Try 10k series resistor from rail plus 10k pot to ground. We want the wiper voltage to go from zero to 10v max. Connect 10uF electrolytic to wiper and ground to smooth things out. Then connect that node to the gate via a 33k to 47k resistor. There should be very little current flow to the JFET gate. Disconnect the 100k from gate to ground. All this is in the gate side of the inout coupling cap. Please see sketch below.
Attachments
Last edited:
I added a power resistor this evening - wanted to make sure this worked
Had it running for about an hour
The bias is only 1.3A
The jfet gets to about 35 deg C , the resistor housing is just shy of 50 deg c and the heatsinks matched the jfet at 38 deg c
sounded good - driven directly from dac so no major volume but enough to hear the quality of sound and to make sure no gremlins
..dB
Had it running for about an hour
The bias is only 1.3A
The jfet gets to about 35 deg C , the resistor housing is just shy of 50 deg c and the heatsinks matched the jfet at 38 deg c
sounded good - driven directly from dac so no major volume but enough to hear the quality of sound and to make sure no gremlins
..dB
dB,
What did you change in order to get the lower 1.3A bias from last power up when the bias was very high?
What did you change in order to get the lower 1.3A bias from last power up when the bias was very high?
yes - only source resistor is now bolted to the heatsink - thanks to this board design, that was a very easy modification to make
Yes, good move to cover all options.
12v or 24v input?
Nice job, The heat is very manageable now 😉
Running with lower bias, did you measure voltage before and after capMx? Curious if its still a 6v drop?
12v or 24v input?
Nice job, The heat is very manageable now 😉
Running with lower bias, did you measure voltage before and after capMx? Curious if its still a 6v drop?
Last edited:
24V into the capmx, 19.7V at the drain
what is curious , during all the manipulation, the gate connection was interrupted on one channel and when I fired it up the Capmx dropped 4V , so it is working as designed but with the larger current draw, it drops further
what is curious , during all the manipulation, the gate connection was interrupted on one channel and when I fired it up the Capmx dropped 4V , so it is working as designed but with the larger current draw, it drops further
Last edited:
Nice! You answered my question before I asked 😀
4v drop is more in line.
Feed it a from a stout preamp and have a good listen!
4v drop is more in line.
Feed it a from a stout preamp and have a good listen!
I need to cobble up a preamp to use on my test bench = many many amps, not so many preamps 🙂
thanks or the guidance thus far
thanks or the guidance thus far
Great work dBel. If you need a stout preamp, look no further than the Aksa Lender. Didn't you get one of those PCBs already?
The other option is to use an OPA454 M2X/Yarra front end like I am using on the LuFo.
The other option is to use an OPA454 M2X/Yarra front end like I am using on the LuFo.
Good stuff dBel! Glad your getting some sound out of yours now and things are stable. Looking forward to hearing what you think after getting a preamp in line.
After testing X’s mod from post #102 I think I’m going to stick to the power resistor route as a means of controlling current. The mod seemed to decrease the current, but not nearly as much as the power resistor.
I ended up soldering (3) 25 watt aluminum resistors in series to get 3 ohms total. They are attached to the chassis heatsink. Total draw for that channel is now 42 watts. Drop at the CapMx is about 4.9v. I think one factor of how much drop there is depends on how stout the SMPS is. I was getting less drop with a beefier 24v SMPS dialed down to 19v. Also, I have the ERSE inductors in the circuit. After 30 minutes of use the inductors were at 40 deg C. So the ERSE inductors do seem to be good to go for this circuit. Ultimately I may play around with the resistance and supply voltage a bit more, but I think the original schematic is the way to go if the power resistors are the correct value and properly heatsinked.
The sound with the AKSA Lender pre is very good. My AKSA build is only using a 24v SMPS. I supposed I could eke out a few more dB with more volts on the AKSA Lender, X? What would it take to drive this little amp to clipping?
After testing X’s mod from post #102 I think I’m going to stick to the power resistor route as a means of controlling current. The mod seemed to decrease the current, but not nearly as much as the power resistor.
I ended up soldering (3) 25 watt aluminum resistors in series to get 3 ohms total. They are attached to the chassis heatsink. Total draw for that channel is now 42 watts. Drop at the CapMx is about 4.9v. I think one factor of how much drop there is depends on how stout the SMPS is. I was getting less drop with a beefier 24v SMPS dialed down to 19v. Also, I have the ERSE inductors in the circuit. After 30 minutes of use the inductors were at 40 deg C. So the ERSE inductors do seem to be good to go for this circuit. Ultimately I may play around with the resistance and supply voltage a bit more, but I think the original schematic is the way to go if the power resistors are the correct value and properly heatsinked.
The sound with the AKSA Lender pre is very good. My AKSA build is only using a 24v SMPS. I supposed I could eke out a few more dB with more volts on the AKSA Lender, X? What would it take to drive this little amp to clipping?
Last edited:
Good stuff chaps.
X, are you planning to amend the PCB to have the option to use your alternative biasing in post #102 as well as the 'power resistor' approach - that would be very handy.
X, are you planning to amend the PCB to have the option to use your alternative biasing in post #102 as well as the 'power resistor' approach - that would be very handy.
Good work Jwjarch, thanks for trying the gate bias mod. I am surprised it was not able to reduce the current to your satisfaction. So I am not sure if it’s worth modifying the layout to do this. The whole idea behind using resistors was to keep it simple. But it sure burns off a lot of heat and we are wasting headroom. Now I am not so sure skipping cascode was worth it. Maybe we should keep that and have more headroom, more power and no burning heat from a resistor.
No problem with re-introducing the cascode from my perspective - to me LuFo Lite is more about a low power derivative rather than a hairshirt solution. That said, is jwj's experience with the post 102 mod just suggesting the values aren't quite right to achieve the target bias?
Last edited:
Yeah, I was expecting more of a decrease in current. The only way I could decrease was by adjusting the 33K resistor to a lower value. I was just experimenting and ended up at 1.5K. But even this didn't get me much below a 100 watt draw on the one channel. Most of the current was still going to the inductors and heating them up a ton. I quadruple checked my wiring against the schematic sketch too. IMO, the power resistor setup has manageable heat/power loss if done correctly.
Overall, I think it depends on your use case, X. For a headphone amp it's a great design. The cascode might be the way to go for speaker use. Either way, a high sensitivity speaker is a must. It may also just make an amazing tweeter amp. I'm hoping to get the second channel built very soon and can give some better feedback on that.
Overall, I think it depends on your use case, X. For a headphone amp it's a great design. The cascode might be the way to go for speaker use. Either way, a high sensitivity speaker is a must. It may also just make an amazing tweeter amp. I'm hoping to get the second channel built very soon and can give some better feedback on that.
You need a negative bias voltage to decrease current
LU is like a jfet it is almost fully conducting with 0V on the gate
LU is like a jfet it is almost fully conducting with 0V on the gate
Yes, that the conclusion I had arrived at, like Ben Mah's choke loaded follower;
25W Single Ended Hammond 193V Choke Loaded 2SK180 L'Amp
25W Single Ended Hammond 193V Choke Loaded 2SK180 L'Amp