Actually Jan does pay his authors. In any case, isn't the fun of writing articles that there will be something left that may still occasionally be read decades after your death? If you want to get rich, it is more effective to become a banker.
I thought Electronics World wasn't the least bit interested in analogue audio anymore (or in anything that I find interesting for that matter). That's one of the reasons why Jan Didden started Linear Audio in the first place.
I'm not that altruistic. I don't care what happens after I die - I won't be around to enjoy it.
But, when I've had articles published, I did get a thrill out of seeing my name in the byline. One time, I was surprised when walking into the local newsagent. The lastest issue was out, and they had made a submission of mine the cover feature. I got quite a thrill out of seeing my own name in 15 mm high letters in every newsagent for a month!
I had several articles in Electronics World (not audio though), before Editor Svetlana started publishing drivel from Ivor Catt and others of his ilk again. Then she stopped mass circulation to newsagents. I emailed her and complained about that, about Ivor Catt, about her taking it away from analogue audio. She sent me a couple of quite long emails saying she was forced to make the changes due to dropping sales. She recognised that she could increase the sales with good audio and good electronics design articles, but she could not find the authors. She asked me to step forward, but I was not going to do it for no pay.
Last edited:
That's debatable, global feedback often adds more distortion harmonics than it cures, and if daft enough to include the OPT it doesn't help that much at all at the frequency extremes where the open loop gain is lower.
An SE triode with local output tube to driver cathode feedback will comfortably out-perform an SE crippled with global feedback.
Guilty, that was me. Yes, I still think that an amplifier must be able to drive any kind of well designed speaker, and play properly any kind of music program. All this in the average domestic listening room, not at your wedding party or a stadium, just to be clear.
Or, we need to stop talk amplifiers, and start thinking about "music reproduction systems" where electronics and transductors are one indivisible unity
Well, I personally agree and this is why I suggested a Williamson with a >30W PP UL output stage. I have an amp like this and I found that it works quite happily with all speakers that I've tried. It struggles a bit when I try to play bass-heavy material on inefficient speakers at high volume, but that is only to be expected.
But some might argue that an amp like this is throwing away some magical property of tubes and trying to play solid-state designs at their own game. Doug Self's suggestion of the "niceness knob" should make clear his views on that issue. He thought tube amps were pointless, but if he ever did design one I'm quite sure it would have low distortion.
Morgan Jones' Crystal Palace is another high power and low distortion amp that might possibly be called Blameless, except it could easily be blamed for a hernia or back trouble.
Maybe 15 years ago I made a P-P KT88 amp with a "magic" knob. The knob was mounted on a 4 gang pot that increased the global negative feedback with 2 sections and increased the line stage gain in front of the amp at the same time with the other two sections. This way the amp's overall gain remained relatively constant over a range from zero to about 15 db of GNFB.
I loaned this amp out to dozens of people over the course of a year asking them to tell me what setting (numbered 1 to 10) on the magic knob they preferred. I worked in a building that designed and built cell phones and other electronic equipment, so the test users were generally technical people of above average intelligence. Many were engineers. Most users were under 35 years old.
The results were not what I expected. The distribution of settings was almost an upside down bell curve with most people preferring the extremes. The maximum feedback extreme was the most popular setting.
One day the amp did not come back to me. An envelope full of cash came back instead, so I made another.......That was the beginning of Tubelab.
I loaned this amp out to dozens of people over the course of a year asking them to tell me what setting (numbered 1 to 10) on the magic knob they preferred. I worked in a building that designed and built cell phones and other electronic equipment, so the test users were generally technical people of above average intelligence. Many were engineers. Most users were under 35 years old.
The results were not what I expected. The distribution of settings was almost an upside down bell curve with most people preferring the extremes. The maximum feedback extreme was the most popular setting.
One day the amp did not come back to me. An envelope full of cash came back instead, so I made another.......That was the beginning of Tubelab.
All distortion is not bad.....depending..... but with tube amps usual triode preamp/PP output, abundance of symmetrical clipping/odd order harmonic emphasis/content, that sounds fake/not very linear response/not pleasing to the ears.....
It's the asymmetrical/even order content that is far more interesting/pleasing.
It's the asymmetrical/even order content that is far more interesting/pleasing.
I disagree, you can get a lower output impedance with local feedback than with global. The local feedback version wins on every metric.
With global (with the daft OPT) your feedback ratio is limited by phase shift and really only works on the mid-range frequencies.
