Biasing a quad (Marshall EL34 100/100)?

  • Thread starter Thread starter Me
  • Start date Start date
Me
Me
Member
Hello! I was thinking about getting some kinda valve socket bias adapter thing (like the one from tubetown.de) to plug my multimeter into to bias my Marshall power amp. I'm used to my Diezel Herbert with the external pots where you have easy banana plug connectors and just bias the pair with the fuse removed to 70mA (for the pair). I've never biased the more awkward Marshall.

So if it's a matched quad can I just stick the quad in, one tube via the socket thing, and adjust til I measure say 35mA on that one el34 and voila, the rest must also be 35mA because they're matched? Pop the socket thing out and directly connect the tube and it's done?

Just wondering as I've seen all sorts of dual or quad socket current measuring devices, but if they're matched, aren't the lot right if one is right? :confused:
 
I test all four. Matched quad can be still be off a little. If the one you test is at 35 one may be at 33 and two may be at 37. Measure all and come to a happy medium.
 
Cool. But I'm right in thinking you can check them one at a time with all 4 in, you don't need some overly complicated 4-socket contraption unless you're lazy or it's your day job?
 
Me":v1pnaayo said:
Hello! I was thinking about getting some kinda valve socket bias adapter thing (like the one from tubetown.de) to plug my multimeter into to bias my Marshall power amp. I'm used to my Diezel Herbert with the external pots where you have easy banana plug connectors and just bias the pair with the fuse removed to 70mA (for the pair). I've never biased the more awkward Marshall.

So if it's a matched quad can I just stick the quad in, one tube via the socket thing, and adjust til I measure say 35mA on that one el34 and voila, the rest must also be 35mA because they're matched? Pop the socket thing out and directly connect the tube and it's done?

Just wondering as I've seen all sorts of dual or quad socket current measuring devices, but if they're matched, aren't the lot right if one is right? :confused:
Click to expand...

You are correct. With Marshall's, there is only one bias for the quad and measured per tube.
 
Me":2hmgo3hd said:
Cool. But I'm right in thinking you can check them one at a time with all 4 in, you don't need some overly complicated 4-socket contraption unless you're lazy or it's your day job?
Click to expand...
No matched quad of EL34 tubes I've ever tested measured exactly the same cathode current...

Just check one EL34 at a time and make sure that the one that measures highest is at or below 70% of max plate dissipation. An EL34 can dissipate 25 watts, so 17.5 watts is 70%.

To get it right, measure the B+ voltage at pin 3 of one of the EL34 tubes. Let's say that measurement was 498 VDC…

The formula is:

(17.5 watts dissipation) divided by (498 VDC) = 0.035 amps cathode current = 35mA

So you would want to make sure that the highest reading EL34 is at 35mA to get max tube life. Of course, you can also listen and bias to what your ears like, but don't get to high above 70% of max plate dissipation.

Steve
 
Cheers! So you're saying just because my Diezel's EL34s are set to 35mA per tube doesn't mean my Marshall should be?

So i need to take the max plate dissipation (25W). Find 70% of this = max plate idling power? Divide by the B+ Voltage (what's that, plate voltage?). So P/V = I = idling plate current? Then I adjust the grid bias so it's a maximum of this idling plate current.

And so if an amp has a lower the plate voltage, the higher the plate current should be to give the same power due to P=IV?

I assume there's no handy way of measuring plate current by measuring the voltage across a series resistor that's already part of the circuit and using I = V/R to save buying any magic adapter to begin with? It's not like I need to bias amps very often.
 
Just read the plate voltage off Pin 3 of one power tube socket. Set DMM to read 2000V DC, black to chassis, red to Pin.

You only need to read this voltage once - or even try to find it in the net. Which pwr amp is it?

