when checking bias with a meter in series...

[Enzo]

When you played your guitar through your amp, you saw the cathode current rise dramatically. I bet you didn;t also check to see what the B+ was doing or the cathode voltage. Your B+ will drop as the current draw increases. This is due to transformer winding resistance, any power supply series resistance, and especially if you have your sag resistor switched in.

Increased current through the cathode resistor of course increases the voltage across it - Ohm's Law. So as the current rises, the cathode voltage rises. This tends to push the tube bias colder.

23w out of 25w is 92%. Pretty hot to me. But if your tube isn;t red plating, you might get away with that if it sounds OK.

So you can generally set the bias somewhat hotter on cathode biased amps. But 100% dissipation would generally mean you were running in class A.

[daz]

Ok, well i guess i'll toss the 450R's and put a couple 500R's i have in. I imagine that should bring it to a pretty reliable state since i never had issue where it is now. What % would you suggest as optimal for tone w/o torturing the tubes?

[Robert M. Martinelli]

....So you can generally set the bias somewhat hotter on cathode biased amps. But 100% dissipation would generally mean you were running in class A.

Enzo,
I agree with everything you said in your last post, but allow me to ( respectfully ) disagree about the above paragraph.

The class of operation is not to be stated in idle conditions, but rather in working conditions, e.g. with signal flowing through the tubes IMHO.

The above is easily demonstrated using the AC30 as an example. The output tubes are ( cathode ) biased at about 10VDC, the +B is about 320VDC, so 310VDC are dropping through the output tubes in idle conditions. The idle current, if memory serves me well, is about 47 mAmps. This means the tubes are biased idle at ( 310*0,047 ) = 14,57 W of plate dissipated power, which is 121 % of an EL84/6BQ5 max plate dissipation. If we look at the sole "idle current" criteria, this would classify the AC30 as a "class A" amp, and this is not the case, as we know.

The rationale for this:
If we'd look at the waveforms on the plates with a scope, we'd clearly see that there is an evident current swing between idle and working conditions, as the current swings from 47 to some 92 mAmps, and it would also be clear that the tubes are on for more than 180 degrees of the signal's period but much less than 360 degrees ( some 250 degrees on and some 110 degrees off, and this places the AC30 with no doubt among class AB1 amps ).

Take the above numbers as approximated figures ( I'm getting old and I don't remember if they're the "dead-on-real-life" values, my two neurons, the only ones left in an otherwise empty space, are on strike right now, asking for an early retirement ), though they should be accurate enough for our purposes.

In a true "class A" amp, the tubes would be conducting during the entire signal's period, the waveforms wouldn't show any cutoff, and the plate current swing would be very little (ideally zero swing-no change ).

All the above statements proof, ( I hope convincingly ) that biasing tubes at, or above 100 % is not to be considered a synonymous of being in "class A".

Cheers

Bob

[Steve Conner]

The bottom line is that if you can get the power stage's B+ current draw to increase by applying a signal, compared to what it is at idle, then the amp is Class-AB. Class-A draws the same current at full audio output as it does at idle.

This is for unclipped output, once the power amp starts to clip all bets are off. Heck, they're all Class-C if you drive them hard enough.

Nerd disclaimer: even-order nonlinearities in the tubes can increase the supply current even in a Class-A setup. RDH4 calls this effect "rectification".

[Enzo]

You are right, and i am aware of that, but I didn;t mean it that way. If you idle at any level, you are not yet operating in ANY class, since there is no signal to conduct.

What I was trying to say, and it came out wrong, was that when you see a cathode biased amp that is biased to 100% dissipation, it was PROBABLY an amp designed to run in class A. SO the bias is not the defining issue, the bias level is a clue to the design intention. Of course your good example is contrary to my point anyway, so I probably ought not to even make the observation.

I really should have made myself more clear, as I often hear people erroneously assuming that cathode bias itself makes an amp class A.

