Noob would like guidance setting bias for new EL34s in a Haze 40w Combo

[bonedaddio]

Hi, new here and hoping for a little guidance.
I'm not new to electrical work, worked as a commercial/residential electrician, but no formal electronics training.

Reason for this post: I recently had an output tube fail in my 1 year old Mashall Haze 40 Watt Combo, it's funny, if turned on either end, the amp sounded fine; but if operated int the normal position it squealed, crackled, thumped, and generally wouldn't work-if you tapped the amp case, it would freak out.
I got a new pair on Groove Tube EL34 Ms, and installed them and the amp works fine.

I need to set bias on the new tubes and I've never done this before.
I found this Haze 40 schematic here on this forum http://music-electronics-forum.com/t16936/, and I've done some reading and research.
-It looks to me like I would check bias values at Con3, between pins 1 & 2 and between 2 & 3.
-The values I'm getting (DCA scale, 200u setting on the dial) are 26.9, and 43.8, respectively.
-If I'm not completely wrong, then I would adjust VR1 and VR2 to get the 39mv value for each (the 39mv is on the schematic at Con3.

Please, some one tell me if I've got this all wrong.

It seems logical that they'd be quite different, as the amp came from the factory with two completely different umatched EL24's in it... TAD "1" and TAD "2" on the labels and they're actually different heights...
Help would be greatly appreciated!
Thanks,

Les

[melvin]

Yes, you should measure the voltage across the pins you specified at Con3. Since you are measuring the voltage at the top of a bias-sensing resistor in the path of a power tube's cathode to ground, this voltage tells you how much current is flowing through the tube. Although the schematic has 1ohm for the value of the resistor, you may want to measure the resistance (with the amp off) across the pins to be sure. Use that value in ohm's law to figure the bias current. The value of 39mV across a 1ohm resistor would give you 39mA. So 39mA is what you should be shooting for.

[bonedaddio]

Thanks so much! With an inexpensive Digital Multimeter set at the 200ohm scale, I'm getting a 1.5 across both sets of pins with the power off. Does that mean that my DMM is not good enough for the job? I get the same 1.5 ohms across R26 and R42, the 1 ohm, .25 watt resistors.
Do I need to be able to set it for a lower scale than 200 ohm? BOY am I a noob!
I just want this bias set correctly, and in future, I'd like to be able to bias my own amps... it doesn't seem like rocket science, but then again, lookit me asking all the probably dumb questions!

Thanks again,

Les

Les

[Enzo]

If you are measuring a 1 ohm resistor, then you should use the lowest resistance scale you have. If that is 200 ohm, so be it. Think of it like a speedometer on a car, if the car can;t go over 80mph, then no point is having a speedo dial that goes up to 300mph. Now touch the probes together. WHat resistance does that show? WHatever it is, subtract it from the resistor measurement. Your probe wires have resistance. SO if you measure a 1 ohm resistaor and your probes measure 0.3 ohms toughing together, and your resistor reads as 1.5 ohms, then subtract the 0.3, and the resistor has a resistance of 1.2 ohms then. Just a made up example.

And something else here. The current through the tube flows through the resistor, and by Ohm's Law causes a voltage across the resistor. From the top and bottom of that resistor pit to the test points, there are traces on the circuit board, and connectors, all of which have resistance. The tube current does not flow through all that. They are just extensions of your probes to the resistor itself. SO you can measure the resistor itself right at the resistor leads, but measuring out at the test pins adds the resistance of those extra things to the resistance reading, but they don;t affect the voltage drop across the resistor.

If the sense resistor inside were 10k or something, then I'd measure it at the pins because the half an ohm they can add is irrelevant to 10k ohms. But when we try to read a 1 ohm resistor, an extra half ohm of junk seriously upsets the reading, but only of the resistance. Not the voltages. In other words, measuring out at the pins will not tell you how close the resistor is.

Try this experiment. Take out a 9v battery, and set the meter for some convenient scale like 20vDC. Measure the battery. You'll get something 9.0v, 8.6v, whatever. Now disconnect one probe wire from the battery and insert a - oh who cares what - a 100 ohm resistor between that probe and the battery terminal. That is a ton more than a half ohm. And what do you get? The voltage reading will be the same, even with the 100 ohms in series with the meter. Since no current is flowing through it, no voltage drops across it. Your meter has something over 1 million ohms internal resistance, so negligible current can flow through it.

What this means is that if I connect my voltmeter to those test points and get 60mv, if I then add a 100 ohm resistor between the test pin and my meter, I still will get the 60mv. That extra resistance is not in the current path and won't upset the reading.

Purists may note, and they'd be right, that even at 10meg ohms, my meter does have SOME incredibly tiny current draw, so if you want to know how much, find out how much 1 ohm is reduced if you put 10 million ohms in parallel with it.

