You are right the op-amp is not appropriate, please don't be anymore irritated than I am when jn points out that copper does not superconduct.
I am an equal opportunity offender..
Hey, when you guys talk all that noise and equationy stuff, the only thing I can do is drool with a dumbfounded look on my face......so I have to look for stuff I can even pronounce...like copper, or helium...duh..
jn
I have successfully used G Ohm resistors and 1000pF or even less caps in this location, and I have another interesting spin on this that I'm trying out now.
whats your opinion on the miniSHARC Scott? i'm considering buying one this week (pretty much set on it). I can do basic IIR crossover in puremusic as well as more advanced EQ and convolutionin my plugins, but I do like the idea of it all being in hardware. if only the designer of allocator would get his act together and release the MAC version hes been promising for years.
Last edited:
Why don't you try an ADAU1781? While limited to 96K sample rates it should be able to be transparent otherwise. And would run for days on a pair of D cells. And you can walk across the hall (I believe) to get help with implementation. Using a PC for the DSP brings a host of grounding, noise (both physical and electrical) and usability issues. USB interfaces don't help much (actually not at all for the most part).
No opinions yet I'm afraid, funny thing I can walk across the street and pick up anything I want but just so much time and too many projects.
I have been working with the Sigma DSP stuff for several years now and the GUI tool is a dream to use for this stuff. The internal path is 28/56 bits so only the most obsessive will complain about resolution. I can't say if its the last word and the SHARC may be more capable if you want to do difficult stuff. Today that would be Dolby/DTS and surround synthesis but for crossovers/eq the SigmaDSP seems completely adequate. If the on board ADC's aren't what you want you can use your own through I2S.
fair enough, no problem, its so little money that i'm probably just going to grab one, i'm completely sorted for DACs and dont feel like changing, so software XO, or a pure i2s in/out DSP are the only choices.
about project overload, I know the feeling, wont stop me adding more though i'm afraid. it really should be considered for official recognition as an addiction this DIY stuff. btw have you looked into the RIAA built into the ES9102 ADC?
Demian: I want as much DSP as I can to allow for expansion, I need all 8 channels. the fact the minisharc has i2s in/out as well as easily accepting an external clock source is appealing, as well as the nice price and accessible dev environment. REW etc. Allocator, acourate or audiolense running on a headless PC is appealing too, but just seems like a pita. so either just running softwaer on my MAC, or hardware is preferred.
if I had some coding chops I would look into brutFIR on a beaglebone black and there is a FIR based crossover software for windows leveraging the i5, i7 etc vector processor that was also in an earlier version running in linux that might be able to be ported too
about project overload, I know the feeling, wont stop me adding more though i'm afraid. it really should be considered for official recognition as an addiction this DIY stuff. btw have you looked into the RIAA built into the ES9102 ADC?
Demian: I want as much DSP as I can to allow for expansion, I need all 8 channels. the fact the minisharc has i2s in/out as well as easily accepting an external clock source is appealing, as well as the nice price and accessible dev environment. REW etc. Allocator, acourate or audiolense running on a headless PC is appealing too, but just seems like a pita. so either just running softwaer on my MAC, or hardware is preferred.
if I had some coding chops I would look into brutFIR on a beaglebone black and there is a FIR based crossover software for windows leveraging the i5, i7 etc vector processor that was also in an earlier version running in linux that might be able to be ported too
Last edited:
The miniSHARC kit and the OpenDRC series have 24bit converters and 32 bit floating point processing path
2x4, 2x8, 8x8 kits and nanoDIDI have 24bit converters and 28/56 bit processing path / Double precision processing
(the 2x4 uses ADAU1701, I don’t know what the rest are using)
Spending hours to meticulously adjust optimum VTA with a specific cartridge, is a pain and I am afraid in practice a waste of time.
Old time records were cut with non standard VTA, but the highest VTA then was 20 d. Most of the European labels were around 15d.
I don’t’ know if modern records are all cut with a fixed VTA but I hear that the modern standard is 25d.
Add to this differences in record thickness and you are set.
