I find most tubes sound best with their plate load set at 2.5rp.
For the 6SN7, with 6.5K rp, this is 16K (roughly).
The distortion figure is not so low at this point. But who said the sonics relate to the distortion? Do you actually like your coffee decaffeinated?
You then select a current - 10mA for the 6SN7 is just fine - and cross 16K this is 160V. Double this voltage and that's your B+.
This is not designing for load line/minimum distortion, I grant you. But it sounds marvellous for most tubes, and is an excellent start point in your quest to find the address and phone number of a vacuum tube.
Cheers,
Hugh
For the 6SN7, with 6.5K rp, this is 16K (roughly).
The distortion figure is not so low at this point. But who said the sonics relate to the distortion? Do you actually like your coffee decaffeinated?
You then select a current - 10mA for the 6SN7 is just fine - and cross 16K this is 160V. Double this voltage and that's your B+.
This is not designing for load line/minimum distortion, I grant you. But it sounds marvellous for most tubes, and is an excellent start point in your quest to find the address and phone number of a vacuum tube.
Cheers,
Hugh
Hugh,
That is a completely subjective evaluation, and I would find it hard to defend any argument that starts with "Distortion sounds good...".
The simple fact is that our output stage, output transformers, and speakers themselves will generate all the distortion you'd ever "want". We don't need to advocate deliberately designing voltage amplifiers with poor response.
4*Rp or 5*Rp would be a minimum in a decent RC coupled stage.
Joel
That is a completely subjective evaluation, and I would find it hard to defend any argument that starts with "Distortion sounds good...".
The simple fact is that our output stage, output transformers, and speakers themselves will generate all the distortion you'd ever "want". We don't need to advocate deliberately designing voltage amplifiers with poor response.
4*Rp or 5*Rp would be a minimum in a decent RC coupled stage.
Joel
Joel,
You are an engineer perhaps? You are very competent at math? You have a very intolerant attitude towards people who don't even bother to reduce measured distortion in their designs?
Furthermore, you don't actually do this stuff for a living, do you? And if you do, you design by scope and meter?
Now, you made a very categorical statement. 4 to 5 times rp. Hmmm. A 12AX7 has a plate impedance of 62K5, from memory. That would mean a plate load of at least 250K, and preferably 312K. Of course, at 4rp, the tube will deliver a gain of 4/5 x mu, which is around 80, not bad from a gain standpoint.
How do account for plate loads commonly used with this tube in successful, commercial designs, around 100-150K? Certainly there are designs which use 270K (Bruce Rozenblit's OTL for one) but most do not, and NEVER in the guitary amplifier world, where sonics are everything......
Look, you have your opinion and I have mine. You comment this is a completely subjective evaluation. I agree unequivocally. But seriously, my friend, is there any other type of evaluation? Why do we like music? Is the reason not purely subjective? Is it not worthwhile trying to be good natured about refuting what you see to be contemptuous, misguided and ignorant attitudes? In the years to come you might even change your mind, people commonly do, and it's a sign of growth, not loss of face....
Frank, thank you for your resounding affirmation. I arrived at this conclusion after serving an apprenticeship in the music industry with tube amps and after building and listening to many tube amps. In one of my designs I use a 6SL7 - rp of 44K - with a 100K plate load. B+ is 340V. This was the best sounding combination I ever found for this tube, and it closely matches the Cary 35W PP amp of about five years back.
Cheers,
Hugh
You are an engineer perhaps? You are very competent at math? You have a very intolerant attitude towards people who don't even bother to reduce measured distortion in their designs?
Furthermore, you don't actually do this stuff for a living, do you? And if you do, you design by scope and meter?
Now, you made a very categorical statement. 4 to 5 times rp. Hmmm. A 12AX7 has a plate impedance of 62K5, from memory. That would mean a plate load of at least 250K, and preferably 312K. Of course, at 4rp, the tube will deliver a gain of 4/5 x mu, which is around 80, not bad from a gain standpoint.
How do account for plate loads commonly used with this tube in successful, commercial designs, around 100-150K? Certainly there are designs which use 270K (Bruce Rozenblit's OTL for one) but most do not, and NEVER in the guitary amplifier world, where sonics are everything......
