AC30 blown OT - how to avoid a repeat?

[AleS]

So my early 2000's Korg/Marshall Vox AC30/6 has a blown OT. I'd done some diagnostics and all voltages seemed fine from an AC30 point of view, ie high: preamp values all as expected/within ±10%, HT at 345 and EL84 plates hit at 334. Dissipation is however at a whopping 18W. The OT went while I was trying the amp fully dimed - while not what I'd consider a regular occurrence or something I'd reach for with any regularity or at all, needless to say it still is not something you'd expect nor want.

Swapping the OT is the easy bit, however my concern is that this might happen again. I had the assistance of a tech looking into it and he seemed concerned by those readings. Voltages-wise, it all seemed fine by me (that's how Voxes run), dissipation wise I'm not too sure - it definitely is above what EL84's are meant to withstand by quite a bit.

Amp run fine for many years prior to that and part of me wonders, would swapping the OT and putting in sturdier tubes (I have a set of matched 6P14P which would go in there) be enough or would the consensus be that some measures should be put in place to lower B+? Swapping the PT seems like the nuclear option and was wondering if a 10 or 20V zener might do the trick, or perhaps looking into a choke with higer resistance might go to some lengths to mitigate this? Or would swapping the cathode resistor help lower dissipation (without affecting B+?)...

Unfortunately I'm not sure about current PT specs. That's true in terms of 5v current limit, meaning a straight swap of the rectifier is likely not an option since any lossier options to the GZ34 draw 50% more current (3A vs 2), but also in terms of primary - if a 250 or 240 was available and I was currently hooked to the 220 that would obviously bring about a couple options (5 and 6.3 supply value dependant of course).

Long story short, I'm evaluating my options and would love to hear yours - AC30's are funny ones as they're the stuff of nightmare from a datasheet point of view, but also tried and tested amps which do run fine in most cases and I'm trying to understand where it's worth standing with this one...!

Many thanks in advance, looking forward to learning more and hopefully be chiming and jangling again soon!

[AleS]

So sorry, this should probably live in the Vox/Hiwatt subforum! If that's the case, could a mod please do the honours? Apologies and please bear with me as I get to grips with the forum

[solderhead]

Are you interested in protecting the OT from overvoltages related to an open secondary tap or a closed secondary tap? ie: leakage inductance or dumping the full inductive energy stored in the OT? the solutions are different for each case. none of them are perfect.

the simplest solutions will involve modding the amp so that there is always a dummy load resistance on the OT secondary (even with no speaker cabinet connected), and MOV across the OT plate and CT on the primary side. Other methods used clamp diodes between plate and ground on the power tubes.

[pdf64]

What's the HT winding resistance of the mains transformer? eg between lugs 4 and 6 of the rectifier socket? Amp deenergised and unplugged from the mains wall power.

What V AC do the 5 and 6.3 winding voltages measure, with the amp idling, and at what measured wall voltage?

[trobbins]

How did you confirm the OPT was 'blown' - was a short circuit or open circuit on a half-primary winding?

How did you calculate 18W dissipation ? For example, 50mA idle through each EL84 would be 200mA through the common cathode resistor - if that resistor was 50R then that is 10Vdc at idle. If Plate was 334V, then plate-cathode is 324V, and plate + screen dissipation is circa 16W (which would be 2W over the nominal max limit).

Operating a quad at high dissipation could expose the OPT to higher fault current level if an EL84 shorted due to lower resistances in OPT and common cathode and PT than say a normal PP.

It's also plausible that valve could repeatedly cause current spikes in OPT and that could cause arcing, which could lead to OPT damage.

That quad offers no way to confirm all four EL84 are idle balanced unless you periodically check with say a bias probe - had you done that? You should do that in the future, or alternatively insert say 1R sense resistor in each individual cathode, and periodically measure idle that way.

The total cathode current can be reduced to nominal average limit dissipation (but also need to allow for worst valve imbalance), and hence reduce idle dissipation, by raising the cathode resistor value.

The GZ34 could have aged, or been stressed from the OPT failing, or from use of standby. Had you included ss 1N4007 with each GZ34 anode to avoide any chance of GZ34 arcing? That is a valid mod going forward.

If the OPT was exposed to anode voltage stress from blocking distortion during fully dimed operation then the modern method to avoid over voltage stress is to add a MOV across each half-winding, or to lesser benefit add blocking diodes to ground, or to even lesser benefit to add some secondary side resistance.

A more onerous protection is to add a fuse in the HT secondary CT - but that has to be carefully selected to provide appropriate discrimination that may help with OPT faults.

Apart from replacing the OPT, care is needed to confirm screen resistors are all still ok, and cathode R and C weren't stressed.

[AleS]

Thanks all for the replies.

PDF64 - not with the amp currently so cannot check as of now but will do!

Trobbins - thank you so much for taking the time for such an exhaustive response, it's hugely appreciated. Will try to address your points below...

How did you confirm the OPT was 'blown' - was a short circuit or open circuit on a half-primary winding?

