It does seem like a well made valve (tube), my first foray into phono stage design using a pentode front end. I am surprised at the difference in performance going to 2V bias, yes both transconductance and noise dropped. Gain dropped a couple of dB as a result which given the large reduction in LF noise was IMO a worthwhile tradeoff.
WRT the problem I am having with my computer based audio measurement set up I guess the next step is some clamp on ferrites on the audio cable from the PC, and perhaps I will ditch the USB power to the Millet Sound Card Interface, and use an analog supply.
WRT the problem I am having with my computer based audio measurement set up I guess the next step is some clamp on ferrites on the audio cable from the PC, and perhaps I will ditch the USB power to the Millet Sound Card Interface, and use an analog supply.
I too have always kept away from pentodes at low levels so I will be very interested in your final conclusions. It is a mystery to me why anybody nowadays bothers with the usual (expensive) suspects when there are so many other valves like these available cheaply. Using both sections of a 6f12p with the pentode trioded also produces easily enough gain in a RIAA stage-particularly with a bit of semiconductor assistance. That could be the "Muscovite Micro"
It seems like the pentode will work, but I suspect some judicious selection of tubes for matching transconductance may be necessary to get good channel to channel match. Overall variations of a couple dB between matched pairs is not a big deal imo, but between tubes it will be. Not sure whether trimming screen current on one channel would work well - something I will experiment with if necessary...
Waiting for the chassis and RIAA network caps to arrive still. Should get here soon, I'm quite excited to see how it all comes together. Worst case I will end up with something a bit different than expected, but it looks quite promising.
Muscovite Micro might well be next, a single tube per channel could be interesting.
Waiting for the chassis and RIAA network caps to arrive still. Should get here soon, I'm quite excited to see how it all comes together. Worst case I will end up with something a bit different than expected, but it looks quite promising.
Muscovite Micro might well be next, a single tube per channel could be interesting.
Interestingly enough a single clamp on ferrite seems to have solved my low signal level noise issue. I've routed the usb power cable, and the balanced send and return audio cable through the ferrite. The meter on the interface now reads 0mV with nothing connected to the I/O whereas before it always read around 7mV.. Eureka
I can actually set a value of 5mV and see it on the meter, now how accurate it is given the most sensitive range is 200mV remains to be seen.. (Not very I suspect)
I can actually set a value of 5mV and see it on the meter, now how accurate it is given the most sensitive range is 200mV remains to be seen.. (Not very I suspect)
D3A Test
The D3A plugs right in without further changes. The particular sample used ran at a plate current of 10mA and a screen current of 2.8mA versus 11mA and 3.3mA respectively. (Noting of course that this will vary from sample to sample.)
The phono stage gain increased by 2dB.. (Note that the 6ж9п at reduced current provided an output level of about 2.13V, not the 2.45V shown in the caption.)
Interestingly noise performance did not improve at all.
See attachments, first is D3A, second is 6ж9п / 6Z9P. I will stay with the Russian tube in keeping with the theme, keeping costs down for those building one and in fact it costs about 1/6th the current cost of D3A. (I have tons of D3A so cost is not an issue since I do not have to buy any.)
The D3A plugs right in without further changes. The particular sample used ran at a plate current of 10mA and a screen current of 2.8mA versus 11mA and 3.3mA respectively. (Noting of course that this will vary from sample to sample.)
The phono stage gain increased by 2dB.. (Note that the 6ж9п at reduced current provided an output level of about 2.13V, not the 2.45V shown in the caption.)
Interestingly noise performance did not improve at all.
See attachments, first is D3A, second is 6ж9п / 6Z9P. I will stay with the Russian tube in keeping with the theme, keeping costs down for those building one and in fact it costs about 1/6th the current cost of D3A. (I have tons of D3A so cost is not an issue since I do not have to buy any.)
Attachments
Updated Schematic
Here is the latest version of the schematic, changes are quite trivial as compared to the previous one posted. (Just boosted the bias in the second stage a bit.)
Overall power consumption is now down to a bit over 20mA as compared to the original at 30mA.
Gain is about 1.2dB lower with the rebias, note that gain will vary somewhat depending on the particular pair of 6ж9п used.
LF noise is reduced by 10dB compared to the original.
The D3A will slot right in without changes, but results in no measurable improvement in performance.
Here is the latest version of the schematic, changes are quite trivial as compared to the previous one posted. (Just boosted the bias in the second stage a bit.)
Overall power consumption is now down to a bit over 20mA as compared to the original at 30mA.
