Scott, I am not sure what you are saying, exactly. However, I found that Siliconix J110's , that I originally spec'd for the JC-2, over 35 years ago, got VERY noisy in the late 70's, and remain so today, because they switched to ion implantation.
In 1990 or so, I returned 100's of Toshiba 2SK147 transistors that did NOT meet worst case spec. How many people have a Quantech noise analyzer? I do, but even Linear Systems tried to lease my unit from me, as they are so few and far between. Many devices just slip through, though Levinson quietly changed from J110's to selected J113's and got reasonable results. I verified this with Ed Oxner. Maybe, you should take it up with him, as he is a consultant for Siliconix, Interfet, and even works with Linear Systems. That is where I first personally met him, over 10 years ago. I will give you Ed's number, if you don't have it.
In 1990 or so, I returned 100's of Toshiba 2SK147 transistors that did NOT meet worst case spec. How many people have a Quantech noise analyzer? I do, but even Linear Systems tried to lease my unit from me, as they are so few and far between. Many devices just slip through, though Levinson quietly changed from J110's to selected J113's and got reasonable results. I verified this with Ed Oxner. Maybe, you should take it up with him, as he is a consultant for Siliconix, Interfet, and even works with Linear Systems. That is where I first personally met him, over 10 years ago. I will give you Ed's number, if you don't have it.
Quantec (actually one guy, almost 90!) offered us their business when one of our fab techs went down there in desperation to get a unit fixed. It was something like 200k for the whole place. Someone should publish the Quantec schematics.
But really I test FETs with a homemade preamp and a soundcard. There are any number of programs to display noise spectra. Just got a bag of NOS Siliconix (1988) J310's that measure good, almost as good as the LSK170's (in my application). Opening my original Intersil data book shows the U310 specd at 10nV/rt-Hz @ 100Hz (same as now). I've misplaced my 1975 Siliconix data book with Ed's stuff in it, unfortunately, but I still have a box of thousands of JFET pairs that we pulled as QC samples in the 70's.
Gettering and annealing can cure most noise from implant damage, it is done. Backside gettering with argon is one way, there are others. A contaminated implanter can happen to anyone but most low-noise op-amps now see at least one ion implant step and there are several examples of 1nV amplifiers these days.
Sorry, I didn't mean to argue I was more upset at someone elsewhere that claimed the 2SK170 sounded "pinched and grainy" in the highs but the LSK170 sounded fine. But to give the idea that the industry moving to more modern processes like ion-implantation means an end to low noise devices is not entirely accurate. The push to tiny geometries and higher current densities has had a stronger effect.
But really I test FETs with a homemade preamp and a soundcard. There are any number of programs to display noise spectra. Just got a bag of NOS Siliconix (1988) J310's that measure good, almost as good as the LSK170's (in my application). Opening my original Intersil data book shows the U310 specd at 10nV/rt-Hz @ 100Hz (same as now). I've misplaced my 1975 Siliconix data book with Ed's stuff in it, unfortunately, but I still have a box of thousands of JFET pairs that we pulled as QC samples in the 70's.
Gettering and annealing can cure most noise from implant damage, it is done. Backside gettering with argon is one way, there are others. A contaminated implanter can happen to anyone but most low-noise op-amps now see at least one ion implant step and there are several examples of 1nV amplifiers these days.
Sorry, I didn't mean to argue I was more upset at someone elsewhere that claimed the 2SK170 sounded "pinched and grainy" in the highs but the LSK170 sounded fine. But to give the idea that the industry moving to more modern processes like ion-implantation means an end to low noise devices is not entirely accurate. The push to tiny geometries and higher current densities has had a stronger effect.
Well, this is all interesting stuff and I don't wish to put a damper on it, but it seems there isn't much addressing my original inquiry. Yeah, I might be able to get samples of the LS devices (has anyone tried to work through their distributors?), but I, personally, would still consider them unobtainium.
It seems like the Pass DIY pearl design might be my best bet. I guess I'll try fixing my malfunctioning EAR 834, with the pearl as the fallback.
I wonder if the golden age of phono preamp design is gone. Seems ironic to me given how turntables and (arguably) tonearms and cartridges have continued to improve.
It seems like the Pass DIY pearl design might be my best bet. I guess I'll try fixing my malfunctioning EAR 834, with the pearl as the fallback.
I wonder if the golden age of phono preamp design is gone. Seems ironic to me given how turntables and (arguably) tonearms and cartridges have continued to improve.
Err.. the Pass Pearl requires the use of unobtainium too.
Seems there's no getting away from 2SK389/170 if you want to build a good solid state discrete preamp. Maybe you can substitute with bjt's but even then the really good ones are "unobtainium".
I would substitute "design", with "good bits" here.
BTW
I'd say your thread is getting good if heavyweight designers like John and Scott are popping in. A large chunk of what they say whooshes over my head but I appreciate the informative insights.
Seems there's no getting away from 2SK389/170 if you want to build a good solid state discrete preamp. Maybe you can substitute with bjt's but even then the really good ones are "unobtainium".
I would substitute "design", with "good bits" here.
BTW
I'd say your thread is getting good if heavyweight designers like John and Scott are popping in. A large chunk of what they say whooshes over my head but I appreciate the informative insights.
john curl said:
They do. Ion implantation without annealing doesn't make any sense.
Its the wafer backside getter that is no so obvious. I have lived (and survived
the days when the getter was accomplished by... scratching the silicon wafer backside At the time, it was pretty good for low noise devices Onvinyl said:
So, what devices do these fantastic phono stages use? And, could you please point me to these cool designs on diyAudio? I guess I was trying to express that it seems that the device manufacturers no longer pursue high sonic quality (except in opamps), making it hard to design a SOTA preamp these days. Just my impression...
cromodora said:
True, but at least it is a fully discrete design. Should I care? Maybe, opamps have gotten good enough that I should not worry about it.
