I was just trying to take some of the embarassement on myself.😉 Total cost would be more, but the package send to Grey would carry lower postage.
I was able to get some of them locally so for now I'm OK. Thank you anyway.
I don't see nothing wrong with helping ea. other in such matters. I did it before as well without even thinking twice. We are all in this together, right?😉
I was able to get some of them locally so for now I'm OK. Thank you anyway.
I don't see nothing wrong with helping ea. other in such matters. I did it before as well without even thinking twice. We are all in this together, right?😉
2SK387
A 2SK387 is n channel power mosfet I believe.
Dual P-FET 2SK 389 BL
http://www.schuro.de/preisl-jap-trans.htm
no charge, no editorializing................
A 2SK387 is n channel power mosfet I believe.
Dual P-FET 2SK 389 BL
http://www.schuro.de/preisl-jap-trans.htm
no charge, no editorializing................
OK. I wonder why I didn't pick up he embarassing part. Maybe I don't have the sense to get embarrassed when I should...
Anyhows, the offer stands for you and Grey should they be needed. And since DumballTheBlack is nowhere to be seen but Fred showed up I will happily send him a bunch as well. I am sure he uses them whenever he can as I recall he fancied Toshiba JFETs if I am not mistaking. If this is an insult I apologize in advance.
From the Schuro prices it seems you'd need 6 of them to justify postage...
Anyhows, the offer stands for you and Grey should they be needed. And since DumballTheBlack is nowhere to be seen but Fred showed up I will happily send him a bunch as well. I am sure he uses them whenever he can as I recall he fancied Toshiba JFETs if I am not mistaking. If this is an insult I apologize in advance.
From the Schuro prices it seems you'd need 6 of them to justify postage...
Jfets in the first stage of the Aleph-X ?
Coming back to Nelson's comment how the current Aleph-X concept could be improved:
"If you have an efficient horn mid/top driver, think about
using a well regulated supply, lower the gain of the amp
from 20 dB to 10 dB, use a JFET CCS instead of a
resistor off the Source of the diff pair's current source,
and use low noise P channel JFETs in parallel on the input."
What are your thoughts: Besides getting it quieter, would Jfets in the input not anyhow sound also better than Mosfets ? Any comments from you Grey (having made some experiences with your preamps)?
Best Regards
Coming back to Nelson's comment how the current Aleph-X concept could be improved:
"If you have an efficient horn mid/top driver, think about
using a well regulated supply, lower the gain of the amp
from 20 dB to 10 dB, use a JFET CCS instead of a
resistor off the Source of the diff pair's current source,
and use low noise P channel JFETs in parallel on the input."
What are your thoughts: Besides getting it quieter, would Jfets in the input not anyhow sound also better than Mosfets ? Any comments from you Grey (having made some experiences with your preamps)?
Best Regards
DumballTheBlack is nowhere to be seen but Fred showed up
Thanks, but I have plenty of them, spice models, and two amps designed with them, and have been designing with then for awhile. I have seen them in any number of amps including Marantz and designs from Nelson Pass and Erno Borbely.
http://www.borbelyaudio.com/index31.htm
Thanks, but I have plenty of them, spice models, and two amps designed with them, and have been designing with then for awhile. I have seen them in any number of amps including Marantz and designs from Nelson Pass and Erno Borbely.
http://www.borbelyaudio.com/index31.htm
Re: DumballTheBlack is nowhere to be seen but Fred showed up
Exactly my thought and I suspected you could be running out. Not so it seems. If nothing else I keep the Aleph-X thread in focus...
Fred Dieckmann said:
Exactly my thought and I suspected you could be running out. Not so it seems. If nothing else I keep the Aleph-X thread in focus...
UrSv, Grey
If timescale is not a problem I am going to the States in March/April for NAB. If you want to send them to me UrSv, I can post them when I get to Vegas😉
If timescale is not a problem I am going to the States in March/April for NAB. If you want to send them to me UrSv, I can post them when I get to Vegas😉
quote:
--------------------------------------------------------------------------------
Originally posted by Fred Dieckmann
A 2SK387 is n channel power mosfet I believe.
Dual P-FET 2SK 389 BL
http://www.schuro.de/preisl-jap-trans.htm
--------------------------------------------------------------------------------
Looking at the toshiba datasheet
2sk389 is a dual N-Channel Junction Type fet
2SK387, P-FET 2SK 389BL...., starts to get messy😉
--------------------------------------------------------------------------------
Originally posted by Fred Dieckmann
A 2SK387 is n channel power mosfet I believe.
