[Lil_JV]

I am designing a MINT-like PCB and am wondering if I could just connect the BW pin on the buffers to the negative rail. It is only going to be used at home with a power supply.

JV.

EDIT: I am going to be using 5-lead TO-220 buf634's. Also is it possible to buffer the virtual gnd as well? If so, will I notice a difference or is it overkill. If I did this would it not be like a META42?

 

[Pars]

I think that Tangent or JMT or someone did a test, and found that the BW was optimal with some resistance in place. IIRC, I though it was 220 ohms, but perhaps more like 470. So no, I think I would leave space and traces for a resistor.

Hopefully Tangent will pipe in <g>.

 

[Lil_JV]

Quote:

I am going to be using 5-lead TO-220 buf634's. Also is it possible to buffer the virtual gnd as well? If so, will I notice a difference or is it overkill. If I did this would it not be like a META42?

 

[guzzler]

KurtW found 200R was the optimum value for distortion figures, half that for each additional buffer to a stack (2 buffers=100R) By buffering virtual ground, you are effectively making a META42 as that's the really big difference between it and the MINT, although the META also used the EL2001 buffers instead of the BUF634. One thing to ponder though: the TLE2426 has a lower output impedance than the BUF634, so it might sound better unbuffered to you. The other option is a TLE in an 8pin case style and install the noise reduction capacitor as well, which isn't available in the TO92 version

g

 

[Lil_JV]

Here is my layout. I know its not very tight but it my first attemp at making a layout and first time using eagle. The only thing not shown are the CRDs for class A biasing. The red are jumpers to connect the buffers to the positive and negative supplies. I decided to just connect the BW and negative pins together since the buffers have heatsinks anyways. The bottom buffer pins (in the layout) are pin 1(BW).

JV.