This amplifier was inpsired by conversation with John Curl in a ‘Blowtorch thread’, as well as by his thoughts ( www.parasound.com/pdfs/JCinterview.pdf ).
The aim was to build a reliable, solid, good sounding amplifier.
Topology is complementary differential, with JFET input, MOSFET cascode and VAS, and BJT multiple pair output stage. The circuit shown might be an inspiration for DIYers in their designs.
A functional sample was build, the circuit is very stable and rugged.
The project is not a commercial project, no PCBs or kits will be sold.
The aim was to build a reliable, solid, good sounding amplifier.
Topology is complementary differential, with JFET input, MOSFET cascode and VAS, and BJT multiple pair output stage. The circuit shown might be an inspiration for DIYers in their designs.
A functional sample was build, the circuit is very stable and rugged.
The project is not a commercial project, no PCBs or kits will be sold.
Attachments
Hi Pavel , before being submerged from requests of PCB and kits that you stated won't be made, can I ask you for completeness the values that are not written of some resistors in the differential stage? And the type of mosfets in the VAS stage.Eventually the supply rails voltage.
Hi pico,
well - there are good reasons not to show them.
2SK170BL/2SJ74BL are matched to within 0.5mA Idss, and Idss is about 9.5 - 10mA. You make a choice of R1 by measurement then, to get Id = 8mA approx. The R5 and R7 are chosen again by measurement, to get appropriate current through VAS Mosfets (check voltage drop across R6, voltage corresponds to Id of input stage + VAS current). You may get something like R1=7R and R5, R7 = 750R for 2SK413/2SJ118 mosfets. Anyway, measurement is necessary. The DIYer should be at least average experienced building this amp. Rails - start with +/-30V and may increase to +/-50V if operating safely.
Last, but not least - voltage at gates of M3,M4 and M5,M6 is set to 13V approx by dividers R75/R76 and R77/R78. Do not forget this, not to thermally burn input JFETs. Before increasing supply rails, consider this.
well - there are good reasons not to show them.
2SK170BL/2SJ74BL are matched to within 0.5mA Idss, and Idss is about 9.5 - 10mA. You make a choice of R1 by measurement then, to get Id = 8mA approx. The R5 and R7 are chosen again by measurement, to get appropriate current through VAS Mosfets (check voltage drop across R6, voltage corresponds to Id of input stage + VAS current). You may get something like R1=7R and R5, R7 = 750R for 2SK413/2SJ118 mosfets. Anyway, measurement is necessary. The DIYer should be at least average experienced building this amp. Rails - start with +/-30V and may increase to +/-50V if operating safely.
Last, but not least - voltage at gates of M3,M4 and M5,M6 is set to 13V approx by dividers R75/R76 and R77/R78. Do not forget this, not to thermally burn input JFETs. Before increasing supply rails, consider this.
Last edited:
this circuit is similar to Lender's circuit - go to
http://www.diyaudio.com/forums/soli...age-large-voltage-swings-lenders-circuit.html
An additional similar circuit I have seen from JLH, but I forget the title and the location here in this forum
Is there are a possible to work without selecting? I. e., I will use jFETs with a very wide range of Idss and a wide range of Vgs threshold voltage.
As I know, it is possible by tube amplifiers, i. e. each tube has in this case their individual bias spindle potentiometer.
Then I can selected, too. But now I select only to identical curve shape (curve passage) of the transfer curve, much more important as Idss and Vgs.
Last edited:
That would put me aside in building such thing , since I don't own any laboratory equipment ...At least my nick would suggest that I "feel" more on cabinet construction rather than elaborated high power amps ! But , since the gramophone is not my style ,the one needs the other 😀
Thanks
Bye
Nonsense. Maximum undistorted output voltage is 6V below supply rail (for 4ohm load). Federmann, I built and measured the circuit. MOSFET VAS has output from drains, and gates are driven from voltage drop across R5 and R7. VAS happily swings almost to supply rails. Please do not comment on circuits you do not understand.
Attachments
Last edited:
You need digital multimeter to measure voltages accross 5 resistors (my 2nd or 3rd post in this thread), and to make a choice of JFETs. Oscilloscope is recommended for any amplifier builder.
It's nice to see real measurements for a change 🙂 is it still that stable with small value caps too such as 0.1mf
Pavel,
R69 and R70. Would there be any advantage not connecting these to the output line but cross coupling as advocated by D Self to help remove charge storage effects and help speed up the driver/output combination ? I wonder if you have any views or reasons for and against.
Pavel,
R69 and R70. Would there be any advantage not connecting these to the output line but cross coupling as advocated by D Self to help remove charge storage effects and help speed up the driver/output combination ? I wonder if you have any views or reasons for and against.
Nice amp PMA, I bet it sounds good. Strictly by listening, not measurements, how do you find MOSFET vs bipolar as VAS and cascode devices?
I have been considering that before I have chosen this solution. The main reason was that this shown here gave smoother distortion residual at high frequencies. Self's type II bias gives a bit lower distortion, but more high order components. That's the main reason why I made it this way (type I by Self).
Federmann, you complained about 10V swing below supply rail. I am telling you that it is 6V (measured, not simulated - forget JFET and MOSFET models). 6V for no limitation, and for distortion below 0.1%.
- Status
- Not open for further replies.