Strange coincidence, possible connection?

This morning per my usual practice turned on my M125's to let them warm up and turned my attention to a few chores. These days they see 10-12 hours on a typical day. Been running like this for over half a year now with zero issues. Today was different. My wife says I hear an odd noise coming from the basement. One of my towers was just screaming, and one of the amps was red plating including both the GZ37 rectifier and a couple of the 6550 power tubes.  

The coincidence is a couple days ago I switched the amps over from UL to Triode mode. Over the months of trouble free operation they were always in UL mode. Could be a coincidence, but seems odd that this should happen so soon after changing modes when the amps have been so behaved over the past few months. Is there any technical reason associated with the Triode setting to suspect these could be somehow related?

WILD speculation here:

Clearly one of the amps went into parasitic oscillation.
Bob's OPTs are not inherently prone to that problem - suggesting it is for some other cause.
Triode Mode effectively pulls the grids out of the circuit.
This will cause the power-tubes to clip much more quickly (at, effectively, half-power).
I do not know how that triode-switch affects GNFB - but if it does, and you had the volume setting "up" on the preamp, I would start there.
If I were to make any sort of global suggestion to tube amp manufacturers using (relatively) high-powered pentodes in their design, it would be to eschew the triode option. It is a silly feature based on myth and received wisdom such that it a customer really does demand the option, offer it as a DIY, with a large (necessary) list of cautions.

Usual question:

Do you have a way to test tubes?
A GZ37 is a very nearly bullet-proof tube.
A 6550 may at least be described as 'rugged'.

It would be nice to know whether the affected tubes are toast or not.

Best of luck with it, and sorry for your loss

First interesting observation... voltage at wall plate today is at the extreme top end of the range for my location at 123v. Another coincidence. My bucker knocks off around 3.5 volts. But maybe this is not enough for Tridoe? ie is the tolerable voltage for these amps lower in Triode mode then UL mode?

Update #2.

Fired up the amp again and noticeable distortion. Either the event took out a speaker driver (not only was the spl level really high it also had this violent oscillating character) or tubes are compromised. I dont have a tester but I do have brand new spare tubes. If its the speaker I'll be gutted.

LeGrace wrote:First interesting observation... voltage at wall plate today is at the extreme top end of the range for my location at 123v. Another coincidence. My bucker knocks off around 3.5 volts. But maybe this is not enough for Tridoe? ie is the tolerable voltage for these amps lower in Triode mode then UL mode?    

From my admittedly limited understanding of what Triode Mode is on a pentode - it removes the grid(s) from the circuit - well, not quite. It connects the screen grid to the plate. Now for a poor, but somewhat timely analogy again not knowing exactly how the VTA handles GNFB in Triode Mode: Nascar just ran its last race with plate-restrictors at Daytona recently. These devices were designed to reduce extremes of acceleration and cut down on extremely high speeds.

https://en.wikipedia.org/wiki/Restrictor_plate

I am not so sure that high voltage is/was anything more than a contributing factor, maybe what pushed an already marginal situation over the edge. When in Triode mode, a pentode is still consuming just as much power, but is dissipating it differently. NFB helps amplifier stability and reduces parasitic oscillation. Is in any way altered in triode-mode???

I would love it were an expert to either confirm, debunk or clarify these altogether speculative ramblings.

One coincidence can be just exactly that, but when I see two stacked on top of one another statistically the odds change. Ever watch the show Mayday? Its almost always this thing happened on top of that which is what ultimately led to the accident. Neither factor alone would have led to the tragic outcome.  

Going to go change-out some tubes now, wish me luck. If the speaker is damaged some amps are going to find a new home on the curb.

I once made a wiring mistake - and the subsequent feedback put 400 watts through a single dome tweeter. The voice-coil blew right out of the front of the speaker into the grille cloth. Red hot.

This was back-in-the-day, and AR replaced it, gratis, with a very nice note 'suggesting' that I be more careful next time. I had sent the old one back with a check, explaining my error - so it was not as if they did not know what happened.

Update #3

Sigh of relief, with fresh tubes installed the distortion is gone. I had a draft email ready in my mind to send to my Tannoy dealer but thankfully it wont be necessary. The driver is a 10" dual concentric so would be expensive to have replaced. Have both amps back in UL mode and everything appears stable again.

