EL34 Tube Overheats

I have a modified ST70 with solid state rectifier and on two occasions I have had a single EL34 tube get red hot and the bias voltage to spike. I have to watch it constantly now. There are individual bias adjustments for each tube. I usually run the bias below 1.5v on each tube. What would cause two separate power tubes to spike in voltage and overheat? I am worried that the tubes could be ruined bow.

Was the main cap replaced? If so it may be the tubes are getting ready to say goodby. I had problems with the bios just before one of my tubes died.

Before we try to diagnose your ST-70, you mentioned it is modified and had individual bias adjustments for each of the 4 EL34 tubes. Can you tell us how many ohms is the resistor to ground of each EL34 tube. There is a resistor from ground to pin 1 and 8. A stock Dynaco uses a 15.6 ohm resistor for each pair of EL34 tubes (left side and right side). When you set the bias to 1.56 volts you are setting the EL34 tubes for 50ma each.

I have seen when Dynaco's are modified where you can set the bias of each tube, 10 ohms is used from the pin 1/8 to ground. You will have to set the bias to .5 volts per tube. Can you verify the resistor value?

If you have 10 ohm resistors, and you set the bias for each tube at 1.5 volts, you will be passing 150ma through the tube which can't handle that plus your power transformer will cook.

Sal

Sailor - All the Caps were replaced. I was wondering if the tubes were going. The two tubes that overheated now have less silver at the top now. Thanks for the response.

Sal - I will check the resistor. When I first got it the bias was at about 1 on each tube. I upped it to the 1.56 but backed it off to around 1.2 because it was running hot. I run a fan on it most of the time, but never had a tube glow red until now. Thanks for the response.

are you using a stock ST70 driver board, a VTA70 driver board, or some other driver board?
Yes, check to see what the value of each cathode resistor is, on the stock ST70 it was a 15.6 ohm shared by each pair of output tubes.
On most amps modified to use a bias pot for each output tube (such as the VTA70 mod) those resistors are 10 ohms.
The correct bias with a 10 ohm resistor on each output tube is .40 to .50 vdc (which is 40-50ma).
Also, modern tubes seem to have trouble going into what I call "bias runaway".
The stock Dyna 70 used 270K "pulldown" resistors (eyelets 1 & 2, 22 & 23).
On the VTA70 mod, most shipped within the past year have been changed to 220K.
On some of them I've even supplied 150K, which definitely works better with the "new" repro Mullard EL34s,
and any using KT88s (MK3 amps use 150K, and many tube amps using KT88s use 100K in that position).
That lower value will tend to keep away the trouble of "bias runaway".

max grid leak resistance for EL34s is usually listed as 500k per tube so if your using 1 grid leak resistor for a pair generally 250k would be considered max but 220k is used since its a close common value resistor. Remember any grid resistor placed on the grid (pin 5) of the output tube will also be included in the grid leak resistance along with the grid leak off the PI plates. Its possible your experiencing bias runaway but normally this will only happen when your playing some loud enough for the bias to jack up. Does it redplate at idle or during play?

tubes4hifi - Thanks for the reply. Here is a picture of the board used "http://www.flickr.com/photos/8486416@N02/7466673200" I still need to pull the base and check the resistor.

dunwichamps - Thanks for the reply I will check the value of the grid leak resistor. I was running the amp loud both times when this happened, and the amp is running the electrostatic panel of some Martin Logan SL3s. I am plugged in the the lower ohm taps. I am thinking the tubes may be on their last leg and this could be contributing to it. I am looking for a tube tester and to buy some new tubes. The last time it overheated there was a burning smell so I need to check the wiring and solder joints.

yea sounds like bias runaway if this happens when its playing loud.

I was wondering if I should put in one of those timer delay boards since this has solid state rectifier. Maybe my tubes are near end of life and that is contributing to the run away. I bought this about four months ago and have played it a lot. I will check all the resister values and adjust bias accordingly and change the grid leak per recommendations here.

Thank you every one for your suggestions and help. I need to get it back up running; I am addicted to the tube sound now.

scuba,
your photo link seems to be a "private, unshared" photo, so can't be accessed.
as for tube testers, having owned 3 or 4 over the past 15 years, the best tube tester is the ST70 !!
If the tube works and holds steady bias, it's good! Tells you more than most tube testers!
Also as the bias changes you can see the tube wearing out, or how it compares to a new one.

Try the picture now, I opened up the permissions on it.

OK, photo is viewable now. So we still have an unknown board, unknown bias resistors, unknown grid resistors, unknown "pulldown" resistors.
It is probably as simple as finding out what the bias resistors are, setting the correct bias (around 40-45ma per tube),
and making sure the "pulldown" resistors are not too large.
All that can be done in about 2 minutes, just looking under the bottom cover.

