against the likelihood of your protection failing, and the expected cost to find and fix either one.

Transformers die two ways. Well, really one way - the insulation on the windings gives up and it gets a shorted turn. Things only get worse from there on. The two ways are how you get to a shorted turn.

One is heat. The wire insulation and insulation between layers and turns within a layer have heat limits. Get them too hot and they either soften and let the wires touch, or they degrade chemically and split/powder and then the turns touch. Or they short to the core.

The other is voltage. Put too high a voltage between two points and exceed the insulation's limits and the insulation punctures, an arc forms, and the arc makes the insulation degrade so the opening is then either thermally intermittent or permanent. Then turns touch, or they short to the core, and death ensues.

Death comes in three forms: burned out turns and a permanent short, intermittent shorts, or a burned-open winding.

You protect against getting too hot by good design and perhaps fusing. You protect against high voltages by using high voltage protection schemes. What you're looking at is one of these.

Things across the windings are designed to limit voltages by "failing" and conducting current when the voltage gets too high for normal operation, but lower than the breakthrough voltage of the wire insulation.

The diode stack from one end of the winding to ground is a (mostly ineffective) attempt to limit voltages on the other half winding by transformer action in clamping one side of the winding between B+ and ground. This fails on transients because leakage inductance prevents the perfect coupling needed to really clamp the off-side voltage.

You can use voltage clamps in the form of TVS devices, Zeners, or MOVS. Well, OK, capacitors that you count on to short when they're overvolted, saving the transformer by giving their own lives. MOVs slowly degrade with each over voltage conduction, so they gradually drop in breakover voltage until they conduct at normal voltage, then they die FAST. TVS devices and zeners are better in this regard, as long as they do not overheat and burn out.

The shunt protection devices on your transformer failed. This is good - the much more expensive transformer did not. Removing them entirely amounts to saying to yourself that "I think these just failed and there was no underlying cause in the amplifier's operation." OK, could be. Could also be that they did exactly what they were supposed to do. Your call.

I like protection devices because replacing them is much cheaper than replacing transformers.

[nsubulysses]

I only wonder about leaving them off because of Wild Bill but he seems too wild.

[R.G.]

The difficulty of predicting what and when will cause a failure makes deciding what protection to use very difficult. Wild Bill is entirely correct in leaving them out if you think that your work will be forgotten when - and if! - the failure syndrome occurs again. It might not ever. It might be next week. The problem is that most crystal balls are very cloudy.

[g1]

They are not caps, can't even remember what exactly they are called. The official factory line was always just to remove them.
If you really want to put something in there, I would suggest the diode route, like some of the new Fenders are going with.

[g1]

Ok, they are thyrector diodes, JPB was kind enough to post this pdf last time around for same topic .

[Enzo]

They are protective devices. Some guys refuse to wear seat belts in their cars, because they don't want to "get trapped in a burning car by a stuck seat belt." You can leave them out, and now your amp has one less seatbelt.

I'd replace them. You can use the same part, I imagine those or something like them can be had at Mouser, or maybe even from Yorkville. Considering that I stock the high voltage diodes like Fender and peavey and others use, I'd install a couple of those instead. Not that I think they are better, but just because I have them. And if I were ordering parts, I'd rather have a drawer full of those diodes than your thyrectors, just for versatility.

I don't have the part numbers on top of my head, but look on a recent era fender schematic like a HR DeVille, or look at a recent era Peavey tube head. One on each side, plate to ground, reverse biased.

[trobbins]

Originally, cost or practical device layout would have meant just using one back-to-back circuit from P to P.

If you can only purchase lower voltage thyrectors, or other similar functioning devices, or want to subtly improve the protection then it would be better to use a back-to-back circuit across each OT primary half-winding (ie. P to B+, and B+ to other P).

I happily use appropriately chosen MOVs - the likely energy going in to a MOV when an amp is cranked with a normal speaker load should be very low compared to higher energy pulses that could cause MOV degradation - even for small 5-7mm disks. One day I'll try and get around to testing the temp rise of such a protection MOV when continuously soft clipping a particularly bad plate waveform.

[nickb]

Ok, they are thyrector diodes, JPB was kind enough to post this pdf last time around for same topic .

Thank goodness they come in "an attractive blue colored fiber tube". I always look for things like that when designing in a new part.

[nsubulysses]

I took the convenient way out as G1 and Enzo suggested and put in a few R3000 diodes from plate to ground, like in the modern Fenders and what not. Thanks for chiming in everyone.

I have never heard of a thyrector before but if it's a selenium rectifier it does not seem too attractive to me, even with it's cool blue color.

[nsubulysses]

Speaking of tricky business the suppressor grid of the power tube is not at ground as is usually the case, but is tied to the control grid negative bias voltage. What effect does this have?