I'm wondering to run safely in Triode mode do I need switch to the alternate taps on my bucker transformer? This would provide closer to 7v in step down.

LeGrace wrote:Update #3

Sigh of relief, with fresh tubes installed the distortion is gone. I had a draft email ready in my mind to send to my Tannoy dealer but thankfully it wont be necessary. The driver is a 10" dual concentric so would be expensive to have replaced. Have both amps back in UL mode and everything appears stable again.

I'm wondering to run safely in Triode mode do I need switch to the alternate taps on my bucker transformer? This would provide closer to 7v in step down.

How about simply welding the switch into the Pentode/UL position? Then put a large rat-trap (set, of course) next to it? I cannot imagine why one would purchase (for a tube device) a high-powered amplifier, then choke it half to death. Like forcing Kiri TeKanawa to sing only in contralto.

Peter W. wrote:

LeGrace wrote:Update #3

Sigh of relief, with fresh tubes installed the distortion is gone. I had a draft email ready in my mind to send to my Tannoy dealer but thankfully it wont be necessary. The driver is a 10" dual concentric so would be expensive to have replaced. Have both amps back in UL mode and everything appears stable again.

I'm wondering to run safely in Triode mode do I need switch to the alternate taps on my bucker transformer? This would provide closer to 7v in step down.

How about simply welding the switch into the Pentode/UL position? Then put a large rat-trap (set, of course) next to it? I cannot imagine why one would purchase (for a tube device) a high-powered amplifier, then choke it half to death. Like forcing Kiri TeKanawa to sing only in contralto.

The M125 offers sufficient power to actually be viable in this mode. To me it did sound more lush/rich. Why I am still interested in understanding if dropping voltage further will address the issue.

Of a certainty, lower voltage (~115 - 118 VAC) would be better.

As I do not know exactly how the Triode option works, and whether GNFB is affected, that is only unsupported certainty - i.e. Opinion.

Could've just been the end of life of those tubes.

corndog71 wrote:Could've just been the end of life of those tubes.

Maybe, but they only had @ 9 months on them.

LeGrace wrote:

corndog71 wrote:Could've just been the end of life of those tubes.

Maybe, but they only had @ 9 months on them.

But you say you've been running them 10 - 12 hours a day, that's well over 2000 hours by now!

Still hoping one of the resident experts will post an answer whether Triode mode stresses the tubes more vs UL. Wish to better understand the technical aspects of how the two modes differ from a tube operating perspective.

Though it is sounding like rewiring my bucker to lower incoming voltage down even further would be a good idea.

LeGrace wrote:Still hoping one of the resident experts will post an answer whether Triode mode stresses the tubes more vs UL. Wish to better understand the technical aspects of how the two modes differ from a tube operating perspective.  

Though it is sounding like rewiring my bucker to lower incoming voltage down even further would be a good idea.

As long as the tube is operated within it's capability/specs, operating it in triode does not stress the tube any more than in pentode or ultralinear mode.

Someone else asked if global NFB would be affected by using triode mode - the answer is yes. Triode mode offers lower gain. And since global NFB simply uses a voltage divider to feed back some of the output voltage back to the front end of the amp it means that with less gain there is lower output voltage - and therefore less to feed back.

Peter W. wrote:I once made a wiring mistake - and the subsequent feedback put 400 watts through a single dome tweeter. The voice-coil blew right out of the front of the speaker into the grille cloth. Red hot.

This was back-in-the-day, and AR replaced it, gratis, with a very nice note 'suggesting' that I be more careful next time. I had sent the old one back with a check, explaining my error - so it was not as if they did not know what happened.

Did that to a KLH 17 once with my old Heathkit AR-15 running in full-clip mode. Ah, dem college days. Henry Kloss wasn't so kind; he kept and cashed the cheque.

Jim McShane wrote:

LeGrace wrote:Still hoping one of the resident experts will post an answer whether Triode mode stresses the tubes more vs UL. Wish to better understand the technical aspects of how the two modes differ from a tube operating perspective.  

Though it is sounding like rewiring my bucker to lower incoming voltage down even further would be a good idea.

As long as the tube is operated within it's capability/specs, operating it in triode does not stress the tube any more than in pentode or ultralinear mode.