Dual P-FET 2SK 389 BL
http://www.schuro.de/preisl-jap-trans.htm
--------------------------------------------------------------------------------
Looking at the toshiba datasheet
2sk389 is a dual N-Channel Junction Type fet
2SK387, P-FET 2SK 389BL...., starts to get messy😉
If I ever learn to type, I'll be downright dangerous. 389 is (I think) what I meant. The dual N JFET...whichever one it is.
Not only that, but some @#$%^& moron spilled something in this keyboard (I'm at work) and the blasted thing keeps skipping letters entirely. I try to catch them when I proof my posts, but can't guarantee that I'll get them all. Space, i, y, and a are the worst.
Blitz,
I'm going to kinda mumble a bit here, thinking out loud. Bear with me.
Regulated supply? No problem. You can go with a minimalist regulator circuit like Nelson's recent paper at www.passdiy.com. Or, if you like, I can stick up a more elaborate one. (Jam will like it, active loads and current sources...all discrete.) Something not to be overlooked is to use a chip regulator with a pass transistor wrapped around it to carry the current. The ARRL book I've got at the house is full of variations on that theme. The chip serves as the brain, and the pass transistor does all the heavy work.
No mistake, JFETs are the ticket for low noise, and some claim they're the best sounding devices available at this stage in history. Assuming that you want to use JFETs for the differential itself, you're going to have to deal with lower gain than with MOSFETs. That's cool. Low gain and low noise are good companions.
I've not fiddled with devices other than the IRF9610s for the front end, but there's no real "magic" there. You could use bipolars if you wanted, too. I've got some MPSA18s (technically low noise devices) on hand which I'd thought about, but then I'd have to turn the rest of the circuit upside down because they're N and I'd need P. Nelson seems fond of the Zetex 550 (450? the P one, at any rate), that might be fun to try.
For that matter, there's a dual P JFET, but the number escapes me.
In case anyone is wondering, the front end runs about 20 dB gain (open loop, one side). If you can get approximately that amount of gain and something around 4V offset across the load resistors, you've got a viable front end--without having to change much else. Oh, and it's P, of course, unless you want to swap things around and go with a P output instead of N.
Back to the JFET front end...the nifty thing to do would be to go <i>all</i> JFETs. The booger is going to be the pass device in the current source. The highest current JFET that springs to mind is a J310, which I think is a 10mA part...but that one's N, isn't it? Need a relatively high current P JFET, and I'm more familiar with the Motorola catalog than Toshiba, et. al. Somebody figure out a relatively high current P JFET that's readily available. Fortunately, a JFET differential won't take much current--5-10 mA at a guess--so it's not as though we need a huge amount of current capability for the pass device.
Uh, lets see, the reference for the current source. Options abound. Go to the passdiy site and look at the DIY opamp paper. About half or two-thirds of the way back, he goes from resistors to current sources and gives schematics for MOSFET, bipolar, and JFET current sources. Use a JFET CCS (any small one will do, don't need a lot of current) to drive the Zener instead of the resistor. Option two--turn that around and use the JFET current source against a resistor. Not quite as steady as a Zener, but quieter. A steady current against a fixed resistance makes a pretty good voltage reference.
You can go with monolithic current sources, and/or monolithic voltage references. Use both together and you're back to the HH-CCS (looks like something from organic chemistry, two hydrogens, single bond to two carbons, and one sulphur).
How crazy do you want to get on the front end current source? We can get pretty wild if you want to take it to extremes and stay discrete. The thing is, for an amp it's not going to be very noticable compared to a preamp or head amp. The further upstream you go, the more critical it is.
The upshot is--JFETs will do quite nicely if you're willing to hack at it a bit to mate it to the rest of the circuit and deal with lower gain.
Grey
Not only that, but some @#$%^& moron spilled something in this keyboard (I'm at work) and the blasted thing keeps skipping letters entirely. I try to catch them when I proof my posts, but can't guarantee that I'll get them all. Space, i, y, and a are the worst.
Blitz,
I'm going to kinda mumble a bit here, thinking out loud. Bear with me.
Regulated supply? No problem. You can go with a minimalist regulator circuit like Nelson's recent paper at www.passdiy.com. Or, if you like, I can stick up a more elaborate one. (Jam will like it, active loads and current sources...all discrete.) Something not to be overlooked is to use a chip regulator with a pass transistor wrapped around it to carry the current. The ARRL book I've got at the house is full of variations on that theme. The chip serves as the brain, and the pass transistor does all the heavy work.
