It looks like we got some 'action' around here. I only want to advise that you look at the circuit tradeoffs carefully. Any of us can design a working circuit. Just like an automotive designer can design a working automobile engine, but would EVERY engine be worthwhile? For example, in engines, we might go back to flathead' like the old Ford V8 engines until 1953 or so. Or the Chevy straight 6 with overhead valves. (they put it in the first Corvettes, why not?) Or the 2CV? How about 3 cylinders, 5 cylinders, 9 cylinders? In principle, each approach would WORK, but would they work WELL, exceeding what could be made with the same time and effort?
This is the question. How do we make an 'optimum' design with the best chance to sound good, measure fairly well, and be easy enough to source the parts for?
In my world, this phono design would be N channel only j-fets on the input, bipolar or fets optional in other stages, DIRECT COUPLED, to remove expensive and controversial coupling caps, POLYSTYRENE EQ CAPS. , and a two gain block phono design. IC's may be very useful for the second stage, especially, because the first stage will absorb all the ticks, pops, and mistracking artifacts. Then, a good clean fast IC would have a fairly easy time, if you chose to use it. Servos are probably necessary, IF you do not use coupling caps. Why not? They are cheaper than an extra high QUALITY coupling cap.
Joachim, you are very prolific with your circuit designs, but please consider the elegance and excellence of each design, and go with the 'best'. This is the secret to good audio design.
This is the question. How do we make an 'optimum' design with the best chance to sound good, measure fairly well, and be easy enough to source the parts for?
In my world, this phono design would be N channel only j-fets on the input, bipolar or fets optional in other stages, DIRECT COUPLED, to remove expensive and controversial coupling caps, POLYSTYRENE EQ CAPS. , and a two gain block phono design. IC's may be very useful for the second stage, especially, because the first stage will absorb all the ticks, pops, and mistracking artifacts. Then, a good clean fast IC would have a fairly easy time, if you chose to use it. Servos are probably necessary, IF you do not use coupling caps. Why not? They are cheaper than an extra high QUALITY coupling cap.
Joachim, you are very prolific with your circuit designs, but please consider the elegance and excellence of each design, and go with the 'best'. This is the secret to good audio design.
In later versions there is only one cap parallel to R20. Yes, i was concerned about PSSR too and it is simpler. I also made a version with folded cascodes and servo that is DC coupled that is not published yet. I am still thinking about a way to include the output cap in the feedback loop. In the inverted version that was simpler by spitting the RIAA.
About your Power Amp ideas i can not say much but distortion in a Power Amp is a much bigger problem then in a low level circuit so i am a bit concerned that you end up with too little loop gain. I thought that we are discussing the Parasound phonostage here but so it goes. I am definetely interested in a simple, good quality poweramp. Maybe it is worth it´s own thread.
About your Power Amp ideas i can not say much but distortion in a Power Amp is a much bigger problem then in a low level circuit so i am a bit concerned that you end up with too little loop gain. I thought that we are discussing the Parasound phonostage here but so it goes. I am definetely interested in a simple, good quality poweramp. Maybe it is worth it´s own thread.
Joachim, ultimate circuit design is NOT a number of 'versions' Different 'versions' come from different design 'oversights' and subsequent alterations. OF COURSE, in the development of ANY serious circuit we have different 'versions' I am on my 3'rd 'version' of the Constellation Audio phono circuit board. And there might be a 4th, if necessary. Actually, the OLD schematics just get in the way, especially when I am troubleshooting a later version. It is the same here. Too many permutations and combinations only confuse the issue. Moreover, sticking to the design GUIDELINES is also necessary.
For example, I could theoretically make an elegant design with paralleled 2SK146-2SJ73 pairs and a complementary T0-5 V fet second or output stage. What good would that do YOU, or anyone else, if I put a schematic up to show everyone? The parts are almost non-existent for the amateur to buy. So, it would be a waste of time.
What I have attempted to try here, is an EXTENSION of the JC-3, making some or all of it discrete, but still following the JC-3 design guidelines, including direct coupling, 2 gain blocks, either all passive or passive-active RIAA EQ. The input noise should not exceed 1nV/rt Hz. The output should be 3V or greater into a 2Kohm load. That is all there is to it.
