decreasing tempco in this Vbe multipl.

[djdamix]

I've build an amp which sounds good too me. The bias current of the output stage is set at about 200mA with the VBE multiplier. It works well, but after a few time outside (winter at -5°C), when I turn it on, I get more than 1A bias current ... It comes from the vbe multiplier, and that's why I would like to decrease its tempco. (The Vbe x is fixed on one of the output transistor).

I've calculated that for 1°C, the current increase in the mosfet of 0,008A.

1 Volt less on ther gate give about 1A less, so I need to decrease 0,008V the gate to get 0,008A less on the output.

Actually this VBE give about 0,029V / °C which is too much. I searching for idea to decrease it ...

Thanks.
David

Attached is the schema of the vbe x

 

[Mr Evil]

You could try placing a Zener diode in parallel with R1, which will limit the maximum output for m the Vbe multiplier. A 9.1V Zener should be about right.

 

[MikeB]

You could try adding a small emitter-resistor to the mje340, a few
ohms should be enough.

 

[BrianL]

See Doug Self's books or his EW/WW articles or possibly his web
site for discussion. Your main problem may be lack of thermal
feedback from the output transistors to the Vbe multiplier.
I believe that Doug also covers the changing of the tempco
of the multiplier. If not there is a JAES paper by Malcom Hawksford
on the topic.

 

[Eva]

Try placing the Vbe multiplier directly on the heatsink just near one of the output devices

The plastic resin of the cases of the TO-247 devices has very high thermal resistance so extreme ambient temperatures make bias setting error unacceptable

 

[peranders]

Why don't you use a mosfet as temp sensing element? I tried with BS170 in my monster headphone amp with good results.

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

The picture shows how I have glued the BS170 on the IRFD120.