Gain is about 1.2dB lower with the rebias, note that gain will vary somewhat depending on the particular pair of 6ж9п used.
LF noise is reduced by 10dB compared to the original.
The D3A will slot right in without changes, but results in no measurable improvement in performance.
Attachments
Unfortunately I really don't know how the EF86 compares noise wise to the 6ж9п, but given it was designed specifically for audio it ought to be at least a little quieter.. The operating current and transconductance are quite a bit lower which might make it difficult to use effectively in the simplistic designs I tend to prefer. (Single stage eq, not intentionally part of the plate load, etc)
Part of the reason for my lack of familiarity with the EF86 is that good ones have always been rather expensive here. The Russian knock-offs for the most part have not been terribly good. I used the EF86 in just one design as the error amplifier for a voltage regulator and the performance was a bit disappointing. I normally use cascode 12AX7A (odd choice that works well for me) or in lower voltage applications I used the 6DJ8 in cascode. My only other exposure to them has been in the many English made variants of the Mullard 5-20...
Part of the reason for my lack of familiarity with the EF86 is that good ones have always been rather expensive here. The Russian knock-offs for the most part have not been terribly good. I used the EF86 in just one design as the error amplifier for a voltage regulator and the performance was a bit disappointing. I normally use cascode 12AX7A (odd choice that works well for me) or in lower voltage applications I used the 6DJ8 in cascode. My only other exposure to them has been in the many English made variants of the Mullard 5-20...
For exactly the same reason I have not been using ef86 either but I have heard a fair number of vintage amps using them which were certainly not quiet but without proper analysis there is no reason to blame the input valve specifically.
12ax7 in cascode with resistive load? High voltage for a preamp! Of course, I forget that you are a GM70 fan-much too scary for me.
On the subject of Russian clones, there does not seem to be much correlation in quality between the valves made for the military only and ordinary consumer ones. 6s3p-ev and 6j9p-e are the former and the ef86 clone(I think) the latter.
12ax7 in cascode with resistive load? High voltage for a preamp! Of course, I forget that you are a GM70 fan-much too scary for me.
On the subject of Russian clones, there does not seem to be much correlation in quality between the valves made for the military only and ordinary consumer ones. 6s3p-ev and 6j9p-e are the former and the ef86 clone(I think) the latter.
The fact that those amplifiers were designed with input sensivities of just a few hundred mVrms for rated power output has a lot to do with the noise issues imho. My designs generally are at least 14dB less sensitive than that for this reason.. Hard to know given the high gains just how noisy the EF86 would be in a lower gain implementation. I did pad down the inputs of several pairs of Leak TL50s to deal with this and other issues - noise performance seemed OK if not stellar.
The 12AX7A cascode is used only in my power supplies - I don't even use the 12AX7A in the audio path these days. They work quite well giving me 44 - 48dB of gain typically and work well in supplies delivering up to 550V output. I had just about abandoned this design, but came back and refined it further for the Muscovite as it just seemed to work better overall than some other things I tried. I've using variants of this design since about 1989, and have posted the design on various threads here over the years. My newer pentode design cannot match it for either noise or regulation performance.
Funny you mention the quality issue, I have had very variable luck with certain types, the 6ж52п being a recent example. It didn't suit the application, and was also amazingly inconsistent from sample to sample - looking at the insides provides some clues. Analogs and Western types in current production tend to be pretty mediocre IMHO, I like the old military and certain civilian types for which there are no exact analogs in the West, they tend to be excellent. The GM70 is an example of a very, very good Russian tube which was not a copy of something made on this side of the iron curtain.
The 12AX7A cascode is used only in my power supplies - I don't even use the 12AX7A in the audio path these days. They work quite well giving me 44 - 48dB of gain typically and work well in supplies delivering up to 550V output. I had just about abandoned this design, but came back and refined it further for the Muscovite as it just seemed to work better overall than some other things I tried. I've using variants of this design since about 1989, and have posted the design on various threads here over the years. My newer pentode design cannot match it for either noise or regulation performance.
Funny you mention the quality issue, I have had very variable luck with certain types, the 6ж52п being a recent example. It didn't suit the application, and was also amazingly inconsistent from sample to sample - looking at the insides provides some clues. Analogs and Western types in current production tend to be pretty mediocre IMHO, I like the old military and certain civilian types for which there are no exact analogs in the West, they tend to be excellent. The GM70 is an example of a very, very good Russian tube which was not a copy of something made on this side of the iron curtain.
