I just found a secret stash of 2sk170 and 2sj74 🙂 .
Enough to share?
I just found a secret stash of 2sk170 and 2sj74 🙂 .
Hopefully you didn't find it in China.....
No. They were purchased directly from Toshiba 5 years ago. I forgot that I had them. I found them while I was looking for my Sony mosfets (still MIA)Hopefully you didn't find it in China.....
Enough to share?
I only have 25 of each left. I might be able to spare a matched pair. I'd need to consider every conceivable amp, I might build over the next 10 years to work out how many I would be willing to sell off.
I only have 25 of each left. I might be able to spare a matched pair. I'd need to consider every conceivable amp, I might build over the next 10 years to work out how many I would be willing to sell off.
Pm sent...
JFets ..
Thank you ZM, I just ordered the matched quad at diyaudio store here, can I just use them? So I see from your tip that besides the bias voltage this circuit work at 32V supply rail. =🙂
I hope to post my F6 pictures soon....
what is exact Ids of JFets you have ?
RC , to bring them to 15V at drains ...... IRFP150 = up to 50W per device , if you keep sink at 50C
increase R7 , R8 to 15K
Thank you ZM, I just ordered the matched quad at diyaudio store here, can I just use them? So I see from your tip that besides the bias voltage this circuit work at 32V supply rail. =🙂
I hope to post my F6 pictures soon....
when you get them , measure Idss and I'll draw what's needed - either simple RC or proper cascodes with small BJTs , to keep them at 15V (+/-) which is safe bracket
I only have 25 of each left. I might be able to spare a matched pair. I'd need to consider every conceivable amp, I might build over the next 10 years to work out how many I would be willing to sell off.
PM sent
Idss and 15V (+/-)?
Hi ZM, sorry for silly questions, you mean to keep the Vds of 2SK170 and 2SJ74 within 15V (+/-)? They seem to be over 15V (+/-) already, or please enlighten me..
Does this has something to do with the 1K resistors at the drain of 2SK170 and 2SJ74 of F5 circuit? 😕
THanks, Chul
when you get them , measure Idss and I'll draw what's needed - either simple RC or proper cascodes with small BJTs , to keep them at 15V (+/-) which is safe bracket
Hi ZM, sorry for silly questions, you mean to keep the Vds of 2SK170 and 2SJ74 within 15V (+/-)? They seem to be over 15V (+/-) already, or please enlighten me..

THanks, Chul
F5 front end is completely different animal , where drain resistors are drain resistors
here , same as in F4 and everywhere you find same buffer stage , buffer is a buffer , and any resistance in drain is having different function (comparing to F5)
so , 15V ......... simple - to keep JFet dissipation at bay
practically anything from 12 to 18 volts , if keeping BL ones in 90-100% Idss area (so, with none or small source resistors)
edit - it's always important to account needed swing of input (thus buffer output , too) in broader picture and construction of buffer
in this case , where swing demand is in sane area , you can put +/- protective envelope around JFet buffer in pretty simple manner ;
in case of F4 , for instance , where both input and buffer output must fly almost from rail to rail , you can't use RC cells to keep buffer drains at bay , simply because sources can't outrun/outfly drains and swing will be limited to at bay level
however , some clever (and still simple) cascoding of drains is way to go , simply allowing cascodes to fold around rails (upper one around positive and lower one around negative ) thus allowing buffer output to swing (almost ) rail to rail
easy to see either in vivo or in LTSpice sim
take F4 V2. (that one with few resistors and uni-diodes added around FE) as example ; think about each part role
here , same as in F4 and everywhere you find same buffer stage , buffer is a buffer , and any resistance in drain is having different function (comparing to F5)
so , 15V ......... simple - to keep JFet dissipation at bay
practically anything from 12 to 18 volts , if keeping BL ones in 90-100% Idss area (so, with none or small source resistors)
edit - it's always important to account needed swing of input (thus buffer output , too) in broader picture and construction of buffer
in this case , where swing demand is in sane area , you can put +/- protective envelope around JFet buffer in pretty simple manner ;
in case of F4 , for instance , where both input and buffer output must fly almost from rail to rail , you can't use RC cells to keep buffer drains at bay , simply because sources can't outrun/outfly drains and swing will be limited to at bay level
however , some clever (and still simple) cascoding of drains is way to go , simply allowing cascodes to fold around rails (upper one around positive and lower one around negative ) thus allowing buffer output to swing (almost ) rail to rail
easy to see either in vivo or in LTSpice sim
take F4 V2. (that one with few resistors and uni-diodes added around FE) as example ; think about each part role
Last edited:
I am getting there but..
Let me put here what I (think I) understood...
1. So the front end of F5 is not a buffer.. the drain resistor is a load resistor and direct coupled to power FET?
2. RC.. you mean simple RC as in a tube circuit, decoupling driver stage from main B+?
3. I still do not know where to put my volt meter probes for that 15V check up.
4. 'keeping BL ones' - what is BL..?
You enlightened me already a little...this is what the ZEN is all about? 🙂
I can use 23V rail but what I have is 32V power source and I hope to use this for F6.