No mistake, JFETs are the ticket for low noise, and some claim they're the best sounding devices available at this stage in history. Assuming that you want to use JFETs for the differential itself, you're going to have to deal with lower gain than with MOSFETs. That's cool. Low gain and low noise are good companions.
I've not fiddled with devices other than the IRF9610s for the front end, but there's no real "magic" there. You could use bipolars if you wanted, too. I've got some MPSA18s (technically low noise devices) on hand which I'd thought about, but then I'd have to turn the rest of the circuit upside down because they're N and I'd need P. Nelson seems fond of the Zetex 550 (450? the P one, at any rate), that might be fun to try.
For that matter, there's a dual P JFET, but the number escapes me.
In case anyone is wondering, the front end runs about 20 dB gain (open loop, one side). If you can get approximately that amount of gain and something around 4V offset across the load resistors, you've got a viable front end--without having to change much else. Oh, and it's P, of course, unless you want to swap things around and go with a P output instead of N.
Back to the JFET front end...the nifty thing to do would be to go <i>all</i> JFETs. The booger is going to be the pass device in the current source. The highest current JFET that springs to mind is a J310, which I think is a 10mA part...but that one's N, isn't it? Need a relatively high current P JFET, and I'm more familiar with the Motorola catalog than Toshiba, et. al. Somebody figure out a relatively high current P JFET that's readily available. Fortunately, a JFET differential won't take much current--5-10 mA at a guess--so it's not as though we need a huge amount of current capability for the pass device.
Uh, lets see, the reference for the current source. Options abound. Go to the passdiy site and look at the DIY opamp paper. About half or two-thirds of the way back, he goes from resistors to current sources and gives schematics for MOSFET, bipolar, and JFET current sources. Use a JFET CCS (any small one will do, don't need a lot of current) to drive the Zener instead of the resistor. Option two--turn that around and use the JFET current source against a resistor. Not quite as steady as a Zener, but quieter. A steady current against a fixed resistance makes a pretty good voltage reference.
You can go with monolithic current sources, and/or monolithic voltage references. Use both together and you're back to the HH-CCS (looks like something from organic chemistry, two hydrogens, single bond to two carbons, and one sulphur).
How crazy do you want to get on the front end current source? We can get pretty wild if you want to take it to extremes and stay discrete. The thing is, for an amp it's not going to be very noticable compared to a preamp or head amp. The further upstream you go, the more critical it is.
The upshot is--JFETs will do quite nicely if you're willing to hack at it a bit to mate it to the rest of the circuit and deal with lower gain.
Grey
AX quality
I've been listening at length to my AX (alpha pcbs) tonight, while occasionally checking and adjusting the dc offset. This is the first night that I've had the chassis warm enough to feel like a class A product; just the right amount of heat and that sweet smell of aluminum cutting oil. Oh yes, what a night. I must say that this amp just keeps getting sweeter all the time! It is so very revealing of the music source. I've been listening to an old Stevie Ray Vaughan cd (Couldn't stand the weather) and I'm hearing through the veil as Stevie taps softly the start up beat on his guitar! I'm hearing mikes being lightly bumped or a stool casually slid an inch or two across the floor..... It really doesn't get any better than this.
Good night everyone,
John
I've been listening at length to my AX (alpha pcbs) tonight, while occasionally checking and adjusting the dc offset. This is the first night that I've had the chassis warm enough to feel like a class A product; just the right amount of heat and that sweet smell of aluminum cutting oil. Oh yes, what a night. I must say that this amp just keeps getting sweeter all the time! It is so very revealing of the music source. I've been listening to an old Stevie Ray Vaughan cd (Couldn't stand the weather) and I'm hearing through the veil as Stevie taps softly the start up beat on his guitar! I'm hearing mikes being lightly bumped or a stool casually slid an inch or two across the floor..... It really doesn't get any better than this.
Good night everyone,
John
Question to GRollins
GRollins, could you post some schematis how to use N channel JFETS as input devices?
GRollins, could you post some schematis how to use N channel JFETS as input devices?
Jfets for the differential pair / CSS
Grey, what about using this guy here:
http://www.schuro.de/Daten/Japanhalbleiter/2SJ109.pdf
It is the P-Typ-Complement of the 2sk389. In the V-Version it can be run with up to 20 mA and is very quiet. And it is a dual-type and therefore already matched.
