The Frugalamp by OS

[ostripper]

by roender (d.self) Figure 6.5 shows the effect on an otherwise Blameless amplifier handling 60 W/8, with 220 ìF rail decoupling capacitors; at 1 kHz distortion has increased by more than ten times, which is quite bad enough.

D. Self took those readings with NFB takeoff "grossly" unbalanced
(right at a emitter resistor) to prove his point. MJ's small
board design errata ? would not produce the same results.
On mine I use 2" shielded wire to take NFB right at OP terminal
to have optimum results.
MJL can still change things with a click of the mouse.
I've aready done my boards so hopefully there are no major
problems with my layout.

An externally hosted image should be here but it was not working when we last tested it.

By AKSA - Another technique using the same principle and saving on decoupling caps is to simply run a 10R resistor and back to back diodes from A to B,

Your suggestion "in the flesh"..
OS

 

[roender]

OS,

We don't talk about NFB take off point, but rail decoupling distortion.
In your PCB layout, and MJL... also, rectified half waves from output stage are polluting the low current stages ground.

 

[ostripper]

good idea..will try it.. still, I do the same on the amps I have
and on the CCS can see no ripple on the LTP (CCS must be
working good)

A very easy mod (cut pattern,one more jumper)
OS

 

[MJL21193]

OS,

We don't talk about NFB take off point, but rail decoupling distortion.
In your PCB layout, and MJL... also, rectified half waves from output stage are polluting the low current stages ground.

Hi roender,
Tell me how yours differs significantly from mine? The low current ground trace is coloured bright red.

 

[MJL21193]

C11 and C12 are bypass caps, no?

I think you are blowing this out of proportion, but I do listen to any good advice I get and use it. In my Patchwork layout, I solicited Eva's comment on it and this "big mistake" was not pointed out.

 

[roender]

C11 and C12 are bypass caps, no?

I think you are blowing this out of proportion, but I do listen to any good advice I get and use it. In my Patchwork layout, I solicited Eva's comment on it and this "big mistake" was not pointed out.

Yes, but that stage in pure classA, free of half waves class B pulses and more than that with a separated power supply, only for VAS and LTP

 

[MJL21193]

Yes, but that stage in pure classA, free of half waves class B pulses and more than that with a separated power supply, only for VAS and LTP

But the trace continues behind it directly to the main decoupling caps! I'm not seeing the distinction.

Here shows my signal ground (yellow) - isolated from the common ground.

Anyway, I need to run off to work. Later.

 

[roender]

In my Patchwork layout, I solicited Eva's comment on it and this "big mistake" was not pointed out.

Yes, it was. Take a look hire:
http://www.diyaudio.com/forums/showthread.php?postid=1668286#post1668286
http://www.diyaudio.com/forums/showthread.php?postid=1668304#post1668304
Eva make the same suggestion, to take the ground for VAS and LTP from star ground directly.

 

[ostripper]

All done..
I see the point, half wave class B pulses cancel out mostly
at midpoint (center ground point) , best place to take the VAS/
CCS ground from
I'm already taking LTP/input from this point..
OS

 

[ostripper]

Did "housekeeping" on whole thread.. All schema's are up
to date. FA1 is working still
Last update to it.."schoolbook amp" is final..

The new one (FA2) should be very close..

An externally hosted image should be here but it was not working when we last tested it.

Go ahead and build them , they will not blow up , FA2 might
"blow your socks off"..
OS

 

[MJL21193]

Eva make the same suggestion, to take the ground for VAS and LTP from star ground directly.

And I did it here. The topology used in my amp will benefit more from such changes, as the PSRR is not as good but in an amp such as this one by OS, there is not going to be a big gain.

BUT, I'm not one to argue with good sense when It is easy to make a change , so here it is:

 

[MJL21193]

Did "housekeeping" on whole thread.. All schema's are up
to date.

Connecting the input to the star ground defeats the purpose of R3.
Why the pot across the differential pair? DC offset will be ultra low (<10mV I'd guess) without that extra complication.

Did you use my method of tinning your boards?

 

[ostripper]

And I am the dummy that makes such a obvious mistake for
all to see...ha ha , AND has to "pants down" correct it in 20
minutes.

We should not fret in our endeavors for with FA1 we have
definitely bested the "quasi" offerings..really ( I have one).

With this one(FA2), I have a feeling it will be the best , cheap
with the ability to be implemented from +- 30 to +- 100 V
(sims great with minor changes to CCS/ VAS Re's). All
the minor haggling over layout.. etc... Time will sort all this out,
I have 18 more 6X6 fr4's.
Did my mouser shopping with Visa tonight.. wed/thurs. amp will
be finalized..pix,too WOW FA1 is freakin' loud
OS

 

[ostripper]

Did you use my method of tinning your boards?

absolutely, I absorb all.
OS

 

[MJL21193]

How's this for service:

I got off early today and decided to dissolve some copper cladding.

Works too. That's always a bonus.
Preliminary tests show it running with a bit of overshoot but that is with 47pF Miller caps (I didn't have any 68pF).
I'll post some scope shots soon.

 

[MJL21193]

20kHz into 8 ohm dummy. No Zobel, no output inductor.

10x magnification:

These were quickly done without a heatsink running on my +/-30VDC supply so they are a bit blurry.

 

[lineup]

Very nice 20 kHz square into 8 Ohm dummy!
Shows a rather fast but stable amplifier, with only a very, very small overshoot.

 

[lineup]

Section A
Waveforms using the Cathode Ray Oscilloscope (CRO)

SQUARE WAVE TESTING

One method of assessing frequency response (and sometimes other characteristics) is to feed a square wave to the input of the device under test and examine its output on a CRO. The square wave is made up of a fundamental frequency and all odd harmonies, theoretically to infinity. A deficiency within the frequency spectrum, from the fundamental upwards, will show a change in the shape of waveform. The test is subjective rather than precise but gives a good indication of the response.

Waveform and Spectrum Analysis
by Lloyd Butler VK5BR

 

[MJL21193]

This is a pretty fast amp and no signs of instability. Look at this 500kHz squarewave:

I have it mounted on a temporary heatsink. I will be trying it on full rail voltage soon but I have a small concern. It doesn't like to clip:

That is a 20k sinewave and is at relatively low rail voltage (30V). This will be a lot worse on higher rails. I think it is that 3rd transistor in the current mirror, or at least that is what the sim shows. I used a 1k resistor to feed it - I figured that the 4.7k was too high.

 

[MJL21193]

I improved the clipping greatly by dropping the current through the VAS from ~10mA to ~2mA. It still sticks but it's much cleaner now.