With two-stage 'local' feedback you can get huge levels of feedback that works across a huge frequency range, and will still be far more stable with any load because you haven't got a huge frequency dependent phase lagging lump of iron right in the middle of the critical path.
@tubelab - interesting story!
LOL - what is your anode resistance compared to a few ohms of the transformer primary? A fraction of a dB in feedback wipes out any effect.
@AudioFreak88 I know what you are saying but with the two-stage feedback (output tube anode to driver tube cathode) I haven't found that to be the case. Perhaps because that part of the circuit is good for a few MHz.
Also even a follower uses local feedback - just over 1 stage rather than 2.
One "blameless" tube amp that I didn't see mentioned - DCPP Amp
We could crowd research this if we wanted to. First we would need to settle on a topology. Getting anyone to agree on anything in here can be like herding cats, which is odd considering we all have a shared interest. We could work on it together and it could become a tube option in the DIYaudio store. We would need to set some guidelines out and decide on a starting point.
I, for one, think that it would need to be a push pull topology using a kt88 or something similar. I love SET but I think that an amp that can push all but the most demanding speakers is a good goal. I do not know about the front end or phase splitter. I think that all modern options should be considered (LED bias, transistor follower, etc.) along with traditional. What ever perform in the most ideal way, stays.
I think that to complete such a project properly we would need to define improvements objectively using a set of measurements. These could include IMD, THD, frequency response, slew, noise and possibly clipping recovery). At the end anyone who wants to continue into the subjective areas are free to do so on their own thread.
Version-ing could be a bit of a problem but we could handle that as we go.
Here is what I propose for a starting point. Please add your own or modify mine and lets see if we can come to a consensus. If not at least we tried.
Front end: open - voltage reg(salas?)
Phase Splitter: open
Output: Push Pull UL Kt88 (can be switchable but prefer to focus on UL)
PS: CLC or custom designed SMPS
We could crowd research this if we wanted to. First we would need to settle on a topology. Getting anyone to agree on anything in here can be like herding cats, which is odd considering we all have a shared interest. We could work on it together and it could become a tube option in the DIYaudio store. We would need to set some guidelines out and decide on a starting point.
I, for one, think that it would need to be a push pull topology using a kt88 or something similar. I love SET but I think that an amp that can push all but the most demanding speakers is a good goal. I do not know about the front end or phase splitter. I think that all modern options should be considered (LED bias, transistor follower, etc.) along with traditional. What ever perform in the most ideal way, stays.
I think that to complete such a project properly we would need to define improvements objectively using a set of measurements. These could include IMD, THD, frequency response, slew, noise and possibly clipping recovery). At the end anyone who wants to continue into the subjective areas are free to do so on their own thread.
Version-ing could be a bit of a problem but we could handle that as we go.
Here is what I propose for a starting point. Please add your own or modify mine and lets see if we can come to a consensus. If not at least we tried.
Front end: open - voltage reg(salas?)
Phase Splitter: open
Output: Push Pull UL Kt88 (can be switchable but prefer to focus on UL)
PS: CLC or custom designed SMPS
This is probably not what you wanted to hear, But I don't this is something that crowd researching will work too well on.
Doug Self was uniquely well qualified to do his blameless work:-
a) He was, by the time he had his work published, a design engineer with decades of experience in designing recording studio products of the highest standard
b) His work was heavily dependent on SPICE analysis, and he had the SPICE skills to do it. Not just the ability to enter in a circuit and see what SPICE does with it, but to really understand the deatil of what's going on.
c) He had access to full range of professional grade test gear/instrumentation.
Note that the bipolar transistor models used in SPICE are very accurate with respect to modelling distortion mechanisms in amplifiers. Tube SPICE models are not accurate in this regard, much as power FET SPICE models also will mislead you when modelling distortion.
His work was not about tweaking circuits by ear, following fashions, or going for "musicallity". His work was about deep engineering analysis and clarity of logic.
Unless you have the full mathemectical analysis skills of a professional engineer, you don't cut it. I have no beef against those who tweak and experiment by ear, but that's not what "blameless" engineering is about.
Unless you have access to professional grade instruments - eg harmonic distortion testing to the highest commercial standards, you don't cut it.
Unless you are familair with the enormous body of technical and engineering literature (peer reviewed journals, professional society proceeding and the like), you don't cut it.
