Thanks Merlin..
Piano3, yes both of my plinths are slate, complete with turntable and arm weigh about 100lbs or 45kgs.. There are 4 layers of 2cm each, the two top cut outs are the standard TD-124 pattern while the lower two have more material left in place to increase mass. I run them layer on layer, but I could at some point go to a constrained layer setup with bamboo, cork, hard rubber or sealed lead. The plinths sit on large hard rubber corks which work quite well as feet for a fraction of the cost for anything intended for the duty.
I have installed one of the ECC84 pairs MKC sent me a few days ago, warming up now.. Will be interesting to listen to this tube - in theory it should provide a bit more gain.. In practice we shall see. Measurements to follow at some point to determine whether they differ significantly in practice from the 6N14P.
Piano3, yes both of my plinths are slate, complete with turntable and arm weigh about 100lbs or 45kgs.. There are 4 layers of 2cm each, the two top cut outs are the standard TD-124 pattern while the lower two have more material left in place to increase mass. I run them layer on layer, but I could at some point go to a constrained layer setup with bamboo, cork, hard rubber or sealed lead. The plinths sit on large hard rubber corks which work quite well as feet for a fraction of the cost for anything intended for the duty.
I have installed one of the ECC84 pairs MKC sent me a few days ago, warming up now.. Will be interesting to listen to this tube - in theory it should provide a bit more gain.. In practice we shall see. Measurements to follow at some point to determine whether they differ significantly in practice from the 6N14P.
Hi,
We're waiting.....................
Just kidding. Those are the Philips we spotted in the background of one of the pics further up the thread then?
Regarding resistors: When the old Holco's were still easily available I once replaced all the resistors of a phono with these. Didn't like it at all.
They sounded much the same way they looked. Dark, subdued.
IOW, the ideal anti-digital glare resistor.
The Roederstein Resista series was very good and so is Beyschlag. That is until you hear a bulkfoil or a naked bulkfoil. Those are the closest to no resistor I can think of.
Anyways, back to the phono-pre....
Ciao,
We're waiting.....................
Just kidding. Those are the Philips we spotted in the background of one of the pics further up the thread then?
Regarding resistors: When the old Holco's were still easily available I once replaced all the resistors of a phono with these. Didn't like it at all.
They sounded much the same way they looked. Dark, subdued.
IOW, the ideal anti-digital glare resistor.
The Roederstein Resista series was very good and so is Beyschlag. That is until you hear a bulkfoil or a naked bulkfoil. Those are the closest to no resistor I can think of.
Anyways, back to the phono-pre....
Ciao,
Hi Frank,
Interesting comments on the resistors, I like the old Holco a lot, and find the Beyschlag and Roederstein to be bright. Oddly enough we agree on the bulk foil and nude bulk foils despite the disagreement on the others. I use horn mids and highs, but I used Magneplanar QR 1.7 before making the switch to ancient technology in the early 2000s. (How does this play to my preferences?)
I can comment on the ECC84, they're not bad at all, viable, but I have the impression having just switched back to the 6N14P that the Russian tube actually does better with low level/back ground details, and things like high hats and the like have a little more shimmer. My other impression is that it is more dynamic sounding. It may be as much a matter of preference.
I have complicated matters by changing arms although I have never made more than one change at a time. I have had some difficulty dialing in the new clone 3012**. I think overall things are improving, but there are a few warts here and there that I can hear - slowly knocking them down, so far nothing demonstrably related to the phono stage design, but you never know. There is some HF hash I am trying to figure out - think it is worse since I changed the arm. Geometry looks pretty close to spot on at this point. I will say there is a good deal more detail, and that while this set up still lags behind the other table and the Muscovite, the gap is smaller now.
(**Issues with VTA, azimuth, geometry.. Pretty much anything worth thinking about. Things are looking closer to optimum and of course the improvements are audible. This is the first arm I have built so I am learning to take nothing for granted. Even had problems with the interconnects from the arm to transformers - 4th iteration is sounding about right. I guess it would not hurt to mention how tough that SPU is, I'm having a lot of klutz moments which if I did not know the cause would be worrisome.)
