Power Amp Schematic

[zinsula]

Hello to all experts

Attached is a schematic I have "backengineerd from a power amp kit, which was assembled by a friend. He bought the kit around 1990 and fired it never up.
He gave me two single channel PCB's, already assebled, but without heat sink and power supply. And without any papers which would had helped me


Rail voltage - no info's, my guess is around ±40 to 50 Volts, according to the data sheets I found, it should be possible with 50 Volts and still good safety margin. Am I right?

Q3, R8, D1 - is this a current mirror? (Sorry, but my elelctronics knowledge is quite limited )

What do you think about the topology? Could it sound good?
I'm not shure wheter I should spend money on the power supply and give it a go.

Thanks to you all!
Tino

[johnnyx]

It looks like the Maplin MOSFET kit. I've built some, and they sound good to me. 100W RMS into 8Ohms with +/- 50v supply, or 150W into 4Ohms.

The design is based on the Hitachi application note for their MOSFETs. Many amplifiers were based on the app note design.

You are correct about the current mirror, I don't know how it compares with one using two transistors, but it works well enough.

I changed the value of R9 so that the collector voltages of TR3 and TR4 are the same. The bias of the whole design varies with the power supply voltage, but it's quiet (low noise) and doesn't thump at switch - on. I like its simplicity.

I have the documentation if you are interested.

[zinsula]

Thank you johnnyx for answering. Yes, since I have nothing documented, it would be nice if you could mail me the documentation. My Mail address in the profile works.

Thanks again, Tino

[JohnW]

Tino,

For such a simple design (not a bad thing) they sound very good. I still remember the first time I heard it (Maplins kit) – I’ve associated the sound with “MOSFET” designs ever since – IMHO you will find it hard to beat with a “commercial” amp anywhere near the price level.

I would recommend using say 20000uF (4 x 4700uF works very well) per rail for the PSU – PSU rejection is not that great.

Good luck, sit back and enjoy

John

[johnnyx]

I have to scan the documentation, so it may take a few days. Glad to be of assistance.

[thanh]

This design use 2 differential stage and Slone only use 1 differential stage . Perhaps this design has less noise than slone. I think like that

[cunningham]

Does this design require a regulated power supply? It probably would work better with one. I bet the PS regection would be better if the differential amp stages had a current source bias...or maybe a current source/mirror. Not suggesting you change anything but if quality is what your after...

[johnnyx]

One mod I intend doing is to regulate the input and drive stages, but leave the output stage unregulated. I have a suitable transformer, but I need to make new PCBs to implement it. This will stabilise bias conditions.
I replaced R3 (input tail resistor) with a 1mA current source, but I get small thumps because the bias now needs a certain voltage to operate.
I would check on the value of R9 though, (from the original documentation as you will see shortly) as I mentioned previously. The documentation from Maplin is incorrect in their description of the differential VAS load; they call it a current source rather than a mirror. Depending on supply voltage, if the voltage across R9 is too high, it cannot pass sufficient current on positive half cycles. Too low and dissipation in the VAS transistor increases, and is imbalanced wrt the other.
I thought about other mods, but it just ruins the beautiful simplicity of the circuit.

[cunningham]

Originally posted by johnnyx
One mod I intend doing is to regulate the input and drive stages, but leave the output stage unregulated. I have a suitable transformer, but I need to make new PCBs to implement it. This will stabilise bias conditions.
I replaced R3 (input tail resistor) with a 1mA current source, but I get small thumps because the bias now needs a certain voltage to operate.
I would check on the value of R9 though, (from the original documentation as you will see shortly) as I mentioned previously. The documentation from Maplin is incorrect in their description of the differential VAS load; they call it a current source rather than a mirror. Depending on supply voltage, if the voltage across R9 is too high, it cannot pass sufficient current on positive half cycles. Too low and dissipation in the VAS transistor increases, and is imbalanced wrt the other.
I thought about other mods, but it just ruins the beautiful simplicity of the circuit.

Nothing wrong with not being as simple as you initially want. If you use a regulated power supply, then forget the current source in place of the emitters' resistor in the diff amp. Use the emitters' resistor and include a currnet source and mirror circuit like I have drawn. This will give you much higer gain and power supply rejection, even if the output transistors are not regulated. As a matter of fact, you could probably eliminate the second diff amp stage and replace it with a single VAS transistor (biased with a current source of course). About the same number of transistors, just aranged differently.

[JohnW]

One worthwhile and simple modification is to add say 33R resistors in series with each of the Emitters of the input transistor pair T1 & T2 (so that’s 2 resistors).

Adding input stage degeneration to an existing design should be done with care as the change OL gain could upset the amplifiers stability criteria. Slew rate will be reduced, but can be restored by increasing input stage current (reducing the value of R3) – something to play with – but watch the amplifiers stability!

This will reduce the OL gain slightly, but will reduce overall distortion and improve the input stage tolerance to RF (from say from an RF polluted input from a CD / DVD player). Typically making the amp sound smoother and less hard.

John