Hi,
@Kevinkr: No need for any apology of any kind. We know and understand each other, don't we?
I'll pass you the details on Fb if you like.
Point is, if there are any members out there with a true passion for valves, for music reproduction and a flair for thinking outside the box, then this is your chance.
The reason no such thing exists is merely for commercial reasons, not performance.
Gentlemen, the floor is all yours,
@Kevinkr: No need for any apology of any kind. We know and understand each other, don't we?
I'll pass you the details on Fb if you like.
Point is, if there are any members out there with a true passion for valves, for music reproduction and a flair for thinking outside the box, then this is your chance.
The reason no such thing exists is merely for commercial reasons, not performance.
Gentlemen, the floor is all yours,
For the 103, the s/n from Johnson noise is about -68.5dB. The SPU 95 gives about -80dB. So...
The difference to copper seems insignificant, fractions of a dB at most. What am I missing?
well expect 6c45p prices to skyrocket. https://www.kickstarter.com/projects/blackie/the-perfect-storm-bass-guitar-preamp
Hi,
We don't care much about that, there are alternatives.
That said, even though I saw some messages from Kevin on FB, I'm not quite sure he wants to go ahead with this.
What I do think is that it can be done but just not for the most idiotically low MC carts out there.
But there I see valves, lots of them. More than my own phono preamp already. With high precision series regs, perhaps valved rectifiers too.
But really, we should start with a well defined design brief for this as I have my ideas on how to tackle this but I do welcome input from others.
Be aware that designing this is hard, really hard. Building it is just not for beginners either as layout is going to dictate performance.
I already envision six series regs and an equal amount of EI xformers. That's would be eighteen valves just for the regs alone. pffft.
Reminds me of J.C Verdier's design...... Sounds fabulous but who can afford that hey?
Anyway, just letting my mind run wild here.
Kevin, you OK?
We don't care much about that, there are alternatives.
That said, even though I saw some messages from Kevin on FB, I'm not quite sure he wants to go ahead with this.
What I do think is that it can be done but just not for the most idiotically low MC carts out there.
But there I see valves, lots of them. More than my own phono preamp already. With high precision series regs, perhaps valved rectifiers too.
But really, we should start with a well defined design brief for this as I have my ideas on how to tackle this but I do welcome input from others.
Be aware that designing this is hard, really hard. Building it is just not for beginners either as layout is going to dictate performance.
I already envision six series regs and an equal amount of EI xformers. That's would be eighteen valves just for the regs alone. pffft.
Reminds me of J.C Verdier's design...... Sounds fabulous but who can afford that hey?
Anyway, just letting my mind run wild here.
Kevin, you OK?
As a teenager I remember seeing the EAR professional 52 preamp in a hifi mag. The 52 referred to the number of tubes in it. TdP never was one for half measures back then!
A tube filled monster would be fun, if only to scratch that teenage itch of wanting one. And the fact that twenty years ago I never managed to get my tube phono stage quiet enough so gave up!
A tube filled monster would be fun, if only to scratch that teenage itch of wanting one. And the fact that twenty years ago I never managed to get my tube phono stage quiet enough so gave up!
Even my simpler designs have not garnered much interest here except for the current crowd, and the general accusation has been that they're; too strange, too complex and too expensive. A few clueless members here have even posted derogatory comments and alleged that I have no idea what I am doing. (Human nature being what it is, this I think is about par for the course here.)
While an assault on the state of the art would be fun I think you have to be prepared for very limited interest. I would also strongly recommend keeping it as simple as possible relative to the overall performance requirements.
Work keeps me very busy, far too busy in fact to really to devote much of my remaining energy to this hobby at this time. I foolishly took on some obligations early this year that I am now having difficulty meeting. I actually live in a state of near perpetual exhaustion due to the aftermath of a recent health issue.
I think realistically speaking you should all proceed without me.
While an assault on the state of the art would be fun I think you have to be prepared for very limited interest. I would also strongly recommend keeping it as simple as possible relative to the overall performance requirements.
