These are Bipolar Junction Transistors, contrary to those Metal Oxide Silicon Field Effect Transistors...
Quasi Version
Someone wanted a quasi version, so I'm having a go....need to pass the time while I wait for the parts!
Sims quite well, even with vertical fets (IRFP240) and very cheap transistors. Some resistor values may need tweaking though. Also there is some potential for themal runaway if the driver transistors get hot. Not sure if that needs to be dealt with or not.
Why are my gate drive waveforms so strange though (top fet gate is top trace, bottom fet gate is bottom trace)?
The beauty of this circuit (if it can be made to work) is that only the main fets need heatsinking. Every other active device has a very easy job and cheap signal transistors can be used.
Current through input stage and VAS can be significantly reduced due to high gain devices at gates of fet (though you are encouraged to experiment).
Even with bootstrap VAS and my weird gate waveforms, simulated distortion is very low and all even order.
This is the first quasi circuit I have tried so I look forward to your help getting it sorted. It's potentially a very easy circuit to build.
Someone wanted a quasi version, so I'm having a go....need to pass the time while I wait for the parts!
Sims quite well, even with vertical fets (IRFP240) and very cheap transistors. Some resistor values may need tweaking though. Also there is some potential for themal runaway if the driver transistors get hot. Not sure if that needs to be dealt with or not.
Why are my gate drive waveforms so strange though (top fet gate is top trace, bottom fet gate is bottom trace)?
The beauty of this circuit (if it can be made to work) is that only the main fets need heatsinking. Every other active device has a very easy job and cheap signal transistors can be used.
Current through input stage and VAS can be significantly reduced due to high gain devices at gates of fet (though you are encouraged to experiment).
Even with bootstrap VAS and my weird gate waveforms, simulated distortion is very low and all even order.
This is the first quasi circuit I have tried so I look forward to your help getting it sorted. It's potentially a very easy circuit to build.
Attachments
You feed the Gates only ~2mA
For big vertikal HEXFET this is on the low side.
I want at least 10mA (or 20mA)
There is a charge current needed into those gates.
This is why driving them in Class A when there is a Gate current all the time
is a good way to get HEXFET to operate.
Say the momentarily gate takes 1mA just for charging.
There will not be a very smooth turning on.
This gives a strange waveform and increased distortion.
Why you do not get so high distortion
is because of effective Feedback correction.
For big vertikal HEXFET this is on the low side.
I want at least 10mA (or 20mA)
There is a charge current needed into those gates.
This is why driving them in Class A when there is a Gate current all the time
is a good way to get HEXFET to operate.
Say the momentarily gate takes 1mA just for charging.
There will not be a very smooth turning on.
This gives a strange waveform and increased distortion.
Why you do not get so high distortion
is because of effective Feedback correction.
GEirin,
Yes, I think that transistor would work, though it might be overkill for the purpose. You will need to tweak the input and vas stage current to suit it.
However, before constructing we need to identify the problems lineup mentioned in post #189. One improvement is to increase R14 and R15 to force more of the transistor output into the fet gate, but then the gate discharge is impeded. A careful compromise is needed.
Lineup, any thoughts on how to improve this? Also, can you please explain how you determined the gate current is limited to 2mA?
Thanks,
Greg.
Yes, I think that transistor would work, though it might be overkill for the purpose. You will need to tweak the input and vas stage current to suit it.
However, before constructing we need to identify the problems lineup mentioned in post #189. One improvement is to increase R14 and R15 to force more of the transistor output into the fet gate, but then the gate discharge is impeded. A careful compromise is needed.
Lineup, any thoughts on how to improve this? Also, can you please explain how you determined the gate current is limited to 2mA?
Thanks,
Greg.
I think he means: "I'll wait for your results..."
(Excuse me, not intending to be unkind)
Excuse me friends.
In swordfishy Quasi circuit Q1 and Q2 runs at 5 mA (4V/750 Ohm)
This is better than 2mA, but I would like to make it 10mA.
We can change R14/R15 to 470 Ohm.
This makes more current pass the need of the Gate.
This is for HEXFETs like IRFP240/IRFP9240
For laterals the circuit should be okay, as it is.
I say swordfishys circuit is good enough.
Nothing wrong I can find.
In swordfishy Quasi circuit Q1 and Q2 runs at 5 mA (4V/750 Ohm)
This is better than 2mA, but I would like to make it 10mA.
We can change R14/R15 to 470 Ohm.
This makes more current pass the need of the Gate.
This is for HEXFETs like IRFP240/IRFP9240
For laterals the circuit should be okay, as it is.
I say swordfishys circuit is good enough.
Nothing wrong I can find.
Yes. But then we should make the VAS use not only resistor+potentiometer.
We should use one VBE-multiplier.
One transistor like BD139 and 2 resistors+potentiometer to set bias.
But this topic is more about using JFET input and MOSFETs
Yes. But this has been a process.
Of learning for me, too.
We are way from the first circuit I published here in topic.
But I say, especially the CCS to achieve stability,
makes it a more reliable and better circuit to build.
Build and not only simulate.
sandyhooker
Look at this post a bit up on page
http://www.diyaudio.com/forums/soli...vas-lateral-output-perfect-4.html#post2562232
Replace 2SK30ATM with one 2SK170BL
AKSA likes this version.
swordfishy is building this amplifier.
The final version depends a bit on his results.
We need some real world testing.
Look at this post a bit up on page
http://www.diyaudio.com/forums/soli...vas-lateral-output-perfect-4.html#post2562232
Replace 2SK30ATM with one 2SK170BL
AKSA likes this version.
swordfishy is building this amplifier.
The final version depends a bit on his results.
We need some real world testing.
danspy
If you build your Fetzilla, you know, I and swordfishy and maybe AKSA
will try to assist you, for best result. If you need.
Congratulation.
You have 8 2SK1058/2SJ162.
Great
I incline towards the quasi version too......
Why, when we know its distortion is inferior?
Because the added distortion from the asymmetric output stage is essentially even low order. This sweetens the amp, makes it more like a tube amp.
Clearly this is not a 'straight wire with gain' and can be derided for failing to deliver text book, low distortion.
Years of building, listening, and observing the audiophile taste have shown me pretty clearly that people talk dry but drink sweet.
People like a bit of warmth to the music, that's why tube amps are so popular. It doesn't have to be much, either, around 0.02% H2 makes all the difference.
Flame suit on.....!!
Hugh
Why, when we know its distortion is inferior?
Because the added distortion from the asymmetric output stage is essentially even low order. This sweetens the amp, makes it more like a tube amp.
Clearly this is not a 'straight wire with gain' and can be derided for failing to deliver text book, low distortion.
Years of building, listening, and observing the audiophile taste have shown me pretty clearly that people talk dry but drink sweet.
People like a bit of warmth to the music, that's why tube amps are so popular. It doesn't have to be much, either, around 0.02% H2 makes all the difference.
Flame suit on.....!!
Hugh
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.