FirstWatt J2

[EUVL]

I have done this umteen times in real circuits decades ago.

Best to try it yourself to be convinced.
Don't take my word.


Cheers,
Patrick

 

[JeffYoung]

Oh, I believe you. It's just that that still leaves me not understanding why it is.

Cheers,
Jeff.

 

[EUVL]

The positive tempco of JFETs are nowhere as bad as BJTs.
They also have less gm / transconductance.


Patrick

 

[Itsmee]

Sorry, my question was academic, I was just thinking about the control loop with two (or more) devices and the best way to control them - as a whole or individually, each has its drawbacks.

As for the J2, the PCB's Iv'e ordered have a single output - hope there's not too much debugging

Oops, Just spotted J3a drain not connected

 

[EUVL]

If you wish to control the two current sources exactly, both for bias and for dynamic current modulation, then you should use separate optocouplers.
They are then only connected at the speaker ouput.
And then they do not need to be matched, whereas the Semisouth's still have to be.

Academic, as you said.


Patrick

 

[Itsmee]

Thank you, it still leaves me wondering about the drains connected together, it changes the loop behaviour.

 

[JeffYoung]

@Itsmee, if you do control them as a whole, you're still only measuring a single set.

 

[JeffYoung]

The positive tempco of JFETs are nowhere as bad as BJTs.
They also have less gm / transconductance.


Patrick

Thanks, Patrick.

This stuff slowly sinks in....

Cheers,
Jeff.

 

[EUVL]

Toshiba MOSFETs are almost perfect penthodes.
Which means that above 4V, the drain current is essentially independent on Vds,
and only dependent on Vgs.


Patrick

 

[Itsmee]

Penny dropped, the answer to my question was obvious with one current source and multiple outputs.
With only half a brain cell, it takes me a while to chew through an idea.

 

[EUVL]

My previous answers were always referring to one current source per Semisouth.
If anything, you need to limit the dissipation of each MOSFET in the current source.


Patrick

 

[EUVL]

Unobtainium made available.
Interest for genuine Toshiba 2SK3497 curve tracer matched pairs & quads


Patrick

 

[darrr]

I can't open links from first post. Maybe someone has the correct ?

 

[JeffYoung]

Yeah, they're dead all right.

Presumably:
FIRST WATT
http://www.firstwatt.com/pdf/prod_j2_man.pdf
6moons audio reviews: FirstWatt J2

 

[JeffYoung]

Probably a stupid question, but:

The 2SK170s in the current source. They're each going to dump their Idss no matter what the other is doing, right? So a 2mA and a 6mA will act the same as two 4mA?

Cheers,
Jeff.

 

[EUVL]

> So a 2mA and a 6mA will act the same as two 4mA?

Yes.

But the one I posted, I use a fixed drain resistor and adjustable current source with an extra source resistor trimmer.
Guess why ?


Patrick

 

[JeffYoung]

I thought it was so that you could use BL parts (instead of GR). But I suppose it also lets you use any part without knowing its Idss?

 

[EUVL]

No, much more important than that.

Think how you calculate open loop gain.


Patrick

 

[JeffYoung]

I do this:

and then plot V(out) / V(ltp-negative-input). But I'm not sure if that's the "normal" way or not....

Cheers,
Jeff.

 

[EUVL]

I meant calculate, not simulate.


Cheers,
Patrick