For example, I could theoretically make an elegant design with paralleled 2SK146-2SJ73 pairs and a complementary T0-5 V fet second or output stage. What good would that do YOU, or anyone else, if I put a schematic up to show everyone? The parts are almost non-existent for the amateur to buy. So, it would be a waste of time.
What I have attempted to try here, is an EXTENSION of the JC-3, making some or all of it discrete, but still following the JC-3 design guidelines, including direct coupling, 2 gain blocks, either all passive or passive-active RIAA EQ. The input noise should not exceed 1nV/rt Hz. The output should be 3V or greater into a 2Kohm load. That is all there is to it.
One last try. Replace the input stage of this one with Godfreys Fet- foldetd cascodes he designed after Johns sugestion. The schematic has a small error. There is a 220 Ohm feedback resistor missing from minus input of the Opamp to ground. Other Opamps can be used to taiste and budget.
Attachments
Joachim, your schematic violates several guidelines. First, it is a summing input. Do you think this is ORIGINAL? I did my first summing input, with the Levinson JC-1AC, 36 years ago. Why don't I use it today? If you don't know, I will tell you: MOST moving coil cartridges don't LIKE a summing input. It OVER-DAMPS them. How do I know? Because I did an A-B test with the SAME preamp to show this, back in 1976. How would an input like this be good for a general phono preamp? Also, it would not ALLOW for moving magnet cartridges to be used, either.
Second, it is CAPACITOR COUPLED. NO COUPLING CAPACITORS! They are unnecessary and redundant in this type of design.
Second, it is CAPACITOR COUPLED. NO COUPLING CAPACITORS! They are unnecessary and redundant in this type of design.
Sure, it only works well with very low impedance cartridges. I was showing it only because of the second stage that uses an Opamp with servo. That is not very original too.
I think i got a bit over enthusiastic. Still i think to make a good sounding phonostage that is affordable is a noble attempt and as far as i can tell from people who heard it you succeded with the JC3. I will make sure that i listen to your JC3 first if i get a chance. Maybe a can learn something.
I think i got a bit over enthusiastic. Still i think to make a good sounding phonostage that is affordable is a noble attempt and as far as i can tell from people who heard it you succeded with the JC3. I will make sure that i listen to your JC3 first if i get a chance. Maybe a can learn something.
John I'd love to some info on this. The cantilever compliance and tone arm mass resonance not believable but, the stylus tip mass/cantilever I don't know. The reciprocal mechanixal/electrical coupling from the motor assembly to the cantilever is weak and I find it hard to believe, listening tests are not adequate, these things are measurable. Current summing input made the cartridge sound like it was "under water" does not cut it.
The argument was made to me years ago in the case of a ribbon mike. That is totally different, the ribbon has an extremely low mass and couples to a very high magnetic field.
Last edited:
Scott, you don't EVEN OWN a MC cartridge. What do you know about it? I make MY living from making audio designs, often phono stages. I first designed an MC preamp, the JC-1 with a LOW feedback summing input. It made the most sense, back in 1973. However, the design INVERTED the signal, so I made the JC-1 AC and the JC-1 DC that did NOT INVERT and used paralleled dual base input, instead. Living with my design, I became aware (that is what makes me a successful audio designer) that there was 'something wrong' with the 'sound' of the MC pre-preamp. So, after due consideration, I realized that I could MODIFY the JC-1AC to either use common base or common emitter input. So, I installed a switch to do this. Then we could have a summing input (1 ohm or so, or 100 ohms. Well, 100 ohms won, and it is my default reference load today for the vast majority of phono cartridges. That is called 'listening' to your designs, rather than just justifying them on technical grounds. What a pain that was, to go in the wrong direction. Cost me plenty. Other people, like the designer of the Magnepan loudspeaker had both the JC-1 and the JC-1AC and he told me the same thing, quite independently, on another occasion. I trust my ears, and those of my associates, even today.