Interesting comments on the resistors, I like the old Holco a lot, and find the Beyschlag and Roederstein to be bright. Oddly enough we agree on the bulk foil and nude bulk foils despite the disagreement on the others. I use horn mids and highs, but I used Magneplanar QR 1.7 before making the switch to ancient technology in the early 2000s. (How does this play to my preferences?)
I can comment on the ECC84, they're not bad at all, viable, but I have the impression having just switched back to the 6N14P that the Russian tube actually does better with low level/back ground details, and things like high hats and the like have a little more shimmer. My other impression is that it is more dynamic sounding. It may be as much a matter of preference.
I have complicated matters by changing arms although I have never made more than one change at a time. I have had some difficulty dialing in the new clone 3012**. I think overall things are improving, but there are a few warts here and there that I can hear - slowly knocking them down, so far nothing demonstrably related to the phono stage design, but you never know. There is some HF hash I am trying to figure out - think it is worse since I changed the arm. Geometry looks pretty close to spot on at this point. I will say there is a good deal more detail, and that while this set up still lags behind the other table and the Muscovite, the gap is smaller now.
(**Issues with VTA, azimuth, geometry.. Pretty much anything worth thinking about. Things are looking closer to optimum and of course the improvements are audible. This is the first arm I have built so I am learning to take nothing for granted. Even had problems with the interconnects from the arm to transformers - 4th iteration is sounding about right. I guess it would not hurt to mention how tough that SPU is, I'm having a lot of klutz moments which if I did not know the cause would be worrisome.)
Attachments
Hi Kevin,
Well, I think your observation re Resista and Beyschlag is correct.
I suspect their brightness was just a welcome treat within the context of the overall balance of that particular preamp.
Whereas the Holcos just seemed to emphasize its character even more.
The reason I mentioned it is that if one only uses a single brand/type of resistor then, given a sufficiently transparent design to begin with, its sonic fingerprint becomes more clear.
Fortunately it doesn't take a handful of bulkfoils for it to become obvious how other resistors can colour the sound.
Same goes for caps etc.
As an anecdote, and knowing you also use tubed regulated supplies, one wouldn't expect these small value filmcaps used in the cascode error amps to be audible yet they are and so are all the tubes used there.
Heck, even the rectifier valve is audible IME.
Back to the ECC84s, the tube rolling seems to turn out a little disappointing then. Maybe the Philips need a longer period to run in, I don't know.
Generally though, over here in Europe people prefer the local NOS ones over the NOS ex-Soviet ones.
Nonetheless it seems that the new Muscovite is a success. Truth be told I had my doubt given the low gain of the valves.
Ciao,
Well, I think your observation re Resista and Beyschlag is correct.
I suspect their brightness was just a welcome treat within the context of the overall balance of that particular preamp.
Whereas the Holcos just seemed to emphasize its character even more.
The reason I mentioned it is that if one only uses a single brand/type of resistor then, given a sufficiently transparent design to begin with, its sonic fingerprint becomes more clear.
Fortunately it doesn't take a handful of bulkfoils for it to become obvious how other resistors can colour the sound.
Same goes for caps etc.
As an anecdote, and knowing you also use tubed regulated supplies, one wouldn't expect these small value filmcaps used in the cascode error amps to be audible yet they are and so are all the tubes used there.
Heck, even the rectifier valve is audible IME.
Back to the ECC84s, the tube rolling seems to turn out a little disappointing then. Maybe the Philips need a longer period to run in, I don't know.
Generally though, over here in Europe people prefer the local NOS ones over the NOS ex-Soviet ones.
Nonetheless it seems that the new Muscovite is a success. Truth be told I had my doubt given the low gain of the valves.
Ciao,
I think it's purely psychological effect. As that German saying goes: "Was kostet nix, ist nix" (What costs nothing, is nothing). As long as old Soviet stock is available for 'peanuts', people will consider those tubes as 'second class'. Wait until the supplies dry out, then you'll see.
