Need suggestion of a good tube book.

[Alan0354]

Hi

I got the Tube amp book by Morgan Jones, I read a little over a hundred pages and I am disappointed with the book. Maybe I got spoiled by the SS power amp book by Bob Cordell. Cordell gives great detail in the strong and weak points of each different circuit configuration, comparing distortion and optimization. So far I see very little with Morgan Jones. there is no theory explanation to speak of either.

I understand this is not supposed to be a theory intense book, I got the RDH4 for that already. But when I read through the mu follower, it is weak to put it kindly. I had to read on line in another article to clarify the explanation. Even after that, read back the Jones' book, still it is confusing. I skip through to the later chapter on power amp and OT, it did not even explain anything on Ultra Linear tap and advantage and disadvantage. I stop after that, it's no point to waste my time. I know they use a tap on the OT for the screen grid, why does that help, how much it help? How does it compare to using triode push-pull that cancel the even harmonics and low on odd harmonics? None is being answered!!!

Is there any better book? I so wish Cordell write a book on tubes.

 

[scopeboy]

The RDH4 explains how ultralinear works. If you missed that part, the original Hafler and Keroes paper "An Ultra-Linear Amplifier" is just a Google search away.

There isn't really much point in analysing the strengths and weaknesses of the different tube topologies. To put it bluntly, by any objective metric they all suck compared to solid-state. It follows that if you chose tube over solid-state, you did it for emotional reasons, as there is no logical reason to do so. Therefore you might as well let emotion instead of logic guide your choice of topology too.

From this point of view the Morgan Jones book is really quite good.

PS: UL gives a little more distortion than triode, but more power output with less grid drive voltage required.

 

[Alan0354]

I am just using the Ultra Linear as one example. In general, I found this book very incomplete. No formula to speak of. Then no analytical details of comparing different configurations. I have a lot of books on electronics, I have to rate the Jones book one of the worst.

Regarding to tube and SS, I stay out of this as I have seen die hard on both side. I do agree SS should give lower distortion, but whether this translates to better sound or not, I'll let other people debate on this.😛. I am planning to do one of each, that's why I studied Cordell also. I pretty much finished the book up to the simulation and some of the thermal management stuff. I pretty have an idea on part of the SS amp, I am still deciding whether I'll go MOSFET or BJT on the power transistors.

 

[SY]

If what you want is comparisons of different topologies (rather than detailed examination of the highest performance topologies, which is what Jones does- at least if you go past the first hundred pages), Crowhurst's "Understanding Hifi Circuits" gives a nice overview.

 

[rayma]

I got the Tube amp book by Morgan Jones, I read a little over a hundred pages and I am disappointed with the book.
Is there any better book? I so wish Cordell write a book on tubes.

For tube circuits you won't find a better single source than RDH4. For specific topics in depth you'll have to google, though.
Wireless World had many excellent articles, including some by JLH and Baxandall. Here are some examples on Self's web site.

The Wireless World Archive

 

[Alan0354]

Thanks both of your inputs. You mean they don't have anything newer since the 50s?

 

[SY]

There aren't many books on buggy whips written since 1910.

 

[rayma]

Thanks both of your inputs. You mean they don't have anything newer since the 50s?

Well, after audio mostly went to semiconductors in the seventies, tubes were left to high end specialty manufacturers.
Not much more was published except for Audio Amateur and Glass Audio.

 

[DF96]

The book you want hasn't yet been written. Have you looked at Merlin Blencowe's site: the valve wizard? That may answer some of your questions.

 

[Alan0354]

To put it bluntly, by any objective metric they all suck compared to solid-state. It follows that if you chose tube over solid-state, you did it for emotional reasons, as there is no logical reason to do so.

I went back and read up some about distortion. Transistor curve is more like pentode, so you are going to have a lot of odd harmonics all the way to 7th and even 9th. But triode curve is power law and predominantly 2nd harmonics with very low higher harmonics. In push pull through OT, 2nd harmonics is cancelled. So I don't understand why you say tubes are much inferior than SS for power amp.

As matter of fact, even in the earlier stages, SS still produce odd harmonics, where tubes mostly 2nd harmonics, which is more pleasing to the ears.

Please explain what I missed.

 

[DF96]

What you may have missed is that all active devices (apart from an ideal JFET) generate distortion at all orders, with a decreasing amount as the order increases. The power law for a triode is 3/2, so it is an infinite series (just like a BJT exponential curve is an infinite series). Even an ideal JFET starts generating all orders once it is in a circuit and its output can affect its effective input.

