Rebirth of the Phoenix

[PHEONIX]

Hello

The is a thread about a high performance amplifier called the Phoenix .

>The values of the freq. comp. components are cc9 = 1nF & rc9 = 56R.
BTW, they don't compensate Q16/16, rather Q22/23.

Are these values for the output stage stability of Q22/Q23.

>Regarding the temp. compensation, it's a bit tricky with this design. The point is that the loading on the VAS output should be kept as low as possible. Even a small capacitance of 10pF has already an effect on the performance. Now imagine what a classical Vbe multiplier will do, a TO126 tranny that is mounted on the heat sink...... Right!
Also, we have issues with the thermal delay.

When you are talking about the output loading of the VAS are you talking about the capacitance of the TO126 tranny would introduce when it is connected to the heatsink (GND). So it forms a capacitor from the VAS stage to ground - is this what you mean ?.

The thermal track diodes of the NJL4281/4302 wouldnt they introduce a similar capacitive loading also.

Regards
Arthur

[Edmond Stuart]

Originally Posted by PHEONIX

Hello

The is a thread about a high performance amplifier called the Phoenix .

>The values of the freq. comp. components are cc9 = 1nF & rc9 = 56R.
BTW, they don't compensate Q16/16, rather Q22/23.

Are these values for the output stage stability of Q22/Q23.

Hi Arthur,

Something went wrong, as the component names have changed in the meantime.
Looking at the latest schematic I mean Q15/16 (again). But actually, it compensates the whole OPS (only when clipping, of course).

>Regarding the temp. compensation, it's a bit tricky with this design. The point is that the loading on the VAS output should be kept as low as possible. Even a small capacitance of 10pF has already an effect on the performance. Now imagine what a classical Vbe multiplier will do, a TO126 tranny that is mounted on the heat sink...... Right!
Also, we have issues with the thermal delay.

When you are talking about the output loading of the VAS are you talking about the capacitance of the TO126 tranny would introduce when it is connected to the heat sink (GND). So it forms a capacitor from the VAS stage to ground - is this what you mean ?.

Yes.

The thermal track diodes of the NJL4281/4302 wouldn't they introduce a similar capacitive loading also.

Regards
Arthur

Regrettably, the data sheet don't specify this capacitance, but I assume it's much smaller. However, this isn't an issue any longer, as I have moved these diodes to next stage (pre-driver), where they can do far less harm. I'm still busy with the temp. compensation. The bad news is that it is over compensated. That's really weird, as the TC of the diodes is lower than the Vbe-TC of the O/P trannies, which should result in under compensation. BTW, you also should have a look at this thread: On Semi ThermalTrak
The good news is that it is probably easier to reduce the compensation of an over compensated circuit than vice versa.

I'm away this weekend, so till Monday.
Cheers,
E.

[PHEONIX]

Hello Edmond

I am away until this Thursday.

Regards
Arthur

[PHEONIX]

Hello Edmond

Q1 an Q2 supply current , but what problem are you addressing with them.

Regards
Arthur

[Edmond Stuart]

Hi Arthur,

Q1 & Q2 are normally off. When the output protection kicks in (which shortens the VAS output), these trannies limits the max. VAS current to about 20mA, otherwise this current will rise to a dangerous 40mA. Also a few other unwanted side effects are suppressed in this way.
In the meantime I discovered an error in the op-amp model. The one we were using so far, is for dual op-amps. Consequently, the supply current is two times as high (10mA instead of 5mA). See also: Spice simulation
So I had to modify not only the model but also the schematic. I also played with the ThermalTrak trannies. Results were very encouraging and I will soon drop the schematic of the new version (please, have a little patient with me).

Cheers,
E.

[Edmond Stuart]

edit > otherwise this current will rise to a dangerous 40mA.
should read as: otherwise this current will rise to a dangerous 70mA.

[Edmond Stuart]

Hi Arthur,

Here's the new version. Regrettably I don't have correct models of TT-trannies NJL4281/4302 (though I do have models of the TT-diodes). So for the moment, I'm still using the Sanken models.
Please have careful look, in particular at D19 and D20. This clever configuration (courtesy of Bob Cordell) greatly reduces cross-conducting during fast transients.
Don't hesitate to put your questions (probably many ).

Cheers,
E.
PS: TDH20k is about 1ppm.
Q16 & Q19, respectively Q17 & Q20 are thermally coupled.

[Edmond Stuart]

More on D19 & D20. This is what Bob Cordell said about it:

"The two “extra” ThermalTrak diodes, D3-4, are used between the emitters of the driver transistors to establish the idle current of the drivers while keeping the impedance between the emitters very small, so that the drivers can operate in push-pull to provide turn-on and turn-off current to the output transistors. Keeping the impedance between the driver emitters very low at high frequencies is essentially what is often done with the speedup capacitor. The ThermalTrak diodes make it possible to do this in a d.c. fashion with the necessary precision because they track the output transistor Vbe’s. "
See also: Biasing/thermal compensation of Thermal Trak transistors

I would highly recommend this scheme.

And below a compilation of some comments, graphs and schematics about the ThermalTrak trannies and diodes.

Cheers,
E.

[PHEONIX]

Hello Edmond

Sorry I have not posted any questions yet, but I am just playing around with the most recent circuit (Pheonix7) to get some idea of all the things that you have done , it has lots of nice tricks. The output stage protection is is really nice the way it pulls the current away from the predrivers q16 and q17 in protection mode . Also the main clamp transistors q13 and q14 are also connected differently than in Pheonix6.

Here is a question though why have you connected the collector of Q1 with the collector of Q13.

More questions will follow.

Regards
Arthur

[Edmond Stuart]

Originally Posted by PHEONIX

Hello Edmond

Sorry I have not posted any questions yet, but I am just playing around with the most recent circuit (Pheonix7) to get some idea of all the things that you have done , it has lots of nice tricks.

Hi Arthur,

Indeed, many trick and many changes. Please take your time to digest them.

The output stage protection is is really nice the way it pulls the current away from the predrivers q16 and q17 in protection mode . Also the main clamp transistors q13 and q14 are also connected differently than in Pheonix6.

That's right. Q13 & Q14 are biased form other sources, i.e. from outside the signal path. As a result, slightly less distortion.

Here is a question though why have you connected the collector of Q1 with the collector of Q13.
More questions will follow.
Regards
Arthur

Similar to Q13 & Q14, Q1 & Q2 also function as an active feedback clamp, this time to limit the VAS peak current. Hence the collectors are tied to the same node (U) in order to provide feedback. Without this FB, currents through the op-amp and Q3...Q12 would rise to unacceptably high levels when the OP protection kicks in.

In the coming days, I will describe step by step the various function and (protective) FB loops, together with graphs of the phase and gain margin.

Cheers,
E.