Tube DIY Asylum

Do It Yourself (DIY) paradise for tube and SET project builders.

Return to Tube DIY Asylum

Message Sort: Post Order or Asylum Reverse Threaded

Thermal runaway?

74.96.187.191

Posted on April 29, 2022 at 10:38:34

Lew
Audiophile

Posts: 10912
Location: Bethesda, Maryland
Joined: December 11, 2000

Last night I had what was for me a bizarre experience. I own and use modified Atma-sphere OTL amplifiers. One of the ways in which my amps are unique is that I built them (from parts supplied to me about 25 years ago by Atma-sphere) such that I can separately bias each output tube. Also, I am using only four output tubes per monoblock, type 7241 triodes. I have had decades of trouble-free operation with these amplifiers. I check output tube bias about once every 6 months to a year. Last night, while in the process of checking bias, one tube was found to be idling at more than 1A DC; I usually bias them at about 450mA, so as to preserve tube life since 7241s are hard to find and very expensive when you can find them. So using my meter to monitor DC current, I set about the job of reducing the bias current. I am pretty sure that I initially rotated the pot in the wrong direction, so that current went up. Then when I realized my error, I tried to dial it back down. This didn't work; bias current kept increasing, up over 2A, and the plates of that tube started to glow red. I quickly shut down the output B+ supply.

Could this be due merely to "thermal runaway" (bias current keeps increasing regardless of the bias supply voltage), or do I need to suspect something else that requires me to open up the chassis and see what's going on? Right now one plan is to replace that relatively new tube with another that is well worn, so as not to put the newer tube at risk, dial back the bias control all the way down, and start up the amp again. If there's something else wrong in the circuit, the bias should be out of control again. If the event was due to "thermal runaway", perhaps there is no real problem that needs fixing. This amplifier was working fine until I set about to re-bias the output tubes, so it seems unlikely that there is a major internal problem, like with the bias supply. Also, the other tube on that same side of the circuit was responsive to bias adjustment.

Hide full thread outline!

RE: Thermal runaway?

trobbins

17:20:54 05/11/22

RE: Thermal runaway?

Lew

20:05:13 05/14/22

RE: Thermal runaway?

7868

00:19:57 05/15/22

...

Show full thread outline!

RE: Thermal runaway?

trobbins

17:20:54 05/11/22

RE: Thermal runaway?

Lew

20:05:13 05/14/22

RE: Thermal runaway?

7868

00:19:57 05/15/22

RE: Thermal runaway?

trobbins

14:23:27 04/29/22

RE: Thermal runaway?

Lew

17:02:39 04/29/22

RE: Thermal runaway?

Triode_Kingdom

11:31:52 04/29/22

RE: Thermal runaway?

Lew

12:53:43 04/29/22

RE: Thermal runaway?

Triode_Kingdom

01:08:33 04/30/22

RE: Thermal runaway?

Lew

10:42:57 04/30/22

RE: Thermal runaway?

Paul Barker

03:13:00 05/01/22

Lew

08:57:03 05/03/22

7868

02:01:49 05/04/22

Lew

08:30:51 05/04/22

Triode_Kingdom

05:52:47 05/05/22

Lew

08:16:35 05/05/22

Lew

07:50:28 05/11/22

Tre'

11:05:38 05/22/22

RE: Thermal runaway?, posted on April 29, 2022 at 11:31:52
Triode_Kingdom Audiophile Posts: 10048 Location: Central Texas Joined: September 24, 2006	This might be a combination of amplifier defect and thermal runaway. I think you'll need to determine why it was idling at more than twice the other tubes. Maybe the answer will become clear when you replace it.

RE: Thermal runaway?, posted on April 29, 2022 at 12:53:43

Lew
Audiophile

Posts: 10912
Location: Bethesda, Maryland
Joined: December 11, 2000

Obviously, the best thing to do would be to open it up a priori. The problem is that the amplifier is now too heavy for me to carry in my old age. If you have ever seen an Atma-sphere MA2, old style, these amps use the same massive and massively heavy two EI type power transformers per chassis. I may end up doing the autopsy on my listening room floor to avoid having to transport it to my basement workshop. It should be very easy to spot a problem, if there is even a problem, once I can see what is going on with the wiring. Actually, I was going to use the 7241 I once bought from you (TK) as the tester, since it has some age on it anyway.

