If you can't get equal distortion on both sides of the wave form, then the trimmer you're using doesn't have enough resistance, or the FET needs too high a bias resistor.
When I biased mine, I wasn't happy with the resolution on my scope. i.e. it seemed like too much of a variance on the trimmer where the wave forms were symmetrically clipping. I ended up getting it close on the scope, and then biasing by ear like described somewhere in this thread. That may seem bass ackwards, but once you hear it happening, it's pretty easy to dial in. I'm sure other people's scopes are more accurate.
I also recommend the LS846. Had the same issue as everybody else in that the original transistors need to be biased with around 20K resistor. I didn't listen too close to the outcome, so maybe that's fine, but the LS846 ends up requiring 3-4K bias resistor.