Why some amps use two triode in parallel?

[JoeM]

There is a balance of theory and try it and see. You don't want to be totally one way or the other.

Ok, tell me what the formula in post #18 sounds like.

[Alan0354]

Ok, tell me what the formula in post #18 sounds like.

What do you mean sounds like, it's about noise, the amount of noise in the calculation!!!!

Theory does not tell you how it sounds, it tells you whether it will do what you are looking for. Just for example, before I challenge about the noise, people here seems to believe you have a big noise advantage using two triodes in parallel, that you actually lower the noise power by half.

I proofed that it is not the case at all, the article support my finding, Merlin jumped in and agree with my assertion. You put parallel triodes for other reasons, not just for the noise performance. That the 68K people automatically put in is actually the major noise generator!!!! This is all theory. So don't say theory is useless. You have to have both.

[JoeM]

ahh, so you do know what that sounds like!

But you missed my point completely

[Alan0354]

ahh, so you do know what that sounds like!

But you missed my point completely

I said post #18 mainly talk about the noise, nothing about the sound. Did you read the post? Where in the post mention about sound?

Re-read post #32 while you at it.

[JoeM]

Alan, respectfully, my point echoes GainFreaks comment in post 29. These are guitar amps, not Mars landers. There is such a history of great classic amps from Fender, Marshall, Vox et al. Build some, experiment and have fun!

[GainFreak]

I'm not saying that you shouldn't know ANY theory however we're talking here basic well known circuits known, tried and implemented for ages. That's why my philosophy concerning topics like this is very simple: check out already known designs, check out how it's done in guitar amps, build it, tweak values and listen to it. What's the purpose of 100 pages of theories? Are you expecting any major revolutions to occur in this field? I'm not. Somebody has figured it out long time before you anyway.

[Alan0354]

Alan, respectfully, my point echoes GainFreaks comment in post 29. These are guitar amps, not Mars landers. There is such a history of great classic amps from Fender, Marshall, Vox et al. Build some, experiment and have fun!

But my original post #13 was talking about one specific point about the claim of -3dB of noise floor by parallel a tube. That has NOTHING to do with the sound. I don't disagree with trying it, theory don't make sound. BUT theory do prove or disprove the claim of the noise performance. This is apple and orange. You can have better sound but no noise improvement right?

Also, it is very apparent that people complain about noise, the hiss of high gain amp. People has been trying to cut noise....and still are. remember the post of rectifier filament and all? Why talk about all the other ways and never change the 68K that is one of the major noise source?

[Alan0354]

I'm not saying that you shouldn't know ANY theory however we're talking here basic well known circuits known, tried and implemented for ages. That's why my philosophy concerning topics like this is very simple: check out already known designs, check out how it's done in guitar amps, build it, tweak values and listen to it. What's the purpose of 100 pages of theories? Are you expecting any major revolutions to occur in this field? I'm not. Somebody has figured it out long time before you anyway.

No, I just see something that is not right and I called it. Not to be blunt, I do believe there are people that actually believe you improve noise by 3dB using parallel tubes.

Yes, the noise theory I said has been around long time before me as it is in all the text books about noise. This has been used years before, invention is not even considered here, just pointing out mistake.

[Alan0354]

How much theory do you think Leo Fender knew? Did he have pages and pages of noise equations?

Not much, but he did not design the parallel triodes thing. He sure did not have to worry about noise as there is no high gain amp in his days. High gain amp is not even around until the mid 70s.

You guys are trying to design cascade gain amp, you better worry about the noise.

[potatofarmer]

I think there should be a subforum here for discussion of Theory and Design. That way, people who don't care, don't have to read it.

Oh! There is such a subforum! And we're in it right now!

Boy, don't I feel silly.

[eschertron]

No, I just see something that is not right and I called it. Not to be blunt, I do believe there are people that actually believe you improve noise by 3dB using parallel tubes.