So the point of my question is does the scenario I described fit the description of thermal runaway, i.e., the tube goes off on its own despite the re-application of bias voltage, once the plate current exceeds a threshold for that tube? Or is this just wishful thinking? Thanks for the response.

RE: Thermal runaway?, posted on April 29, 2022 at 14:23:27
trobbins Audiophile Posts: 189 Location: Melbourne Joined: March 26, 2012	It seems to fit a gassy valve developing, with grid voltage going more positive by its own grid current. If your amp can handle it then you could confirm by swapping the suspect tube and confirm the good swapped tube biases ok, but the suspect one continues to act badly. A gassy valve isn't going to come back to normal.

RE: Thermal runaway?, posted on April 29, 2022 at 17:02:39
Lew Audiophile Posts: 10912 Location: Bethesda, Maryland Joined: December 11, 2000	Thanks. I'm planning to do exactly that experiment.

RE: Thermal runaway?, posted on April 30, 2022 at 01:08:33

Triode_Kingdom
Audiophile

Posts: 10048
Location: Central Texas
Joined: September 24, 2006

You must be thinking of someone else. I don't remember ever seeing a 7241, let alone selling one.

I know what you mean about weight. I'm rebuilding a pair of 211 SET monoblocks that I started in a different form about 20 years ago. They are rapidly approaching 80 lbs each. My workspace is really limited, so I frequently have to carry them up and down a flight of stairs to the bench. This is definitely not fun.

RE: Thermal runaway?, posted on April 30, 2022 at 10:42:57

Lew
Audiophile

Posts: 10912
Location: Bethesda, Maryland
Joined: December 11, 2000

I am guessing that my Atma-sphere monoblocks weigh somewhere between 80 and 100 lbs each. When I did formerly transport them to the basement workshop, I would remove the cosmetic but heavy alu faceplate and side pieces first. That probably reduces weight by 10 lbs, maybe. Then I remove all power tubes, lest I drop the amplifiers while they bear maybe $800 to $1K worth of 7241s. Then I can lift them off their stands, but, like you said, walking down the stairs to the basement is a scary proposition. I probably need a dumb waiter/elevator built into my house. (The original version of these amps as supplied by Atma used six 6C33Cs per chassis. I converted to four 7241s, for about the same power out.)

Almost forgot to mention: I could have sworn I bought one used 7241 from you, maybe 6-8 years ago. I wrote "TK" on the box. I am not doubting your memory of not having owned a 7241; I will have to take another look at that tube and its box.

RE: Thermal runaway?, posted on May 1, 2022 at 03:13:00

Paul Barker
Audiophile

Posts: 911
Location: North Yorkshire
Joined: November 17, 2002

Does the amp have cathode bias? when you get the amp upside down, check the bypass cap for that valve isnt drawing current.

Aged 64 here, where did the time go since vsac? Working as a heating engineer until Im 66 but now its very arduous and taking years off the end of my life. When I was an apprentice the old guys were respected. Today youre told off for being too slow by young guys with 4 years experience which is a very long time these days. Age impacts on my work life and my love of boatanchors.

There is an EE smps expert on youtube that claims he can make smps's that are great for audio. I wish I could afford his fees. Meanwhile I'll muddle on lifting heavy choke input power supplies. What doesnt break you makes you stronger!

Some days I feel alone holding a candle for choke input, so last century, but so much better, all be it 4 times the weight.

Grid bias, posted on May 3, 2022 at 08:57:03
Lew Audiophile Posts: 10912 Location: Bethesda, Maryland Joined: December 11, 2000	And I've now determined that's where the problem lies. The driver tube for that power tube tests ok on my Hickok, so it may be the potentiometer, which means the amplifier has to be moved to my basement. Meanwhile the tube that momentarily passed 2A is likely ruined.

