Question in RDH4

[Alan0354]

I have a question in page 549 chapter 13 equation 8

M = (Emax-Emin)/2EC1 =2.82 sqrt(P0RL) (8)

But

P0=(1/8)[(Emax-Emin)^2]/RL (4a)

=> 8P0RL=(Emax-Emin)^2 => (Emax-Emin) = 2.828sqrt(P0RL)

You see the denominator is missing? It should be:

M = (Emax-Emin)/2EC1 =2.82 sqrt(P0RL)/2EC1

Can anyone verify this, I don't want to say the RDH4 is wrong.

[Alan0354]

Another question on page 561 regarding the Fig.13.20 and equation (1) to (8)

In Fig. 13.19 and Fig.13.20 The Q point is Eb1 and Ib1. But in the equation (1) to (8), only Eb and Ib are used. I take it Eb1 and Ib1 are Eb and Ib. RSH4 make a typo on this, is this true?

[Alan0354]

Anyone?

[Merlinb]

I have a question in page 549 chapter 13 equation 8

You see the denominator is missing? It should be:

M = (Emax-Emin)/2EC1 =2.82 sqrt(P0RL)/2EC1

Yes, it appears you are correct, there must be an Ec term in the denominator.

[Alan0354]

Yes, it appears you are correct, there must be an Ec term in the denominator.

Thanks for your time. How about the first question?

That's the problem with a lot of books, they have mistakes.........likely typo. But these are very important because if things don't make sense, it's very hard to move on and ignore it. That's the reason when I study a subject, I buy 4 to 5 text books so I can verify the equations. Tube stuffs are too old, kind of hard to find books.

[Austin]

It is good that tubes are forgiving.

[Alan0354]

But you still want to learn the right thing. Nothing is worst than having the wrong information. That's the reason I always go through the derivation step by step on the important formulas as you cannot trust the books. The only books that is likely mistake free are the ones that has many revisions, mostly used in the lower division college classes or some more popular classes. I remember I studied a very popular phase lock loop book by Roland Best, there were so many mistakes I wrote a bad review on Amazon. Dr. Best actually wrote to me on Amazon to offer me a revised copy soon to be published at the time.

Particular on old books like this one, I believe it's written in the 40s. Electronic theories have moved a few quantum leaps forward since that time.

[Steve Conner]

I really think the math in RDH4 is ancient history and not worth learning, except in so far as it might help you understand more modern concepts. You can get a long way with just Ohm's Law and Murphy's Law.

For instance, with screen voltage conversion, you can get a rough idea from eyeballing the tube curves. If you want a curve for 350Vg2, it will be about halfway (ok, maybe one-half to the three-halves power!) between the ones for 300 and 400. Certainly within the tolerances between different brands of tube.

What's most important in musical instrument amp design is the non-linear behaviour (and failure modes! ) under heavy overdrive, which was never covered in great detail as engineers of the tube era saw it as an undesirable thing to be minimised.

Tube hi-fi guys like Morgan Jones and George Anderson like to do all this math, but I think they're missing the point. If they really wanted an amp scientifically optimised for high performance, they should have started with transistors.

[Austin]

"they should have started with transistors." BLASHPHEMY!!!!!!!

[J M Fahey]

Where have you been living the last 50 years?
Or ... are you reading this on an ENIAC ???
Or ... to keep this within the Audio realm: no 15W tube amp in the world matches the specs of a humble $2 TDA2050.

[Austin]

Where have you been living the last 50 years?
Or ... are you reading this on an ENIAC ???
Or ... to keep this within the Audio realm: no 15W tube amp in the world matches the specs of a humble $2 TDA2050.

In my van.

Hey my laptop mabe kinda slow but it does not have any tubes in it.

The 2050 is cool. I think tubes are cooler though and don't you? If things were as easy as plugging into an IC why wouldn't everybody do it? Oh wait... Most people do... Still, I think one 12ax7 could drive one 8417 and vastly outspec a 2050 with a very simple power supply and a very few resistors and caps. Way less than the 2050 has going on inside...

[J M Fahey]

Can't argue on "cool" because it can't be measured or quantified.
As of specs, a "Golden Years" (50's. early 60's) tube designer would have Killed to get TDA2050 specs with his amps.
And I bet that 1 to 10 Million IC output power amps are sold for every tube amp out there, with the sole exception of Guitar Players who, yes, crave on Distortion.
Oh well.