[Steve Conner]

What I was trying to say, and it came out wrong, was that when you see a cathode biased amp that is biased to 100% dissipation, it was PROBABLY an amp designed to run in class A.

Well, that seems reasonable. After all, if the tubes were idling at 100%, and signal could increase the dissipation, it would go above 100%, which is bad.

But when we see an AC30 idling at 125%, lord knows what to think. I say Class-OMG.

[Robert M. Martinelli]

....But when we see an AC30 idling at 125%, lord knows what to think. I say Class-OMG.

OMG!

Steve, what I said about Enzo as a teacher days ago applies to you as well!

It's waaay easier for people who want to learn something new if hard-to-grasp concepts are presented by a knowledgeable person in a hilarious way. ( not to mention the sheer amusement found in reading a good joke )....

You know what? I'll prepare a sticker saying "True O.M.G. class amp" and I'll stick it on my '64 AC30! I swear!

My two neurons want to thank you as they're still laughing!

Cheers

Bob

[daz]

Ok, now that you freaking genius techs have had your little party, can i ask you something? (seriously, you guys amaze me with your knowledge...just wish i understood 1/2 of it !)
Anyways, i'm going to throw either 500R or 560R's in it today if i don't have an answer by the time i leave work. So i just want to know whether it's not a good idea to leave it at 100% and if not, what would the highest percent you think i should set it? I can get some more resistors locally with a bit of a drive, but i have a set of 500's and 560's. As i said right now it's at 100% with the sag resistor out. It's biased pretty cold when i switch it in, but i'd prefer to have it biased right w/o the sag R, as thats the feel i'd prefer if i could only have one. I like the way it sounds right now, and i haven't had any issues even with the couple rare instances the old band got together and played. But with what you guys are saying i'd rather have peace of mind if the way i'm running it really is bad.

By the way, is it just me or is the forum extremely slow? It took it a full two minutes from when i clicked reply to when the typing field actually came up.

EDIT: this is truly a letdown. I replaced the 450R per side with 560R per side, and now it's running at 85% in no sag mode. But it really lost it's edge. It just doesn't sound nearly as good. Very disappointing. But i fear going back to the 450's because thats 100%. Like i said, i never had a problem like that. But after this thread i can't help but worry that it may be the reason the first amp like this i built (now owned by a friend) blew it's output tranny. It was running 100% too.

[Robert M. Martinelli]

Hi Again Daz,

I guess you moved directly to 560s to be on the safe side, but why don' t you give the 500s a try? When trying to find a "sweet spot" changes have to be applied gradually....

Anyway, I've got some questions for you :

1 - can you describe what changed with the sound when you switched to 560s?
( less/more lively? Less warmth? Unpleasant sound/unwanted distortion? Other changes? )

2 - Are you using matched output tubes ?

3 - Is the OT properly sized? ( do you have the OT specs ? )

Cheers

Bob

P.S. Yep, the forum looked slow in the past days, things seem to be OK now......

[daz]

1 - can you describe what changed with the sound when you switched to 560s?
( less/more lively? Less warmth? Unpleasant sound/unwanted distortion? Other changes? )

Thinner and a bit less harmonically rich...not as round and lively sounding. i guess i shouldnt have made it sound as bad as i did in the last post. It's not far worse. But it did lose that fine line between good sounding and exceptional.

2 - Are you using matched output tubes ?

yes. winged C's

3 - Is the OT properly sized? ( do you have the OT specs ? )

Yes. Its a 3.2k 50 watt marshall replacement type from hammond.