Thanks, Chul
F5 front end is completely different animal , where drain resistors are drain resistors
here , same as in F4 and everywhere you find same buffer stage , buffer is a buffer , and any resistance in drain is having different function (comparing to F5)
so , 15V ......... simple - to keep JFet dissipation at bay
practically anything from 12 to 18 volts , if keeping BL ones in 90-100% Idss area (so, with none or small source resistors)
edit - it's always important to account needed swing of input (thus buffer output , too) in broader picture and construction of buffer
in this case , where swing demand is in sane area , you can put +/- protective envelope around JFet buffer in pretty simple manner ;
in case of F4 , for instance , where both input and buffer output must fly almost from rail to rail , you can't use RC cells to keep buffer drains at bay , simply because sources can't outrun/outfly drains and swing will be limited to at bay level
however , some clever (and still simple) cascoding of drains is way to go , simply allowing cascodes to fold around rails (upper one around positive and lower one around negative ) thus allowing buffer output to swing (almost ) rail to rail
easy to see either in vivo or in LTSpice sim
take F4 V2. (that one with few resistors and uni-diodes added around FE) as example ; think about each part role
Let me put here what I (think I) understood...
1. So the front end of F5 is not a buffer.. the drain resistor is a load resistor and direct coupled to power FET?
2. RC.. you mean simple RC as in a tube circuit, decoupling driver stage from main B+?
3. I still do not know where to put my volt meter probes for that 15V check up.
4. 'keeping BL ones' - what is BL..?
You enlightened me already a little...this is what the ZEN is all about? 🙂
I can use 23V rail but what I have is 32V power source and I hope to use this for F6.
Thanks, Chul
Matched Jfet back in stock.
Can these be used in the F6 build? I understood the F6 was a Mosfet amp... excuse my ignorance 🙂
If not what are the best transistors to buy for this build?
I am based in EU and in the process of ordering everything for an F6. I am planning to get the F6 kit, univ psu v3, keratherm insulators, (and potentially the matched grade B Jfets) from the DIYaudio store, the chassis from hifi2000 in italy, and rest of the parts locally. Does anyone know a good EU based online store for the various parts?
Last edited:
I would recommend h_a for supplying the well matched output mosfets.
Hifi Collective in the UK has a good range of "Audio grade" components for the rest. Also Farnell;Rs or Conrad.
Hifi Collective in the UK has a good range of "Audio grade" components for the rest. Also Farnell;Rs or Conrad.
Last edited:
Thanks. So I should order 2 matched pairs of IRFP240 mosfet from h_a.
Could someone tell me why JFETS are being advertised in the diyaudio store as being a perfect match for a stereo F6 build? Can these be used instead of the MOSFETS? Is there a different sonic signature? Are the input and output transistors different? Cannot find information on this.
Could someone tell me why JFETS are being advertised in the diyaudio store as being a perfect match for a stereo F6 build? Can these be used instead of the MOSFETS? Is there a different sonic signature? Are the input and output transistors different? Cannot find information on this.
Thanks. So I should order 2 matched pairs of IRFP240 mosfet from h_a.
Could someone tell me why JFETS are being advertised in the diyaudio store as being a perfect match for a stereo F6 build? Can these be used instead of the MOSFETS? Is there a different sonic signature? Are the input and output transistors different? Cannot find information on this.
Inputs are the jfets and mosfets are the output devices.
Firstly a big thank you Nelson for your F6 gift and for all you do for us 🙂
This amp being my first pcb build did give me a WTF moment when it took some 20 seconds or so to bias up hehe.
For anyone that finds the amp as I initially did lacking dynamics, particularly evident on plucked guitar strings and piano where the start of the note lacked bite-sounding soft, changing R3 100R feed back resistor for something perhaps up to 1K (yet to try other lower values) and biasing up (in my case 650mv across lowered value .33R source resistors) has given me much of what I'm looking for.
Please, I'm not suggesting others should bias up beyond recommendations.
Still some tweaking to be done but I really do think Nelson has given us something special, thanks again.
I've looked at R4 18R anyone care to enlighten me?
Regards,
asx
This amp being my first pcb build did give me a WTF moment when it took some 20 seconds or so to bias up hehe.
For anyone that finds the amp as I initially did lacking dynamics, particularly evident on plucked guitar strings and piano where the start of the note lacked bite-sounding soft, changing R3 100R feed back resistor for something perhaps up to 1K (yet to try other lower values) and biasing up (in my case 650mv across lowered value .33R source resistors) has given me much of what I'm looking for.
Please, I'm not suggesting others should bias up beyond recommendations.
Still some tweaking to be done but I really do think Nelson has given us something special, thanks again.
I've looked at R4 18R anyone care to enlighten me?
Regards,
asx
I'd appreciate that..
My local time is maybe 8 hours ahead of yours so my reply to your post may be late. I am off for holiday till next Monday in a place where internet is not stable, so please take your time in 'zenlightening' me... 🙂
Chul
I'll make some sketch tomorrow , now is time to zzzzzz
My local time is maybe 8 hours ahead of yours so my reply to your post may be late. I am off for holiday till next Monday in a place where internet is not stable, so please take your time in 'zenlightening' me... 🙂
Chul
changing R3 100R feed back resistor to 1K I've looked at R4 18R anyone care to enlighten me?
The (R3/R4+1) value sets the closed loop gain. By raising R3 to 1k, there will be virtually no NFB left,
and much higher gain.
- Home
- Amplifiers
- Pass Labs
- The diyAudio Firstwatt F6