Hi friends,
I would like to build a class AB with a 10mA quiescent current from +/-100V supply, and 100mA output current into 1k load? Will a moderate gm HEXFET like irf610 /9610 (1A / V) do or should I select a low gm lateral MOSFET pair? I need good slewrate and bandwidth of 1 MHz. I can take distortion upto about 2-3% at full power.
Please suggest few devices that may be suitable. Thanks in advance.
I would like to build a class AB with a 10mA quiescent current from +/-100V supply, and 100mA output current into 1k load? Will a moderate gm HEXFET like irf610 /9610 (1A / V) do or should I select a low gm lateral MOSFET pair? I need good slewrate and bandwidth of 1 MHz. I can take distortion upto about 2-3% at full power.
Please suggest few devices that may be suitable. Thanks in advance.
Yes feedback is from point of load. MOSFET stage within loop.
But would I get the high bandwidth of 1 or 2 MHz, closed loop with gain of 50?
I think IRF610/9610 because although the gfs of these devices is near 1A/V, its lesser when Vgs near Vth (3-4V).
Am I right?
Aiming at 10mA, would I fall short of 3V (threshold) at the gates? This is my actual doubt.
But would I get the high bandwidth of 1 or 2 MHz, closed loop with gain of 50?
I think IRF610/9610 because although the gfs of these devices is near 1A/V, its lesser when Vgs near Vth (3-4V).
Am I right?
Aiming at 10mA, would I fall short of 3V (threshold) at the gates? This is my actual doubt.
Last edited:
200 Vp-p at 1 MHz is a difficult slew rate. IMO, most of the usual audio circuits you see won't do it. Combine that with 100 mA and layout without oscillation becomes difficult. Getting the devices to shut off quickly enough will be difficult. Global feedback may be difficult because of phase shift. I'd start by looking at old and new oscilloscope drive circuits and see if those can be adapted to what you want. If less than full power is needed at the high end, it will be an easier task.
I have spiced a variant that can put out 200Vp-p with low distortion.
But only tested with 1 kHz, so far.
The load is 1kOhm.
So, it is +/-100 mA out.
It was not easy. My third attempt however was successful.
In a real circuit, as has been said, you must have perfect PCB.
To avoid parasitic capacitance and such.
We are talking here RF, radio frequencies.
Anyway, the upper bandwidth of my amplifier in SPICE
is more than 10 MHz
at a gain of x100.
It is current feedback using IRF610/9610 both for VAS and for OUTPUT.
Idle bias is 10mA.
Supply is +/-110 Volt.
So, one slight problem is that IRF610/9610 are specified 200 Vds
There are HEXFET that takes more, I know.
I post it here.
Maybe it can give you some ideas.
Attachments
look at output Cds - at 1 MHz, 100 V you'll be puting ~ 50-100 mA into the Mosfet device output C alone
possibly some of the VAS/predriver bjt would be better
even if you parallel 2x 2SC3503/1381 you will be way ahead of Mosfets - but SOA would need to be watched - more info about the application, signal, load would help
possibly some of the VAS/predriver bjt would be better
even if you parallel 2x 2SC3503/1381 you will be way ahead of Mosfets - but SOA would need to be watched - more info about the application, signal, load would help
Last edited:
it is hard guesing whats important without further application information (strong hint)
I assumed anyone specifying 10 mA Class AB wouldn't want to waste as much current and more power driving the Mosfet parasitic Cds than they are acutally delivering to the load at 1 MHz - why wouldn't you use Class
A from the start if that much inefficiency were not a problem
I assumed anyone specifying 10 mA Class AB wouldn't want to waste as much current and more power driving the Mosfet parasitic Cds than they are acutally delivering to the load at 1 MHz - why wouldn't you use Class
A from the start if that much inefficiency were not a problem
Last edited:
Agree that BJT may be a better option. The MOSFETs will follow the gate if you can drag/force/drive/scare it with sufficiently low impedance. That can be more complicated than it looks because the switching of the device will couple back through the gate, plus errors due to PCB impedance. If you can't get either type of device turned off fast enough, shorted supplies and damage are inevitable.
CH
CH
True that feedback resistor will effect a current feedback design.
But this is no issue.
You chose a resistor and that is that.
I am sure you HAVE TO use current feedback.
For a bandwidth upto 10 MHz, at least.
Specs, specs, specs... DC precision, noise, small-signal bandwitdh, capacitive load driving capability, source impedance--just to name a few.
There are commercial modules approaching the performance you're asking: PA107U_E.pdf, PA85U_O.pdf and PA98U_M.pdf
Samuel
There are commercial modules approaching the performance you're asking: PA107U_E.pdf, PA85U_O.pdf and PA98U_M.pdf
Samuel
Do you have any idea whre to buy those Apex OP-AMPs (either in CH, AT or IT)?
thanks
effebi
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.