gainclone with output stage

[steven344]

I wonder if its possible to add a output stage to a lm3875 or lm3886,Sub the gainclone chip for Ic1d in the schematic and use the transistors in the rest of the circuit

[Ouroboros]

Why would you want to? The design you showed is a crude version of the old Texas Instruments output stage, and only needs an op-amp to drive it.

One thing you could do with a 'gainclone' iif you really wanted to make an amp capable of driving lots of current into a low impedance load would be to use the amp as the 'control amplifier' together with a pair of really big output transistors in a 'current dumping' configuration.

[steven344]

Yes thats what I mean,Use a lm3875 to drive Complementary Sziklai Output stage like the one in the schematic below,It should work if it works with the op amp below

[Ouroboros]

No! Adding an output stage like that defeats the purpose. The o/p stage shown in the schematic only needs a few mA of drive, so using a 3875 or similar doesn't do anything useful. Now adding a simple PNP and NPN, together with the appropriate passive components to make a 'Quad style' current dumping stage is a much more interesting concept. (My opinion only of course!)

[steven344]

In the circuit do you think there would be any crossover distortion ,I intend to use it only for a sub woofer.I have tried gainclone for sub woofer but dont feel they can supply current,dont get me wrong I think they are great for mid and highs like a sat sub system

[Ouroboros]

If you look at the schematic, you'll see that Q1 and Q2 have no forward bias, so the amp is relying on the global NFB to reduce the crossover distortion. The original Texas app-note design from way back had a proper biasing for this part (I can't find a schematic on the web, but in the UK it was originally publisehed in Practical Wireless magazine in the 1970s). The o/p stage has a gain of x4 (roughly) and that type of o/p stage can have a tendancy to oscillate.

For a low-frequency amp only, the NFB probably operates fast enough to minimise distotion, but crossover distortion is hard to remove by a simple overall NFB loop. Adding a conventional Vbe multiplier between the bases of Q1 and Q2 will help, but quiescent current stability will be hard to manage unless you also add a couple of low-value resistors to the output transistors as well.

[lineup]

In my attached circuit
LM3875 provides current alone below +- 2 Ampere output.
At about 2 A output (0.7 Volt across R1/R2) the power transistors start to conduct
and help IC with current.

At 5 Ampere output LM3875 will output 2 A and transistors 3 A.

This is theoretical idea of operation.
To make it work in a real circuit, may need some additional considerations and components.

I have seen a circuit like this published in Elektor long ago.
That circuit used a TDA2030 audio amplifier chip.

[Ouroboros]

Yeah, that sort of thing can work well. I think usually there's a capacitor between pin 4 of the chip and the output (rather than the direct connection), and a dummy load (perhaps 100 Ohm) from pin 4 to 0v.

The only issue would be to select the value of the sense resistors to bias the output transistors just below the point at which they start to draw current.

[Vigier]

Probably you mean this schematic. It uses cheap IC's and transistors.

Power:
2x6V/1x12V = 5W @ 4Ohm, 85mVeff input
2x6V/1x12V = 10W @ 2Ohm, 85mVeff input
2x12V/1x24V = 40W @ 4 Ohm, 211mVeff input
2x12V/1x24V = 80W @ 2 Ohm, 211mVeff input
2x18V/1x36V = 100W @ 4 Ohm, 330mVeff input

Grtz, Joris

[Bazukaz]

I have built very similar schematic(probably even the same) as a DIY kit several years ago.The sound quality was not so good.