Fresh simulations have shown using a VBE multiplier transistor in very good thermal contact with the output pairs provides very good bias stability from 10C to 55C. The quiescent current starts from its lowest of 12.1mA per pair at 10 Celcius to 16.58mA per pair at 55 Celcius. I think, this can be used to avoid having to use a thermistor and making things more complicated.
The actual physical circuit is shown as my avatar.
Why am I asking?
At present, the VBE multiplier is installed between output transistor 4 and output transistor number 5. The transistors on the sides of the multiplier are however, not output devices, but the drivers. So, the VBE multiplier senses the drivers' temperature. I think, this may be contributing greatly to the amplifier's bias instability issues.
The actual physical circuit is shown as my avatar.
Why am I asking?
At present, the VBE multiplier is installed between output transistor 4 and output transistor number 5. The transistors on the sides of the multiplier are however, not output devices, but the drivers. So, the VBE multiplier senses the drivers' temperature. I think, this may be contributing greatly to the amplifier's bias instability issues.
Last edited:
Why not attach the Vbe multiplier transistor directly to a power transistor, and avoid the problem? A TO126 type that can be clamped (metal face down) on top of a convenient power transistor, just by fitting on the same 3mm mounting bolt, would seem a simple and effective way of over-riding any influence of the drivers and good enough to test your suspicions at least.
Last edited:
I am worried about doing precisely what you wrote because I might tighten the bolt and break the feable and small BD135 transistor. I will have to spread the pressure using some stainless steel washers.
Yes, there is some risk of over-tightening but you can buy or form your own Belleville washers from plain, large dia. washers and these should distribute pressure more evenly over a wider area. If the pressure applied is so great that it risks crushing a TO126 semi though, then depending on the heatsink+insulator flatness, it would also be a risk to larger format, moulded epoxy semis.