Hi -- according to the picture, the board uses 4 tubes. Is this some version of the Wellbourne Labs upgrade for the ST70? The Wellborne is the only PCB that I have seen that uses 4 tubes. Out of curiosity, what are the 4 tubes?

Just my two cents worth.

I can see the 4 tubes are 6922, which is also what the Curcio mod uses, but this is NOT a Curcio or Welbourne mod,
as I'm familiar with both of those. The Welbourne uses a 6U8 (ughg!) and 5751 tubes.

I am away from where the amp is located, but I should be back Sunday night or Monday. I will post results as soon as I get a chance to pull the bottom. I do not know who the maker is of the board, but will look for markings when I open it up.

New Pictures of everything is located at - http://www.flickr.com/photos/8486416@N02/sets/72157630387243290/

I see a lot of spaghetti and a lot of stuff not on the original schematic. This is heavily modified!

On the two tubes that went red plate the bias resistor is cooked. The resistors on the two remaining bias resistors are 15ohms. I get the math now using ohms law (I = V / R). So my voltage should be .75 (750mv) with a 15 ohm bias resistor on each tube. If my bias resistor was shared then I would halve this number. Now the really strange stuff:

- all my tubes are missing pin 5 (I think but looking at schematic it must be pin six missing)
- pin 1 is not connected to anything on every tube
- pin 4 had a 30 ohm resistor before going to the transformer on every tube

On the pull down resistor I see a 22 ohm resistor off of pin 5 to 6 (one of these pins is not is use - see above bullet) and them wires are going from pin 6 to the unknown board. It looks like it is connecting to the white and green cap and green resistor that you see on the top picture of the amp (example right side right of the right 6922 tube up front). Then there is a series of three blue resistors. I will have to look at it later tonight to determine if this is where point 21, 22 and 23 is (or 1, 2 and 6). I am assuming if these are on this board then they may have accounted for that.

Looks like my first task will be to replace the two fried resistors. When the amp last went red plate it sounded fine, if that means anything at all. Looking at the two tubes that went red plate (happened about a week apart) the silver at the top of the tube is less than the other two tubes, so I am guessing best case I have just lost some life on them. I do not see any other burnt electronics. Could the tube socket get damaged, because the pin 8 seems lose?

There is a Repair sticker on the bottom for some place in Oregon (503-284-8693). I am calling them later today.

I think I read the resistor wrong on pin 5 to 6 it must be a 1k (brown black red and not red red black) and it goes to the resistors and Cap I mentioned before.

looking at the pictures the resistor I thought were 22 and 23 to 21 may not be the ones.

the 1K resistor from pin 5 to pin 6 is correct. 15 ohm cathode resistors are fine, so yes, set your bias to read 750mv across each one of those.
Setting bias is usually interactive, which means you'll have to keep re-adjusting a few times as changing one will influence the others until they are all very close to the same amount.
Pin 1 and pin 8 on each output tube should be tied together (screen grid and cathode) then thru the 15 ohm bias resistor to ground.
The 33 ohm resistors on pin 4 tied to the OPT should be fine, that's just added protection.

Thanks tubes4hifi this has been a great learning experience (and some re-learning) - What is the bottom line impact of not applying the negative bias to the grid (pin 1 - grid no. 3)? I need to replace the two burned out resistors and tie in pin 1 before I can test it out. I am hoping I have not fried the two tubes that were on the burned resistor circuits. I think I was running the amp with one burned out resistor for a while without it red plating. I had the bias turned back on that tube though. I definitely fried those resistors by having the bias up too high. Weird part is I had been running the amp for sometime with the high bias, but I think I was running the volume much lower.

Added pictures of the top of the circuit board and noticed transistors (link at the start of the post). I guess I have a Hybrid - LOL

Will a 1/2 watt resistor be OK? I calculate the watts at .0375 when I have the bias correct on 15ohms per tube. However should I be as low as 1/8 on the idea that you want the resistor to burn out if you get a run away bias, so nothing else fries? However the amp was still playing when the the resistor burned and it did not seem to stop anything. I do need to check the resistor when I pull it out to see if it is allowing any current through.

I recommend that you change all 4 of those resistors to 10ohm 1% wire-wound resistors. The wattage is not that important. You can get them from Mouser for about a buck apiece.
http://www.mouser.com/ProductDetail/Vishay-Dale/RS01A10R00FE12/?qs=NmlBeBNZel40XkxyugdQpeY%252bm6ctSkJqvwtDFt6GE1k%3d

hawaii.ken - Thanks for the replay. Funny you recommend that. I went to Radio Shack and picked up some half watt 15 ohm and got some one watt 10 ohm since there was no one watt 15 ohm in stock. That would make my bias voltage half a volt then. The 10 ohm ones are metal film blue color resistors.

10ohms would give you .50v at 50ma per tube.

I recommend 36ma to 44ma per tube.