Someone else asked if global NFB would be affected by using triode mode - the answer is yes. Triode mode offers lower gain. And since global NFB simply uses a voltage divider to feed back some of the output voltage back to the front end of the amp it means that with less gain there is lower output voltage - and therefore less to feed back.

Perhaps not the right question, as stress is not a metric with associated values. Are the tubes exposed to identical voltage parameters under both conditions? If yes, then why do they sound different?

Don Sachs added a triode option to my HK amp when he rebuilt it and while I understand the concept of reducing power output it seems unnecessary at least as far as my needs go. Nice feature but i just dont get it...we used power soaks on big Marshalls but this is obviously a different concept. The soak let us crank the amp but not destroy the apartment or have the cops show up.

wildiowa wrote:Don Sachs added a triode option to my HK amp when he rebuilt it and while I understand the concept of reducing power output it seems unnecessary at least as far as my needs go. Nice feature but i just dont get it...we used power soaks on big Marshalls but this is obviously a different concept. The soak let us crank the amp but not destroy the apartment or have the cops show up.

The purpose of the triode switch/triode connection was NOT to reduce power. Changing from ultralinear or pentode to triode changes the gain (and therefore the amount of global NFB), it changes the distortion spectrum of the amp, it changes a whole lot more than just the power. In fact the loss of output power is not desirable - but it's unavoidable when you change to triode.

All those changes have some effect on the amp's sound.

...incoherent mumblings....

Guys and gals, once upon a time, there was only the triode. It serves, in the tube world, much as the hit-or-miss gasoline engine serves in the 4-stroke engine world. Useful, better than nothing, but not exactly a paragon of efficiency!

https://www.youtube.com/watch?v=Z5X0bIh2Up8  

Then along came the power-pentode.

https://www.youtube.com/watch?v=4XORCX_cnSg  

A bit of a different proposition.

Now we have "triode mode" making a pentode "behave more-like-a-triode"

https://www.hemmings.com/magazine/hmn/2008/04/Cadillac-V-8-6-4/1610045.html  

Once more - a power pentode is NOT a triode. Operating it such that it mimics a triode (will never be more than an emasculated tetrode in actual service) will do all sorts of things - not necessarily fatal, but certainly contrary to the proper function and actual design of the tube. Want another analogy - going off-label with a drug. Might be benefits and there will definitely be differences - and possible consequences.

But, what one does, effectively and in most cases (not sure of VTA, just the general case) is connect the screen grid to the Anode - with a resistor to reduce/eliminate parasitic oscillations. Which, from the description given, seems to be what happened here.

Leading to the logical conclusion: if one wants a triode amp, beg, borrow, purchase, build or steal a triode amp.

Here's an informative presentation on the topic:

http://www.audiodesignguide.com/New2A3/ETF06TS.pdf

Peter,

You analogy reminds me of the Ford 2.9L V6 Cologne motor in my Granada in 1983 that used to shut down 3 cylinders if not needed at cruise. the EEC module at that time could not provide a smooth transition either and the project was abandoned. The only lasting part of this aborted attempt was the continued use of a split inlet manifold system and two upside down vane air metering units that filled with water on a frequent basis...

Back on topic, on most UL amp designs we normally see a resistor in the feed to the screen grids, possibly to limit disipation, but on this Dynaco the UL winding is directly connected to the screen. Has anyone done a study to see the effect of adding a resistor on tube life, power and distortion etc?

Steve

Thank you!

Note that the thrust of this white-paper was the re-purposing of vintage, but very common TV-focused tubes for audio use. Not the emasculation of audio-focused tubes because one can.

Schlangen presents a lot of graphic, illustrative information about the entire topic - triode, tetrode, pentode, variations on operation, load lines, and more - then concludes with examples of TV pentodes connected as triodes for audio applications.

VTA M125 "emasculation" occurs through connecting the screen grid to the plate through a 100R resistor as a grid stopper to eliminate oscillation, just as you indicated earlier.

I personally prefer the tone, response, and power of my M125s in pentode mode. I have built a couple of tube guitar amps with pentode/triode switches, but again, prefer the pentode tone, feel, and punch.

Returning to the OP's original reasons for posting, those tubes have seen a lot of use, so perhaps one or more failed from use. It would be good to check the grid stoppers to make sure one or more isn't open. I haven't found any readily available online references that state operating a pentode in triode mode puts more stress on the tube, but maybe Bob or Ray will chime in...