F4 power amplifier

[jeffreybehr]

Originally Posted by Zen Mod

1. sort of , but not exactly
2. it's not constant vs. frequency
3. read #2

TYVM Zen. So replacing Cs 1 and 2 is not a good idea. So then I'd add the new hi-pass coupling cap before the input-load resistor and use 47K as the load value? If so, even this idiot can do that.

But Cs 1 and 2 definitely are carrying the signal; yes? If so, it makes sense to use the highest-quality caps there that we choose to afford. How about for Cs 3 and 4?

[Zen Mod]

C1 & C2 are more DC level shifters , than signal carrying caps ; for better sleep , you can bypass them with something as 10nF MKC (polycarbonats) ; ;that's , at least , my fave type for bypass

C3 and C4 are there in bias circ and no need for them to be of any better quality than proper industrial type

[elektroj]

Originally Posted by jeffreybehr

1. Are Cs 1 and 2 signal-coupling capacitors?

Yes, they are.

2. If so, what is the load impedance for filter-point calc?

In simplified form, without going any deeper (ignoring TL431 dynamic impedance, Cgs of output stage transistors and all that sort of stuff) the formula is as follows:
f=1/2pi (R6 II R7) (C1+C2)

3. Can the cap-value be changed to create a hi-pass filter in the audible range for a speaker that requires one?

Yes, but you will get 6dB/oct slope only.
(I've done it in the past to similar output topologies, btw)
If your filter frequency isn't critical to the 10th of an octave, just use the formula above and be done with it.

@Zen Mod - C3 and C4 serve several purposes, but they are in the signal path. Do the AC signal path analysis.

[jeffreybehr]

Originally Posted by elektroj

Yes, they are.

In simplified form, without going any deeper (ignoring TL431 dynamic impedance, Cgs of output stage transistors and all that sort of stuff) the formula is as follows:
f=1/2pi (R6 II R7) (C1+C2)

Yes, but you will get 6dB/oct slope only. ..which is exactly what I want.
(I've done it in the past to similar output topologies, btw)
If your filter frequency isn't critical to the 10th of an octave, just use the formula above and be done with it.

@Zen Mod - C3 and C4 serve several purposes, but they are in the signal path. Do the AC signal path analysis.

elektroj, TYVM. The formula I've been using for decades for RC filters is F = 159155/RC where F = Filter Point in Hertz, R = resistance in Ohms, and C = capacitance in microFarads. I have no idea how to read your formula. Is the equivalent simply that C currently = 220uF and R = 10K-Ohms? That makes the FP 0.07Hz. So for a 72Hz filter, I'd use 0.22uF caps?

NP answered my e-mail last nite: "I would leave those caps as they are because they are used as part of the supply system for the Jfets.

Presuming that this is the stock F4, the input impedance is 47.5K ohm, and you can place a capacitor in series with the input using that value for a filter."

I don't like the idea of 'lytic caps in series with the signal, but what the hell do I know about it? (The answer is 'virtually nothing'.)

So...replace Cs 1 and 2 with much-smaller, highest-quality caps, or add a cap at the input and improve the quality of Cs 1 and 2*?

And it's more complicated than the above. I'll be using 2 amps and biamping vertically. In the low-frequency channel--that is the one with the hi-pass filtering and feeding the upper-bass/lower-MR driver and the internal bass amp--probably I'd simply add a cap at the frontend. The speaker rolls off this signal starting at 500Hz (1st-order). The upper-frequency channel would get no additional filtering but would get improved Cs 1 and 2. At least that's the 'plan' before I've even had my hands on one amp.


* as in replace them with BlackGates.

[Zen Mod]

Originally Posted by elektroj

........
@Zen Mod - C3 and C4 serve several purposes, but they are in the signal path. Do the AC signal path analysis.

whatever you say

[elektroj]

Originally Posted by jeffreybehr

elektroj, TYVM. The formula I've been using for decades for RC filters is F = 159155/RC where F = Filter Point in Hertz, R = resistance in Ohms, and C = capacitance in microFarads. I have no idea how to read your formula. Is the equivalent simply that C currently = 220uF and R = 10K-Ohms? That makes the FP 0.07Hz. So for a 72Hz filter, I'd use 0.22uF caps?

jeffreybehr, I presented the formula in its classic form:
f=1/2pi x R x C
where C is expressed in Farads and R in Ohms.
If you do the math, you will arrive at the same simplified formula you've been using for Ohms and microFarads.
Your values are correct. 0.22uF will give 72Hz :-)

=========

EDIT- Oops, I was referring to F4 schematic dated 11/8/06! Now I see that actual F4 schematic r0 6/4/07 look different

[elektroj]

Originally Posted by elektroj

EDIT- Oops, I was referring to F4 schematic dated 11/8/06!
http://www.6moons.com/audioreviews/firstwatt6/hero.jpg
Now I see that actual F4 schematic r0 6/4/07 look different

jeffreybehr, I am sorry, it will be, indeed, difficult to define -3dB point with C1 and C2 since they're no longer working into fixed 10k load as in older schematic. It looks that you should stick to filter at the input.

[jeffreybehr]

Originally Posted by elektroj

jeffreybehr, I am sorry, it will be, indeed, difficult to define -3dB point with C1 and C2 since they're no longer working into fixed 10k load as in older schematic. It looks that you should stick to filter at the input.

OK; hi-pass filter at the input it'll be. 0.056uF will give me 60Hz, right in the lower part of the range.

TYVM all.

[luvdunhill]

So, I'm considering adding a small pair of subs to my setup (DX4 in TP-1 with bridged F4s). What's the easiest way to do this? I have plenty of gain and current in the F4s to drive the additional drivers (8 ohm most likely), but not sure the best way to go about this.

[luvdunhill]

I'm wondering if I don't want something very similar to the B5.

Take the output from Impasse preamp and split it, first goes to (a new amp?) that powers the corner horns full range, and the second goes to the F4 which power the subs and has a passive crossover...

I keep starting at the manual for the F4 and there are a lot of possibilities...