Look, you have your opinion and I have mine. You comment this is a completely subjective evaluation. I agree unequivocally. But seriously, my friend, is there any other type of evaluation? Why do we like music? Is the reason not purely subjective? Is it not worthwhile trying to be good natured about refuting what you see to be contemptuous, misguided and ignorant attitudes? In the years to come you might even change your mind, people commonly do, and it's a sign of growth, not loss of face....
Frank, thank you for your resounding affirmation. I arrived at this conclusion after serving an apprenticeship in the music industry with tube amps and after building and listening to many tube amps. In one of my designs I use a 6SL7 - rp of 44K - with a 100K plate load. B+ is 340V. This was the best sounding combination I ever found for this tube, and it closely matches the Cary 35W PP amp of about five years back.
Cheers,
Hugh
AKSA said:
Hi Hugh.
I'm not an engineer, and no I don't make and sell amps for a living, which apparently will be all the justification you need to ignore me. And yes, I do "design by scope and meter", among other things - don't you? My ability to guess at voltages by grabbing the wire with my hands is very poor, and I'm also not very good at hearing the ring on a 100Hz square wave. Maybe you fare better.
Yes.
Under what level of NFB? And what you define as "successful" may differ greatly from my idea. Also, I would not bring guitar amps into this - they are not relevant - with a response of only 100Hz-6kHz you don't need to be very close, or even "right" to make one work.
Did I use those terms? Did I say I felt that way? Nope.
This is not a philosophical debate Hugh. There is no mystery about what distortion is, its negative effects, how to measure it, and how to avoid it. 2Rp is needlessly low and will introduce distortion.
If you like that, then you'll be happy, and I'm happy for you!Cheers
Joel
ps. your average guitar speaker is so poor in low frequency response that in the way old days when they had 25Hz line frequencies (and other odd ones too I'm sure), they needed next to no PSU filtering because the speaker itself couldn't reproduce the 50Hz hum! That's why my old '37 Gibson had about 24uF of capacitance in the whole amp.
> ...My CDP puts out 6V p-p...
You've measured that?
The standard home hi-fi CD player puts out (for digital-max signal) 2VRMS, 2.8V peak, or 5.6V peak to peak. This is a nice compromise between a high enough level to drown the noise of any output buffer and the desire to stay inside +/-5VDC in the mixed digital/analog parts. Small differences in buffer design may give 20% more or less, so 5 or 6 volts p-p is reasonable.
It also meshes with the much older "-10dB" standard where "0VU" was about 0.3VRMS and peaks were about 16dB higher. Nearly all my post-1975 cassette decks output 0.316V at 0VU (I think Dolby-chip reference design had some to do with that). The VU meter being about 16dB low on very-short peaks, they put out 2VRMS for a few cycles per song.
Walkman-type CD players put out less (limited to less than 1V RMS at max signal by 3V batteries).
I usually like to see excess gain in the chain, would not design so the ab-max output of the source barely drives the power amp to full power. But I rarely use the excess gain so it is not an unreasonable design.
You've measured that?
The standard home hi-fi CD player puts out (for digital-max signal) 2VRMS, 2.8V peak, or 5.6V peak to peak. This is a nice compromise between a high enough level to drown the noise of any output buffer and the desire to stay inside +/-5VDC in the mixed digital/analog parts. Small differences in buffer design may give 20% more or less, so 5 or 6 volts p-p is reasonable.
It also meshes with the much older "-10dB" standard where "0VU" was about 0.3VRMS and peaks were about 16dB higher. Nearly all my post-1975 cassette decks output 0.316V at 0VU (I think Dolby-chip reference design had some to do with that). The VU meter being about 16dB low on very-short peaks, they put out 2VRMS for a few cycles per song.
Walkman-type CD players put out less (limited to less than 1V RMS at max signal by 3V batteries).
I usually like to see excess gain in the chain, would not design so the ab-max output of the source barely drives the power amp to full power. But I rarely use the excess gain so it is not an unreasonable design.
Joel,
I cannot vehemently disagree with your comments. However, our views are very different, so it's probably better to agree to disagree, since a wise man knows what he says, and a fool says what he knows.
No, with posts as pungent as yours, it will not be possible to ignore them! Your directness is arresting, to say the least....