Work keeps me very busy, far too busy in fact to really to devote much of my remaining energy to this hobby at this time. I foolishly took on some obligations early this year that I am now having difficulty meeting. I actually live in a state of near perpetual exhaustion due to the aftermath of a recent health issue.
I think realistically speaking you should all proceed without me.
some noise and distortion results for a few 6C45 (6S45)
A friend who is a hollow-state or bust type, having given up on me, designed and built his own MM RIAA preamp with tubes and transformers, chokes in the PS, etc.
He had felt that stepup transformers for MC, if sounding good, were too expensive. He asked me for a stepup stage and I responded with a JFET/Boxall bipolar folded cascode, and he sniffed at it. But now he is trying to get satisfactory results with tubes. He built up a stage with two 6922 and the four triodes in parallel, and has apparently a lot of high frequency noise, possibly due to oscillation.
I took a break from a sand-state design and dug out some 6S45s. A sloppy layout of a single one, initially with no cathode resistor, and a lowish plate resistor allowed operation from a passively-filtered ~80V supply (the guy plans on using a transformer in the plate(s) ).
The environment here is sufficiently noisy that the results may be pessimistic. In order to avoid some hum pickup I filtered the Audio Precision at 400Hz, which of course cuts out a lot of 1/f noise too.
I eventually migrated, for the purposes of measurement without requiring a higher voltage supply, to a very simple LED-biased PNP current source of 10mA, which will as well contribute some noise. But I wanted to get data on the mu and plate resistance better than eyeballing the curves on the Sovtek datasheet.
The results are that the mu is pretty close despite the lowish plate current and voltage to the nominal specification. By the time I was looking at this the guy asked me to put 10 ohms unbypassed in the cathode, so the inference about mu is somewhat subject to error. However, loaded by the Ap input of 100k, the gain for the first sample was times 45.2 (+33.1dB). The input-referred average noise density for a 400Hz - 22kHz filter was 2.3nV/sq rt Hz.
Two more tubes were tested. Both had slightly higher noise, and somewhat higher gain. Referred to the input they were about 2.4nV/sq rt Hz. These measurements were made with a 200 ohm R to common and a 20k R in series from the Ap generator, and the thermal noise of the 200 ohms was root-difference-of-squares removed to get the input-referred voltage noise densities.
As expected the very-lightly-loaded plate resulted in low distortion, with the best number at the crossover point of noise versus distortion down to about 40ppm, virtually all second.
The grid current was about 11uA, which will not please some but I suspect is not too serious. Of course paralleled tubes will make things proportionally worse.
Oh, I also looked at noise with 5mA and 15mA, using the same simple sand I source. In both cases it was worse that 10mA for one tube.
However, the microphonics are most interesting, so high Q that once excited, the residual ringing can be seen for several seconds as it decays into the noise floor.
I opined that for an MC stepup about 1nV/sq rt Hz should be about the maximum tolerable. The 1/f with 22Hz HP made things worse but not terrible. But it's clear that some paralleling is going to be needed---for midband noise maybe 5 tubes, and with the emphasized low frequency noise, a considerably larger number. I told him that to get into JFET territory he might need 100 tubes, but this may be slightly pessimistic.
BTW, solving simultaneous equations for a couple of configurations, I extracted transconductance for the first tube as about 20mA/V, and plate resistance as about 2.3k, this again at 10mA.
Just thought I would share these results before I clear the bench and get back to other pursuits.
A friend who is a hollow-state or bust type, having given up on me, designed and built his own MM RIAA preamp with tubes and transformers, chokes in the PS, etc.
He had felt that stepup transformers for MC, if sounding good, were too expensive. He asked me for a stepup stage and I responded with a JFET/Boxall bipolar folded cascode, and he sniffed at it. But now he is trying to get satisfactory results with tubes. He built up a stage with two 6922 and the four triodes in parallel, and has apparently a lot of high frequency noise, possibly due to oscillation.
I took a break from a sand-state design and dug out some 6S45s. A sloppy layout of a single one, initially with no cathode resistor, and a lowish plate resistor allowed operation from a passively-filtered ~80V supply (the guy plans on using a transformer in the plate(s) ).
