much less crazy idea

[Freddie]

Complementary feedback pair

Where can I find more information about the CFP outputstage. The only info I've found about it, is this site.

http://www.dself.demon.co.uk/cfp.htm

/Freddie

 

[Duo]

KSD401 should at least be HFE class Y not to load your current mirror to hard.

Sonny [/B]

Um , what does hfe class y mean??

 

[Freddie]

Hmm..

After reading some about the CFP outputstage it appears to me that the value of the collector resistor (R11,R12) for the drivers is really important. But how do I determine the optimum value?

 

[Freddie]

Duo,

Um , what does hfe class y mean

Look in the datasheet under hfe classification.
Hfe class Y means that hfe is 120-240.


http://www.fairchildsemi.com/ds/KS/KSD401.pdf

 

[sonnya]

The compound output stage works this way.

You driver are a emitterfollower with a currentsense resistor inserted into the collecter. Your common emitter power BJT connected to the currentsense resistor will amplify the current through your driver transistor.

So your R11 and R12 (collector resistor) sets the currentgain or how hard you wants to load you driver transistor.

But R11 and R12 in large value will add a dominating pole to you outputstage. On the other hand a low value will let you driver transistor work harder for a given load.

Its your decission!?

Sonny

 

[Freddie]

Its your decission!?

OK.. I think I will just start to build the circuit and when it's finished I can play around with different values.

Thanks for the help.

/Freddie

 

[sonnya]

Good idear!!! Then you can compare phase margin as well with different values of R11 and R12.

Sonny

 

[Duo]

Okay, I see what you mean by class Y, so basically, the KSD 401 will work in here, ye, ne?? And, I'll do like freddie and just make the amp and experiment, I think that's best.

 

[Duo]

Also, how do I figure out what's going on with bias of the VAS, I know that if it's wrong, I'll have an offset at the output of the whole amp... I'm not sure how it's supposed to achieve bias with the current mirror in there. Does it simply balance itself when the offset from the output comes to the LTP and causes a bias change? or is it independant of that?

 

[Duo]

Well, could anybody answer my previous question about bias of the VAS, also. Anymore ideas to make this amplifier better? I found out that with the transformer I have, I'll have rails of +/- 80 volts, where my original supply voltage was to be about 65 or 70.
I will change my resistances where needed to compensate for voltage change, but I'm having trouble how I perform the VAS bias! Please, can anyone help me with that part?

 

[gromanswe]

It balances somewhat to zero.

Then you do what we call a trim.
The goal is to get no offset at output,
and to have exactly the same current
in the 2 LTPs.
That can require some changing of resistors.
And changing the current source supplying LTP.

Hope you can trim it!
figure out how?
What resistor to change?
You are smart, aren't you?
Paper and pen, and caculator.
No time limit!!!!!!! Like in school.
You Your Self see if you got all rights!!!!

gro

tips draw up the current flows first.
messure and write in currents on your circuit drawing

messure volt/R=current
all currents going into circuit must come out at bottom.
sometimes current splits in 2 smaller,
and 2 currents can join to a bigger.
so some adding and subtraction gives the answer

 

[Duo]

Okay, I kinda thought the design balanced automatically. Since that's the truth, I'll have no trouble balancing it, that'll be very easy! Thanks...

 

[gromanswe]

How did trim GO?

Well, could anybody answer my previous question about bias of the VAS, also. Anymore ideas to make this amplifier better? I found out that with the transformer I have, I'll have rails of +/- 80 volts, where my original supply voltage was to be about 65 or 70.
I will change my resistances where needed to compensate for voltage change, but I'm having trouble how I perform the VAS bias! Please, can anyone help me with that part?

I just wonder if you did it

gro

 

[Shaun]

I have your PCB [RE: much less crazy idea]

Duo

Believe it or not, I have a pcb for your amp! My amp is about 99% the same as yours!

You can view it here:
http://users.pandora.be/airborne/blameless.html


I will have revised PCB's for the amp by the end of this month. As for the ones in the picture, I have 2 PCB's left. If you're interested, email me.

Regards
Shaun

 

[Duo]

Wow, how on earth did you find pcb's that close? This was designed with all kinds of duu- hickies and oddball circuitry, I can't believe it has all that on it. Does it actually have spots for the currnet sinks and mirrors in the LTP and all that?

 

[Duo]

Cool, you're interested in flying too? I love to fly, wish I could afford a pilot's licence but that's a bit heavy on the money for now. What kind of stuff do you fly?

 

[Shaun]

That's not my website!

I've only ever flown once!

The site belongs to Luc Henderieckx, who helped me develop the amp.

The circuit is based on Douglas Self's Load Invariant amp. I believe that you would have probably come to the same circuit if you had continued to improve your design.

I'll try and attach an image of the schematic (my first time). Excuse the low quality, please.

Shaun

 

[e96mlo]

Some questions:

1. Why do you have parallel transistors in the diff. stage?

2. It seems like you have a limiting transistor on the VAS. What for?

3. Wouldn't the current limiter be on all the time because of the resistor from base to supply?

/Marcus

 

[halojoy]

Some questions:

1. Why do you have parallel transistors in the diff. stage?

2. It seems like you have a limiting transistor on the VAS. What for?

3. Wouldn't the current limiter be on all the time because of the resistor from base to supply?
/Marcus

If you double numbers of transistors you half the noise.
So 4+4 transistors would give 1/4 of the noise in one transistor.
I dont remember if it is -3dB or -6dB in signal/noise (SNR)
for every doubling.

That limiting transistor has to be normally off.

halo

 

[Shaun]

1. They're not meant to be parallel. Think of them as alternatives. Use either MPSA56 or 2SA970 (I think); they have different pin-outs, so I made allocation for both options on the revised PCB.

2. The Vas needs current-limiting for when the VI protection kicks in. But this current limit is set higher than the standing current of the Vas class A stage, so it will not normally come into play.

3. No. (a) There is a voltage divider formed by the other resistors
(b) These are not the final resistor values.
(C) This is not the final protection circuit. I would like to
have triple-slope SOAR protection as per last year's
Electronics World.

Shaun