Upupa Epops said:In theory yes, with BJT also yes. But with lateral mosfets is efficiency lower (around 30 % ). It is result of my practical experience...
I was speaking about my push-pull class A with 2SC5200/2SA1943 BJTs.
For supply voltage 2 x 27V it gives 17Vrms voltage output, i.e. 36W into 8 Ohms. Quiescent current for this class A operation is 1.5A. So the total power is 54V * 1.5A = 81W, output power is 36W, and efficiency is 44%. This is also an example from praxis.
From Leach's page http://users.ece.gatech.edu/mleach/lowtim/output.html
Maybe he didn't use a big enough gate stopper resistors. My A75 uses longer leads to connect the driver to the output stage than my Leach and has no oscillation problems.
Maybe he didn't use a big enough gate stopper resistors. My A75 uses longer leads to connect the driver to the output stage than my Leach and has no oscillation problems.
Hi,
for your 100w classA you will need a minimum of 2.5A for Iq and preferably you should add on 100mA to ensure you stay out of the knee at low Ie. You will also need about 49Vdc to 50Vdc on the rails giving 255W to 260W of dissipation.
This only allows you to stay in class A for a load of 8r or more.
Most 8ohm speakers have a minimum impedance of 5.5r to 6.5r, some are worse.
Decision time:- do you want to stay in classA for the minimum impedance of your chosen 8ohm speaker?
If you choose 6r then bias needs to be increased to about 3.33A +0.1A giving a dissipation of 342W.
Either of these options will not work on a 0.5C/W sink. My guess is that you would need somewhere between 0.07 to 0.15C/W to keep the transistors inside a sensible Tc thus allowing sufficient current before hitting the derated SOAR limit. Thus arriving at my suggestion that mighty big sinks are required.
for your 100w classA you will need a minimum of 2.5A for Iq and preferably you should add on 100mA to ensure you stay out of the knee at low Ie. You will also need about 49Vdc to 50Vdc on the rails giving 255W to 260W of dissipation.
This only allows you to stay in class A for a load of 8r or more.
Most 8ohm speakers have a minimum impedance of 5.5r to 6.5r, some are worse.
Decision time:- do you want to stay in classA for the minimum impedance of your chosen 8ohm speaker?
If you choose 6r then bias needs to be increased to about 3.33A +0.1A giving a dissipation of 342W.
Either of these options will not work on a 0.5C/W sink. My guess is that you would need somewhere between 0.07 to 0.15C/W to keep the transistors inside a sensible Tc thus allowing sufficient current before hitting the derated SOAR limit. Thus arriving at my suggestion that mighty big sinks are required.
Upupa Epops said:Andrew, pushpull class A have typical efficiency 25-30 %, so dissipation will be cca 350 W. If you use 3 pair of these transistors, with summary dissipation 1.5 kW, safety coefficient will be in this case very good. Stay on earth....
PMA said:
Is it the Vgs of the mosfet that makes the efficiency better in a BJT output stage compared to the mosfet power stage?
\Jens
Dear Bob,
First, I don't familiar with A75, what is that?
Do you mean your Leach Amp is using Mosfet ?
I usually put a 120 ohm resistor with gate of each Mosfet and then pararell them. A pair of zener also put between the "Gate" and "Output".
I also make use of magnetic ringe to prevent oscillation.
Thomas
Attachments
Thomas,
The A75 is a Pass/Thagard MOSFET Class A design, see the legacy projects section of www.passdiy.com. It uses 220R gate stoppers, but no ferrite beads. I have a 100MHz scope and see no sign of oscillation.
My Leach amps are stock exactly as published on Professor Leach's boards. I may try a mosfet version using the boards Jens designed with suitable changes in the Vbe multiplier and gate stoppers. Lots of other projects in line (not all audio) first, though.
The A75 is a Pass/Thagard MOSFET Class A design, see the legacy projects section of www.passdiy.com. It uses 220R gate stoppers, but no ferrite beads. I have a 100MHz scope and see no sign of oscillation.
My Leach amps are stock exactly as published on Professor Leach's boards. I may try a mosfet version using the boards Jens designed with suitable changes in the Vbe multiplier and gate stoppers. Lots of other projects in line (not all audio) first, though.
100W classA
Hi,
My transistor numbers were a bit conservative.
Some calcs gave the following
4 pairs MJL21193/4 running on 70degC sink can drive 100W into 8ohms (Tc max=92degC). Similarly 8 pairs 2sa1943/c5200.
If you want to run 6 pairs 2sa1943/5200 the sink must be kept down to 55degC (Tc max=75degC).
All these options can drive 4ohms to full power at a phase angle of 60degrees, but drive voltages above about 20Vpk to 24Vpk will revert to classAB.
Hi,
My transistor numbers were a bit conservative.
Some calcs gave the following
4 pairs MJL21193/4 running on 70degC sink can drive 100W into 8ohms (Tc max=92degC). Similarly 8 pairs 2sa1943/c5200.
If you want to run 6 pairs 2sa1943/5200 the sink must be kept down to 55degC (Tc max=75degC).
All these options can drive 4ohms to full power at a phase angle of 60degrees, but drive voltages above about 20Vpk to 24Vpk will revert to classAB.
Ahh. You're looking for the speed of MOSFETs and output stage efficiency - enter the Simple Killer amp topology, which can produce 50W rms in Class A on 31V supplies and dissipate only 110W which is quite manageable with a 0.36C/W (300mm Conrad) heatsink for 40C rise, in a 25C ambient = 65C.
Or it's larger version with 4prs IRFP240/9240 and 'modulated bias' cascode which can do Class A 100Wrms from 43V supplies,
dissipating only 220W needing 2 x 300mm heatsinks. It's ready to go for <64V supplies.
Cheers,
Greg
Or it's larger version with 4prs IRFP240/9240 and 'modulated bias' cascode which can do Class A 100Wrms from 43V supplies,
dissipating only 220W needing 2 x 300mm heatsinks. It's ready to go for <64V supplies.
Cheers,
Greg
Thomas,
The "someone" who said about didn't like the sound of mosfet in the Leach amp was the designer himself, Professor W. Marshall Leach Jr.
I assume you have visited his web site already.
Unfortunately no, I have never tried to covert the Leach amp into Class A operation because I have no way to dissipate the heat in my house, and it would require too much change to the original design.
Have you actually owned a Leach amp yet?
Even though driver stage of Leach amp, up to Q16 and Q17, is in Class A operation, if you have not already using a Leach amp yet, I would suggest you to go for something simpler Class A design, such as those of Nelson Pass or even as what you have mentioned before, Krell.
The "someone" who said about didn't like the sound of mosfet in the Leach amp was the designer himself, Professor W. Marshall Leach Jr.
I assume you have visited his web site already.
Unfortunately no, I have never tried to covert the Leach amp into Class A operation because I have no way to dissipate the heat in my house, and it would require too much change to the original design.
Have you actually owned a Leach amp yet?
Even though driver stage of Leach amp, up to Q16 and Q17, is in Class A operation, if you have not already using a Leach amp yet, I would suggest you to go for something simpler Class A design, such as those of Nelson Pass or even as what you have mentioned before, Krell.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.