Back to 6N14P..
Keep rollin' kevinkr! If you need any assistance with Russian (stuff or language), let me know.
Not sure what the difference is here between the ECC84 and 6N14P and we are talking about pretty subtle things I think. The 6N14P has significantly different construction details, plate structure is completely different, the grid wire is gold plated and the cathodes are shaped differently (rectangular and larger), shield structure and some other things appear different. Clearly a functional equivalent, but close in construction quality to the Soviet military surplus.
Generally native and Western European NOS has been preferred to current production over here as well. Very few are familiar with the really good Russian military surplus types and a few domestic types here and there for which there are sometimes analogs here in the west. It came to me as a big surprise just how good the GM70, 6S3P-EV, and 6S4P-EV are for example. I've not heard much better from any device, and since those discoveries I have been looking for others. It is also only in recent years tubes like the D3A, C3G, E810F/7788 not to mention others have caught the attention of people over here. None of these would have been found in hifi back in the day - too expensive, and at least over here unknown.
Gain for this cascode with a plate current of 10mA and a 12K load is 36dB, a 6DJ8 would give you about five dB of additional gain and the equivalent noise resistance is half - not that significant actually. (300 ohms vs 600 ohms) It's still more than quiet enough that the dcr of my SUTs and cartridge are the dominant broadband noise source at the input. Much higher gains are possible with elevated plate voltages or somewhat lower operating currents - in both cases obviously with higher load resistances. I will be looking at other operating points with a view to extracting more gain linearly from this device.
I try to use a variety of different parts where I feel they will do the best job. Caps generally are the most problematic passives as you can see from my struggles with the Muscovite Mini. I do have a mix of different brands of metal films in the design and changed the cascode load resistors to low inductance precision wire wound resistors. Now I'm not 100% sure about them, but so far I have kept them - there seems to be a slightly glassiness at times that was not there before, this is what I have associated with these in critical circuit locations before..
Shifts in operating points in both stages mean that I can get away with 2W resistors where 4W was required before. This gives me more flexibility in choices. Regardless they will run warm...
Pretty much it seems when you reach a certain point every component choice you make has an impact. I don't think anyone following this thread would be surprised to hear me say this. This viewpoint is by no means universal, and working in ATE for years taught me that passives have a lot of imperfections some of which might be relevant to audio designs.
Generally native and Western European NOS has been preferred to current production over here as well. Very few are familiar with the really good Russian military surplus types and a few domestic types here and there for which there are sometimes analogs here in the west. It came to me as a big surprise just how good the GM70, 6S3P-EV, and 6S4P-EV are for example. I've not heard much better from any device, and since those discoveries I have been looking for others. It is also only in recent years tubes like the D3A, C3G, E810F/7788 not to mention others have caught the attention of people over here. None of these would have been found in hifi back in the day - too expensive, and at least over here unknown.
Gain for this cascode with a plate current of 10mA and a 12K load is 36dB, a 6DJ8 would give you about five dB of additional gain and the equivalent noise resistance is half - not that significant actually. (300 ohms vs 600 ohms) It's still more than quiet enough that the dcr of my SUTs and cartridge are the dominant broadband noise source at the input. Much higher gains are possible with elevated plate voltages or somewhat lower operating currents - in both cases obviously with higher load resistances. I will be looking at other operating points with a view to extracting more gain linearly from this device.
I try to use a variety of different parts where I feel they will do the best job. Caps generally are the most problematic passives as you can see from my struggles with the Muscovite Mini. I do have a mix of different brands of metal films in the design and changed the cascode load resistors to low inductance precision wire wound resistors. Now I'm not 100% sure about them, but so far I have kept them - there seems to be a slightly glassiness at times that was not there before, this is what I have associated with these in critical circuit locations before..
Shifts in operating points in both stages mean that I can get away with 2W resistors where 4W was required before. This gives me more flexibility in choices. Regardless they will run warm...