In fact this idea is used in fancier test equpment. Let's say I have a power supply on my bench, and I want to send 5VDC into some digital circuit at several amps. 10 amps of 5v is not uncommon in computer stuff. Now basic power supply units on a bench have a meter that measures what the output voltage is - AT THE SUPPLY. But 10 amps is a lot of current. If the wires out to the circuit I am powering and the connections have even a 1/10 ohm of resistance, Ohm's Law tells us that there will be 1 volt dropped along it all. SO in order to get 5v at the circuit, I need to adjust the power suppy to 6v to compensate for the 1v loss. The meter in the unit doesn;t really tell me what I need to know.

A nicer supply doesn;t measure the voltage it is making at the source, it adds two extra sense wires, running out to the ends of the feed wires - in other words they run right out to the point where the supply hits the circuit. The meter is connected to those wires. Now those wires have resistance also, but the 10 amps is not flowing through them, only the tiny current through the meter, so there is no voltage lost. Because the voltage that has to be at the supply to make exactly 5v at the circuit is not the issue, only the net result matter, this works out. Cool, huh?

There is also a similar setup for making accurate resistance measurements, especially for very low resistances... like 1 ohm. If any of this is interesting, look up "four-wire" measurements.

Frankly, this is not precision rocket surgery. I just assume the 1 ohm resistors are pretty close, close enough anyway. If the setting is off a couple milliamps, no harm will come to anything. Bias is NOWHERE near as critical a thing as so many people tend to think.

[Steve Conner]

Thanks so much! With an inexpensive Digital Multimeter set at the 200ohm scale, I'm getting a 1.5 across both sets of pins with the power off.

No... It means you need to adjust the bias with the power turned on! (And the standby switch set to play mode, if the amp has one.)
You won't measure anything with the amp turned off or in standby, because there isn't any current flowing to measure.

You set the meter to its lowest DC volts range: millivolts (mV) if it has that. Not ohms.

And "careful now" as Father Ted would say. To bias an amp you have to work on its guts while it's live, and it contains high enough voltages to kill you pretty dead, so take all the usual precautions. I can't be bothered typing them all out, but you should know them. One hand in pocket, don't stand on a damp concrete floor in your bare feet, use a screwdriver with an insulated handle to twiddle the bias adjusters, have someone else standing by to help you if you get shocked, etc.

[Enzo]

Steve, he is checking the resistance of the 1 ohm sensing resistor as instructed by Melvin in powst #2. And I weighed in on afterwards. Hasn't gotten to the millivolt readings yet.

[Steve Conner]

Oh, why did someone tell him to do that? It just confuses things.

The zero error in the meter, lead resistance, etc. add up to a bigger error than the tolerance of the sensing resistors. His bias would be less accurate applying the "correction" than if he just adjusted for 39mV.

[bonedaddio]

I don't have words enough to express my gratitude! It was all good information, I read EVERY word, and here's my evening so far:
1. Tested the resistance through the leads, no resistor or circuit, and got .5 ohms! Bullseye Enzo!
2. Set bias without elecrocuting myself... used to work on 440V equipment, so I already have some respect/fear/loathing for "joining" myself to a circuit in any way! Set it as close as I could to 39.5.
3. After I setup to play the amp... it wouldn't work. The effects loop in was fine, I could plug in and play the power amp & speaker, but no preamp!
4. Pulled the amp out of the cabinet AGAIN so I could really see the tubes, and preamp tube #1, under the metal cover, was burned white on top! Replaced it with an old spare 12ax7a, and bingo, it worked! That's one power amp tube and one preamp tube failure with in the same week... amp is really close to 1 year old!!
Seems a little odd... and what kind of preamp ECC83/12AX7s would anyone suggest? The amp is a Marshall Haze, and up to this point, it's been an awesome amp!!

Thanks again so much,

Les

[Enzo]

There are several current production 12AX7s, and old stock 12AX7s are readily available. Your Haze will always sound like a Haze. ANy differences between tubes will be relatively subtle. But that is OK, nothing wrong with going for those subtleties. Tone is totally subjective, so try as many types as you can in there and listen for the differences. And don;t be surprised when sometimes there are no differences.


Tubes have that silvery;/black stuff on part of the glass, usually the top on a 12AX7. When that turns white, it is not burnt. That stuff is called the "getter", and when it turns white, that means air got in the tube. In other words the glass is cracked. Is part of the little glass point snapped off?

[bonedaddio]

Thanks, Enzo, right again... the tube base shows a crack, I guess it just decided to crack at this particular time??!! Could well be my fault due to removal and reinstallation a prior time. The design of the Haze has the tubes almost inaccessible unless you pull the chassis out of the amp, you can do it by sticking your head in with the amp on it's side and using a light, but I'm old and my near vision is terrible. I probably damaged it.
Sigh. Another lesson learned, just pull the damn chassis next time.

Again, Thanks! If I wanted to get further into this kind of thing, can you suggest a starting point?

Les

[melvin]

rocket surgery! Anyway, sorry to have everyone cleaning up after my advice. I was biasing something the other day that had 33 ohm bias sensing resistors in it. so I thought... well, you get the idea. Sometimes I don't know how I get by.