I agree and I hope they continue their discussion, but it doesn't seem to be moving in the direction I would wish.
Ok,
some:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=81575&highlight=
http://www.diyaudio.com/forums/showthread.php?s=&threadid=77602&highlight=
http://www.diyaudio.com/forums/showthread.php?s=&threadid=88929&perpage=25&highlight=&pagenumber=2
http://www.diyaudio.com/forums/showthread.php?s=&threadid=36260&highlight=
http://www.diyaudio.com/forums/showthread.php?s=&threadid=81788&highlight=
http://www.diyaudio.com/forums/showthread.php?s=&threadid=12397&highlight=
to name a few. there are a lot others. Use 'search'.
Among the best phonostages I actually heard are blue amp model 42, arbiter phono stage, lyra connoisseur (maybee 4-2?)
Rüdiger
some:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=81575&highlight=
http://www.diyaudio.com/forums/showthread.php?s=&threadid=77602&highlight=
http://www.diyaudio.com/forums/showthread.php?s=&threadid=88929&perpage=25&highlight=&pagenumber=2
http://www.diyaudio.com/forums/showthread.php?s=&threadid=36260&highlight=
http://www.diyaudio.com/forums/showthread.php?s=&threadid=81788&highlight=
http://www.diyaudio.com/forums/showthread.php?s=&threadid=12397&highlight=
to name a few. there are a lot others. Use 'search'.
Among the best phonostages I actually heard are blue amp model 42, arbiter phono stage, lyra connoisseur (maybee 4-2?)
Rüdiger
It may just be a question of semantics, but I wouldn't consider an opamp-based phono stage to be discrete simply because it happens to include one or two transistors as a headamp. At best, I'd call it a hybrid; neither fish, nor fowl.
You can complain all you want, but the sad fact of the matter is that the best JFETs are the ones that are being phased out. The 2SK170 is, I believe, still available, although the 2SJ74 already seems to have left town. Don't say I (and others) didn't warn you. If you want the '170, buy them now. They won't be around much longer.
I've got some of the Linear Systems devices here and have no complaint with them. No, I don't have some of the fancy test gear that John and Scott have, but I'm pretty well equipped for a DIY-type guy. (Ladies...Ladies!...Please take a number and wait in line. I'll get to you as soon as I'm able.) The problems always seem to be of the "I heard it from a friend" nature, which makes them difficult to pin down. The only substantive problem that I'm aware of is that they don't have the P-ch complements. That one you can't argue with. If, however, you're willing to build a design using only N-ch devices, I don't see why they wouldn't do the trick.
I've done several phono stages. Some discrete. Some tubes. One using the AD797 and the...oh, hell, it slipped away from me just as I reached for the memory trace...one of the JFET input AD chips. 843? 845? Something like that. I've got another fully discrete, balanced, zero feedback design on a back burner at the moment. I'm in what I hope are the finishing stages of work on an all-FET update of John Curl's JC-3 circuit and will then decide whether to go for a revision of that circuit with bipolars in various positions or turn up the heat on the phono stage.
Grey
You can complain all you want, but the sad fact of the matter is that the best JFETs are the ones that are being phased out. The 2SK170 is, I believe, still available, although the 2SJ74 already seems to have left town. Don't say I (and others) didn't warn you. If you want the '170, buy them now. They won't be around much longer.
I've got some of the Linear Systems devices here and have no complaint with them. No, I don't have some of the fancy test gear that John and Scott have, but I'm pretty well equipped for a DIY-type guy. (Ladies...Ladies!...Please take a number and wait in line. I'll get to you as soon as I'm able.) The problems always seem to be of the "I heard it from a friend" nature, which makes them difficult to pin down. The only substantive problem that I'm aware of is that they don't have the P-ch complements. That one you can't argue with. If, however, you're willing to build a design using only N-ch devices, I don't see why they wouldn't do the trick.
I've done several phono stages. Some discrete. Some tubes. One using the AD797 and the...oh, hell, it slipped away from me just as I reached for the memory trace...one of the JFET input AD chips. 843? 845? Something like that. I've got another fully discrete, balanced, zero feedback design on a back burner at the moment. I'm in what I hope are the finishing stages of work on an all-FET update of John Curl's JC-3 circuit and will then decide whether to go for a revision of that circuit with bipolars in various positions or turn up the heat on the phono stage.
Grey
Do not despair.
There are plenty of the excelllent 2SK170/2SJ74 to buy for input stages, and you will easily find MOSFETs for output if that is what you need.
Borbely audio is a reliabe source.
]http://www.borbelyaudio.com
Borbely also has some very nice phono preamps that might fit your bill.
An all out balanced phono preamp for ex could be Borbely's
now improved 320 kit (dual for balanced). Schematics is not yet updated but the source resistors are now gone for lower noise
Sigurd
There are plenty of the excelllent 2SK170/2SJ74 to buy for input stages, and you will easily find MOSFETs for output if that is what you need.
Borbely audio is a reliabe source.
]http://www.borbelyaudio.com
Borbely also has some very nice phono preamps that might fit your bill.
An all out balanced phono preamp for ex could be Borbely's
now improved 320 kit (dual for balanced). Schematics is not yet updated but the source resistors are now gone for lower noise
An externally hosted image should be here but it was not working when we last tested it.
Sigurd
Paul Ebert said:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.