Here's a first: using classic measurements of THD and IM, and assuming your global NFB is limited to 20+/-1 dB, and assuming a straight 100/200W into 8/4 Ohms, what would you consider to be acceptable measurement results open and closed loop?
I mentioned the NFB factor only to let you know there's not much of a backup out there, you have to get it right more or less in open loop.
This is just another way of saying it like John, "use as little feedback as possible", and the amount is my own thing, the reason being that I feel amps generally sound better WITH global feedback, and to my ears, 20 dB is the maximum above which global NFB will no longer audibly improve anything. Bear with me on this point.
No funnies, now, all right?
I mentioned the NFB factor only to let you know there's not much of a backup out there, you have to get it right more or less in open loop.
This is just another way of saying it like John, "use as little feedback as possible", and the amount is my own thing, the reason being that I feel amps generally sound better WITH global feedback, and to my ears, 20 dB is the maximum above which global NFB will no longer audibly improve anything. Bear with me on this point.
No funnies, now, all right?
I'm unaware of any amp with a constant 20dB (or any other constant amount) of feedback over 20-20k. And for a design engineer, a lumped THD figure is quite incomplete, at least since 1968. 8 ohm resistive loads are not the intended purpose of most amps, so it looks to me that you've thrown enough artifice and restriction to make the question meaningless.
I get it - you have no idea of what to say, hence the philosophy speech.
For cryin' out loud, SY, is it that hard to say: my amp, in open loop regarding global NFB, and at rated power into 8 Ohms, has a THD of x.xx %, and at 50 kHz, this is y.yy%?
I didn't ask for your analysis of my amp or my sex life, just a simple, easy question, and you let me worry about the philosophical side of things. You may prefer to rock'n'roll straight off, I prefer to take it easy, one step at a time. I'm that kind of a guy, I first like to make sure I've squeezed what I can from it under lab conditions, and I'll evaluate it later with a 4 Ohm load in parallel with 1uF with 10 kHz square waves.
Fine, let the global NFB be equal to 20 dB, +/- 1 dB, at 1 kHz; let the nominal load impedance be an 8 Ohm laboratory resistor, and let the nominal voltage across that resistor be 28.3 Vrms or 40 Vpeak. It is not necessary for you to understand why I would like to hear such results, unreal as they may be, just provide your opinion of what you'd think of as good results, or let it go.
If you put enough false assumptions in a question, it becomes meaningless, as do any answers. It may not please you to have them pointed out, but I think that's better than starting a pointless debate. Perhaps you'd do better to ask what measurements design engineers actually do to ensure that their amps work as intended?
Never mind SY, forget it.
I think my question was quite precise, if not the first time, then in my last message. If you don't like it, that's tough luck, but you have to face the simple fact that not everybody thinks along the same lines as you do (or I, or anybody else). You have no idea why I was asking just that particular question, but you know it's pointless?
So - never mind, forget it.
Last edited:
Sy,
He meant 600 watts @1 ohm use as much negative feedback as necessary , give it your best shot ...
"Best shot"?
That also is not at all scientific, Wayne, you old rogue. You should know by now that here, only science works.
Being a tube guy, I tend to avoid 1 ohm speakers.
What ..? You need Carver to show you how , his tube amp has a 1 ohm tap..
"Best shot"?
That also is not at all scientific, Wayne, you old rogue. You should know by now that here, only science works.
Yeah we should have Carver send an amp over to Sy , I think too much flea power has killed his zeal ..
On another note we have one of John's halo amplifier's runRning in now for a few weeks, while its protection circuit didn't like the Big bembeh , it's doing very well on the Maggie's (3.6) listening session last weekend was impressive , this is a very good amplifier sonically , I'm now tempted to try the JC-1's , well if I get one of the kids to pay their way thru College ..
Cheers ...
People on here used to be (wrongly) accused of always using THD, which our (absent) friend told us was meaningless. Now people are being criticised for refusing to use THD.
2% THD might be acceptable. 0.02% THD might be unacceptable. It all depends. Perhaps I should clarify for those who are confused by numbers, feedback, Ohm's Law etc. : I am not implying that higher THD is necessarily an advantage!
Does that answer the question?
2% THD might be acceptable. 0.02% THD might be unacceptable. It all depends. Perhaps I should clarify for those who are confused by numbers, feedback, Ohm's Law etc. : I am not implying that higher THD is necessarily an advantage!
Does that answer the question?
Well,
Have you tried a mono signal into head phones one in each side from different sources or through different "signal paths and components" playing the same music?...Then reverse the HP on your head and compare...Just for fun of corse..Direct comparison...
I think our sonic short term memory is not so good..
Regards
M. Gregg
Have you tried a mono signal into head phones one in each side from different sources or through different "signal paths and components" playing the same music?...Then reverse the HP on your head and compare...Just for fun of corse..Direct comparison...
I think our sonic short term memory is not so good..
Regards
M. Gregg
Wave, what's wrong? Why the ridicule? What's wrong with classic measurements? They don't tell the whole story, to be sure, you could even say they tell relatively little of the whole story, but is that reason enouugh to make someone asking about it a pariah?
What's happening here? Did I chance upon a question which the tube group, as they are the only ones showing signs of life (as well as DF96) and agitation, ask a question which must never be asked? Because the answer might be something like - yeah, it has 5% of THD, but it's the NICE 5%?
Will any of you, or rather CAN any of you tell why is my question so repulisve to you?
What's happening here? Did I chance upon a question which the tube group, as they are the only ones showing signs of life (as well as DF96) and agitation, ask a question which must never be asked? Because the answer might be something like - yeah, it has 5% of THD, but it's the NICE 5%?
Will any of you, or rather CAN any of you tell why is my question so repulisve to you?
Quite agreed.
Since you were the only with the decency of giving a fair answer, let me tell you why I asked in the first place. After a lot of experimenting, I managed to get the 1 kHz THD, at 100W/8 Ohms and no global NFB, to 0.08%. On a 20-20,000 Hz level, this was 0.245%, and at 50 kHz, THD was 0.39%.
The amp has a cascoded FET front end, buffered second stage and cascodes in the second stage. Output devices are Motorola/ON Semi MJL 3281/1302. Open loop power bandwidth was 94 kHz/8 Ohms.
I probably overcompensated it, meaning there's still a lot of work to be done, but overall, I thought the initial static distortion results were rather good. For a moment, I though almost too good to be true, but upon several bouts of checking and rechecking, I kept getting the same results.
Nevertheless, I thought I might check here as well. Always good to hear other people's experiences.
I am certainly not claiming that the amp will sound great, or anything like that, despite its promising results in basic and not so basic tests, it is after all an ongoing project. I am waiting for the delivery of those trannies now, they should be here by the end of next week.
Anyway, thank you for offering your view, and thanks go to M.Gregg as well.
Oh, BTW, I am well aware that more than its absolute value, the distribution of the harmonics and their THD rates are far more important. But that fact should not prevent anyone answering the basic question.
- Status
- Not open for further replies.