So... pushed by the very nice sounding moamps' HGC I built JFET based version of a HGC. The buffer and its current source are cascoded, JFETs are BF245C and the rest is same as in Mr. Rasmussen's HGC. But... that's not it! Although the sound is (to my ears) better (not so agressive) than GC, there is no "air" and "naturallity" I've heard with moamps' GC. In fact; I think that mentioned improvement is due more to HF filter than to the implemented buffer. 
Attachments
Buffer
You have to try a high gm jfets like 2sk370 ( V or GR ) or 2sk389.If you want to have 0V dc at output add a source resistor to T2, call it Rs. Rs = R5. And in my opinion a cascoded circuits isn't the best sounding ones. You could try put volume control after buffer. A lot of stuff about jfets design :
www.borbelyaudio.com
You could try an tube buffer. There is harder mechanical assembly, but only tubes ( low power triodes ) have stright transfer curve, low input capacitance. And their more beauti than plastic jfets.
You have to try a high gm jfets like 2sk370 ( V or GR ) or 2sk389.If you want to have 0V dc at output add a source resistor to T2, call it Rs. Rs = R5. And in my opinion a cascoded circuits isn't the best sounding ones. You could try put volume control after buffer. A lot of stuff about jfets design :
www.borbelyaudio.com
You could try an tube buffer. There is harder mechanical assembly, but only tubes ( low power triodes ) have stright transfer curve, low input capacitance. And their more beauti than plastic jfets.
jfet 0 temmperature coefficient
It is possible to get 0 temperature drift, especially when you use a monlithic pair ( 2sk389 ).
There is a formula for 0 temperature coefficient.
Vgs = Vp + 0.63V ( where Vp is the gate pinch-off voltage ).
Vp - you need to measure it for every one jfet.
www.borbelyaudio.com/ae599bor.pdf
It is possible to get 0 temperature drift, especially when you use a monlithic pair ( 2sk389 ).
There is a formula for 0 temperature coefficient.
Vgs = Vp + 0.63V ( where Vp is the gate pinch-off voltage ).
Vp - you need to measure it for every one jfet.
www.borbelyaudio.com/ae599bor.pdf
Hi Vuki,
Experience I had buffering ‘clone is a bit different. But it seems our experiences with bare GC are different too… I have not found that the original (bare) GC’s sound is aggressive. However, I have encountered that the aggressiveness can occur with added buffer and for some time I had problems with this.
I am not sure I can, looking at your schematics, recognize where possible problem can come from, but your buffer’s cascode doesn’t look good to me. And I definitely would not use 470uF caps after the regs, definitely much smaller should go here. And finally, though I can not comment your choice regarding the JFETs, it wouldn’t be bad, before you give up, to check some device that is known as good sounding device.
Also, if you did not do that till now, I would humbly suggest you to check my GC page.
Pedja
Experience I had buffering ‘clone is a bit different. But it seems our experiences with bare GC are different too… I have not found that the original (bare) GC’s sound is aggressive. However, I have encountered that the aggressiveness can occur with added buffer and for some time I had problems with this.
I am not sure I can, looking at your schematics, recognize where possible problem can come from, but your buffer’s cascode doesn’t look good to me. And I definitely would not use 470uF caps after the regs, definitely much smaller should go here. And finally, though I can not comment your choice regarding the JFETs, it wouldn’t be bad, before you give up, to check some device that is known as good sounding device.
Also, if you did not do that till now, I would humbly suggest you to check my GC page.
Pedja
Hi Pedja,
thanks for the advice. I don't think that GC could generally be described as agressive, but compared to 300B SE or Mini-Aleph it was, and I found that tubed HGC sounded much closer to those amps and much better than bare GC. JFET buffer did help the sound somewhat, but didn't get it close to tubed version. So... I will probably try different JFET but maybe you should try tubed buffer
thanks for the advice. I don't think that GC could generally be described as agressive, but compared to 300B SE or Mini-Aleph it was, and I found that tubed HGC sounded much closer to those amps and much better than bare GC. JFET buffer did help the sound somewhat, but didn't get it close to tubed version. So... I will probably try different JFET but maybe you should try tubed buffer
Vuki,
I put your buffer into spice (with appropriate BF245C model now) and definitely, it won’t work the way you made it. Such cascode is ok for current source, but not for source follower. Your signal JFET has working voltage close to 1V, which is totally unacceptable. In reality, since tolerances in manufacturing of this type of device are not tight, your voltages should not be the same as shown here, but they’ll not differ much.
Possible options are:
1. Refer the gate of T1 (J3 in my schematic) both to positive rail and to output thus making voltage divider, in which case the later resistor will need bypass cap;
2. Put 9V battery in series with R1 (in your schematic, in mine it is R10);
3. Simply remove that buffer’s cascode (for start).
Pedja
I put your buffer into spice (with appropriate BF245C model now) and definitely, it won’t work the way you made it. Such cascode is ok for current source, but not for source follower. Your signal JFET has working voltage close to 1V, which is totally unacceptable. In reality, since tolerances in manufacturing of this type of device are not tight, your voltages should not be the same as shown here, but they’ll not differ much.
Possible options are:
1. Refer the gate of T1 (J3 in my schematic) both to positive rail and to output thus making voltage divider, in which case the later resistor will need bypass cap;
2. Put 9V battery in series with R1 (in your schematic, in mine it is R10);
3. Simply remove that buffer’s cascode (for start).
Pedja
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.