• Guys, we've spent considerable money converting the Les Paul Forum to this new XenForo platform, and we have ongoing monthly operating expenses. THE "DONATIONS" TAB IS NOW WORKING, AND WE WOULD APPRECIATE ANY DONATIONS YOU CAN MAKE TO KEEP THE LES PAUL FORUM GOING! Thank you!

Marshall query

Sean

Goldie's Man in London
Joined
Jul 15, 2001
Messages
2,021
I've already discussed some of this with NOCD off-list, and I'd appreciate any other opinions, I don't get this.

Basically, my 50w 2204 from 1978, RICH-mastered, generates about 350v on the plates, and I've cranked up the plate current to max 47mA. Even this only gets about 15.5w at idle.

I first spoke to Dave about this yesterday, when the bias measurements were 367v/ 38mA, later they fell to 358v/ 33mA, today they were 345v/31mA. I cranked up the current to max 47mA, and now the speakers "chirp" like a bird when I turn up the volume. Stop laughing.... :)

Thing is, the amp still works, I can play through it (though I've kept it low, it's late here).

Any ideas what's going on, anybody? Is it going to blow or anything?

Yours nervously....
 

Ed Rafalko

Les Paul Forum Member
Joined
Jul 15, 2001
Messages
6,287
Sean, your BIAS voltage is going from 40 to 31? Bad! BAD Sean!
I think they don't like to be under 30, for that amp.
But I'm no expert, see whjat NOCD has to say.
 

Sean

Goldie's Man in London
Joined
Jul 15, 2001
Messages
2,021
Hilarious! It's now 12.45 at night here, I've just measured it again, and now the readings are 360v, 52mA, which means it's gone from 15.5w to 18.5w at idle in the last 90 mins!

Current seems to move up/down with the voltage, I thought they moved against each other..... curiouser and curiouser......
 

NewOldCokeDave

Active member
Joined
Jul 16, 2001
Messages
4,945
What is more important than the bias voltage is plate dissipation.

360*50=18W

Don't use bias voltage as a measurement to go by. Use idle current, instead. Which can be measured thru the cathode.

Every single tube is going to have a different relevant biasing point so bias voltages are irrelevant.

-NOCD
 

NewOldCokeDave

Active member
Joined
Jul 16, 2001
Messages
4,945
Sean the observation you are making is exactly right. More bias voltage, less idle current. Think of bias voltage like a faucet in reverse, more you turn it on, the less will come out.

The change in house voltage can be accounting for the changes in the amp..

remember a change of one volt in the UK means a change of ~2 volts. Here it can be as much as 4 volts..

-NOCD
 

Sean

Goldie's Man in London
Joined
Jul 15, 2001
Messages
2,021
I don't understand why bias voltage is irrelevant.

Dissipation is a product of voltage and current. For some reason my Marshall has a really low plate voltage, so I have to jack up the current so that the plate dissipation is a decent measure, like your example, 360v x 50mA = 18w.

I can't see how you could determine dissipation without measuring *both* plate voltage (pin 3 to ground) and plate current (across 1ohm resistors, pin 8 to ground).

I *bet* I've just made a horse's ass of myself :)
 

BrianGWN

Great 'Double White' North ~ Electronics Specialis
Joined
Jul 15, 2001
Messages
1,401
buy isss

Whether a tube or a solid state device, the case physics (size, shape, material) will play a big role is determining how much heat it can tolerate before it ends up self destructing. The static power dissipation is thus the product of the voltage across the device and the current going through it, and typically there are also maximum allowable voltages and currents which if any are exceeded the device can end up :dead:

With the bias issue, I think what NOCD is getting at is that tubes are by no means consistent in terms of their exact performance, static or dynamic. It's the same with solid state bipolar transistors, there's a stated range of performance but it's not an exact science, to build devices with extremely predictable repeatable tight specs would be prohibitively expensive.

With specifically the tube bias issue, the amp design may allow the negative grid bias to be set somewhere between say -32v up to maybe -40v but that only gets you somewhere in the ball park. If you want to specifically pay attention to the idling power dissipation then as NOCD mentions you would adjust for a certain plate current in milliamps. Whether you get to that setting with the particular tubes you have at a negative grid bias of -32 versus -36 versus -40 volts doesn't matter as much.

I understand some of the tube suppliers started some kind of rating system so that apparently you could get a replacement set of power tubes that would behave a little more predictably. I don't know if any of that holds true in practice...