As for the 500R's, i DID try them before. I had the 450R's in there for many months. Then i decided i'd try running cooler after my friends amp i mentioned blew his OT. So i stuck 560R's in his and mine. Then i felt it didn't sound as good and stuch the 500's in mine. I wasn't sure at that point how far i was from the way it sounded with the 450's. then the other day after a month or so of 500R's in it i went back to the 450's, and when i did i realized how much tone i'd lost. Today when i went to the 560's i lost that tone again. 500's wont make a difference, so i know going to those isn't worth the hotter running. If i'm gonna chance running hot i may as well use the 450's and sound great. The 500's and 560's both sound about the same. but like i said, it's not a huge loss. It's just that this amp is at the point where i've had it sounding amazing, and any loss that removes that edge even tho subtle seems more than subtle, if that makes any sense.

[Robert M. Martinelli]

Hi daz,
I already had an idea about the changes in sound you now described, but I didn't say because I didn't want to exert any influence on your judgment.

What I'm going to say now could sound like an heresy to someone ( the "matched tubes' lovers" ), so I beg your ( and the real gurus ) pardon.

I strongly suggest you to take those matched tubes out the amp, find a pair of "reasonably close" ( but not matched ) tubes and stick them in.

Why am I saying this?

Because I suspect that, by biasing the tubes colder with the 560 Ohm resistors, you are suffering a loss of even order harmonics - this makes the sound more sterile ( less warmth ).

Push pull designs like yours already have a natural tendence to cancel even order harmonics by design, and the more the two legs of the PP are symmetrical, the more this effect is pronounced. Add this to the presence of a couple of good quality "perfectly" matched tubes, and the even order harmonics are more likely to be canceled.

Biasing the tubes on the hot side helps you recover some of these "nice" harmonics, and this could explain why the tone suffers with a colder bias.

In other words, by biasing the amp hotter you take away some of the the "coldness" imposed by the output stage symmetry. Unfortunately, to find a tone you like, you have found that you have to bias the currently installed tubes at 100 %.

The reason why I'm suggesting you to used "unmatched" ( I'd better say "reasonably close", say within 5-8 mAmps ) tubes is that their "non symmetry" causes some more even harmonics to pass through the output stage and reach the speaker. At that point you' ll probably find that you can bias the tubes colder ( say around 70-80 %, 18-20 W, which is a reasonable value for almost any EL34 AB1 amp out there in the real world ) without losing too much warmth/liveliness.

Hope this helps

Best regards

Bob

[daz]

I've been thinking about that actually because at one point i have mismatched cathode bypass caps and it sounded really good in that way. i no longer have because i had some motorboating and was told i should use much bigger caps because the 47uf and 22uf i was using were too small and the motorboating i was occasionally experiencing was due to that. So i putt 220uf's in there. But with 220's i didn't think mismatching them would matter because at that size all frequencies are likely allowed to pass. But i've been thinking i should try maybe a 100uf on one side and i may try that today.

that said, i'm of course talking about AC signal. But you are talking bias current. So if i try that, what about this idea....how about if i just put say a 600R on one side and leave the 560R on the other? If that would work as well i'd rather do that instead of always having to order tubes that aren't matched and worry about making sure just HOW mismatched they are or arent. Come to think of it, i suppose it's easy enough to try by simply adding 50 or 200R to one side. I'll try it today and let u know. Thanks.

[daz]

I added a 68k to one side and it basically sounded like it does when i intentionally imbalance the PI by making the plates more asymmetrical. That would be that it sounds darker/muddier. It takes some highs away which is good, but it doesn't liven things up in any respect including harmonics. So thats a no go. Guess i'll just leave it as is. It's good as is, but i hate knowing i've had it sounding better but can't leave it like that.

EDIT: I tried what i had sdaid i didn't think would work, IE: AC imbalance via different sized caps. As i said it worked before but i didn't think it would work with caps as big as i had to go to rid it of that motorboating. But it did. i added a 220uf to one side making it 440uf. so 440 on one side and 220 on the other. Didn't so much thicken things up, but it took the hard edge off and added the harmonic complexity i was missing to some degree. Sound pretty darn good like this. Think i'll try a 100uf in place of the 440 tho and see how that works. In any case, things are much better like this.