...thought I read that "underwater" comment somewhere before:
http://www.audioasylum.com/audio/tweaks/messages/72510.html - do go up thread to see some more "audiophile luminaries" comments
like Scott I doubt the electromechanical model coupling supports the theory of a audible frequency response change due to electrical loading mechanical damping with MC - but we would need some detailed cart/coil magnet geometry and suspension damping numbers to see if that suspicion is true - or only true for "high output" MC as Risch suggests in the link
Last edited:
You HAVE TO TRY IT! I first believed that summing type input was the best way. It is in my patented JC-1 design, that started being produced in 1973. Makes sense, and it seemed to work with modest summing, but with true common base input, it failed with most cartridges. Now, PLEASE PROVE ME WRONG. You will then save Constellation Audio at least $1000 OEM in parts for each unit, IF you can show me that a summing input works just great. BUT if you cannot PROVE IT, then please suspend your OPINIONS, because they have no substance.
for some the goal is to link engineering principles to results
provide some numbers showing clear changes beyond established DBT audibility thresholds in either measurements or modeling - or show DBT ABX subjective listening test results acceptable for a peer reviewed publication and I'll accept that you've "proved" the proposition
reference to anecdotal comments by people claiming DBT is "too insensitive" doing sighted listening doesn't impress - particularly when Risch, Jcarr both seem to say virtual gnd MC preamps are subjectively great with low impedance/low output cartridges
so where is the dividing line – can we put numbers on it – can we actually turn medieval guild style “mastery” into modern audio engineering knowledge?
it is entirely possible that modeling or measurements could convince me - I admit I don't have the numbers - and I haven't seen detailed enough descriptions of MC engineering designs to derive them - links to the detailed design specs of MC cartridges would be very welcome
I really don’t understand the use of iron armature for “moving coil” – if the iron is operating linearly I would call such construction “variable reluctance” – whether the coil is on the moving iron part seems a pointless distinction
my model of “moving coil” is a open coil moving in a relatively coil current invariant free air spatially structured magnetic field from the permanent magnet/pole pieces - I'd expect highest linearity would come from "flat" mag field and orthogonal coil axis to that the small angle sine approximation is good – I’d guess this is “low output MC”
if there is such a fundamental transduction principle difference between high/low output MC then I would suggest they require/deserve different preamp designs
provide some numbers showing clear changes beyond established DBT audibility thresholds in either measurements or modeling - or show DBT ABX subjective listening test results acceptable for a peer reviewed publication and I'll accept that you've "proved" the proposition
reference to anecdotal comments by people claiming DBT is "too insensitive" doing sighted listening doesn't impress - particularly when Risch, Jcarr both seem to say virtual gnd MC preamps are subjectively great with low impedance/low output cartridges
so where is the dividing line – can we put numbers on it – can we actually turn medieval guild style “mastery” into modern audio engineering knowledge?
it is entirely possible that modeling or measurements could convince me - I admit I don't have the numbers - and I haven't seen detailed enough descriptions of MC engineering designs to derive them - links to the detailed design specs of MC cartridges would be very welcome
I really don’t understand the use of iron armature for “moving coil” – if the iron is operating linearly I would call such construction “variable reluctance” – whether the coil is on the moving iron part seems a pointless distinction
my model of “moving coil” is a open coil moving in a relatively coil current invariant free air spatially structured magnetic field from the permanent magnet/pole pieces - I'd expect highest linearity would come from "flat" mag field and orthogonal coil axis to that the small angle sine approximation is good – I’d guess this is “low output MC”
if there is such a fundamental transduction principle difference between high/low output MC then I would suggest they require/deserve different preamp designs
Last edited:
Not necessarily, folks at MIT took those RHOM transistors and an old RCA 44 ribbon mike and made it transformerless. There was a dramatic reduction in the tendency to pop the ribbon also that hey claimed was the damping. They tried to get us to make a version for MC carts, I don't think they thought it through. From what I have read the high frequency peak in a MC is the sylus tip mass/cantilever resonance (I could be wrong) which is decoupled from motor damping somewhat to a lot depending on construction.
I had to laugh at someone thinking they could patent a virtual ground in 1979.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.