None of this has anything to do with home constructing tube amps, going for particular types of sound you can get with tubes, and enjoying the results, and having friends and relations admire what you've done. All of which is prefectly valid and worthwhile thing to do. Things I have enjoyed doing myself.
Here's an analogy: I also have enjoyed hotting up cars. That's another nice hobby to take up - lots of enjoyment. But I would not for the minute think I can do what engineers in the labs of GM, Mercedes, etc do, and engineer from the ground up a new engine for volume manufacturing.
Cround researching works well, when there is a large problem needing a lot of work, the folk leading it are firmly in control, and the crownd of volunteeers are each prepared to do their own tiny allocated piece of it in just the way requested. Is that a picture of us here?
Last edited:
What I think the folk on this forum could do, that would be very valuable, is crowd research what causes unpleasantness in reproduced sound.
It has been known at least as far back as the 1940's that THD has only a very poor correlation with the level of distortion as precieved by the ear.
Intermodulation distortion has a better correlation, but even it doesn't serve all that well.
For example, it is well known that a Class B amp showing the same THD or intermod as a Class A as measured on instruments, will in fact sound significantly worse.
There has been a lot of research done on this since the 1950's, but the topic is by no means pinned down.
Perhaps folk might be willing to construct amplifiers deliberately designed to provide variable amounts of specific types of distortion, and survey friends as to what the audible effects are.
I recall a member of this forum constructed and amplifer having adjustable feedback done in such a way as the volume remained constant. He aonly labelled the knob 0 to 10, and told nobody what it did. He lent it to friends and workmates as I recall, and asked them what their preferred position of the knob was.
The result was surprising. Most people prefered maximum feedback, but a few preferred none. Hardly any preferred moderate amounts. This is actually valuable knowlege. Even more valuable if he had investigated why.
Just the sort of thing members of this forum could do.
I think Keit's second last post pretty much sums up the dissonance between the concepts of "tube" and "blameless". You could design a powerful, low distortion PP amp capable of driving any traditional speaker (and I think MJ's Crystal Palace topology is a great example of this) but then you might as well have made a solid-state one. The difference between tube and solid-state is the very thing that Self would consider to be "blame".
And yes, resistance of the OPT windings is significant and global feedback will compensate it and improve speaker damping. This is Valve Amp 101, in any textbook since the 50s. I've also experienced it with my own gear. I have a Williamson type amp with 4 and 8 ohm taps on the OPTs. Irrespective of the actual impedance of the speaker, I get the best results when it is connected to the same tap that the feedback is taken from.
And yes, resistance of the OPT windings is significant and global feedback will compensate it and improve speaker damping. This is Valve Amp 101, in any textbook since the 50s. I've also experienced it with my own gear. I have a Williamson type amp with 4 and 8 ohm taps on the OPTs. Irrespective of the actual impedance of the speaker, I get the best results when it is connected to the same tap that the feedback is taken from.
Last edited:
That was me, and I briefly discussed it in post #27.
I think investigating "why" would have been a daunting task, since even though most of the test subjects were technical people, many had no knowledge of the principles of audio and sound engineering.
More important would have been to keep tighter controls on the users demographics, listening preferences, biases, and the type of audio equipment they normally used.
At the time of this experiment I was working in a Motorola plant where cell phones and two way radios were designed and MANUFACTURED. We were developing the first generation Nextel walkie talkie cell phone, and typically working 60 hour work weeks, so the intended "controlled experiment" became uncontrolled chaos rather quickly.
Early in the experiment I actually went to a few users residence and connected the amp up and introduced the user to tube amps in general. After the first few months, I lost control of the experiment. There were weeks at a time when I didn't know who even had the amp. There was a log book that followed it around for testers to log their experience. A fellow engineer collected the data and ran the stats.
At that time there were a few thousand technical people employed at that location, and about 1000 of them were engineers. Within 10 years about 90% of them had been laid off. Nextel no longer exists, Motorola sold the phone brand to Google, who kept the patents and sold the rest to Lenovo. The plant is now a ghost town, and I left earlier this year after a 41 year career. That large sample of "test subjects" is lost forever.
Unfortunately, that group was not the typical demographic of people found here. The median age was probably under 30, and from my earlier visits, the typical program material being used was also not the norm here. The common program material I saw was rock, rap, techno, or dance music, home theater or a mix of all of these. Most tube amp users here prefer a bit different material, and are probably a bit older than 30.
