• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Feedback - global or local?

Thanks for the useful inputs. Let's simplify things to a 3-stage voltage amplifier with line-level output. The voltage gain of each stage is 10. In the A version let's degenerate each stage by 6 dB local negative feedback so the the gain will be 5 for each stage, 5 x 5 x 5 = 125 globally. In the B case let's apply 18 dB global negative feedback. Do we get the same distortion at the same output level? And more importantly: do we get the same FFT harmonics spectrum in the two cases? Somewhere I read that GNFB reduces the net harmonic distortion, but the higher order spectral component will increase over 2nd harmonic. So there will be a spectral imbalance that is unpleasant to human hearing. Is it true?
First errata on my last post: I also tried modelling the E83CC with the GNFB and the modelling reported to be inferior to UL+CFB.

Regarding your question: FWIW: The FFT did not display major differences in modelling. A deviation of the optimum impedance results in a greater change of FFT than between GNFB and CFB. I certainly measured less phase shift with CFB.

An OPT cannot not easily be modelled accurately in LtSpice (if at all) with the result of inaccurate theoretical observations. A different OPT used will have different end results. I fully expect to see different results once built.

But the McIntosh shows a high frequency range extending to 100Khz. GNFB often curtails the high frequency range in order to have unconditional stability. The Pye HF10 Mozart, the Philips EL3516 tape recorder, the Cossor 522 and the Cossr 540 all used CFB.

What brought me to building a SE amplifier with CFB is a remark by George, Tubelab, that cymbals sounded like cymbals. Also both McIntosh and the Quad II appear to have very long lives of their output tubes.

The EL506 was the first output tube I bought as a teen for PP amplifier and I have a pair of them out of an estate sale (spares Lowther 18). They are the latest iteration of the 7868 / 7591 from the era and the EH7868 have more in common with the Brimar EL506 than the EH 7868 has with the RCA 7868. (magnoval base and low value g1 resistor).
 
Last edited:
Somewhere I read that GNFB reduces the net harmonic distortion, but the higher order spectral component will increase over 2nd harmonic.
When you have an amplifier that only generates second-order distortion, applying negative feedback will convert some of that into third and higher order distortion. That holds for both local and global feedback. It is rarely a problem, because amplifiers that only generate second-order distortion are very rare.

See https://linearaudio.nl/sites/linearaudio.net/files/Farren feedback1938.pdf for details. It's an article from 1938 that refers to an even older article by Feldtkeller.
 
  • Like
Reactions: EL506
In tubes, distortion is very much proportional to signal level. This means nearly all the distortion is produced in the output stage. Hence local feedback will reduce this much less than global feedback will. Hence global feedback will reduce the distortion much more than local feedback around each stage.

Cheers

Ian
With due respect Ian: In the modelling 've done the E81CC - 7868/7591/(EL506) CFB + UL (or UL+GNFB) shows better end results than when the E81CC is replaced by an E83CC. It seems to indicate that the greater non-linear distortion of the E81CC offsets some of the non-linear distortion of the output tube.

I've noticed this before: When an adjustable floating paraphase is adjusted for perfect drive signal to the following stage(s) then the ignal at the speaker has more distortion that when the floating parahase is adjuted for best output signal to the speakers. In my case it resulted in almost halving the amount of distortion. The downside is adjustment is required when tubes are changed (or swapped around) because when no adjustment is made distortion can be double or even triple from when it is a perfect drive signal.

To answer zintolo: I look for a balance of distortion and phaseshift. FWIW:There are some papers that suggest that phase shift is at least (if not more) important than distortion.
 
Last edited:
With due respect Ian: In the modelling 've done the E81CC - 7868/7591/(EL506) CFB + UL (or UL+GNFB) shows better end results than when the E81CC is replaced by an E83CC. It seems to indicate that the greater non-linear distortion of the E81CC offsets some of the non-linear distortion of the output tube.
This is a well known technique/effect which has been discussed elsewhere on this forum. I have mentioned previously that the EMI REDD47 preamplifiers use exactly this technique in combination with local NFB around the output stage plus global NFB.

Cheers

Ian
 
With due respect Ian: In the modelling 've done the E81CC - 7868/7591/(EL506) CFB + UL (or UL+GNFB) shows better end results than when the E81CC is replaced by an E83CC. It seems to indicate that the greater non-linear distortion of the E81CC offsets some of the non-linear distortion of the output tube.

I've noticed this before: When an adjustable floating paraphase is adjusted for perfect drive signal to the following stage(s) then the ignal at the speaker has more distortion that when the floating parahase is adjuted for best output signal to the speakers. In my case it resulted in almost halving the amount of distortion. The downside is adjustment is required when tubes are changed (or swapped around) because when no adjustment is made distortion can be double or even triple from when it is a perfect drive signal.

To answer zintolo: I look for a balance of distortion and phaseshift. FWIW:There are some papers that suggest that phase shift is at least (if not more) important than distortion.

