power rail calculation

[lt cdr data]

I have a schematic in front of me.

closed loop gain from feedback resistors is calculated at

52.59, which is 34.42db.

input sensitivity is quoted at .775 mv.

so that means, full output volts is 40.76 volts rms.

into 8 ohms, that gives 207.64 watts

peak volts are thus 40.76 x 1.41 = 57.6 volts.


output transistors are 2sc3858 with 0.22 ohm emitter resistors.

what happens if those emitter resistors are altered, changing local feedback?

and finally, what do the power rails need to be for this amplifier please?

thanks

[fizzard]

I would go with 70 volts or so. Simulate and see how it works anyways. It's easier to change in simulation than real life.

Is 2SC3858 the only output device? In other words, is it quasi-complementary?

You can effectively lower the emitter resistance by paralleling more output devices. By my calculation you're losing 1.6 V to your emitter resistor (based on two output devices, one per rail). You could cut this to 0.8 V by using four output devices, 0.5 V by using 6 output devices, etc.

Depending on how many stages of current gain there are you'll lose about 3 volts of voltage headroom for each stage, plus the emitter resistor drop.

I wouldn't lower the emitter resistor values unless your match your transistors very well.

[lt cdr data]

Its got 4 pairs of output devices, in a triple darlington configuration.

I was also wondering, power into 4 ohms is given as 330 watts.

bias is 45ma per transistor.

so if it has, and I think it has, a 1kva transformer, what's limiting the power( current) to the speakers, at 4 ohms?

[fizzard]

Originally posted by lt cdr data
Its got 4 pairs of output devices, in a triple darlington configuration.

Yeah so around 70 V, maybe more, would work fine.

Originally posted by lt cdr data
Its got 4 pairs of output devices, in a triple darlington configuration.

I was also wondering, power into 4 ohms is given as 330 watts.

bias is 45ma per transistor.

so if it has, and I think it has, a 1kva transformer, what's limiting the power( current) to the speakers, at 4 ohms?

The saturation voltage (the voltage "eaten" by the output transistors) goes up as the output current goes up.

[unclejed613]

definitely keep the emitter resistors as-is, especially with 45ma of bias current per device. while this isn't real high, it is on the high side of average. most amps leave the factory with output device bias currents between 15-30ma, unless they use the distortion analyzer method of setting the bias, in which case the output bias currents average between 30-60ma leaving the factory. in addition to providing emitter degeneration (aka "local" feedback), the emitter resistors keep paralleled output devices from current "hogging", where one transistor (the one with the highest beta) will sink or source most of the current in the output stage, leading usually to it's early demise. this happens in output sections when the transistors have no emitter resistors to balance the currents between them. parallel strings of transistors also decrease the output resistance, but beware you don't overload your driver transistors. each paralleled output device soaks some current from the driver, so you need more robust driver transistors as you add parallel outputs. that's why some high power PA amps use the same type transistor as a driver as is used in the output section (i.e. 10 parallel sets of mj15024/25 outputs and the driver transistors are also mj15024/25 transistors).

[AndrewT]

Hi,
200W into 8r0 should theoretically give near 400W into 4r0.
A 600VA 50+50Vac (45Vac is a little low) transformer should allow this amp to deliver full power.
You may find a slight increase in performance by using a bigger transformer.
The sinewave current into a resistive load will be around 14.1Apk into 4r0.
Transient current into a reactive load can be upto three times this i.e. approaching 42Apk (10Apk per output pair). Design for this level of current into a 4ohm reactive load. The output stage should also be designed to supply full power current for a few seconds into a 2r0 load. (28.2Apk=800W into 2r0).
Fuse the supply rails with F8A.
Use a soft start on the primary side and fuse it with T3A.
I recommend +-40mF smoothing to ensure good low frequency bass power with minimal rail ripple/sag.

Optimal ClassAB bias for 0r22 emitter resistors is about 15mV to 25mV.
This amounts a bias current of 68mA to 113mA/pair giving a total output bias of 275mA to 450mA. This will need a very big heatsink to keep those devices cool. The voltage amp and drivers take a further quiescent current of 10mA to 100mA depending on design.