Pretty much it seems when you reach a certain point every component choice you make has an impact. I don't think anyone following this thread would be surprised to hear me say this. This viewpoint is by no means universal, and working in ATE for years taught me that passives have a lot of imperfections some of which might be relevant to audio designs.
For those who don't know it the power supply and filament supply for the Muscovite Mini II are exactly the same as I used in the Muscovite Mini. I recommend a stereo supply as I implemented.
Here is a revised schematic showing what I did in this supply with the exception that I have two separate filament windings for the 6W6GT pass tubes.
Note that only a single power transformer and rectifier is required and two is probably beyond the realm of reason for this design.
The filament regulator is one of the dc regulators advertised on eBay as having a 40uV output noise performance. Probably not that good, but it is good - well under 1mV - I measured 200uVpp on one of mine. This is quiet enough I think. I used a 50VA 9V toroid and a conventional bridge rectifier into a 22000uF 25V electrolytic and directly into the regulator. The supply is floating, but not elevated. There is a pair of 330 ohm 1/2W resistors to audio ground (potential divider).
Here is a revised schematic showing what I did in this supply with the exception that I have two separate filament windings for the 6W6GT pass tubes.
Note that only a single power transformer and rectifier is required and two is probably beyond the realm of reason for this design.
The filament regulator is one of the dc regulators advertised on eBay as having a 40uV output noise performance. Probably not that good, but it is good - well under 1mV - I measured 200uVpp on one of mine. This is quiet enough I think. I used a 50VA 9V toroid and a conventional bridge rectifier into a 22000uF 25V electrolytic and directly into the regulator. The supply is floating, but not elevated. There is a pair of 330 ohm 1/2W resistors to audio ground (potential divider).
Attachments
My dabbling with Mills MR200 is at an end, I have endured several weeks of some fairly nasty sound which was strongly confirmed to be phono pre dependent tonight when I swapped the Mini II for the Muscovite with the better TT and ended up hearing pretty much what I was hearing in the other set up. I would describe this as hard, nasal, fuzzy and murky sounding all at the same time - sorry for the subjectivity here, but that is the best I can come up with. (I was wondering if I had somehow damaged my Meister Silver cartridge or was having an arm problem - hence the swap out. Happy to report the pre is completely to blame.)
For the moment the Russian Micas and Teflon EQ caps shall remain, but I am thinking that conventional polystyrene or polypropylene may prove to be a better combination and will try some soon. I'm not finding most Russian parts to be my cup of tea, this seems to mostly not apply to the tubes however.
I am now using 40yr old 13K 1% 2W TRW glazed metal film milspec RN75 resistors in place of the Mills and the things sound a lot more right. I've lugged piles of these TRW resistors from place to place for 30yrs now, guess I will never run out.. lol
Compared to the Muscovite this design is quite a bit less dynamic in a direct comparison.
For the moment the Russian Micas and Teflon EQ caps shall remain, but I am thinking that conventional polystyrene or polypropylene may prove to be a better combination and will try some soon. I'm not finding most Russian parts to be my cup of tea, this seems to mostly not apply to the tubes however.
I am now using 40yr old 13K 1% 2W TRW glazed metal film milspec RN75 resistors in place of the Mills and the things sound a lot more right. I've lugged piles of these TRW resistors from place to place for 30yrs now, guess I will never run out.. lol
Compared to the Muscovite this design is quite a bit less dynamic in a direct comparison.
Attachments
I removed the 1000pF micas across the 0.01uF and replaced them with a series pair of 2200pF 5% teflons. This resulted in a significant improvement in sound quality and the RIAA EQ looks pretty good with selected parts.
You can change R3 from 12.1K to 13.0K with a very slight increase in gain, if you do so R6 should be 200K + 3.01K - 3.32K in series. This could net you better than +/- 0.1dB RIAA with very carefully selected parts.
You can change R3 from 12.1K to 13.0K with a very slight increase in gain, if you do so R6 should be 200K + 3.01K - 3.32K in series. This could net you better than +/- 0.1dB RIAA with very carefully selected parts.