The environment here is sufficiently noisy that the results may be pessimistic. In order to avoid some hum pickup I filtered the Audio Precision at 400Hz, which of course cuts out a lot of 1/f noise too.
I eventually migrated, for the purposes of measurement without requiring a higher voltage supply, to a very simple LED-biased PNP current source of 10mA, which will as well contribute some noise. But I wanted to get data on the mu and plate resistance better than eyeballing the curves on the Sovtek datasheet.
The results are that the mu is pretty close despite the lowish plate current and voltage to the nominal specification. By the time I was looking at this the guy asked me to put 10 ohms unbypassed in the cathode, so the inference about mu is somewhat subject to error. However, loaded by the Ap input of 100k, the gain for the first sample was times 45.2 (+33.1dB). The input-referred average noise density for a 400Hz - 22kHz filter was 2.3nV/sq rt Hz.
Two more tubes were tested. Both had slightly higher noise, and somewhat higher gain. Referred to the input they were about 2.4nV/sq rt Hz. These measurements were made with a 200 ohm R to common and a 20k R in series from the Ap generator, and the thermal noise of the 200 ohms was root-difference-of-squares removed to get the input-referred voltage noise densities.
As expected the very-lightly-loaded plate resulted in low distortion, with the best number at the crossover point of noise versus distortion down to about 40ppm, virtually all second.
The grid current was about 11uA, which will not please some but I suspect is not too serious. Of course paralleled tubes will make things proportionally worse.
Oh, I also looked at noise with 5mA and 15mA, using the same simple sand I source. In both cases it was worse that 10mA for one tube.
However, the microphonics are most interesting, so high Q that once excited, the residual ringing can be seen for several seconds as it decays into the noise floor.
I opined that for an MC stepup about 1nV/sq rt Hz should be about the maximum tolerable. The 1/f with 22Hz HP made things worse but not terrible. But it's clear that some paralleling is going to be needed---for midband noise maybe 5 tubes, and with the emphasized low frequency noise, a considerably larger number. I told him that to get into JFET territory he might need 100 tubes, but this may be slightly pessimistic.
BTW, solving simultaneous equations for a couple of configurations, I extracted transconductance for the first tube as about 20mA/V, and plate resistance as about 2.3k, this again at 10mA.
Just thought I would share these results before I clear the bench and get back to other pursuits.
Just because I haven't commented much doesn't mean that I'm not very interested in what you have to say and what you're doing. I doubt I'm the only one in that camp. I encourage you to continue.
When I designed the EO phono pre, I was well aware that the complexity and unfashionable mix of technologies would severely limit the number of people who would try to build it. Nonetheless, I like to think that the articles on it were at least a little bit useful to a few people.
Hi,
@bcarso: Thank you for your contribution. Most of us are aware of the difficulties this kind of design brings along.
And sure enough you can't beat sand devices here in many aspects.
But that's not really the point of the exercise.
The point is (to my mind at least) to create a stage that can cater for MC's and is ultimately musically pleasing by using valves. Well mostly.
Having designed a MC capable add on which isn't too bad (if I may say so) but was designed to a price level and no USSR readily available valves at that time, I think we can do better nowadays. For pete's sake the thing is old, really old.
You can only do so much with valves to reduce their "Eigen Noise" so to speak but it will be valves and some of us like that as we feel it's closer to how we think music should be reproduced.
IOW, to me music trumps specs. Not that I want lousy specs mind you.
@KevinKR: I have the utmost respect for your feelings. Just let me tell you this; I have to fight too. You're the best though.
Could go on for ever...........
Cheers,
@bcarso: Thank you for your contribution. Most of us are aware of the difficulties this kind of design brings along.
And sure enough you can't beat sand devices here in many aspects.
But that's not really the point of the exercise.
The point is (to my mind at least) to create a stage that can cater for MC's and is ultimately musically pleasing by using valves. Well mostly.
Having designed a MC capable add on which isn't too bad (if I may say so) but was designed to a price level and no USSR readily available valves at that time, I think we can do better nowadays. For pete's sake the thing is old, really old.
