Those pics of the circuit board are old I have put some covers over the longer leads now.
I've had another quick look and I think I've got everything connected correctly around R13/14/15/16 / C9 etc. I also checked it against Mike's layout.
I was concerned that the layout might be an issue but Mike's is also pretty tight (it fits in a 1590B enclosure).
Thanks for the tips. I'm also building a Bassman emulation based on your bassboy schematic.
@ samhay & hex_void
Yep bias voltage is the issue. Thanks for you advice. I originally thought it must be to do with biasing but because Mike's vero used 470k bias resistors also I assumed it couldn't be that.
I've used 10k resistors and 100uF caps and it's stable now.
I've read through RG's article on biasing before, but I must admit I don't follow it.
In this circuit most of the opamps are connected straight to Vref, there are 1M and 10k resistors connecting to Vref on the first and last opamps. So that would suggest using 10k or 1k resistors.
Looking at the supply current (http://www.ti.com/lit/ds/symlink/tl062.pdf
Multiply that by 10 and divide into 9volts and half, you get 18k.
I know I've got it working now but I want to understand why (or how to avoid this in the future
Now that I have it working, I'm thinking of doubling the supply voltage with a charge pump. The original unit runs of 15v so 18v will give me some more headroom.
I'll need to have the charge pump on a separate veroboard, so I was thinking should I include a Vref buffer with 470k/1M bias resistors?
Thanks again for everyone for putting up with my noob understanding.