Attachments
Hi Frank,
200K plus a small trim resistance is the correct value for the network as implemented. The reason I went with a relatively high network impedance is that the load impedance reflected back to the first stage is frequency dependent, and given the high source impedance particularly with the original pentode based front end measurable gain modulation occurs in the first stage, which just serves to make the EQ a little less accurate. The noise hit of a couple of dB is not the limiting factor in circuit performance. This is actually a very quiet pre-amp. (measured and subjectively)
Hi Piano,
There were no measurable effects in the frequency domain, I think the issue probably revolves around the linearity of mica dielectric at audio frequencies. I have spent 4 decades avoiding mica caps in the audio path because I heard a distinct sonic signature wherever I used them. There was so much hype around these Russian micas that I felt I had to try them. Next time I do linearity experiments I will look for changes in spectral content, and I may make another attempt at a pentode front end at some point based on the specific findings from those measurements.
This project (and the predecessor upon which it was based) was largely an exercise in rebuilding manual dexterity and strength, an opportunity for some fun and learning were also intended to be part of the adventure. I probably choose too many new things to try all at one time, and interestingly as I peel away the experiments and go back to the things I found worked in the past the results have become more consistent with my (high) expectations. I admit I spent a lot of time looking at passives some 25 yrs ago and developed some preferences particularly in the area of film capacitors based on measurement and listening tests. (Some blind)
The big question of course is that if you don't try new things from time to time and assume that what you knew in the past remains true in perpetuity you can't grow. I most particularly wanted to increase my exposure and try new things. Experimenting with high transconductance tubes, cascode and pentode based front end topologies, Russian tubes, and some Russian made capacitors has paid off in new knowledge and new ways of solving specific problems. It has also taught me that some of my long ago work is still a good foundation for what I am doing today.
200K plus a small trim resistance is the correct value for the network as implemented. The reason I went with a relatively high network impedance is that the load impedance reflected back to the first stage is frequency dependent, and given the high source impedance particularly with the original pentode based front end measurable gain modulation occurs in the first stage, which just serves to make the EQ a little less accurate. The noise hit of a couple of dB is not the limiting factor in circuit performance. This is actually a very quiet pre-amp. (measured and subjectively)
Hi Piano,
There were no measurable effects in the frequency domain, I think the issue probably revolves around the linearity of mica dielectric at audio frequencies. I have spent 4 decades avoiding mica caps in the audio path because I heard a distinct sonic signature wherever I used them. There was so much hype around these Russian micas that I felt I had to try them. Next time I do linearity experiments I will look for changes in spectral content, and I may make another attempt at a pentode front end at some point based on the specific findings from those measurements.
This project (and the predecessor upon which it was based) was largely an exercise in rebuilding manual dexterity and strength, an opportunity for some fun and learning were also intended to be part of the adventure. I probably choose too many new things to try all at one time, and interestingly as I peel away the experiments and go back to the things I found worked in the past the results have become more consistent with my (high) expectations. I admit I spent a lot of time looking at passives some 25 yrs ago and developed some preferences particularly in the area of film capacitors based on measurement and listening tests. (Some blind)
The big question of course is that if you don't try new things from time to time and assume that what you knew in the past remains true in perpetuity you can't grow. I most particularly wanted to increase my exposure and try new things. Experimenting with high transconductance tubes, cascode and pentode based front end topologies, Russian tubes, and some Russian made capacitors has paid off in new knowledge and new ways of solving specific problems. It has also taught me that some of my long ago work is still a good foundation for what I am doing today.
Hi Kevin,
My mistake. I went back to the schematic on the first page. Since then the component numbering has changed.
IIRC it was Thorsten Loesch who praised the silver micas in his preamp some years back.
Maybe he had some extraordinary ones, maybe not.
To me there are very few components that do not have a sonic character.
Unfortunately it is very, very time consuming to find out what's doing what not to mention why...
So, from your experiments some questions arise which I hope you can answer.
For starters, is the cascode stage an advantage in this position?
What about these high transconductance valves? Are they to be preferred over the classic high mu ones a la ECC83?
