Triple EF BJT power stage.

[roender]

Hello all,

What do you think about this final EF stage?
What should I cut or keep? The main goal is stability, high input impedance and low distosion at 20Khz

[wojtek5001]

What is R1 and R11. They will dissipate lot power. Isn't better to simply connect emiters by resistor, like in many others amp?

[roender]

Originally posted by wojtek5001
What is R1 and R11. They will dissipate lot power. Isn't better to simply connect emiters by resistor, like in many others amp?

It is a very simple way to discharge the base junction at HF and to prevent negative base impedance

[wojtek5001]

I think, that resistor between emiters Q1 and Q3 (optionaly pararell with cap) will do the same, like in Douglas Self amp. You should also insert some small resistors in series of output transistor bases (especialy when using cap mentioned in previous brackets )

[roender]

The cap trick don't work at high freq. and base stopers increase phase shift

[AndrewT]

Hi,
the Cordell thread seems to have confirmed that BJT output stages need about 15mV to 25mV across each of the emitter resistors to stay at or near optimum bias to help hide crossover distortion.

They also agree it is better to err on the high side rather than the low side (variations with junction/ambient temperatures) to stay above the 15mV lower limit for all operating conditions.

Your 100mA and 0r1 put you too far below the 15mV.

If you follow their advice then either bigger Re or some big sinks to dissipate more Iq.

Parallel pairs with low value Re will require more precise matching than with higher Re values.

I don't follow the diamond arrangemment of the emitter resistors on the drivers. I have to leave that to the experts.

[lineup]

Originally posted by wojtek5001
What is R1 and R11. They will dissipate lot power.
Isn't better to simply connect emiters by resistor, like in many others amp?

I have to agree.
So if this is good,
why don't you replace R5 with 2 resistors to supply lines?

Now, if it is good in place,
it is most probably good in the other place.
So:
Replace R1 / R11 with one resistor between emitters.
Just like R5.

The current variations caused by R1 and R11 which are loading the drivers
can NOT be good.
Imagine +/-20 Volt across those resistors ......

Besides, I think PSRR (supply rejection ratio) will be worse with those resistors.


Questions:
- What Supply voltage, you had in mind?
- What Output power you had in mind,
- And into What Load?


Regards
lineup

[CBS240]

Hi

What kind of front end is biasing this circuit?
It seems to me that R1 & R11 would cause Q1 & Q3 to conduct asymetrical current(referenced to audio voltage) and the input stage would have to correct for this extra error. If these resistors were replaced with CCS's, the impeadance above the CCS bias setpoint is very high and would reduce this extra error current load. I agree with wojtek5001 about the resistor/cap. Why wouldn't this work at higher frequencies? The point of this cap is to prevent cross-conduction at high frequencies by discharging the bases to turn them off. R5 could have a parralel cap as well. Base resistors will not cause as much phase shift as the extra darlington transistor will, as opposed to a two stage darlington. Maybe reduce the gain slightly but insure more equal current sharing. Vbe multiplier may be helpful in controling output bias setpoint.

Yes Lineup, PSRR would be worsened....

[lumanauw]

Hi, Roender,

R1 and R11 are interesting. Electrocompaniet 25WclassA also uses these resistors. In the later Electrocompaniet amp, R1 and R11 are replaced by CCS.

It is a very simple way to discharge the base junction at HF and to prevent negative base impedance

How do you know it can prevent "negative base impedance"?

[roender]

R1 and R11 is a well known practice in HF amplifiers, where is a must to use BE resistors.
CCS instead those resistors is a very good idea (will improve PSRR also).
The base current in predrivers is to small to replace r5 with cross resistors like in final stage.
The target power is 100w in 8ohm, power BJT's will be NJL0302/0281