The two windings are shorted.

How did you calculate 18W dissipation ? For example, 50mA idle through each EL84 would be 200mA through the common cathode resistor - if that resistor was 50R then that is 10Vdc at idle. If Plate was 334V, then plate-cathode is 324V, and plate + screen dissipation is circa 16W (which would be 2W over the nominal max limit).

My apologies, I hadn't checked my notes - 334 is plate-cathode.

Operating a quad at high dissipation could expose the OPT to higher fault current level if an EL84 shorted due to lower resistances in OPT and common cathode and PT than say a normal PP.

It's also plausible that valve could repeatedly cause current spikes in OPT and that could cause arcing, which could lead to OPT damage.

That quad offers no way to confirm all four EL84 are idle balanced unless you periodically check with say a bias probe - had you done that? You should do that in the future, or alternatively insert say 1R sense resistor in each individual cathode, and periodically measure idle that way.

All four EL84's test fine, but agreed it's likely at least one had been arcing causing the problem in the first place or had seen a high current surge it couldn't handle. Neither of those I believe would leave traces in the tubes as such, so hard to tell. And no, I had not been periodically checking measurements so I have no history to back any of this up unfortunately.

The total cathode current can be reduced to nominal average limit dissipation (but also need to allow for worst valve imbalance), and hence reduce idle dissipation, by raising the cathode resistor value.

Raising the 50ohm cathode resistor was on my radar, either with a 75 or 100. Would using a 5V4 rectifier and therefore losing an extra 15/20V off the B+ help as well? Would it be a case of either/or in your opinion?

The GZ34 could have aged, or been stressed from the OPT failing, or from use of standby. Had you included ss 1N4007 with each GZ34 anode to avoide any chance of GZ34 arcing? That is a valid mod going forward.

The GZ34 blew too, as did both the HT and mains fuses, though they did so after some time, not immediately. Backup diodes are 100% on the to do list!

If the OPT was exposed to anode voltage stress from blocking distortion during fully dimed operation then the modern method to avoid over voltage stress is to add a MOV across each half-winding, or to lesser benefit add blocking diodes to ground, or to even lesser benefit to add some secondary side resistance.

Had never heard of this but it's very encouraging. What value would I be looking for in the case of the MOVs?

A more onerous protection is to add a fuse in the HT secondary CT - but that has to be carefully selected to provide appropriate discrimination that may help with OPT faults.

I will need to double check what the fuses are and do as there's more than one, so that might be taken care of (and might have taken care/saved parts of the amp away from the OT!)

Apart from replacing the OPT, care is needed to confirm screen resistors are all still ok, and cathode R and C weren't stressed.

Fair point. How would I go about testing them aside from checking resistance and capacitance? Given the expense and ease of access I might be tempted to swap them out anyway, certainly if considering raising the cathode R.

In short, my inclination at the moment would be to do the following:
- replace OT
- raise cathode R from 50 to 75
- replace the dead GZ34 with a 5V4 - assuming the latter's 175mA capacity is adequate for the 5 preamp valves and 4 power tubes? Alternatively, a 10 or 20V zener?
- add backup 1N4007 on the valve socket pins
- replace the EL84's with sturdier 6P14P's
- add MOV's across the new OT half windings
- refrain from having a go at full volume just because, just in case

My hunch/hope is that should see me ride into the sunset with a happy AC30 for years to come. How does that sound to the good people of the forum?

Thanks a lot in advance once again, this has already proved very educational for me, for which I am very grateful.

[trobbins]

Can you clarify if both OPT primary winding sections (Anode to B+, and B+ to Anode) were shorted? It's not simple to appreciate why that happened along with GZ34 failure and both HT and mains fuse - and my first thought would be both sides of the output stage were conducting heavily, but unsure if that was initiated by a fault(s) or heavy blocking forcing both valves from each side conducting to their max on alternating cycles for long enough to cause OPT winding damage, which may then have cascaded to GZ34 failure and then fuse operations.

I checked out my son's VOX30 CC2 (circa 2007) a few years ago. That amp included a HT secondary CT fuse - a 0.5A time-delay (preferably use an IEC60127-2 fuse, not a UL fuse). That amp runs an 82 ohm common cathode resistor on its 'warm' setting.

It sounds like you didn't measure EL84 cathode currents to determine your 18W value?

Your somewhat high HT of 345V is worth clarifying as to why. As pdf64 suggested - what is your 6V heater voltage when loaded? Do you have a somewhat high mains voltage? Jumping to another rectifier may be a little knee-jerkish, as the 5V4 may raise HT a lot sooner than a GZ34, and its worthwhile doing a peak current assessment.

I installed 2x series MOV's, each of 330Vdc 1mA rating, across each half-primary winding in my son's amp. A single MOV with about 660Vdc 1mA rating would be an equivalent approach. Some background info on that selection is in https://www.dalmura.com.au/static/Outpu ... ection.pdf.

I don't have any info to comment on 6P14P's.

It's always fun to try and determine fault causes, and to recover from a bad fault and make an amp more bullet proof - especially when it's not your own amp!

[Stevem]

To avoid blowing any OT due to having the secondary side open ( like when a speaker(s) blow simply wire a 250 ohm 10 watt resistor across the output jack.

[solderhead]

I installed 2x series MOV's, each of 330Vdc 1mA rating, across each half-primary winding in my son's amp. A single MOV with about 660Vdc 1mA rating would be an equivalent approach. Some background info on that selection is in https://www.dalmura.com.au/static/Outpu ... ection.pdf.

That's a great article. I notice that it's got a Dec 2024 publication date, but I could swear that I read that article almost 10 years ago. Did you write that article several years ago and recently update it in 2024? Looking at my purchase records, I ordered MOV back in 2017 when we were discussing this at Ampage and I could swear that I read that article back then.

Fwiw I was working on amps with higher B+ so this is what I ordered back then: (the TDK/Epcos part is available in small quantities)
https://www.newark.com/epcos/b72214s062 ... dp/78Y9823

I ended up ordering the 14mm discs in the hope of getting a little more power dissipation before device degradation occurs.

On a related note, I also put these on the 120VAC mains:
https://www.newark.com/littelfuse/v14e1 ... dp/58K7242

I like to use a loading resistor on the OT secondary to protect amp heads from inadvertent unloaded conditions -- things like a user forgetting to hook up the speaker cab, or somebody inadvertently tripping on a cable and causing a disconnect. Granted, this isn't really necessary in the OP's situation (combo amp) but for separate head/cab setups I think it's a worthwhile safety measure.

Most of the things we've been discussing are covered in RG's series on Immortal Amp Mods. Since the OP is doing a deep dive into amp protection, RG's series of articles is worth reading. I can't find the original thread at Ampage but google lists the articles that he wrote later on for Premier Guitar magazine:
https://www.premierguitar.com/the-immor ... lifier-mod
https://www.premierguitar.com/the-immor ... -mods-pt-2
https://www.premierguitar.com/the-immor ... -mods-pt-3
https://www.premierguitar.com/the-immor ... -mods-pt-4

[trobbins]

All my articles are 'living', and I don't include a revision list as I typically end up doing many many updates over a few years. The date on an article is the latest 'update' date. I always seem to find something new to add 'just around the corner'.

I recall chasing down RG's articles, and emailing him on a few topics many years ago.

[R.G.]

Wow. That's reaching back a ways into history!

To make a PT substantially immortal, (1) use a properly chosen MOV across the primary, (2) fuse every winding, (3) make sure that you don't feed the primary an AC voltage higher than the trannie was designed for; this implies that you go find out what that highly sought after vintage PT was rated for. If you really, really want to do belt and suspenders, put a 150F or so thermal cutout on the iron and have this break the AC line.

To make an OT substantially immortal, (1) suppress transients across the primaries; (2) fuse each winding section, generally one on each half-primary; (3) don't let the power amp oscillate, whether from unloaded operation, or from wiring-induced oscillation. (1) and (3) have some subtleties involved.

There are really only three ways guitar amplifier transformers die, and they all eventually involve breaking down the winding wire insulation layer. A transformer is made of only three things: iron, copper, and insulation. The iron will work OK as a transformer up to temperatures of 770C, where it loses its magnetic properties; the copper is good up to its melting point, 1084C. It's the insulation giving up that kills them by shorting turns and letting >HIGH< currents flow un-impeded.

Insulation is rated for the temp it can withstand before not insulating any more, and for the voltage it can withstand. The commonest rating used to be 105C, but most new commercial transformers now use class 130C or better.

The three ways they die are (1) overheating till the insulation gives up, (2) mechanical abrading/damage wearing holes in the wire insulation, and (3) voltage puncturing the wire insulation so a short can happen between wires.

Overheating happens when a winding is forced to carry too much current, heating the wires and melting the insulation. This can be done by shorting or soft-shorting a winding. It can also happen when the primary is fed so much voltage that the iron is saturated and no longer limits primary current. To prevent this, don't over-volt the primary and current limit the windings. Fuses are good for this, as is just clever, thoughtful, and in-depth design. A sleeper is old-design transformers from the era when line voltages were smaller; in the USA, they were as low as 110 to 112Vac; today, we get 125, and old transformers struggle with it.

Another sneaky way to overheat is to have a power amp oscillate at very high frequencies, making the OT carry RF at very high levels. This is one danger of the unloaded tube amp.

Mechanical abrasion was a problem before transformers were impregnated with varnish and the like. The magnetic fields inside the windings wants to move the wires around. If they're not glued in place, they can wear the insulation through. All modern transformers are either impregnated, or use wire so big it can't move. Mechanical damage to the outside of the coil is also possible.

Transients punching a hole in the insulation happens from the primary with AC line transients, or from secondary transients caused by inductor flyback. On the primary, use a MOV, TVSS, ZNR, RC snubber or the 3KV reverse diodes to each end of the primary. On the secondary, don't open the load when the amp is carrying signal.