Yes, I USE a scope and meter constantly, although I don't DESIGN with them. I tried gauging voltage a few times with my fingers, but the experiment did not bode well so I bought a Fluke DMM.
And yes, I do carefully study the 100Hz square wave on my CRO, in fact at much higher frequencies as well. And I do X-Y traces to determine phase shift, I find this an excellent indicator of how something sounds.
But I do feel quite strongly that the distortion spectrum is as important as the THD. If we have two amps with 0.05% THD, and one has only H2 and H3 and the other a whole range extending to say H8, then the H2/H3 amp will always sound better. It is certainly true that a heavily plate loaded triode will induce higher distortion, but most of it will be H2 and H3, and I would suggest that this can actually mask the bad sounds of a a distortion spectrum angled towards the higher harmonics.
That said, tubes by and large do not produce higher harmonics when there is no nfb. They might produce a little H5, not much beyond it, unlike their SS brethren. I tend to design for hybrid circuits, so I'm always trying to mask the higher harmonics produced by the nfb SS circuits. This might better explain my arguments.
Cheers,
Hugh
I cannot vehemently disagree with your comments. However, our views are very different, so it's probably better to agree to disagree, since a wise man knows what he says, and a fool says what he knows.
No, with posts as pungent as yours, it will not be possible to ignore them! Your directness is arresting, to say the least....
Yes, I USE a scope and meter constantly, although I don't DESIGN with them. I tried gauging voltage a few times with my fingers, but the experiment did not bode well so I bought a Fluke DMM.
And yes, I do carefully study the 100Hz square wave on my CRO, in fact at much higher frequencies as well. And I do X-Y traces to determine phase shift, I find this an excellent indicator of how something sounds.
But I do feel quite strongly that the distortion spectrum is as important as the THD. If we have two amps with 0.05% THD, and one has only H2 and H3 and the other a whole range extending to say H8, then the H2/H3 amp will always sound better. It is certainly true that a heavily plate loaded triode will induce higher distortion, but most of it will be H2 and H3, and I would suggest that this can actually mask the bad sounds of a a distortion spectrum angled towards the higher harmonics.
That said, tubes by and large do not produce higher harmonics when there is no nfb. They might produce a little H5, not much beyond it, unlike their SS brethren. I tend to design for hybrid circuits, so I'm always trying to mask the higher harmonics produced by the nfb SS circuits. This might better explain my arguments.
Cheers,
Hugh
Nick,
Yes, as you say.
What you look for is two coincident straight lines at roughly 45 degrees, arranged for the same scale where possible.
As you sweep through the frequency range up to about half a meg you should see the two coincident lines diverge slightly into a very thin ellipse, and then start rotating bodily as phase shift really kicks in.
You want the coincident lines to remain one to at least 40KHz. This is a potent test of the audio qualities, best I've seen.
Cheers,
Hugh
Yes, as you say.
What you look for is two coincident straight lines at roughly 45 degrees, arranged for the same scale where possible.
As you sweep through the frequency range up to about half a meg you should see the two coincident lines diverge slightly into a very thin ellipse, and then start rotating bodily as phase shift really kicks in.
You want the coincident lines to remain one to at least 40KHz. This is a potent test of the audio qualities, best I've seen.
Cheers,
Hugh
I got those Hammond 156C chokes in the mail today (it took way longer than expected) and a couple of hours after opening the package the amp was up and running with its new driver stage
Havent´had the time to listen so much yet, but the first impression was good. Nothing missing in the frequency extremes as far as I could hear. I´ll have a closer look at the frequency responce as soon as I get a sinewave generator.
Joel, I´m on your side in that "cheap choke war"
Havent´had the time to listen so much yet, but the first impression was good. Nothing missing in the frequency extremes as far as I could hear. I´ll have a closer look at the frequency responce as soon as I get a sinewave generator.
Joel, I´m on your side in that "cheap choke war"
More X-Y questions - what do you connect to the 2 scope probes? The input to the amplifier is X, and the output from the amplifier is Y? And then adjust gain/attenuation on the scope so that the slope is 45 degrees? And what signal do you feed the amp when you make the sweep - does it matter if it's a sine wave or a square wave?
Thanks,
Saurav
Thanks,
Saurav
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.