You could also assume that it is between 450V and 480V - and simply just set each tube to 30mA, which would be 54% of max plate dissipation at 450V and would equal 58% @ 480V - which means 30mA is within a safe area. 54-58% is IMO perfectly fine and allows the tubes to live a pretty safe life :D



here's a chart for you: idle current vs. plate volts between 55 and 60% max plate dissipation for 25 Watt tubes
 

Attachments

  • el34.gif
    el34.gif
    9.8 KB · Views: 3,614
Ah, good old Olaf, you always have answers! I should just PM you if I have a technical question! :lol: :LOL:

I take it if I can find any resistor leading to each grid I can calculate the grid current by measuring voltage across that and knowing what value R it is? Or if I take the HT fuse out can I measure the current across that to be 4x30mA = 120mA, although this might be a bit risky if I have probes and they keep coming off or reattaching while doing the measurement I'll keep biasing and disconnecting the grid all the time - if the grid floats I assume the tube will just conduct lots! Maybe not else an amp without an HT fuse would blow its tubes!? :confused:
 
Dude... Just measure the voltage on pin 3 of one of the power tubes. Then divide that number into 17.5 to get your bias setting. Not hard.

You don't need to go finding resisters and fuses and doing complicated math.
 
FourT6and2":3lswdvu0 said:
Dude... Just measure the voltage on pin 3 of one of the power tubes. Then divide that number into 17.5 to get your bias setting. Not hard.

You don't need to go finding resisters and fuses and doing complicated math.
Click to expand...
That calculates what the bias current needs to be. The maths bit was to see if I can set the bias without having to buy any adapters or sockets or anything. I'm being a cheapskate and seeing if I can do it for free using just a multimeter rather than have to spend another £40 on something from a UK shop, or buy something from germany for about €30 and wait for the longer post. Either way I can still have an evening in the pub for that much! If there's a single series resistor leading to the grid I should be able to in theory if I'm not mistaken.

Take this example with a schematic I just found on t'internet for a JCM800 power section (stolen from ampbooks.com):
power-amp-schematic.gif

Looks to me like the feed from the phase invertors come in on the left via the caps. Then there's a resistor from this down to -40V, and another single resistor on each of these halves from this point to the grid of each tube. So the grid resistor is 5k6 yeah? So the grid current has to flow entirely through this component. If I want 30mA grid current from Olaf's "safe bet" (or that's what the voltage on pin 3 is divided by 17.5), then:

V = IR (Ohms law)
V = 30 mA x 5k6
V = 168V

So if I can find that 5.6k resistor in the amp and can adjust the bias til I get 168V across it while idling (without killing myself) then job's a good-un and I've saved £30 on a socket adapter? I can also then check the grid current of the other 3 by measuring the voltage across the other 3 grid resistors (or in this 50W schematic, other 1) without turning it off and on multiple times and moving a socket adapter around. Or does that not work and I have to buy an adapter for some reason I can't think of?

I'm guessing the bias pot and the HT fuse is connected somewhere off the -40V bit of the diagram. I'm guessing it must be if you can adjust a DC idling current when everything else is connected off a DC-blocking capacitor? If the HT fuse feeds only the -40 V point in the diagram above then when there's 30 mA flowing into each tubes grid there would be 60mA flowing across the fuse (or 120mA if it was a quad). Hence if you take the HT fuse out and bridge that point with a multimeter on current setting you should be able to bias based on total grid current for all tubes through the fuse (I think that's how external Diezel Herbert biasing must work). :confused:
 
Why don't you do this instead?

See the wires that connect pin 8 of each of your power tubes to ground? Remove those wires and put 1 ohm, 1%, 2 watt resistors in place. Now you can just fire up the amp and measure across that resistors on the mV setting of your meter. Measure from pin 8 to ground. So say you get a reading of 35 mV, then you have a current of 35 mA. Just follow my other instructions and you'll be fine…

Steve
 
Yeah, don't follow that plan of mine it's a load of bollocks!!! I want to measure the plate current, not the grid current. Twat! :doh:

I've managed to find a circuit diagram. There's no nice resistors to measure plate current across. The only thing I'd be able to isolate to measure current across is the 16 ohm winding of the output transformer. This would be point A to the anode in the diagram above (or pair of anodes in my case).

The plate voltage is 166V. If I go for Olaf's slightly more conservative 60% plate dissipation to make the tubes last longer that gives me 15.6W / tube.

I=P/V
15.6W / 466V = 33.5 mA per tube.

Without a tube socket doodad I can't check per tube. I can check the current across both halves of the output transformer though. Since I can't change the relative bias between the 2 tubes I guess it's tough what it is per tube anyway and I have to assume they're matched well enough, but I'm playing for the safer 60%, not 70% so that should be ok?