Alternatively, the complement of the 2sk170:
http://www.schuro.de/Daten/Japanhalbleiter/2SJ74.pdf
For the current source: Why not use two matched 2sj74 which could run in the V-Version with up to 20mA each. I have paralleled some 2sk170V in my preamp in the current sources and it works very well.
What would be your estimate on components values around these devices ? Or do we need SPice-Models in any case ?
Best Regards
Grey, what about using this guy here:
http://www.schuro.de/Daten/Japanhalbleiter/2SJ109.pdf
It is the P-Typ-Complement of the 2sk389. In the V-Version it can be run with up to 20 mA and is very quiet. And it is a dual-type and therefore already matched.
Alternatively, the complement of the 2sk170:
http://www.schuro.de/Daten/Japanhalbleiter/2SJ74.pdf
For the current source: Why not use two matched 2sj74 which could run in the V-Version with up to 20mA each. I have paralleled some 2sk170V in my preamp in the current sources and it works very well.
What would be your estimate on components values around these devices ? Or do we need SPice-Models in any case ?
Best Regards
Re: 2SK387
2SK389 is a dual <B>N</B>-Channel JFET and the complementary part is the dual <B>P</B>-Channel JFET 2SJ109😉
Sorry Fred,Fred Dieckmann said:
2SK389 is a dual <B>N</B>-Channel JFET and the complementary part is the dual <B>P</B>-Channel JFET 2SJ109😉
Complementary part of J310
The problem is I can't find a complementary part (i.e. P-channel JFET) of the J310. (or J309) The latter are N-channel.
If I did find it I would build a KWAK-CLOCK with complementary JFET's to see if it has any benefit....😕
Hi Grey,GRollins said:
The problem is I can't find a complementary part (i.e. P-channel JFET) of the J310. (or J309) The latter are N-channel.
If I did find it I would build a KWAK-CLOCK with complementary JFET's to see if it has any benefit....😕
A quick progress report,
I've nearly populated the boards from John.
It looks like it came out of the factory (better not be to neat it won't look like diy!).
I've used pcb terminatal blocks for all the wiring connections on the boards to enable screw in wiring. This is enable a fair amount of experiementation without damaging the tracks.
Tonight if time permits I'll fine tune the matching of the diff pair.
I think I will run with Grey orginal CCS to start with and go from there.
Will post some pic soon.
Ian
I've nearly populated the boards from John.
It looks like it came out of the factory (better not be to neat it won't look like diy!).
I've used pcb terminatal blocks for all the wiring connections on the boards to enable screw in wiring. This is enable a fair amount of experiementation without damaging the tracks.
Tonight if time permits I'll fine tune the matching of the diff pair.
I think I will run with Grey orginal CCS to start with and go from there.
Will post some pic soon.
Ian
fine tuning?
Hi Macka,
Tell me more about the FET fine tuning process. I really like your idea of using pcb terminal blocks. I used wire connectors but think your idea is better.
John
Hi Macka,
Tell me more about the FET fine tuning process. I really like your idea of using pcb terminal blocks. I used wire connectors but think your idea is better.
John
Bricolo wrote:
hmm. i think low noise. I´m very intersted in this point too, i hope some of the knowledged ones may write an abstact about jfets. I want to know what types to use, not only japan types because not easy to obtain. what types from fairchild or onsemi would be good?
hmm. i think low noise. I´m very intersted in this point too, i hope some of the knowledged ones may write an abstact about jfets. I want to know what types to use, not only japan types because not easy to obtain. what types from fairchild or onsemi would be good?
Well, that has been done already by Mr. Borberly:
http://www.borbelyaudio.com/ae599bor.pdf
http://www.borbelyaudio.com/ae699bor.pdf
Best Regards
http://www.borbelyaudio.com/ae599bor.pdf
http://www.borbelyaudio.com/ae699bor.pdf
Best Regards
Just a word about 'matched' JFETS.
Although it is convienient to use a matched pair, one should read the specs carefully... In the marority of the cases matching is only as close as 2 -5 %.
In this application one would like to achieve matching as closely as possible. As dual fets in one package are produced but (normally) not trimmed I feel that these devices are less usable.
I believe that is one of the reasons why Mr. Pass choses to select the input pair from single devices.
If you carefully study the grahps from FET's, you typically find a value for Id where all temperature curves more or less cross in one point. In my view the input pair can be selected among the suitable devices where the value of the temp independant Id is close to the DC operating point and then match devices as closely as possible. Then, as already indicated by BrianGT use direct thermal coupling between the pair.