Tracking power supply

[ionutgaga]

This is the positive half of tracking power supply. It represent a buck converter with voltage control mode.
U4 is integrator of square wave 600kHz, resulting in triangle wave with 5.6 v pp.
U3 is error amplifier, with 2 pole and 2 zero compensation network (R12, R11, C15, C14, C35, C14).
R14 determine the amplification of input signal (A=R11/R14=15). R48, conected at -15v introduces a offset in output, coresponding +7.2v at idle (Vcc x R11/R48=7.2v).
U5 is pwm comparator, feeding U1 fast optocuplor (6N137). Because the switching f=600kHz is high (for best transient response), Q12, Q13, M2, M3 form a buffer with sufficient current capability to drive M1 Mos Fet (IRFP260N).
L1, C5, C6, R3 is part of buck converter filter, with aprox impedance of 4.4 ohm.

[ionutgaga]

The prototype:

[Workhorse]

Ion,

This is something known as Class-TD[Lab Gruppen Style]

Have you tried the feedback directly into comparator, this will save you from lot of troubles arising from loop stability comprising error amp....[You have to change the phase also to implement it and also the amplitude of the triangle must be reduced to about 600mV to achieve high open loop gain]

regards,
Kanwar

[Workhorse]

A link to Fredos post, explaining how to eliminate error amp...

http://www.diyaudio.com/forums/showt...77#post1104977

[ionutgaga]

Feedback before or after LC? Because is ease in one way (before) but not compensating resonance of filter. Also the transient response is different.
Maybe is your turn to present some of your work.

[Workhorse]

Originally posted by ionutgaga
Feedback before or after LC? Because is ease in one way (before) but not compensating resonance of filter. Also the transient response is different.
Maybe is your turn to present some of your work.

Feedback After Filter......Take a look at this.. example

[Workhorse]

The Front end section of Monorail Class-TD....
Front opamp is absolute value converter, then comparator...

[Workhorse]

Hi ION,

What about your other prototype???

Kanwar

[ionutgaga]

From my vision, the front end of monorail is too simple. Because RC input for speed-up high frequency interfere with absolute dc converter (non linear in+/-0.65v), because you sense output signal, not floating dc power.
Power disipation on class AB+tracking power, is P=7.5V * I(out), just because power final is polarised with aproxx 7.5 v, for normal operation (+/- ripple).
In monorail, power disipation it's double (2 * 7.5v * I(out)), just because H bridge...

Look at my proposed front-end (sensing diferential voltage, absolute value converter)...

I tryiet to simulate D1200 feedback in OrCAD: no positive result.
Any idea?

Regards,
Ion

[Workhorse]

Originally posted by ionutgaga
From my vision, the front end of monorail is too simple. Because RC input for speed-up high frequency interfere with absolute dc converter (non linear in+/-0.65v), because you sense output signal, not floating dc power.
Power disipation on class AB+tracking power, is P=7.5V * I(out), just because power final is polarised with aproxx 7.5 v, for normal operation (+/- ripple).
In monorail, power disipation it's double (2 * 7.5v * I(out)), just because H bridge...

Look at my proposed front-end (sensing diferential voltage, absolute value converter)...

I tryiet to simulate D1200 feedback in OrCAD: no positive result.
Any idea?

Regards,
Ion

Hi Ion,
I sense VAS signal not output....You absolute value converter takes care of small amplitude signals very well...
The power dissipation could be halved if polarised to 3.5V .....just same disspation as half bridge...but inturns you will get double the slew rate of half bridge....
I have not used Feedback in my monorail tracking amp, but i relied on Feedforward compensation by adjusting the amplitude of Triangular wave in proportion to power rail[not tracking, but the idle one] fluctuations....so simple....

What was your gain setting in sim....it had to be around 5X only...to get good results and manipulate the sensing signal also...to tally the pwm gain with amp gain....