It's part of a "ring of two" CCS. This type of CCS is liable to oscillation. You could eliminate it in favor of a simpler one transistor CCS. A base stopper for Q7 might help.
An externally hosted image should be here but it was not working when we last tested it.
I'm back with some changes
I replaced the output Bjts with Mosfets and make a few changes as well to supply the circuit with a dual 55 volts
how about this circuit??
it sounds great without any oscillation like with bjt
any mistakes?
thank you in advance
I have no doubt that outputs keep blowing out. With no base stopper resistors, no degeneration and no current limits for the small signal devices.....ect. Simply replacing BJT's with mosfets is unlikely to be optimum due to the simple fact that they are not similar parts and have different characteristics. With no gate stopper or gate Zobel or Zener protection I will be surprised if the mosfets don't fail either. True that there is more flexibility in bias for mosfets due to their lower Gm, but there are some fundamental issues with the schematic, particularly the output stage. For example, there is no need for flyback diodes with mosfets because the internal body diode, part of the die, suffices for this purpose. Mosfets make great oscillators and thus need local dampening.
Last edited:
I'm sorry
I forgot to inclute gate resistor in schematic
But I use 470 ohms resistors for each mosfet
Sorry about that
But I don't have any more power bjts on my hand and the oscillation was driving me crazy too
here it is
again, sorry about power output changes
An externally hosted image should be here but it was not working when we last tested it.
again, sorry about power output changes
A couple of thoughts are - 0.5 volts across base emitter of Q7 seems not enough to turn it on sufficiently to conduct 10 milliamps - further if you have set the resistor values of the voltage multiplier/biasing transistor R7,R8 accordingly then through starving of current the voltage difference between bases of Q16 and Q17 will be further apart.
You really need to make the voltage multiplier adjustable by incorporating a trimpot. The structure should be adjustable around the 3- 4 volt level. I suspect the voltage drop between bases Q16 and Q17 is greater and the cause of excessive dissipation in your output stage.
I note you have changed R9 from 100R as per the original circuit to 62R without a corresponding reduction in the value of R10 (presently 33k) so you are asking Q9 in your CCS to deliver more current without increasing the base current feed.
From Ohms law a 10 milliamp current will cause a 620 millivolt drop across R9 and you need to find a lower value of R10 to provide that.
If the voltage amplifier stage is not RC decoupled from the outputs it might pay to try a value lower than what is strictly necessary say 10k - if that is not to your taste you could start with 22k.
It is a fair bet that the cascade transistor arrangement suffers in the same way and you might apply the same thinking with R20 (also presently 33K).
With regard to fault finding one approach to in circuit measurements is to limit the current flow through the output stages.
One way to do that having fuse holders in the supply rail lines is a wire-wound resistor say 10 watts and 100R in place of the fuse - these can be soldered in place. You can then deduce the current draw from Ohms law.
If there is a major fault such as a full rail voltage the resistors will get very hot but you can switch off without causing damage. If the current draw is large but not seriously so there may be no fault and the suggested trim pot around Q5 would make possible a quick reduction.
If otherwise the next step would be to measure transistor base to emitter voltages which should be around 0.6 volts if working correctly.
Last edited:
In a Class AB amplifier the output stage draws fluctuating and substantial amounts of current and there is a need to minimize this influencing the voltage amplifier stages which operate in Class A where the current needs to be constant.
The voltage amplifier stage takes the form of an operational amplifier which has the properties of rejecting supply rail anomalies however some designers take further measures such as a series resistor and capacitor to ground to decouple the voltage amplifier.
The Leach amplifier is a case in this point see
Leach Amp Plans - Part 1
If this strategy is not adopted and the current through the base of the constant current transistor in your amplifier is on the low side, it is arguable that running more than what is needed does more to keep that stage working under the influence of the extreme current fluctuations on the output stage.
Some designs include a cap of a few hundred pico farads around the constant current source. From memory Nelson Pass' A40 design uses that strategy. In fact he goes so far as to replace the resistor with a FET arranged as second constant current source.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.