Same for long tail splitter.
A good example was the power amp VTL from venerable Manley; they have a trimmer in one section of splitter to trim the THD at the output; normally with a very good selection of tube the test is easy ( you need a distortiometer rand a good generator)
This must be done at different level so you chan fix a proper setting where the THD is reasonable low; it is always a dynamic problem

Walter
 
Tests have shown our ears distort a considerable amount but our ears are very sensitive to phase shifts.
Fred Terman does not agree,
 

Attachments

  • Terman on Phase 6W 30C.jpg
    Terman on Phase 6W 30C.jpg
    83.9 KB · Views: 86
  • Like
Reactions: tubetvr
To illustrate the effect that the distortion of a preamplifier tube can have on the distortion of the combination "preamplifier tube + power triode":

The first diagram shows power output against distortion of an AD1 driven with a undistorted input signal. The power output is 4.2 Watt at 5.5 % total distortion.

The second diagram shows power output against distortion of the combination ABC1 + AD1. The power output now is 4.8 Watt at 4 % total distortion.

Note that the total distortion in the first diagram is for the largest part made up of second harmonic distortion, while the distortion in the second diagram contains a much bigger part of third (and fourth) harmonic distortion.

AD1.jpg


ABC1 + AD1.jpg
 
  • Like
Reactions: EL506
To illustrate the effect that the distortion of a preamplifier tube can have on the distortion of the combination "preamplifier tube + power triode":

The first diagram shows power output against distortion of an AD1 driven with a undistorted input signal. The power output is 4.2 Watt at 5.5 % total distortion.

The second diagram shows power output against distortion of the combination ABC1 + AD1. The power output now is 4.8 Watt at 4 % total distortion.

Note that the total distortion in the first diagram is for the largest part made up of second harmonic distortion, while the distortion in the second diagram contains a much bigger part of third (and fourth) harmonic distortion.

View attachment 1031435

View attachment 1031436
I'll take the second over the first, any day. It's the first watt that is important.

Modelling shows that with cathode feedback the 2nd can be larger than the 3rd. Only at the "perfect" load goes the 2nd below the 3rd. And how many speakers present a "perfect load? Not to forget that most distortion in the OPT is 2nd.

Phase shift: Read the article many years ago, never saved it and am not going to look for it now. Some people are sensitive to it, others not, YMMV.
 
In older literature you often read that the human hearing doesn't detect phase shift/distortion. What surpises me again and again is that this statement often is not backed up with proof/literature.

Attached a more recent master thesis on the influence of phase distortion on sound quality that I found on my PC (conclusions on pages 59 to 61).
 

Attachments

To illustrate the effect that the distortion of a preamplifier tube can have on the distortion of the combination "preamplifier tube + power triode":

The first diagram shows power output against distortion of an AD1 driven with a undistorted input signal. The power output is 4.2 Watt at 5.5 % total distortion.

The second diagram shows power output against distortion of the combination ABC1 + AD1. The power output now is 4.8 Watt at 4 % total distortion.

Note that the total distortion in the first diagram is for the largest part made up of second harmonic distortion, while the distortion in the second diagram contains a much bigger part of third (and fourth) harmonic distortion.
What frequency did you make the measurements at? What was the load? Is it the same across the entire audio bandwidth?

Cheers

Ian
 
They are not mine. The diagrams are on pages 267 and 268 of Book V from this Philips series on tubes. The book is in Dutch though.

PHILIPS' TECHNICAL LIBRARY - Series on ELECTRONIC TUBES

The book doesn't clearly specify at what frequency or frequencies the measurements were made, but on pages 262 to 266 there is a description of an installation to measure power output and distortion, which uses an oscillator and filter for 500 Hz.

The book explains that the ABC1 has to deliver 35 Veff to drive the AD1 to full power. This creates relative high second harmonic distortion in the ABC1, which cancels in part the second harmonic distortion of the AD1.
 
Last edited:
Let's think about typical, simple two stage SE amplifier. At full power the voltage amplifying stage generates some 0.5 % THD while the output stage generates 5%. Then we apply 12 dB (4-times) GNFB. The result is ~1.25% THD at full output power.

Second option; we apply 6 dB local NFB in both stages (12 dB total). The result is 2.5 % THD at full output power.

Which one is better...?
This sounds a good proof that GNFB is generally better than local distributed NFB.
So, I don't understand this discussion going on.
What is wrong in this proof ?
 
  • Like
Reactions: ruffrecords
This sounds a good proof that GNFB is generally better than local distributed NFB.
So, I don't understand this discussion going on.
What is wrong in this proof ?
A lot since it does not take into account that the driver can offset the non-linarity of the output tube. Neither does it take into account the effects of an output transformer. As I mentioned before "it all depends" and a blanket conclusion cannot be made.
 
This sounds a good proof that GNFB is generally better than local distributed NFB.
So, I don't understand this discussion going on.
What is wrong in this proof ?
Depends on your definition of better. In terms of its ability to reduce distortion the answer is generally it is better. There are certain special exceptions where by design or accident a pair of successive stages produce near equal and opposite distortion but other than in that special case it is better at reducing distortion.

Of course, little in life comes for free and in this case the price is increased difficulty in maintaining stability and maintaining stability over three stages is a lot more difficult than round a single stage. This problem is exacerbated by the fact it is hard to dc couple tube stages (which adds unwanted poles and xeros to the transfer function) and it is just as hard to maintain NFB down to dc. Life for transistor designers is a lot easier. Dc coupling between stages is easy and mainting NFB down to dc is not hard either. The result is they have no stability problems at low frequencies. Hence their only stability problems are at high frequencies and these they tend to club to death with a dominant pole at a few Hz and vast quantities of open loop gain. No such luxuries for tube designers.

Cheers

Ian