RE: Grid bias, posted on May 4, 2022 at 02:01:49
7868 Audiophile Posts: 82 Location: Wairarapa Joined: March 4, 2022	If the bias pot is a carbon one then that's where I would look first. Don't like them for that purpose.

RE: Grid bias, posted on May 4, 2022 at 08:30:51

Lew
Audiophile

Posts: 10912
Location: Bethesda, Maryland
Joined: December 11, 2000

I originally built the amplifiers in the late 90s with parts supplied by Atma-sphere. Back then I was more of a novice in the art of circuit design and understanding, so I had help from a local guy. We used milspec pots to bias the output tubes (six per chassis, because the build permits six 6C33Cs per channel that each can be individually biased). Since then, although I have made many upgrades to the circuit, at least in my mind, I have not had to mess with those pots. I doubt they are carbon, but we shall see. Assuming at least one of them is bad, I would welcome a suggestion about what brand or type is most likely to remain stable and last a long time.

RE: Grid bias, posted on May 5, 2022 at 05:52:47
Triode_Kingdom Audiophile Posts: 10048 Location: Central Texas Joined: September 24, 2006	How is your bias circuit wired? I always design so that an open or intermittent wiper will cause output current to decrease.

RE: Grid bias, posted on May 5, 2022 at 08:16:35
Lew Audiophile Posts: 10912 Location: Bethesda, Maryland Joined: December 11, 2000	Need to identify the problem before I speculate. I'll let you know.

Follow-up, posted on May 11, 2022 at 07:50:28

Lew
Audiophile

Posts: 10912
Location: Bethesda, Maryland
Joined: December 11, 2000

I don't quite understand why bias current drifted upward over time, although I have a theory. But the bottom line is I found no problem with the bias circuit. I am going to install some output tubes and re-check on the bench, but it might be just a tube that went bad.

Yes, with tubes installed, all voltages are as expected and stable, and the bias supply is fine. I think for some reason the plate current went up and this damaged the tube beyond salvage in some way that caused the grid to suck current. (With the bad tube attached to the bias supply, I observe that the bias voltage sags significantly [becomes less negative] but only on that bad tube; bias to other tubes is unaffected.)

RE: Thermal runaway?, posted on May 11, 2022 at 17:20:54
trobbins Audiophile Posts: 189 Location: Melbourne Joined: March 26, 2012	Did you confirm that the fault moved with a particular tube, or always occurred at the same location no matter what tube was in there?

RE: Thermal runaway?, posted on May 14, 2022 at 20:05:13

Lew
Audiophile

Posts: 10912
Location: Bethesda, Maryland
Joined: December 11, 2000

No, it was the tube. There is nothing at all wrong with the bias circuit. The only question in my mind is why did the bias current drift upwards over time sufficiently to send that one tube over the edge? I have an idea it is something I did; I have undone it. Now I will put the amp back in service and check bias more frequently to be sure I cured the issue, hopefully before another tube goes up in flames, metaphorically speaking.

RE: Thermal runaway?, posted on May 15, 2022 at 00:19:57

7868
Audiophile

Posts: 82
Location: Wairarapa
Joined: March 4, 2022

Big bore tubes need to be run cool and have a grid reisitor that is not too high in value. Cathode bias can have a larger grid resistor than fixed bias due to the self current limiting of the cathode bias.

There may have been heating up of g1 increasing the current through the tube - the new versions of the 7591 & 7868 made in Rusia are well known to be prone to this in unmodified amplifiers. The original tubes could handle far higher grid resistor values than the Russian replacements. Jim McShane wrote about this.

RE: Follow-up, posted on May 22, 2022 at 11:05:38
Tre' Industry Professional Posts: 17303 Location: So. Cal. Joined: February 9, 2002	That sounds like the grid is partially shorted. Tre' Have Fun and Enjoy the Music "Still Working the Problem"

Page processed in 0.035 seconds.