Bob Cordell's Super Gain Clone PCB (LM3886) and a stripped-down version: Compact3886

Jump to Latest
S

slungu

Member
Joined 2010
What power supply are you guys using for the SGC ? I have here 2x18VAC that I would like to use - would this still be possible, even if I would not get a very high voltage out of it - i'm listening in rather small rooms and max 3m away from the speakers so max power is not the highest priority as long as the amp dehaves at the lower rail voltage ( should come out around 25V if I'm not mistaken ).


Regards, Stefan
 
Mark Johnson

Mark Johnson

Member
Joined 2011
Paid Member
Did you see this section of post #1?

Mark Johnson said:
WHAT POWER TRANSFORMER WILL YOU NEED?

You'll need a power transformer rated at least 200 volt-amperes ("VA"), with dual independent secondaries, and with primaries that are compatible with the AC mains in your country. Using the rule of thumb

DC_supply_voltage =approx= [(Trafo secondary rating in VAC rms) * 1.414] - 3.1V,

and consulting the LM3886 datasheet's Absolute Maximum Ratings, I conclude that you can use any 200VA transformer you like, as long as its secondary's output voltage is between 14.5VAC and 31.5VAC (rms ratings at full load). The LM3886 datasheet provides a helpful graph that I have copied below as Figure 7. Applying the above rule of thumb, we get Figure 8. These predict the maximum output power at the onset of clipping (namely, when THD rises to 0.1%).
Click to expand...

Eyeballing Figure 8, I think it predicts something like (25W into 8 ohms), (32W into 6 ohms), (44W into 4 ohms) when the transformer secondary is 18VAC rms at full load.
 
00940

00940

Member
Joined 2003
Actually, that page was always pretty bare. As far as I can see, Bob has two relevant measurements in his book (I only have an used copy of an old edition though).

One is a thd vs frequency at 40W/8r. A bit more than 0.001% at 1khz slowly raising to 0.01% at 20khz. This is in line with a simple lm3886 best measured performances but it is obviously dependent on implementation and should be assessed for a particular pcb.

The other is a comparison of the klever klipper on/off, showing distortion vs power. Without the soft clipping, thd starts quickly to raise past 50W. With the soft clipping, it starts to raise at about 25W going to 1% at about 50W. This should be less sensitive to implementation.
 
4

454Casull

Member
Joined 2004
Mark Johnson said:
Yes, Bob included several measured performance curves in his book. He also has a web page named Super Gain Clone here.

Maybe the web page once included all of those figures, before the book was published? You could check the Internet Archive / Wayback Machine to find out.
Click to expand...
I mean for the completed amp. The layout's pretty important too, not just the circuit. Unless your PCB is an exact duplicate of what Bob used?
 
J

jimk04

Member
Joined 2018
Paid Member
So I am getting myself a little bit prepared for the boards.
I have this big honking 23v toroid. It has 6 outputs at 5A.
I am going to try these psus as i was included in the group buy, and now i have a use for them! ( I have the CRC psu boards, not just the rectifier boards.)

LT4320 based active rectifier

using 2 windings per psu pcb and one psu per channel, so kind of a dual mono?

Should give me a good start I hope!

The power amp will be fed from my Salas DCG3 preamp that I think I set to 2x gain. Currently using a tpa3255 class d power amp with 20db gain.
 
Last edited:
Mark Johnson

Mark Johnson

Member
Joined 2011
Paid Member
jimk04 said:
Gain = 21 then by my maths.
Click to expand...
You're about 5% too high; try it using pencil and paper instead of a calculator. The arithmetic is easy, and pencil+paper eliminates "fat finger" data entry errors.

If you're deriving from first principles, keep in mind that SGC is a cascade of two inverting amplifiers. Don't used memorized expressions that apply to non-inverting topologies.
 
J

jimk04

Member
Joined 2018
Paid Member
I just used the R values from the schematic and admittedly used a calculator, although I think I hit the right keys.
What I dont understand is if R6 is part of the equation for gain, and yet R6 can be omitted if not using the klipper?

I am by the way very new to electronics!
Your last paragraph meant nothing to me.!
 
Mark Johnson

Mark Johnson

Member
Joined 2011
Paid Member
Here is the pencil and paper method of gain calculation. It gives answers that are about 5% below yours. More importantly, pencil+paper matches the gain number Bob Cordell wrote in his book: page 686 of the 2nd edition, page 541 of the 1st edition.

_
 

Attachments

  • Image1.jpg
    Image1.jpg
    127 KB · Views: 777
You must log in or register to reply here.
Top Bottom