Hey, good question! I’m not sure and I couldn’t find the info in the H2n manual.
Doing some poking around though, I think my 2nd assumption was off for the H1n. The manual says the input gain is from -infinity to + 39 dB. So it does provide trimming! @circuitghost - can you confirm this? i.e. if the input knob is at 0, then there is no incoming signal?
Thanks for exploring this with me 
happy times in the green, orange, and red zones to you all!
Yeah, I haven’t looked at the manual but the specs on the H2n website does mention:
AUTO GAIN: Automatic input gain adjustment (3 levels)
Don’t know if they describe that more in the manual.
@circuitghost 's method seems quite good. Turn it down until you don’t hear (or see) any clipping. 
They do describe auto-gain a little, seems like a simple feature. For me, I just leave the H2N gain at 0 and then adjust from the SiX. This seems best to me since the SiX has plenty of output and the H2N has no attenuation function.
Going to play around with going just less than clipping. On many of my tracks, I went too conservative with the levels and have to boost them in a DAW later
This is correct. At zero, it records nothing.
Awesome a mystery that was never really a mystery is solved! