Transformer and valve protection

[JB]

I'm gathering parts for a JTM 45 build.  I already had some old but never used GEC KT66's and a friend recently gave me an original Radiospares Deluxe output transformer, as used in the early Marshall amps.  With near irreplaceable parts like that I started to think about what safety features to build in to protect them, without altering the tone if possible.

I'm thinking of things like:

1/ Secondary fusing - single fuse or both transformer legs?
2/ Silicon diodes in series with the GZ34
3/ OT primary flyback diodes
4/ Resistor on the bias pot(s) to protect against wiper opening

What are your opinions on the above and are there other tips worth including? 

[kagliostro]

Both transformer legs (for CT transformers)

K

[HotBluePlates]

1/ Secondary fusing - single fuse or both transformer legs?

Agree with Franco: fuse both legs if you're gonna do it, and don't put a fuse in the center-tap of the power transformer. Use a single HT fuse like Marshall between the OT center-tap and the first filter cap.

2/ Silicon diodes in series with the GZ34

This doesn't protect anything, but does keep the amp playing in the event of a shorted tube rectifier. So think of it as a way to keep the amp performing even after a failure.

3/ OT primary flyback diodes

I've never used them, but that doesn't mean they're not a good idea. 3x 1N4007's (or similar) in series is typical to give the flyback diodes a 3kV total rating. They mostly help if you play the amp with no speaker load attached, or speakers blow.

4/ Resistor on the bias pot(s) to protect against wiper opening

Always a good idea.

... are there other tips worth including?

I asked a question like this on this forum 10 years ago (or more). Basically, I wanted to know every typical failure mode for an amp so I could incorporate some way to prevent it as well as some external test point to identify it once it happened. As I recall, we never settled on "typical failures" and the only external test points I wound up with were bias points.

Why? Cause in the absence of abuse and catastrophic failure of parts, an amp may live fine for many decades. And the nature of "abuse" can be hard to predict when it is defined as "wrong-use".

Hopefully some others will have some good idea to incorporate.

[kagliostro]

Ciao HotBluePlates

This doesn't protect anything, but does keep the amp playing in the event of a shorted tube rectifier. So think of it as a way to keep the amp performing even after a failure.

I was thinking that the diodes in series with the tube has the same effect of diodes in series

so we obtain a higher reverse voltage ability that may be a form of safe, isn't so ?

this of course is to be added to the quality of avoid AC going on the circuit

---

One other idea will be fuses at the cathode of the final tubes (with or without the small led circuit that warn you if the fuse blows)

K

[stratele52]

Install many fuses is good , but install right value , if too high , no protection at all .

What Marshall do ;

www.drtube.com/schematics/marshall/jcm800pw.gif

[sluckey]

2/ Silicon diodes in series with the GZ34

This doesn't protect anything, but does keep the amp playing in the event of a shorted tube rectifier. So think of it as a way to keep the amp performing even after a failure.

A shorted GZ34 can take out a PT. And the resulting AC voltage on the B+ rail will likely damage some filter caps.

4/ Resistor on the bias pot(s) to protect against wiper opening

The JTM 45 doesn't have a bias pot. However, later Marshalls did. But the bias circuit that Marshall uses doesn't require a protection resistor on the bias pot because if the wiper (or pot) opens, the bias voltage goes max negative. That protection resistor is a good idea on Fender style bias circuits.

[HotBluePlates]

I was thinking that the diodes in series with the tube has the same effect of diodes in series

so we obtain a higher reverse voltage ability that may be a form of safe, isn't so ?

Yes, that's true.

But he's taking a known-good circuit with a GZ34, and adding silicon diodes in series. The GZ34 already has a high enough reverse voltage rating to work in the amp.

A shorted GZ34 can take out a PT. And the resulting AC voltage on the B+ rail will likely damage some filter caps.

I didn't think about the a.c. on filter caps. Thanks for pointing that out!!

It seems like the PT damage might be secondary to the filter cap failure, because the positive peak voltage is already being passed to the filter caps. But the negative voltage will certainly kill the caps.

This could be a brain teaser thinking of all the ways to kill an amp...

[sluckey]

This could be a brain teaser thinking of all the ways to kill an amp...

Absolutely. And adding layer upon layer of protection adds more complexity to a device. I have seen some equipment where the protection and monitor circuits were far more complex (and less reliable) than the circuit they were monitoring and protecting.

[jjasilli]

The vintage solution is a 10Ω 1/2W resistor from cathode to ground for ea power tube(s).  These resistors serve as a fuse and are easy to install, or to replace if blown due to an overcurrent condition.  (Also, you can also read bias across it, but adjust for 1 decimal place to the right,as a 1Ω bias sense resistor is usual these days.)

[kagliostro]

The vintage solution is a 10Ω 1/2W resistor from cathode to ground for ea power tube(s).

That is a vintage solution that is new to me

Thanks for sharing 

K

[JB]

I think you're right, there is the danger of over engineering the protection and ultimately reducing reliability.  This amp is only for my use so I hope the no load condition never arises - hence maybe not bother with the flywheel diodes.  10R high watt resistor across the speaker jack normal contacts would probably be sensible - thought doesn't cover the other end being unplugged or a faulty lead!

I think the no. one safety feature is the right fusing, in the right place.  I think the early JTM 45's fused the P.T. centre tap.  The re-issue has a B+ fuse between the rectifier and first filter cap.  I'm thinking of fusing both secondary windings into the rectifier.  Anyone bother with heater fusing?  Or known of a heater short causing PT damage?

I think I'll try the series rectifier diode trick.  I've seen a GZ 34 arc-over in a hi-fi amp - looked pretty whilst I was reaching for the mains plug!  No damage to amp, or rectifier.  Has anyone tried this though?  Does effectively pre-rectifying the feed to the rectifier do anything to the much coveted sag?

The original amp doesn't have adjustable bias but I was going to put dual pots in so I can balance up unmatched KT 66's.  But I'll arrange in a way that doesn't cause bias loss on pot-failure.  And use decent pots.

The 10R cathode resistor trick sounds good. There'll be bias resistors at that point anyway so why not.

How about dual bias diodes - one from each half of the secondary then combined?  Would cover for diode failing open circuit and hence bias supply collapsing.


 

[Willabe]

Does effectively pre-rectifying the feed to the rectifier do anything to the much coveted sag?

No, all the B+ current still has to go through the rectifier tube so it will still sag.


             Brad     

[HotBluePlates]

How about dual bias diodes - one from each half of the secondary then combined?  Would cover for diode failing open circuit and hence bias supply collapsing.

I've never seen a bias diode fail.

I have seen a failed bias filter cap, but the result was simply hum, which was not stopped when the phase inverter tube was pulled. You'd think that a.c. instead of negative d.c. would cause an overheated tube, but it seems the time the voltage swings negative offsets the heating of the voltage swinging positive.

That said, I've only ever seen that happen once (in a Fender Princeton Reverb).

[stratele52]

I have seen a failed bias filter cap, but the result was simply hum, which was not stopped when the phase inverter tube was pulled. You'd think that a.c. instead of negative d.c. would cause an overheated tube, but it seems the time the voltage swings negative offsets the heating of the voltage swinging positive.

That said, I've only ever seen that happen once (in a Fender Princeton Reverb).

I see that one months ago in a Gibson Duo Medalist

[jjasilli]

10R high watt resistor across the speaker jack normal contacts would probably be sensible - Hi W resistor not needed if you boost value.  See:  http://ampgarage.com/forum/viewtopic.php?t=16086&highlight    Or Hiwatt used a dead short:  http://mhuss.com/SmallBox/page4.html

[silverfox]

Is it possible to use a current sense on the transformers to detect a surge. Of course this would require incorporating digital parts like OP Amps... I don't know if this can be done or it is likely the circuit would trip on bass frequencies.

Couple this with an AC detection circuit in the power supply to sense rectifier failure. A diode with the positive end connected to the negative power supply rail would conduct if an AC wave passed through the rectifier. Feed this to a Crowbar circuit to energize a relay that locks out the AC feed.



Silverfox.

[jjasilli]

My understanding is that the new powerful hi-fi tube amps are doing something like that:  http://www.mcintoshlabs.com/us/Products/pages/ProductDetails.aspx?CatId=amplifiers&ProductId=MC2301

Current draw & bias circuits are internally "monitored" and the amp will shut itself down if there's a problem.

[silverfox]

I don't know how reliable Mean Time Between Failure is. Just thinking about it. There are a lot of electronic devices that run for years and years without failure and have a relatively high component count. Typically, after a 30 day burn in period, a product will perform years beyond the calculated failure point based on component count. Consider the Desktop computer or video game properly taken care of.

I'd have to see some data rather than rely on a market driven computer analysis based in profit. Not criticizing, just an observation.

Silverfox.