Sorry Kevin, I hadn't looked at the PSU carefully; nice idea if a bit extreme! What about cascading two? (although I'm one to talk-my variable EQ phono has a separate choke input supply for each stage;I wouldn't use such brute force methods now). I've got lazy and use semiconductors in the PSU's now. Probably I should be using transistors as the amplifying devices in preamps too. I find the engineering/physics arguments overwhelming and cannot disagree but you know what; having been a record collector since I was a boy, I still find the records which were generated by valve equiptment to be much clearer and detailed, not just early stereos but monos and the later 78's. This opinion is always shared by my musician colleagues at listening evenings who wouldn't have any idea how an amplifier works. I realise this is a highly contentious and complex issue and I never get involved in any of the heated discussions particularly as my brain agrees 100% with the scientists.
The 6j52p is the most ridiculously crooked valve I have ever seen. It would be hard to imagine two could ever be the same. It's a pity since these specs should be useful for something; maybe a cathode follower for driving headphones.
The 6j52p is the most ridiculously crooked valve I have ever seen. It would be hard to imagine two could ever be the same. It's a pity since these specs should be useful for something; maybe a cathode follower for driving headphones.
The cascade is less elegant, but obviously can provide a lot more gain if required.. Feedback arrangements can get a little problematic unless you don't mind dissipating some power..
I briefly used a differential error amplifier (LTP) until I realized I was throwing away at least half of the available gain, in a cascade this would work well and solve the issue with feedback, a pentode could get the OLG way up there.. Then you would need a very precise low noise reference which is doable.. Hmmm...
I've got eight 6ж52п and they definitely suffer from QC issues.. Triode connected they might be interesting, I have at least a few that should be usable for something. In pentode connection they need so much plate voltage and current to get them up into the transconductance range that attracted me to them in the first place that they are not really useful. Parametrically speaking they are scattered all over the place.. I suspect they might be OK as a CCS under a long tailed pair running at pretty high currents.
I briefly used a differential error amplifier (LTP) until I realized I was throwing away at least half of the available gain, in a cascade this would work well and solve the issue with feedback, a pentode could get the OLG way up there.. Then you would need a very precise low noise reference which is doable.. Hmmm...
I've got eight 6ж52п and they definitely suffer from QC issues..
Triode connected they might be interesting, I have at least a few that should be usable for something. In pentode connection they need so much plate voltage and current to get them up into the transconductance range that attracted me to them in the first place that they are not really useful. Parametrically speaking they are scattered all over the place.. I suspect they might be OK as a CCS under a long tailed pair running at pretty high currents.
I've got a pile of 6ж52п / 6Z52P (6J52P) which I have measured, unfortunately they bear little resemblance to the D3A parametrically speaking. I bought them originally because I thought they were a cheap alternative to the D3A, they're not - most are basically rubbish.
The 6ж9п / 6Z9P despite the significantly lower rated transconductance is much more similar to the D3A to my surprise, and much better made than the 6ж52п as well.
I'm going to evaluate the 6ж52п in triode and see if they are viable for use in a head amp. (pre-pre)
The 6ж9п / 6Z9P despite the significantly lower rated transconductance is much more similar to the D3A to my surprise, and much better made than the 6ж52п as well.
I'm going to evaluate the 6ж52п in triode and see if they are viable for use in a head amp. (pre-pre)
Yes, I think the 6j52p should be consigned to triode use. They have quite a high power rating for their size-if you believe the spec sheet at all! I doubt the ratings would have specified a total of about 11W if they had been intended for the military; probably just to give the Soviet tv repairman something to do. If you want a cheap high gm pentode then the small power valve 6e6p-e/6e5p is really nicely made or there is the slighly bigger brother of the 6j9p-e, the 6j11p-e.
Kevin, I have been planning to get round to the same test since I want to get my Technics 305MC back into service. I have been using MM for many years now because of the necessity of changing styli but I am going to put a second arm on my Garrard 401 and run another phono pre in parallel (RIAA only,so much simpler)
Sowter's MC transformers are really good-I used them in my father's phono pre-but it would be interesting to see if it was possible to do without and stay thermionic.
Kevin, I have been planning to get round to the same test since I want to get my Technics 305MC back into service. I have been using MM for many years now because of the necessity of changing styli but I am going to put a second arm on my Garrard 401 and run another phono pre in parallel (RIAA only,so much simpler)
Sowter's MC transformers are really good-I used them in my father's phono pre-but it would be interesting to see if it was possible to do without and stay thermionic.
I have a Garrard 401 I will be starting to restore shortly, the Muscovite Mini will be used with that turntable as well as my second TD-124.