How about noise? Does the cascade + high transconductance really help overall resolution (lower noise floor)?
After you answered those you know I have another 999 qq lined up in the pipe line.....
Anyway, thanks,
My mistake. I went back to the schematic on the first page. Since then the component numbering has changed.
IIRC it was Thorsten Loesch who praised the silver micas in his preamp some years back.
Maybe he had some extraordinary ones, maybe not.
To me there are very few components that do not have a sonic character.
Unfortunately it is very, very time consuming to find out what's doing what not to mention why...
So, from your experiments some questions arise which I hope you can answer.
For starters, is the cascode stage an advantage in this position?
What about these high transconductance valves? Are they to be preferred over the classic high mu ones a la ECC83?
How about noise? Does the cascade + high transconductance really help overall resolution (lower noise floor)?
After you answered those you know I have another 999 qq lined up in the pipe line.....
Anyway, thanks,
Hi Kevin, I actually tried to arrange some capacitor linearity experiments on a quiet day at an electrodynamics experimental course I was attending (having brought all kinds of specimens with me!) but unfortunately could not generate any enthusiasm amongst the lecturers/lab technicians. I would be very interested to know of any results that you might generate at these kind of signal levels.
Hi Frank,
Some of my design experiments are just curiosity expressed and desire to not use the same solution to a given problem over and over.
For many years I just paralleled up 12AX7 to get transconductance up and rp down, one day it dawned on me that there were probably more modern triodes or pentodes (triode connected) that would give me the same or better performance but with considerably less miller capacitance, and much less mechanical complexity. My first effort was with the D3A and for fun I used gyrator loading with that in an effort to get the gain reasonably close to mu.
The D3A triode connected has the same transconductance as approximately 30 12AX7 in parallel and a rather low plate resistance as well. Miller capacitance is equivalent to a few 12AX7 in parallel. Noise performance may be 20dB or more better than the 12AX7. Mu is typically around 80 so slightly greater than the 5751 I generally used in lieu of the 12AX7.. The D3A triode connected is amazingly linear as well.
My move to cascode input topology resulted from my observation that even my MC cartridges and their transformers seemed to interact adversely with the high input capacitance of my triode connected D3A phono stage. I confirmed this by designing a pre-amplifier with low input capacitance, (the Muscovite) and also using a solid state pre-pre designed by a friend to isolate the input capacitance of the phono pre from the cartridge and SUT. In both cases the improvement in performance was noticeable in the upper octaves above a kHz or so.
As an added benefit these low rp high transconductance tubes can generally drive lower impedance passive RIAA networks effectively which picks up a few dB of additional SNR. Generally a single or several 12AX7 will have to drive a significantly higher impedance RIAA network to minimize load impedance generated response errors. Some tweaking of cap values is possible in lower impedance networks to compensate for this, but in so doing you become somewhat more vulnerable to parametric variations tube to tube.
The pentode with its very high rp is even more vulnerable to this issue than the cascode which generally has a substantially lower rp (although much higher than the triodes upon which it is based)
In my last three designs the dominant noise source has been the noise generated by the cartridge/SUT combination rather than the pre-amp itself.
There are so many ways to skin a cat and of course this is only the latest route I have followed. I'm certain that other approaches can achieve equally good results with equal care. All I can say is I'm getting better at this and my recent designs represent real progress compared to the older ones.
Some of my design experiments are just curiosity expressed and desire to not use the same solution to a given problem over and over.
For many years I just paralleled up 12AX7 to get transconductance up and rp down, one day it dawned on me that there were probably more modern triodes or pentodes (triode connected) that would give me the same or better performance but with considerably less miller capacitance, and much less mechanical complexity. My first effort was with the D3A and for fun I used gyrator loading with that in an effort to get the gain reasonably close to mu.
The D3A triode connected has the same transconductance as approximately 30 12AX7 in parallel and a rather low plate resistance as well. Miller capacitance is equivalent to a few 12AX7 in parallel. Noise performance may be 20dB or more better than the 12AX7. Mu is typically around 80 so slightly greater than the 5751 I generally used in lieu of the 12AX7.. The D3A triode connected is amazingly linear as well.