Half the Tx'er is 16.3Ω, the other half 15.7Ω. So 16Ω +/- 1.9%. Again it's a matched quad, I can't adjust each tube individually so trying to set it exact is impossible anyway. 1.9% off 60% should be ok though by my reckoning?

So 2 tubes off each half = 67mA from output Tx centre tap to anodes. V=IR (Ohms law). V = 67mA * 16Ω = 1.072V (across each half of the output transformer).

Trying to measure that across a pair of probes when you don't have a great contact isn't gonna be amazing. How to the bias adapters work? Do they break the anode connection so you can measure plate current directly? Or do they measure V across say a 1Ω resistor? If it has a 1Ω resistor in to measure current across I'll still be closer in final usage than I would with one of those. If it breaks the connection and you put a meter in series then it's probably not gonna be as good.
 
sah5150":1ap4zphz said:
Why don't you do this instead?

See the wires that connect pin 8 of each of your power tubes to ground? Remove those wires and put 1 ohm, 1%, 2 watt resistors in place. Now you can just fire up the amp and measure across that resistors on the mV setting of your meter. Measure from pin 8 to ground. Just follow my other instructions and you'll be fine...

Steve


So say you get a reading of 35 mv,or .035mv.That converts directly to milliamps.
Click to expand...

Cheers Steve! You posted while I was working out that above^

If I measure the DC resistance of half the output Tx'er with it off, then I can pairwise bias them using the 16Ω output transformer instead of a 1Ω resistor per tube. Should work since It's idling and DC so the AC impedance shouldn't make a difference.

P.S. If anyone is trying to use this thread in future and wants to save some time the output transformers are the ones towards the back of the amp, the outputs of it are towards the middle of the amp, and centre tap of the output Tx is white, and so I want 1.072 V from the white to black, and 1.072 V from white to red (in an ideal world with a perfectly matched quad of tubes).

I'm full of beer: lets play high voltage amp poker!... :thumbsup:
 
(Just the small question of which pot is bias and which is balance for the phase inverter. You can't tell from this side of the board. Let's wiggle 'em and see! Ah-ha, the outer ones!)
 
Here we go! I made a video of me trying my theoretically correct biasing method. Should work for any vaguely similar amp with no test points if you don't have a bias adapter.

Still uploading but when it's done...
 
Like Steve said, the easiest way is to replace the wire on pin 8 to ground with a 1ohm resistor. Like the photo below. Then you can just measure from pin 1 or 8 to ground to get your bias reading. Otherwise, you're gonna have to get one of those socket adapters. All that other stuff is just making it more difficult than it needs to be.

3cf66ad13be81e78b80961e531220d30-d327guz.jpg
 
Leffe :D
(sorry, can't watch the whole vid . . .)

Just get a TT adapter or use the "shunt" method if you should have a good DMM.
 
I reckon it's already done with my method unless you can see any reason it shouldn't work though. I'm not an amp expert but ohms law must still apply to DC in it? Why use a soldered in 1 ohm shunt resistor to each tube when I can use the output transformer (which is already there) as a 16 ohm shunt resistor to a pair of tubes and double the current (since it's a pair)?

Sure, I don't know the current to each tube individually then, but I can't adjust it individually anyway. If as you suggest Olaf that 55% to 60% plate dissipation is good, and Steve says 70%, that's quite an allowed tolerance (30-37mA), and as long as they're around 60-65% then none can draw too much or too little assuming they're sensibly matched to within a few %? It was suggested earlier to find a happy medium anyway where they all average 70%. So this method would just be averaging the pair for me, so I could even go for the 74mA per pair if I don't want a more conservative 65-70mA.

Or am I missing something? I'm not trying to be difficult it just seems pointless putting an extra 1ohm shunt resistor in circuit when I can use the resistance of an existing component that's already there (as long as it has a linear I vs V relationship to DC, which it should since there's no AC when idling so no Z considerations). :confused: Maybe I'm a genius, maybe I'm a retard, I'm fine with being a retard as long as I know why I am! :lol: :LOL:
 
Shunting means to read the current directly at the OT. DMM set to read 200mA DC, red to middle of OT black to PIN 4 of one tube.
Attention: up to 500V on your DMM. Only reliable if the inner resistance of your DMM is low (good multimeter).
 
You must log in or register to reply here.
 
Back
Top