Exposing the VTA myth? [English]
George
The current noise spectral density of the spontaneous fluctuations in the base current can consist of several contributions. There always is shot noise and 1/f noise and, occasionally, there is generation-recombination (GR) noise and burst noise, or so-called random telegraph signal (RTS) noise. The shot noise spectral density contribution is given by the well known 2*q*Ib relation, hence the value that Scott mentioned: SQRT(2*1.6E-19*2E-6)~0.8pA/rtHz. The 1/f noise in bipolar devices is often discussed in terms of mobility fluctuations. Depending on the base current origin (charge carriers recombining in the space-charge region between emitter and base, or charge carriers injected from the base into the emitter) the 1/f noise can have different expressions. Finally, burst noise and GR noise in bipolar transistors are often associated with traps or GR centers near the emitter-base space-charge region.
Sorry, I can't make it more simple than it is. You may quibble about but usually only 1/f noise qualifies, due to it's spectral distribution, as "excess noise". Shot and GR noise spectral distribution are, in a first approximation, flat.
Well, welcome to the 'club' Bob Cordell. I am in strong agreement with you on the discrete part, AND I also use an OPA134, but in the JC-3 as a second stage. I am glad that you chose that op amp, because I was nervous that it might not make the grade, but it does, in that location.
The Constellation phono stage that is just coming out (finally) is yet another layer above yours in complexity. It is probably too complex, with all fet construction, with both paralleled Njfets and Pjfets, and even a complementary Jfet power supply buffer. It costs a lot to make as well, perhaps too much.
The circuit was originally designed as an ALL OUT, no expenses barred, design, and I addressed it that way, but with the availability of Pch Jfets almost impossible, today, I might have taken a slightly more compromised approach, if I had to do it over.
The biggest difference between your and my design is the use of IC's in the audio path, and yet I use the same op amp as you in the JC-3, in the same location, so I apologize for saying anything about your phono preamp. It is, at least, 'buildable' and it should perform very well.
The Constellation phono stage that is just coming out (finally) is yet another layer above yours in complexity. It is probably too complex, with all fet construction, with both paralleled Njfets and Pjfets, and even a complementary Jfet power supply buffer. It costs a lot to make as well, perhaps too much.
The circuit was originally designed as an ALL OUT, no expenses barred, design, and I addressed it that way, but with the availability of Pch Jfets almost impossible, today, I might have taken a slightly more compromised approach, if I had to do it over.
The biggest difference between your and my design is the use of IC's in the audio path, and yet I use the same op amp as you in the JC-3, in the same location, so I apologize for saying anything about your phono preamp. It is, at least, 'buildable' and it should perform very well.
I am an equal opportunity offender..
Hey, when you guys talk all that noise and equationy stuff, the only thing I can do is drool with a dumbfounded look on my face......so I have to look for stuff I can even pronounce...like copper, or helium...duh..
jn
Ain't my world either. And, havent followed analog phono preamps in several decades. Its all a flash-back experience for me. -RNM
Thanks for that. If one looks at the somewhat dated data in Motchenbacher et al. one can see that the base current noise is always higher than the shot noise expected from the base current magnitude, which of course corroborates the notion that other things are going on. And it's not "lumpy", compared for example to JFET voltage noise curves also shown. These have gotten much better since those data in general --- see Danyuk's recent curves in a short piece about currently available SM JFETs, and his references, here: Measurements Rate SMT Low-Voltage n-JFETs Under Consistent Conditions | Power content from Electronic Design
I think again, though it has been mentioned, the important thing to account for about MM/MI cartridges and a few high-output MC is the inductance. This is why estimates that plug in the cartridge resistance and multiply by the parallel noise from the amplifier and the termination resistor fail to reproduce the overall noise voltage at the input. Even though RIAA EQ shapes the overall noise spectrum drastically, and aural acuity helps some more, the high-frequency noise from the parallel noise and rising cartridge impedance is still significant. This is why bipolar preamps are suboptimal, and as mentioned are also much more susceptible to RF rectification and envelope detection.
