• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Need your help: Change from el34 to 6550 ends in a desaster

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

I am running for some years now the ul40s2 from Menno with the autobias:

Autobios improved

Everything is nice and quiet. I am running this amp in ultralinear.

I have just pulled the el34 and plugged 6550 from svetlana in withou changing anything else.

i got a huge hummmm...music could be played, but the humm was a lot. I plugged in a matched quad of kt88 of the treasury series...same behavior.

This is issue no 1. No idea why this can happen.

Worse, I plugged in the el34, and zhe amp was not quiet anymore. slight hum, but the channel as well is now not as loud as the other anymore (i did this experiment only on one channel). I did not change anything in the autobias module or anything.

What happened here ?

Thanks for your support
This is like calling a doctor on the phone and saying it hurts here.

You need to get a voltmeter and see what exactly is happening to the bias voltage and what your other voltages are at those tubes.

Be quick and shut the thing down before tubes go into thermal runaway. My guess is that the bias voltage is wrong due to a failed part in the circuit, but don't overlook other possible causes.

You can always look for burned parts, too.

However, fixing these kinds of things requires good detective skills, knowledge of the circuit, and at least a VOM.
Disabled Account
Joined 2010

What makes you think that you can change from one tube to another without a rebias?

It is reliant on some preset value to set the bias on the tube. So my guess is you have possibly damaged the op Tx, or something has been damaged in the cathode circuit..

Ie unless I am missing something there should be a way to set the bias current reference. Has this changed or is it set to high? It would be different for different tubes. So yes it will self adjust for tube ageing however there should be a preset for the circuit to adjust the value for a given tube. Not one value fits all tubes:eek:

So in a nutshell how does the circuit know what value to bias the tube at if its set for EL34? can it read the lable on the side of the tube..or is there something I'm missing here?

I have no Idea whats in the circuit so I'll have a guess along similar lines..

I'll give an example..if you change from EL34 to 6550 you will probably need more negative bias voltage to shut the tube down, so looking at your circuit I assume the tubes get a variable negative supply via R13/R20 220k and this is derived via a cathode shunt that is fed to some sort of comparator, can the circuit supply enough neg bias for a 6550..is the preset for the comparator correct it was set for EL34?

Here is another guess..

I see some jumpers that say 0-90V is this the amount of bias available or a setting for the cathode shunt? is it set via the jumpers for getting a more accurate setting of the autobias..ie its easier to set 0.1v on 0-10V than 0-90V...

M. Gregg
Last edited:
Ok, I found that one of the auto bias circuits for one tube did not work anymore, so one tube got a neg. grid voltage of -80v, while the rest got -35v. the 6550 needs about -45v, so well in the set range of -90v.

the auto bias was set for the el34 to 50ma. sure, the 6550 could get more like 80ma, but besides this, should it not work with 50ma as well ? Not following your point.

The key question though: Why did this all happen at all ? And if I now plug in the next 6550...what will happen than ?
Last edited:
Do you have a schem. for the autobias board? It may have a failed part or two.

50mA should be good for the 6550 but the total dissipation really depends on the B+ also, so if you have too high a B+ for 6550's that will cause problems even if the Bias is set lower.

Last edited:
Disabled Account
Joined 2010
I have seen,

Output tubes flashover and blow cathode resistors apart..if this was to happen, what will happen to the cathode shunt and anything connected to it, if the shunt survives and the board fails. It is also quite common to see Op amps go faulty with the correct voltage out with the wrong polarity..

As mentioned running 6550 at lower idle is OK...however can the board output enough neg bias (Ok we know it should but can it?)to control the tube..is the bias supply on before the B+?

Without a schematic its hard to say what has happened...what protection is there for run away with this circuit? ie what would happen if you lost the supply to the bias board?(fusing for the Output Tx)or does it fail in a safe mode with high bias voltage and shut the tube down?

It would be interesting to see what has failed on the board..

M. Gregg
Last edited:
Ex-Moderator R.I.P.
Joined 2005
regarding grid connections, are those tubes the same.... 'internally' ?
I think Mr. Menno chose EL34 because of the 'internal difference' (for his new cathode feedback designs)

apart from that, wouldn't EL34 also be able to run on lower trafo impedance load ?
if so, maybe it does, maybe it doesn't
Well...I just changed the svetlana el34 to an canada fuller el34...same phenomen: I get a strong hum...but this time i can go back to the svetlana el34 without an issue. But why do I get the humm ? The svetlana has one pin not even there...but the should a el34 not be a el34 ?
Make sure your tube sockets are making good contact. I recently tried some GE 6550's in my Conrad Johnson MV75. As soon as I turned it on one of the tubes started to glow. I turned it off before anything was damaged then looked for the cause. I found that the GE's tube pins were smaller in diameter than the previous tubes & weren't making good contact. After I bent the socket contacts in a bit everything was OK.
tinitus makes a good point. Make sure the amp is unplugged before you do anything. It's also a good idea to wait a while before touching anything & take a voltage reading on each connector first. Take it easy on the socket connectors or they may break. So as tinitus states, please be careful.
The "autobias" board consists of a bias supply (voltage tripler) and four individual bias servo circuits with a common reference to define the required current. If there are no issues with sockets etc. then you have damaged one of those bias servo circuits. Each one looks like this. Since it is all surface mount then repair is a PITA but probably worth the hassle. Check that 10 Ohm current sense resistor first.



  • BServo.jpg
    69 KB · Views: 227
Last edited:
check your power supply diodes....my bet is that they over heated and now have leaks. When you do tube family hopping from one family to another...you need to find tubes in that family that have an Ip value about the middle of what the bias range for your amp is. so for example....one amp I have can use 6l6's, el-34's, 6550's, kt-90s, and kt-88's if there Ip value is more than 20 but less than 30...I can stretch it a bit and use Ip's of 32, but over time as the tube gas up...the Ip values will climb. Ip is the bias value that was obtained during the burning process under real world test conditions using a matching circut like this one. There are tubes out there that are all over the place in Ip. The higher the Ip, the more current is going thru your power supply.....amp runs hotter.....power supply diodes get stressed. Use a diode tester or use a multimeter and check the diodes for reverse voltage leaks. make sure your amp is fully discharged or you will fry your multi meter.


  • Modified tube matcher.jpg
    Modified tube matcher.jpg
    191 KB · Views: 176
Ok, here is now an interesting observation and i would love to hear your thoughts:

I integrated an electronic choke into the psu of the power amp. In theory the ripple of the simple psu of the ul-40s2 which is described here


Gives easily 10v ripple or more. With my svetlana el34 matched quads, only a very small audible hum came out of the speaker. the el34 run with 50ma in class ab. When i changed now to other sets of tubes, loud humming as desribed. svetlana back in...normal noise level.

Now, withe the electric choke explained here


...absolute silence with the svetlana el34 and small rest humm with the other tubes, but not at all comparable with the original loud hum.

So...I guess with other tubes, i do not have the required symetry to cancel out the big ripple of the original psu. But why is this the case ? The other tubes have been sold as matched quads as well and we have the auto-bias module
Autobios improved in place.....?
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.