Back to the measurements... If you are attempting to determine the cathode current and thus plate current by having a 1 ohm resistor in series with the cathode and measuring the voltage across it using a good digital DVM, there's some other factors that can affect your readings. Since V=IR or I=V/R, 40ma of current is only going to produce 40 millivolts across that resistor, don't know how well a given DVM is going to consistently read that voltage. Then, as things heat up or cool the value of the 1 ohm resistor could possibly drift a bit. It's possible you might see a voltage reading across that resistor drift up slightly and in fact the current hasn't changed at all, the resistance could have drifted up slightly if the resistor gets a little warmer and thus the voltage drop across it increased.
To be absolutely sure you are getting accurate precise current readings you probably have to resort to some kind of jig that I heard of that plugs into the original sockets of the output tubes that then connects a true milliampmeter in series with the cathode of the power tube(s).

I'm still wondering about your plate supply voltage only being around 350V, isn't that rather low, or is that normal with that particular model amp?

Are you having fun yet?... ;)
 

NewOldCokeDave

Active member
Joined
Jul 16, 2001
Messages
4,945
I should have been clearer.. I meant *measuring* bias by using voltage method is irrelevant.. Sorry for the confusion!! You need to be concerned about the disspation which is the *plate* dissipation which is the product of the current and the voltage thru the plate. which can be n=measured thru the cathode..
( A side tidbit, you are also picking up screen current too, about 3 - 5 ma on top of the plate current thru the cathode.)

Given the plate volatge is at a certain level say 400VDC and the plate current is set at 40ma, then the net product at idle is 16W.
This is what I was trying to drive at...


The most common analogy of idle current is like the idle on the car, except your setting your idle at about 60 - 70 % of full power..

Is this getting any better as far as an explanation?? :)

-NOCD
 

NewOldCokeDave

Active member
Joined
Jul 16, 2001
Messages
4,945
Oh yeah as Brian said, there are physical limitations, you couldn't run the tube at 1VDC and run the current at 18 AMPS.. That's where the design maximums in tube manuals come into play..


-NOCD
 

Sean

Goldie's Man in London
Joined
Jul 15, 2001
Messages
2,021
Thanks guys, this is great stuff.

Re the low voltage at pin 3; I called Marshall today to ask them if this is normal. They agreed with NOCD (of course!), they think it's probably a diode problem. I won't pretend I understand what they were talking about, but I'll have another look at the diode stuff you mailed me, Dave, maybe I can check it out myself.

Apparently if someone wants their Marshall serviced or repaired, you can send it to Marshall! They suggested booking it in for a free quotation, so I might just do that (they're based about 90 mins from London). Show them what a *real* master is all about! :)

And, yes, BG, I am having fun, this is cool stuff, I just wish I'd done this stuff in school or something... :)
 

NewOldCokeDave

Active member
Joined
Jul 16, 2001
Messages
4,945
I think what I told Sean was to measure, from ground before and after each diode in the power supply to see if the voltages were consistent, before and after, with each respective half of the transformer. If there was something wrong before the diode it's the transformer, if it is after then it's probably the diode.


Sheesh, Marshall agress with me!?!?!? I'd better call it a day then!!

-NOCD :lolspin :wow :fiend
 

BrianGWN

Great 'Double White' North ~ Electronics Specialis
Joined
Jul 15, 2001
Messages
1,401
Sean, I'll have to check my copy of the Marshall book for the schematic. If a high voltage supply of +350 really is too low to be considered acceptable, maybe we will have to do a group effort here to go "in search of the missing voltage"...

If that schematic or others with a similar output stage have sample DC voltages, I should be able to get back to you with a good guess at what AC voltages you should be getting on the input side of the power supply diodes. There is a fairly predictable relationship when it comes to AC to DC conversions in traditional "full wave" power supplies. The ratio in theory is 1.414 (square root of 2) minus some overhead such as the 0.6 volt loss in the diodes. In other words 10 volts AC should get you about 14 volts DC, 100 AC would get you about 140 DC. If you are getting 350 DC then the corresponding AC measured at the input to the diodes should be around 250 volts AC.
 

Sean

Goldie's Man in London
Joined
Jul 15, 2001
Messages
2,021
NOCD strikes again! Your suggestion of moving the selector from 240VAC to 220VAC for the UK ACvoltage moved the plate voltage from about 350v to 388v! Suddenly I have 388v x 52mA, approx = 20w at idle, and I'm turning the plate current down! Nice one Dave:)

Little thought, some of the schematics in the Doyle book suggest about 430v on the plates for a 50 watter (much higher for 100 watters, like 560v), but the sideways chart on page 249 referring to a 50w with EL34s suggests that 380v is correct. All very confusing...

So I'm going to leave it as it is. I think I have enough to make the amp sound it's best, and anything after that goes to a tech.

MORE, MORE, MORE! :)

Thanks guys, I'd love to get one of you alone with a Marshall for 30 mins to explain how it all works, fascinating stuff.
 

BrianGWN

Great 'Double White' North ~ Electronics Specialis
Joined
Jul 15, 2001
Messages
1,401
Sean, while your at it make sure you check out the thread titled "negative feedback?", you may enjoy the ideas and issues kicked around there. ;)
 
Top