JLH10W- MOSFET output

Alex Nikitin@1992,2003 x-pro come in
attachment.php
 
Last edited:
Update
The asymmetry seems to be worse if the quiescent current is too high.
By reducing the current the simulation showed a better balance! (1 to 1.2A for 10W)...........Premature clipping (internally) not seen when power supply at 40V and driving 16W into 8 ohm load where Iq=1.5A.
Its good to see a departure from the old TO3 BJT formula. However, it does seem rather high dissipation for TO220 power transistors and suggests a hefty heatsink or two with best quality insulation material. Otherwise, its on my list for an original style matrix board build soon. I plan to use somewhat higher RDSS types IRF530 (not the N suffix variety). Any comment?
 
I originally proposed this as a 10W output, and the quiescent current for that is about 1A from 30V, so not so bad for a TO-220.
In general I prefer TO-3 for high powers; as they usually have better thermal characteristics, but recent TO-220's are improving. IR have largely done away with metal can FETs though.
Regarding the choice of device, I recommend higher current devices not for the current handling but the higher gain.
 
I looked up Nelson's designs. As I thought, he only used a single input device (FET) and the output pair (the other design you mention is the same but with parallel outputs).

From my simulations the original configuration (bipolar input and driver) has lower distortion.
 
OldDIY -
You've shown it's been done before. None quite match my design, but I'm sure if you keep digging you could well find it.
The circuit in your post #34 is closest, but I suggest a bootstrap is better unless you take a CCS transistor to a higher supply voltage.
The circuit in post #35 is not how I implemented driver transistors but does not meet my objective of keeping to 4 transistors.
 
shaan
Some observations:
1. In the BJT version the bandwidth is ~2MHz, in the Lateral one it is ~5.5MHz(expecting oscillation and gate stoppers).
2. The rise time and fall time are respectively one sixth and half of that of the BJT version.
3. The BJT version had 0.2% distortion at 10VPP into 8ohm; the Lateral has 0.16%, with decreased higher order harmonics.
4. With +-20V supply, the BJT version has maximum voltage swing of 33V(+-16.5V), the Lateral has 30V(+-15V).
5. Clipping seems as good as the BJT one.
6. Maximum current output. This is where I am surprised. The circuit seems to be able to deliver current more than 2.5 times Iq. Initially I thought I made a mistake. But after many hours of experimenting no errors were found. I ran the sim at 1.6A of Iq, both the FETs can reach 4A when the load is down to 2ohm. I expected assymetrical clipping, as usual with the BJT version. At 1.6A Iq, the voltage at the gates of the FETs is ~2.7V.
JLH 10 Watt class A amplifier
257161d1325069042-jlh-10-watt-class-amplifier-nudoz-png