My move to cascode input topology resulted from my observation that even my MC cartridges and their transformers seemed to interact adversely with the high input capacitance of my triode connected D3A phono stage. I confirmed this by designing a pre-amplifier with low input capacitance, (the Muscovite) and also using a solid state pre-pre designed by a friend to isolate the input capacitance of the phono pre from the cartridge and SUT. In both cases the improvement in performance was noticeable in the upper octaves above a kHz or so.
As an added benefit these low rp high transconductance tubes can generally drive lower impedance passive RIAA networks effectively which picks up a few dB of additional SNR. Generally a single or several 12AX7 will have to drive a significantly higher impedance RIAA network to minimize load impedance generated response errors. Some tweaking of cap values is possible in lower impedance networks to compensate for this, but in so doing you become somewhat more vulnerable to parametric variations tube to tube.
The pentode with its very high rp is even more vulnerable to this issue than the cascode which generally has a substantially lower rp (although much higher than the triodes upon which it is based)
In my last three designs the dominant noise source has been the noise generated by the cartridge/SUT combination rather than the pre-amp itself.
There are so many ways to skin a cat and of course this is only the latest route I have followed. I'm certain that other approaches can achieve equally good results with equal care. All I can say is I'm getting better at this and my recent designs represent real progress compared to the older ones.
Hi Piano,
At some point I may devise an experiment and see whether or not I can measure something meaningful. At this point I can say it sounds significantly better without those caps, but that I do not have the full story.
I still have a pair of 330pF mica in the step network portion of the RIAA and suspect they are doing a little less damage there since they are not quite as directly in the audio path. I plan to replace these with polystryrene films once I get my hands on some more.
At some point I may devise an experiment and see whether or not I can measure something meaningful. At this point I can say it sounds significantly better without those caps, but that I do not have the full story.
I still have a pair of 330pF mica in the step network portion of the RIAA and suspect they are doing a little less damage there since they are not quite as directly in the audio path. I plan to replace these with polystryrene films once I get my hands on some more.
Hi Kevin,
A long overdue reply but one that you may find interesting or amusing (or both) but is in in no way a criticism; one I'll try to formulate carefully:
IMHO the design flaw that occurs when adhering to a strict passive RIAA network is that not all of the interfering actors are taken into account. This negligence then results in a less accurate RIAA curve at the output.
Again, this may not be the case as far as your designs are concerned. I'm just trying to figure out the why's.
As an example (and it may seem a tad extreme) a good friend of mine designs and builds phono stages in a more "holistic" way. A way which makes analyzing his designs quite puzzling.
He designs his phono pre in such a way that everything included is part of the RIAA correction.
From Miller capacitance of the input valves to the volume pots' impedance, the value of the cathode resistor decoupling caps, change a thing and you'll deviate from the RIAA curve.
He's also taken every precaution in the book and beyond so that the valves he uses won't age easily. (He knows how)
Not tweaker friendly but, and these are his words, when done like this it ends up the way I feel it should be. IOW, accurate throughout the useable life of the valves he uses. IOOW well over 20 years.
No need to mention, his power supplies are, to put it mildly, extreme.
To cut a long story short, I often receive the somewhat sneering remark that designing a phono pre is easy. Understood is easier than an amplifier.
It is not. It is extremely hard to do and by doing so everyone who takes a stab at it should be more than just a little respected for their efforts.
Hats off to you for presenting your designs. With some luck and good health we'll be discussing this in Belgium at he ETF.
Who knows what happens between a trappist or two.....
Most respectfully,
A long overdue reply but one that you may find interesting or amusing (or both) but is in in no way a criticism; one I'll try to formulate carefully:
IMHO the design flaw that occurs when adhering to a strict passive RIAA network is that not all of the interfering actors are taken into account. This negligence then results in a less accurate RIAA curve at the output.