[Austin]

Can't argue on "cool" because it can't be measured or quantified.
As of specs, a "Golden Years" (50's. early 60's) tube designer would have Killed to get TDA2050 specs with his amps.
And I bet that 1 to 10 Million IC output power amps are sold for every tube amp out there, with the sole exception of Guitar Players who, yes, crave on Distortion.
Oh well.

The 8417 is very clean and needs only a very small signal to go full power. The super tube of the 1980s.. I paid at least 15 bucks for mine and that was lucky as they can fetch much more than that. The loudest and cleanest possible tube in the THD Univalve amp I have. I bet a hifi made with them would really excel.

[J M Fahey]

Yes, very clean indeed ... for a tube, that is.
Datasheet claims 2.5% distortion, in an Ultra-Linear amp.
Don't even want to imagine the specs on a conventional one.
The excellent Dynaco Mark VI amp, which was designed around them, applied a lot of high level design technology to be able to claim less than 1% distortion.
And 10Hz to 40 KHz frequency response ... at 1 W level.
Now, having one of them is beyond cool, no doubt.
Almost as cool as having a 1926 Bugatti parked in front of your Van:

Mmmmmhhhhh, I'm DROOLING !!!!
What a BEAUTIFUL car !!!!

[Alan0354]

I really think the math in RDH4 is ancient history and not worth learning, except in so far as it might help you understand more modern concepts. You can get a long way with just Ohm's Law and Murphy's Law.

For instance, with screen voltage conversion, you can get a rough idea from eyeballing the tube curves. If you want a curve for 350Vg2, it will be about halfway (ok, maybe one-half to the three-halves power!) between the ones for 300 and 400. Certainly within the tolerances between different brands of tube.

What's most important in musical instrument amp design is the non-linear behaviour (and failure modes! ) under heavy overdrive, which was never covered in great detail as engineers of the tube era saw it as an undesirable thing to be minimised.

Tube hi-fi guys like Morgan Jones and George Anderson like to do all this math, but I think they're missing the point. If they really wanted an amp scientifically optimised for high performance, they should have started with transistors.

I think it's important to learn the math if possible. There is a lot of insight going through the process rather than just play with it by experience. There are things that is intuitive, but there are things that is not. It is my experience that you can make the non intuitive part become intuitive by pain stakingly going through the steps. I know because I never have a formal education in EE before and I had a successful career in EE for 20 years getting by with good dose of common sense, ohm's law. But I changed, I spent the last 11 years studying 2 to 3 hours a day 6 days a week, making up all the requirement of BSEE and beyond. Changing my approach from common sense to theoretical, it pay off big time.

For example, pcb layout and/or component placement and wiring of tube amp. A lot of people consider it's black magic, in the older days, most follow what was told by experienced people which they learn from older people. Then as speed goes up, EMC books popped up and give you a cook book approach. I spent 3 years studying all the advanced calculus, studied 3 books on electromagnetics. I can now starting to see how the circuit behave by predicting how the current propagates, how magnetic field effect the components etc. The point is making the part that people consider non intuitive and make it intuitive, you see more.

I also spent 3 years studying microwave, going through every steps of the design pain stakingly, all the derivation, smith chart and all. I have done circuit design that after putting onto pcb, the graph I plotted out from the pcb is almost identical to the prediction. Most of the RF engineer don't want to pay the piper, skip the tedious part of Smith Chart, they resort to just copy the sample circuit from the manufacturer only. It's the insight that you get from the detail that really help.

I am new in tubes, after I got the RDH4 thanks to Enzo, I really read and picked out parts that I consider important and really get into it. I just used what I learned about how to use the load line to get the most unclipped output, setting the RL and B+ by dropping resistor to get the plate voltage I want to get the best sound. Also looking at the even harmonic distortion by the load line. I just did this, change the resistors to lower the plate voltage, optimize the dynamic range and heard a definite improvement in the sound this past weekend one time through from calculation, the voltage reading is within 10V from calculation. I know this mean nothing to expert like you, but this is a starting point for me.

I do skip 80% of the RDH4, I just use my experience to pick out things that is important from solid state design and really go into it.