I will explain one of the early tests that I witnessed. I took the amp to a friends house (an electrical engineer) that had never heard a tube amp before, but was curious, and probably wanted to build one more for novelty value and snob appeal than anything else. Him, his wife, two young kids, and two cats were present.
I hooked the tube amp in place of the front channels of his multi gigawatt 5.1 channel home theater box. We played with it in 2 channel mode, with and without the coffin sized sub. Early on we had an action movie playing with the amp in the zero feedback setting. The sub was engaged. The sound was not as well controlled as with the HT box, but this was not noticed by the users. When the fighting started and the guns started blazing, both cats ran from the room and hid somewhere. This had never happened before. We played some music including some CD's that I had brought to show off the amp, then I left.
The users kept the amp for the weekend, and returned it on Monday. They preferred the maximum feedback setting for all program material, but chose not to build a tube amp. They all agreed that their existing HT box was "better". Incidentally the guy had built a car stereo system in his Nissan 360Z sports car. He thought it sounded great....including the rattling license plate from the big a$$ sub, but I couldn't stand to ride in that car.
We had an audio test chamber in that plant. It was developed by an ex recording engineer who had a few gold records on his wall. It's purpose was to recreate any acoustical environment that our products would see in the world. The engineers would take the Nagra to an environment (say the assembly line in a Chrysler car manufacturing plant) and make a recording. Then we could recreate a matched SPL test situation. I knew many of the people who used to work in there, and we "recreated the Pink Floyd Pulse concert that came to Miami....without the drugs, rain, and laser show. None of them liked the tube amp either!
I think that recreating the same experiment today would require one, or preferably, more than one amp being tested against a "blameless" standard amp. Since a "blameless" tube amp hasn't been defined yet, it would probably be a solid state amp. Passing an amp (any amp) around from user to user in their own home, does gather real world data about user preferences, but does not make for a well controlled experiment. There are too many variables. Perhaps a similar experiment should be done at one of the larger audio gatherings like Burning Amp?
Tubelab, I appologise for not naming you - I remembered reading about it, but could not remember just where.
Yes, indeed it would be daunting, if one was to try and get it all pinned down like double blind medical experiments are supposed to be....
But it would be extremely valuable to know a few things about :-
a) who prefered max feedback
b) who prefered no feedback
c) who prefered middle feedback.
Perhaps the ones who prefered full feedback like to listen to certain types of music, and the ones who wanted no feedback liked to listen to some other genre.
I've noticed that some of my friends like stereos that have perceotable distortion. I asked why. The answer in a couple of cases was that their experience was that their husband/wife/teenage son keeps the volume low, so they don't have to put up with music they don't want. When they had another system that had calean sound, they turned it up and caused anoyance.
When FM radio started in the UK, many of the public were exposed for the first time to low distortion music (less than 1% THD!) over a band 50 Hz to 10 kHz for the first time (see note below), after decades of 50Hz to 5KHz AM. Housewives in droves disliked FM! They thought the treble was some sort of unwanted noise that shouldn't be there. "My old Murphy never made those tinny noises!". Nowadays, you'd be hard put to find someone who likes AM radio that isn't deaf.
The moral of the story is this: people like what they are used to.
My wife, given half a chance, turns our subwoofer off. I've never got a sensible reason out of her why she does that. But clearly she doesn't like deep bass. And equally clearly no deep bass means no realism.
And your amp: No disrespect to your amp or your workmates, but maybe the ones who prefered no feedback weren't after no feedback or tube sound at all, maybe they liked the speaker resonance than can be just perceptively greater with the reduced damping that comes with no feedback. Or maybe just because the extra THD matched their own.
I'm particularly interested in the small number who liked middle amounts of neg feedback. Why is this minority different?
NOTE: In case someone says Keit is talking nonsense again: Yes, I know the technical standards for FM broadcasting call for an audio bandwidth 50 Hz to 15 kHz. That was the case in the USA, Australia, and all other countries that I know about. But in the UK when FM started, the BBC had long been in the practice of renting 10 kHz studio-transmitter links from the British Post Office, who acted as the phone company. Their best practice for AM was links flat to 10 kHz. When FM started, the British Post Office said in typical government monopoly we-know-best style, in effect "Flat to 10 kHz is what we do, and flat to 10 KHz is what you get." And on AM, even though the BBC spent good money providing a 10 kHz audio bandwidth transmitted, almost universally radios and radiograms made in the UK provided only a 5 kHz audio bandwidth, often not even that good.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.