Yes, the noise theory I said has been around long time before me as it is in all the text books about noise. This has been used years before, invention is not even considered here, just pointing out mistake.

Alan, your question was answered, with the math, in post #14:

The physical basis behind it is that audio signal processed by 2 parallel gain elements will be in phase and add up, while noise is generated by each of them individually, will NOT be in phase (precisely because being statistically random is inherent to *noise*), or, to be more precise, phase will be changing ar random all the time, so sometimes it will add, sometimes it will substract, the average is that you *lose* sqrt 2 or 3dB for 2 elements, 6dB with 4 and so on.

3dB less noise for each transistor/tube count doubling.

The two in-phase signals give 2X, or 6dB gain. The noise from each tube is root-mean summed to be SQRT(2) louder. 2 divided by SQRT(2) is equal to SQRT(2), or 3dB. If the gain of the two do not match, then there will be less than 2X gain, but there will also be a less noise. In the real world, the noise improvement may not be 3dB, but it'll be close.

[Alan0354]

Alan, your question was answered, with the math, in post #14:




The two in-phase signals give 2X, or 6dB gain. The noise from each tube is root-mean summed to be SQRT(2) louder. 2 divided by SQRT(2) is equal to SQRT(2), or 3dB. If the gain of the two do not match, then there will be less than 2X gain, but there will also be a less noise. In the real world, the noise improvement may not be 3dB, but it'll be close.

How is post #14 answer my question? The article in post #14 only confirm I was right.


No you don't look at the tube alone to claim noise reduction, it is only a small part. In really life, you are likely not even getting close to 3db lower. There is never an ideal setup that you can ignore the grid current and internal plate resistance.

If you just join in, read the posts first.

[Leo_Gnardo]

The two in-phase signals give 2X, or 6dB gain. The noise from each tube is root-mean summed to be SQRT(2) louder. 2 divided by SQRT(2) is equal to SQRT(2), or 3dB. If the gain of the two do not match, then there will be less than 2X gain, but there will also be a less noise. In the real world, the noise improvement may not be 3dB, but it'll be close.

Thank you eschertron!

Sure saves reading dozens of textbooks. Besides I'd have to brush up on my Laplace transforms, & I'm lazy.

Next subject: why bumblebees can't fly.

[Alan0354]

Thank you eschertron!

Sure saves reading dozens of textbooks. Besides I'd have to brush up on my Laplace transforms, & I'm lazy.

Next subject: why bumblebees can't fly.

He better explain what he said because it is not close to -3dB in real life as in the article and my presentation. Not even close. Even Merlin jumped in and agree that. So don't put away the book yet!!! Apparently, he did not read any of the presentations.

I don't just read about noise, I spent over a year full time working on optimizing noise for low noise design, it was my job, not just talking. I pretty much laid out in post #13 and #18. I actually wrote the spread sheet for noise calculations in the attachments in #19 for work. You will find this paper that JMF linkes http://users.ece.gatech.edu/mleach/papers/Parallel.pdf. It goes through the exact same model and calculation like what I did in post #19.

Anyone come in and give a simplistic answer on noise must not have actually done noise work before. people actually worked on noise know that it is never simple in noise calculations. That's where the Noise Figure circle and all the optimizing of impedance for noise!!! Everyone has to go through like my spread sheet or in the article.



I studied Laplace Transform for designing closed loop control systems. But I ended up using Bode Plot alone to do all my closed loop feedback designs. I found it worked better as you can see every pole and zero and compensate each one.

[tubeswell]

Re: the noise, parallel triodes allow you to configure a gain stage with 1/2 the plate resistance and 1/2 the cathode resistance that you would for a single triode, all other things (HT and AC load) being equal, in order to get the same gain as the single triode. Being able to halve the load resistance immediately reduces resistor noise. But you also have a lower output impedance with a parallel stage, so more bandwidth is preserved through the AC load, which is a higher S:N ratio, so there is less noise in that regard as well.