If we pursue the approach for MM that Bob shows in his LA article, of realizing the 75us final rolloff (or close to it) by a much smaller input termination, that is, using the cartridge inductance and termination R as the last lowpass, then the issues of reciprocity, which it seems are still controversial, come into play. This is where test records would seem to be essential. Of course if we like how it sounds, who cares?
When Marsh did his common-base MC preamp, published in Audio Amateur many years ago, again terminating the cartridge in a much lower impedance than would be typical, he reported liking the sound a lot. I think Scott Wurcer somewhere said he found it rather less to his liking. Chacun à son goût.
It will be good to see the FET updated/upgraded version from Borbely et al coming up in L.A. publication. Mine was all bipolar. I had owned several pre-pre... even John Curl's pre-pre. And Denon step up transformers. I had a comparison to others. About time this design got updated for those still playing their music on phonographs using MC. Good to know Borbeley is still active, too.
[Dont know about Scotts tastes and what he listened with... My ref speakers where a tri-amped system with RTR electrostatic tweeters direct drive from tube amp plates, Quad electrostatics for the midrange and 2 18 inch Hartly subs in transmission line enlosures each the size of refridgerators.] -RNM
[Dont know about Scotts tastes and what he listened with... My ref speakers where a tri-amped system with RTR electrostatic tweeters direct drive from tube amp plates, Quad electrostatics for the midrange and 2 18 inch Hartly subs in transmission line enlosures each the size of refridgerators.] -RNM
Last edited:
Hi John,
Thanks, John. Actually been in the "club" a long time, but indeed I have never made a commercial phono pre. I'm glad you agree with me on some of the choices I made.
It sounds like the Constellation will be quite a preamp. Am I correct in understanding that the whole signal path is discrete? If so, that does indeed qualify as quite complex, certainly in parts count, if not functional complexity. Will we be able to see the Constellation at RMAF this Fall?
Cheers,
Bob
In some of my research for the VinylTrak preamp I tried to see if there was evidence of enough electromagnetic coupling in an MM cartridge to make any difference. I must admit that I was unable to see anything. Here's what I did. I injected a small AC current into a cartridge that was completely unloaded. I set the current so that it produced something on the order of 5-50mV across the cartridge. I swept the frequency and looked at the signal amplitude across the cartridge to see if there was any evidence of an amplitude perturbation at or near the frequency of the cantilever resonance. I could not see anything. Maybe this testing approach was just flawed or not sensitive enough.
Cheers,
Bob
Not me, JC has mentioned his dislike of virtual ground inputs. As for excess noise I prefer not to put any noise you don't expect under that umbrella. Only that as outlined in that survey paper from LIGO.
Was this test done with the stylus in free air or was it on a vinyl record (or something else)? Cantilever resonance is certainly real, but may not be the dominant mechanism for the system's high frequency resonance.
Thanks,
Chris
LSK489 Application Notes
Some of you may be interested in seeing the application notes I wrote for Linear Systems on their new LSK489 dual N-channel JFET. The 16-page app note can be found on their web site under Literature/Application Notes/LSK489 Ultra Low Noise JFET.
Noise sources in JFETs are discussed and numerous application circuits are shown. One of the key advancements in the LSK489 is process improvements that have reduced the G-R noise contribution. The LSK489 has lower transconductance and higher voltage noise than the LSK389, but very substantially reduced capacitances.
As such, the LSK489 is a very good choice for MM phono preamps, and is the device I use in the MM input stage in the VinylTrak preamp.
Cheers,
Bob
Some of you may be interested in seeing the application notes I wrote for Linear Systems on their new LSK489 dual N-channel JFET. The 16-page app note can be found on their web site under Literature/Application Notes/LSK489 Ultra Low Noise JFET.
Noise sources in JFETs are discussed and numerous application circuits are shown. One of the key advancements in the LSK489 is process improvements that have reduced the G-R noise contribution. The LSK489 has lower transconductance and higher voltage noise than the LSK389, but very substantially reduced capacitances.
As such, the LSK489 is a very good choice for MM phono preamps, and is the device I use in the MM input stage in the VinylTrak preamp.
Cheers,
Bob
- Status
- Not open for further replies.