Again, this may not be the case as far as your designs are concerned. I'm just trying to figure out the why's.
As an example (and it may seem a tad extreme) a good friend of mine designs and builds phono stages in a more "holistic" way. A way which makes analyzing his designs quite puzzling.
He designs his phono pre in such a way that everything included is part of the RIAA correction.
From Miller capacitance of the input valves to the volume pots' impedance, the value of the cathode resistor decoupling caps, change a thing and you'll deviate from the RIAA curve.
He's also taken every precaution in the book and beyond so that the valves he uses won't age easily. (He knows how)
Not tweaker friendly but, and these are his words, when done like this it ends up the way I feel it should be. IOW, accurate throughout the useable life of the valves he uses. IOOW well over 20 years.
No need to mention, his power supplies are, to put it mildly, extreme.
To cut a long story short, I often receive the somewhat sneering remark that designing a phono pre is easy. Understood is easier than an amplifier.
It is not. It is extremely hard to do and by doing so everyone who takes a stab at it should be more than just a little respected for their efforts.
Hats off to you for presenting your designs. With some luck and good health we'll be discussing this in Belgium at he ETF.
Who knows what happens between a trappist or two.....
Most respectfully,
Hi Frank,
It's been a while since I looked at this thread..
First I just wanted to report that I replaced the remaining pair of Russian Micas (330pF) with a pair of Siemens 2% 330pF 160V polystyrenes with a pretty significant improvement. Overall cleaner presentation...
In terms of how I do my equalization I do try to take into account all of the possible interactions both with the driving stage and the succeeding stage. Since I don't intend these designs to be one offs (what I intend and reality may not match.. ) I try to design things so that parametric variations tube to tube and over tube life have minimal effect on the accuracy of the RIAA, not to say that there is no effect of course.
Interesting and unfortunate that you run into that attitude, I've been fortunate not to have really encountered that level of cluelessness - as you know designing really good phono stages is the culmination of a great deal of understanding and experience. I was designing good power amps long before I designed my first good phono stage. Always something new to learn which is why I have tried so many new things lately.
Not to be snarky or anything but I have heard so much bad sound (most especially at high end shows) in recent years that I am not surprised that there is little appreciation of the effort required to design a good phono stage let alone anything else. Stratospheric prices for things designed by people who can't engineer their way out of a paper bag, of coarse this is not even the whole story since we are interfacing with human senses. The lack of respect for science and engineering, and a certain arrogant self-assurance these days seems to play a large role in this. There must be a reason why many of my friends are engineers, scientists or makers.. (Not to mention the fact that no one else seems to have a clue about what I am chattering about a lot of the time..)
It's been a while since I looked at this thread..
First I just wanted to report that I replaced the remaining pair of Russian Micas (330pF) with a pair of Siemens 2% 330pF 160V polystyrenes with a pretty significant improvement. Overall cleaner presentation...
In terms of how I do my equalization I do try to take into account all of the possible interactions both with the driving stage and the succeeding stage. Since I don't intend these designs to be one offs (what I intend and reality may not match..
) I try to design things so that parametric variations tube to tube and over tube life have minimal effect on the accuracy of the RIAA, not to say that there is no effect of course.Interesting and unfortunate that you run into that attitude, I've been fortunate not to have really encountered that level of cluelessness - as you know designing really good phono stages is the culmination of a great deal of understanding and experience. I was designing good power amps long before I designed my first good phono stage. Always something new to learn which is why I have tried so many new things lately.
Not to be snarky or anything but I have heard so much bad sound (most especially at high end shows) in recent years that I am not surprised that there is little appreciation of the effort required to design a good phono stage let alone anything else. Stratospheric prices for things designed by people who can't engineer their way out of a paper bag, of coarse this is not even the whole story since we are interfacing with human senses. The lack of respect for science and engineering, and a certain arrogant self-assurance these days seems to play a large role in this. There must be a reason why many of my friends are engineers, scientists or makers.. (Not to mention the fact that no one else seems to have a clue about what I am chattering about a lot of the time..)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.