Question in RDH4

[Austin]

Have you read the other great books at Technical books online yet?

[Steve Conner]

Ha! Another sacred cowburger from my grill.

I agree with JM's sentiments. It makes no sense to use tubes in the 21st century. Therefore, if we are using tubes, it must be for some emotional reason that isn't susceptible to logic. So, using math to refine a design based on illogical assumptions just seems wrong.

When comparing tubes to transistors, the giant stumbling block is the output transformer. Even if the tube were a perfect amplifier, the OT degrades the signal considerably. (The transformer design chapter of RDH4 is fascinating.) Philips tried speakers with 300 ohm voice coils for direct drive, but they didn't catch on.

I plead guilty to using manufacturers' sample layouts for RF circuits, however I have had success making my own layouts for DC-DC converters. I am also responsible for EMC in our products.

[Alan0354]

Have you read the other great books at Technical books online yet?

This is good stuff. I am downloading a few of the tube books as I type here.

[Alan0354]

Ha! Another sacred cowburger from my grill.

I agree with JM's sentiments. It makes no sense to use tubes in the 21st century. Therefore, if we are using tubes, it must be for some emotional reason that isn't susceptible to logic. So, using math to refine a design based on illogical assumptions just seems wrong.

When comparing tubes to transistors, the giant stumbling block is the output transformer. Even if the tube were a perfect amplifier, the OT degrades the signal considerably. (The transformer design chapter of RDH4 is fascinating.) Philips tried speakers with 300 ohm voice coils for direct drive, but they didn't catch on.

I plead guilty to using manufacturers' sample layouts for RF circuits, however I have had success making my own layouts for DC-DC converters. I am also responsible for EMC in our products.

I had been designing with feeling and emotion for a long time even on all the modern stuffs. I am not down playing the feel and emotion part as that is still the most important in design. You don't design by numbers and theory. BUT as I concentrate on theory in the last 10 years or so, I see stuff that common sense and feeling don't give me.
I was converted by a very intelligent scientist that I work with for over 10 years. I had so many I ideas in designs that I made it work and incorporated into the system, but I cannot explain it in writing and explain it, we worked together and he help me explaining my design in theory and words, it was so impressive. He himself is a prominent scientist in the field of mass spectrometer and he has a lot of inventions that have been patented, published in scientific journals. I really see the other side of creativity, the side that comes from mixed theory and emotion. I was lucky to work with a great mind and had the chance to see this in another way.

RF and EM is my real passion in electronics, I actually spent a lot of time plotting Smith Chart by hand, feel how the curve goes, then use simulation to verify. Talk about not intuitive, the most important thing about RF is the Smith Chart, it has nothing to do with common sense, it is like learning a foreign language. I always joke about dancing on the Smith Chart. It is months of playing around the Smith Chart!!! As for EM, all electronic circuit are EM wave, current and voltage is ONLY the consequence of the boundary condition of EM wave traveling in wave guide ( transmission lines) be it a wire above ground plane or coax. These are all starting from observation in the original Maxwell's equation and interpreted by calculus. In fact, the whole topic of Vector Calculus was mainly to interpretate the Maxwell's equations and observation. You see the EM wave propagation, you'll see the EMC and microwave circuit much easier. These are not by simple common sense. Actually, calculus is the language of science, it is like English for other people. A lot of the advanced topics can only be explained by calculus. You'll find a lot of explanations for difficult subjects fall short at the end.

[Alan0354]

I have a new question. In page 549 of chapter 13, equation (2) to (4b)

How is equation (4b) come from the others? It stated:

Power Output = 1/8( Emax - E min)(Imax - Imin) = 1/8 RL (Imax - Imin)^2 = 1/8 [( Emax - E min)^2]/RL = 1/2 EbIQ(1-Emin/Eb)(1-Imin/Ib)

Anyone has the steps of the derivation?

[Jazz P Bass]

My head hurts reading that!

[Alan0354]

Have you read the other great books at Technical books online yet?

Took a long time to download all the books!!!! Now I have more reference to double check and compare.

[Alan0354]

I have another question:

This is in page 34, pentode in triode mode. It is my understanding transconductance gm= ip/vg when plate voltage Vp is constant ( vp=0). Where ip is the change of plate current and vg is change of grid voltage. These are consider AC parameters.

Same definition screen transconductance is gs= is/vg when plate voltage Vp is constant.

Triode transconductance gt= ( ip+is)/vg = gm + gs(2)

As you can see, these are all AC parameters, not the steady state current.

Based on this, gm/gt= ip/ik Which is AC parameters. But the book gave gm/gt= Ip/Ik which is using DC parameters. Why?

The rest of the equations in this page are all using steady state currents. As you all know the DC current has nothing to do with the change in current (AC). I don't see how this is correct. I took a quick scan in the books I downloaded from Austin, I am sure my definition of transconductance is correct.

[Alan0354]

What happen? Nobody study these theories?

[Steve Conner]

Well as I just tried to explain at length, they are pretty much useless for guitar amp design.

In your first question: I think they substituted Imax = 2Iq, Emax = 2Eb, etc. which are reasonable assumptions for a single-ended Class-A amplifying stage.

In your second question, it is just linearisation around the operating point. If the tube is sitting there with 90mA plate current and 100mA cathode current, I think it would be reasonable to believe that if the cathode current changed to 101mA, the plate current would increase to 90.9.

[Steve Conner]

Well as I just tried to explain at length, they are pretty much useless for guitar amp design.

In your first question: I think they substituted Imax = 2Iq, Emax = 2Eb, etc. which are reasonable assumptions for a single-ended Class-A amplifying stage.

In your second question, it is just linearisation around the operating point. If the tube is sitting there with 90mA plate current and 100mA cathode current, I think it would be reasonable to believe that if the cathode current changed to 101mA, the plate current would increase to 90.9.

[Alan0354]

Thanks for the reply. I was thinking about the second question. I was thinking even though Ip/vg is not the same as ip/vg, but it is reasonable to assume is/ip = Is/Ip. So there is a relation of the ratio. With that:

gm/gt= ip/ik implies gm/gt= Ip/Ik because the ratio is a constant.

I think this is what you imply.

[Steve Conner]

Yes, that's right.

You get a similar situation with transistors. Beta is the DC current gain, hfe is the small-signal current gain, but everyone uses them interchangeably because they work out the same, for the same reasons as the tube example.

At least at low frequencies where the device capacitances don't shunt any signal current.

[Alan0354]

Yes, that's right.

You get a similar situation with transistors. Beta is the DC current gain, hfe is the small-signal current gain, but everyone uses them interchangeably because they work out the same, for the same reasons as the tube example.

At least at low frequencies where the device capacitances don't shunt any signal current.

Thanks for the reply. I am still working on the other question, I try substitute Emax=2Eb etc. It will not give me the correct equation. That's what I first tried before.

I thought the question is related to screen dissipation which is important. I know these are not helpful in clipping, but we do use guitar amp in linear( clean ) mode. Then all these equation apply.

I skipped over 80% of the book, mainly concentrate on bias setup, output power, distortion prediction, screen dissipation and amplifier classes.

[Steve Conner]

Well, all I can think of is try substituting Emax=2Eb-Emin, Imax=2Iq-Imin instead. In other words, the signal is symmetrical about the operating point.