My original plan was to use the 6ж52п / 6Z52P in the front end of the mini, but I was not even able to get decent curves on my uTracer with this tube, and my Spice model of it indicated that getting the high transconductance of which it is theoretically capable required very high plate currents and with required load resistance for the target gain a ridiculously (and dangerously) high plate supply voltage as well. It did not look promising.
Still waiting for a bunch of parts from abroad to kick this off, hopefully some of them will arrive this week. I've splurged on Russian teflons for the EQ, with padding done by those oddly good sounding Russian micas. (We'll see if that is really true)
I've got a new power supply design prompted by our discussion yesterday, the pass tube is still EL34, but the error amplifier is a cascade pair of 12AX7A based differential amps, OLG is well over 66dB, and both output impedance and noise should be rather low. I'll share the schematic here soon with the understanding that the simulations its based on may prove to be ridiculously optimistic, perhaps it won't even work. It is interesting enough to investigate at least so I do plan to build and listen to the full fledged mini powered by it. Design wise it looks quite similar to one of the U.S. Govt recommended regulator circuits from the mid 1955s with some changes.. (Only so many ways under the sun to do this - not inspired by that design, just ended up that way.)
My original plan was to use the 6ж52п / 6Z52P in the front end of the mini, but I was not even able to get decent curves on my uTracer with this tube, and my Spice model of it indicated that getting the high transconductance of which it is theoretically capable required very high plate currents and with required load resistance for the target gain a ridiculously (and dangerously) high plate supply voltage as well. It did not look promising.
Still waiting for a bunch of parts from abroad to kick this off, hopefully some of them will arrive this week. I've splurged on Russian teflons for the EQ, with padding done by those oddly good sounding Russian micas. (We'll see if that is really true)
I've got a new power supply design prompted by our discussion yesterday, the pass tube is still EL34, but the error amplifier is a cascade pair of 12AX7A based differential amps, OLG is well over 66dB, and both output impedance and noise should be rather low. I'll share the schematic here soon with the understanding that the simulations its based on may prove to be ridiculously optimistic, perhaps it won't even work. It is interesting enough to investigate at least so I do plan to build and listen to the full fledged mini powered by it. Design wise it looks quite similar to one of the U.S. Govt recommended regulator circuits from the mid 1955s with some changes.. (Only so many ways under the sun to do this - not inspired by that design, just ended up that way.)
My Linn Sondek was consigned to a cupboard after restoring my 401. The most dramatic improvement I managed was in the manner of arm mounting. I have not used super-arms but only SME3009 and an upgraded Rega 300. On top of the heavy wooden plinth I put a 120mm square sandwich of 10mm plywood with thick and dense old style computer mouse mat material. This arm mounting platform seems to decouple the arm from the motor unit very well and provide the right amount of damping for the arm itself. It was just an experiment and I was probably just lucky. The bass is terribly clean and tight for such modest arms.
Russian capacitors; where would we be without them!
You must be one of the very few forum members who regularly build valve regulators. Is it merely aesthetic-which is a perfectly good reason anyway- or have you found that the high frequency performance is better? Because my first build was a mic amp which was occasioned by an emergency(in session failure of commercial amp) my whole system is entirely differential from cartidge to speaker output. When I build a second system for my other speaker it will be SE and I am sure I will not be able to resist valve regulation-if only for the fun of it.
Russian capacitors; where would we be without them!
You must be one of the very few forum members who regularly build valve regulators. Is it merely aesthetic-which is a perfectly good reason anyway- or have you found that the high frequency performance is better? Because my first build was a mic amp which was occasioned by an emergency(in session failure of commercial amp) my whole system is entirely differential from cartidge to speaker output. When I build a second system for my other speaker it will be SE and I am sure I will not be able to resist valve regulation-if only for the fun of it.
Hi Kevin,
For the Muscovite micro. Have you considered 6F12P? I have no experience with them. I bought a few to try out, but have never got around to it. I don't know if they are rubish. But, perhaps one could get away using only one valve?
Thanks for some great threads. My 15 year old son has asked for a record player and I found a old Thorens TD318 for him. I might be tempted to build one of the muscovites for him or even better let him build one with me around when power has to be switched on for the first time
Mogens
For the Muscovite micro. Have you considered 6F12P? I have no experience with them. I bought a few to try out, but have never got around to it. I don't know if they are rubish. But, perhaps one could get away using only one valve?
Thanks for some great threads. My 15 year old son has asked for a record player and I found a old Thorens TD318 for him. I might be tempted to build one of the muscovites for him or even better let him build one with me around when power has to be switched on for the first time
Mogens
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.