People often say that triode is mainly 2nd and pentode is mainly 3rd but the truth is that both produce more 2nd than 3rd, and more 3rd than 4th etc. Give the triode a high impedance anode load and its distortion will go down but it will still produce all orders.

About the most distorting active device is the BJT. Some people don't seem to realise that a bare BJT produces 2nd order distortion at the same percentage as the input signal in peak mV. Yes - put a 1mV signal into a BJT and get 1.4% 2nd (plus all higher harmonics too, at lower levels) out. The reason BJTs are usable in high quality amps is that the circuit ensures only small signals reach the base or uses the BJT in a current-driven mode.

 

[SY]

In push pull through OT, 2nd harmonics is cancelled. So I don't understand why you say tubes are much inferior than SS for power amp.

In a transistor amp, there's no need for an output transformer. So the output stage is commonly run as an emitter follower. Since the gm of a transistor is high, the feedback inherent in an emitter follower is likewise high, thus the distortion is low. Likewise source followers. As an example, note the experimental comparisons in distortion between source followers and cathode followers documented in Jones's book.

If an amp is properly engineered, the distortion is low enough that the order doesn't matter much.

 

[Alan0354]

In a transistor amp, there's no need for an output transformer. So the output stage is commonly run as an emitter follower. Since the gm of a transistor is high, the feedback inherent in an emitter follower is likewise high, thus the distortion is low. Likewise source followers. As an example, note the experimental comparisons in distortion between source followers and cathode followers documented in Jones's book.

Actually this the very problem with crossover distortion using SS followers. The low output impedance ( high gm) creates a voltage divider effect. During the transition, you going from both transistor driving to only one transistor driving, the output impedance change and create the crossover distortion. Tube is high impedance, this minimize the voltage divider effect at the transition. So valve output is superior to SS in the crossover distortion department.

This is well explained in page 101 to 103 in Bob Cordell's book. Tube does not have this problem.

 

[DF96]

The low output impedance ( high gm) creates a voltage divider effect. During the transition, you going from both transistor driving to only one transistor driving, the output impedance change and create the crossover distortion. Tube is high impedance, this minimize the voltage divider effect at the transition.

As I said in the other thread, the opposite is true. It is the low output impedance of source/emitter followers which means that their crossover distortion is not at high as it otherwise might be.

Think about how voltage dividers work.

 

[Alan0354]

What you may have missed is that all active devices (apart from an ideal JFET) generate distortion at all orders, with a decreasing amount as the order increases. The power law for a triode is 3/2, so it is an infinite series (just like a BJT exponential curve is an infinite series). Even an ideal JFET starts generating all orders once it is in a circuit and its output can affect its effective input.

People often say that triode is mainly 2nd and pentode is mainly 3rd but the truth is that both produce more 2nd than 3rd, and more 3rd than 4th etc. Give the triode a high impedance anode load and its distortion will go down but it will still produce all orders.

Triode has relatively much lower amplitude in odd harmonics than pantode and transistors. 2nd harmonics can be cancelled out in push pull valve stage with OT, but odd harmonics added. That's the whole point of using triode output stage for low distortion. Transistor is very similar pentodes that has more odd harmonics.

 

[Alan0354]

As I said in the other thread, the opposite is true. It is the low output impedance of source/emitter followers which means that their crossover distortion is not at high as it otherwise might be.

Think about how voltage dividers work.

I am thinking exactly about the voltage divided effect. As I added later in the last post, the crossover distortion is explained in page 101 to 103 in Cordell's book on SS. The tube has high output impedance. So even with step down OT, the impedance is still not low. So if you use voltage divide to calculate, the higher the output impedance, the less effect on the voltage developed across the speaker. This is the whole concept of current source. The higher output impedance of the current soruce, the less effect on the current with changing of the load impedance.

With SS output, you always have a distinct kink at the crossover.

 

[DF96]

You still have it back to front. Almost all speakers are designed to work from a voltage source. Raising the source impedance makes any non-linearity in that impedance more of a problem, not less of a problem as you seem to think.

The main reason for using triodes as output devices is the low output impedance, which reduces the need for global NFB.

 

[Alan0354]

I read about 300 pages of the book already, I still need to read the last part on different brand amps. I still have to say I would review it as two and half to max three stars out of five. I have no reservation of giving Cordell's book five stars. Big difference in the presentation. The only topics that is better is on the capacitors, OT is ok, but still not much on crossover distortion characteristics and optimization. We definitely can use a better book.