You can only do so much with valves to reduce their "Eigen Noise" so to speak but it will be valves and some of us like that as we feel it's closer to how we think music should be reproduced.
IOW, to me music trumps specs. Not that I want lousy specs mind you.
@KevinKR: I have the utmost respect for your feelings. Just let me tell you this; I have to fight too. You're the best though.
Could go on for ever...........
Cheers,
I'm not trying to talk anyone out of valves, which I enjoy and respect btw.Hi,
@bcarso: Thank you for your contribution. Most of us are aware of the difficulties this kind of design brings along.
And sure enough you can't beat sand devices here in many aspects.
But that's not really the point of the exercise.
The point is (to my mind at least) to create a stage that can cater for MC's and is ultimately musically pleasing by using valves. Well mostly.
Cheers,
But as I did some bench testing and got pretty decent data, I thought it might be of interest. I found the thread by searching for "6C45" and "noise" after I had built the test setup and done the measurements.
I knew about the microphonics from when I used the tubes as mere cathode followers to buffer and drive a simple highpass, followed by another follower, for an active crossover, but hadn't observed them with higher gain and with the Ap until now. The nice thing about high-Q resonances: it's hard to excite them, relatively speaking.
Brad
Good luck selecting 5 sufficiently matched 6C45s. I would prefer some mild hiss to the sonic dive paralleling brings but of course others have different sonic priorities.
WoW, that seems Very good to me for tubes ! I know it's using a HPF though, but hiss would be nice & lowish, for a MM anyway. I know this about MC's, but that spec would be great for MM's
So it "appears" that below 400Hz is where the noise increases. Filtering out PS nasties shouldn't be a problem ?
Hi ZeroD,
Not power supply nasties actually, but 1/f noise which is significant in all tubes and when you are dealing with signal levels in the range of a few hundred uV @ maximum groove modulation this is going to be a problem.
I had actually hoped the 6C45 would be under 2nVrtHz, and that no more than 4 would be required. Paralleling them and getting good match between tubes would require a pretty large sample size.
The D3A might actually be a little quieter, but this tube is disappearing rapidly it seems. (Perhaps more will be found at some point)
Within narrow limits possibly a slight reduction in filament voltage would help with 1/f noise as would very careful selection of operating point.
My enthusiasm for the 6C45 has waned a bit with Brad's very careful and methodically performed measurements. It seems that it would be well suited to a MM level application where SNR would be quite acceptable.
While I can't hear above 12kHz, I am cursed with a very low threshold on the LF end of the spectrum (actually a benefit ) and the LF noise would be a problem for me and my system which gets down into the low 30s, and has usable output to at least 28Hz. (Not to mention what RIAA EQ does to LF noise)
I appreciate all of the positive comments about my participation and contributions here, it is nice to know who your friends are..
I will see what I can do over time, right now I'm struggling with too many commitments. (One will be gone soon)
Not power supply nasties actually, but 1/f noise which is significant in all tubes and when you are dealing with signal levels in the range of a few hundred uV @ maximum groove modulation this is going to be a problem.
I had actually hoped the 6C45 would be under 2nVrtHz, and that no more than 4 would be required. Paralleling them and getting good match between tubes would require a pretty large sample size.
The D3A might actually be a little quieter, but this tube is disappearing rapidly it seems. (Perhaps more will be found at some point)
Within narrow limits possibly a slight reduction in filament voltage would help with 1/f noise as would very careful selection of operating point.
My enthusiasm for the 6C45 has waned a bit with Brad's very careful and methodically performed measurements. It seems that it would be well suited to a MM level application where SNR would be quite acceptable.
While I can't hear above 12kHz, I am cursed with a very low threshold on the LF end of the spectrum (actually a benefit
) and the LF noise would be a problem for me and my system which gets down into the low 30s, and has usable output to at least 28Hz. (Not to mention what RIAA EQ does to LF noise)I appreciate all of the positive comments about my participation and contributions here, it is nice to know who your friends are..
I will see what I can do over time, right now I'm struggling with too many commitments. (One will be gone soon)