RC time constant of bootstrap

[AndrewT]

How does one estimate the roll off frequency of the bootstrap capacitor?

Using ESP's project03 fig1

http://sound.westhost.com/project03.htm

the 100uF cap is feeding a pair of 2k2 resistors.
Could it be the load the cap is driving is as simple as 2k2//2k2 giving an RC time constant of 100*10^-6 * 1100 = 110ms?

What other factors/components come into determining a suitable capacitor value?

[jcx]

for CCS use usually the impedance at the load is high or is effectively boosted by the bootstrap so the time const is nearer the R to the power supply times the C

in the ESP circuit the bootstrap C is connected in the wrong place, connecting to the BD139 emitter directly is better - no need to loose the 0.22/8 Ohm divider ratio
care should be taken so that there is little parasitic C associated with the bootstrap components

[pacificblue]

The RC time constant to charge the capacitor will be formed by the upper 2k2 resistor, the capacitor itself, and the circuit formed by the NFB loop and the load to ground. NFB||load = ~load and in relation to the 2k2 resistor the load is so small that it is sufficiently exact to assume RC = ~2k2x100µ.

The RC time constant to discharge the capacitor is formed by the capacitor itself, the lower 2k2 resistor, Rbe of the BD139 transistor, the 0R22 resistor and the NFB loop||load circuit. The discharge time constant depends on the BD139's operating point with ~(2k2+Rbe)x100µ.

[AndrewT]

I kinda guessed it could not be as simple as I proposed.
Thanks.
Any other offerings, particularly simple ways to model what Pacific shows is becoming quite complex, or simply alternative strategies to analyse what is happening.

[kaos]

The open-loop plot of an amp with a bootstrap will exhibit a peak at some low frequency, typically below 200Hz. Some people who deploy bootstraps actually tune for this as 2nd harmonic distortion tends to increase below this point, giving the bass a “livelier” more tube-like quality. A simulation program is very handy for looking at this sort of thing.

[pacificblue]

Originally posted by AndrewT
what Pacific shows is becoming quite complex, or simply alternative strategies to analyse what is happening.

Not really. All you have to do is follow the path to ground.

Charge


Discharge


If you want it more exact, you have to add the alternative paths through the NFB circuit and through the 10r resistor in parallel to the 5µH coil and the Zobel, but I think their influence is negligible.

Originally posted by kaos
The open-loop plot of an amp with a bootstrap will exhibit a peak at some low frequency, typically below 200Hz. Some people who deploy bootstraps actually tune for this as 2nd harmonic distortion tends to increase below this point, giving the bass a “livelier” more tube-like quality. A simulation program is very handy for looking at this sort of thing.

Surprising. Usually the bootstrap circuit has a frequency roll-off below audible frequencies. The capacitor should maintain a stable voltage bias and therefore needs a high RC time constant. In the P3 e. g. it is below ~0,73 Hz.

[destroyer X]

maybe i have an inverter diode from my eye to brain.

Thank you very much.

regards,

Carlos