It seems like it would be a worse scenario for secondary emissions since the suppressor grid would be more negative than normal. Or is this some sort of local feedback arrangement since the suppressor grid and control grid would both see the signal?

[Enzo]

I can't open the schematic, but if the suppressor grid is tied to the bias SUPPLY, it does not see the signal.

The suppressor grid repels any electrons trying to bounce off the plate and back to the screen. By grounding it, it is 400v more negative than the plate and so the electrons are more attracted to the plate. Making the suppressor grid even more negative, it becomes even more effective at keeping those electrons where they need to be.

It is not a bad thing.

[potatofarmer]

Tying the suppressor(s) to the bias supply will reduce plate dissipation by diverting electron flow towards the screen instead. See Steve Conner's and Merlin's posts in this thread: http://music-electronics-forum.com/t18504/

[nevetslab]

Ashdown BTA-400 Output Transformer

I have an early Ashdown BTA-400 Bass Amp, which uses 8 KT-88 power tubes, with the transformer design being of the Ultra-Linear style (taps for the Screen grids) that presently 'barks' loudly on bass string transients. I'm presently in dialog with one of their technical staff on the problem, as we've discovered the amp of theirs in our rental inventory is lacking circuit upgrades. From their first email to my contact, I learned of the following

A BIT OF HISTORY, ABOUT 200 UNITS INTO PRODUCTION THIS FAULT YOU ARE DESCRIBING STARTED TO HAPPEN AND WE TRACED IT TO THE OUTPUT TUBES AND THE DRIVER BOARD VARIOUS MODIFICATIONS WERE INPLEMENTED TO CURE THE PROBLEM ALONG WITH SOME IMPROVEMENTS THAT HAD BEEN REQUESTED.

THERE ARE 3 STABILTY NETWORKS ON THE OUTPUT TRANSFORMER CONSISTING OF CERAMIC CAPS AND 1 K 5 WATT RESISTORS THE CAPS WERE UPRATED TO 6300V TYPES FOLLOWING A FEW FAILIURES WHEN PEOPLE FORGOT TO PLUG IN A SPEAKER. ONCE FAILED THESE WILL FLASH OVER ON TRANSIENTS AND CAUSE A BANG IN THE SPEAKER BUT NOT SHOW UP ON THE SCOPE AND A DUMMY LOAD.

ANOTHER CAUSE OF THE PROBLEM YOU ARE DESCRIBING IS A DEFECTIVE OUTPUT TRANSFORMER, WE HAVE HAD A COUPLE OF THESE THAT WORK OKAY UP TO A CERTAIN LEVEL AND THEN GO SHORT CIRCUIT CAUSE OF FAILIURE NOT 100% ESTABLISHED.

They are using a string of three high voltage diodes from the plates to ground in addition to the RC networks. One of the networks, series R-C set to be across the two plates isn't installed.

Anyway, the question I have has to do with the beginning stages of output transformer adjacent turns arcing, once the insulation system has initially failed (for whatever reason...fault during winding operation, thermal failure, over-voltage failure). It makes sense not to experience voltage breakdown on a resistive load on the output, as you're just simulating the 'nominal' speaker impedance in the 150Hz region, but with normal speakers, we're really loading the amp with an impedance curve, ranging from that 'nominal' value to maybe 10 times or more that value, where there are narrow impedance resonances on both sides of the cabinet port tuning. So, the output voltage will also swing much higher, with the primary side load to the tubes also increasing proportionally.

When I hear abrupt loud barking from the speaker,, even at moderate levels, and seeing oscillation on the grid driver circuit to the power tubes, could that be indicative of internal arcing within the output transformer? Thus far, it hasn't gone catastrophic and failed. But, I have experienced high mains current draw when I had nearly no output voltage from the amp briefly. My experience with failed output transformers is after they've died.

[trobbins]

That's interesting feedback from the manufacturer.

Overloading UL (without causing any overvoltage protection or stress) may be part of the barking observation - it will be interesting to see what more comment comes from the manufacturer.

The use of diodes (and thyrectors) are going to cause abrupt protection - is output feedback used in the amp? The diodes are not directly across the portion of winding that is generating an over-voltage - which is not a great technical application.

Insulation breakdown in an OT is likely to be very specific to the construction. I would doubt that turn-turn breakdown would occur with modern winding wire. Some common techniques for increasing creepage and clearance may be too expensive to deploy - such as margin tape at the ends of each layer, or designing in enough margin, or sleeving the ends of each layer, or vacuum potting the windings.

Loudthud presented some excellent oscilloscope traces of anode voltages with a speaker load and a good strum from the guitar.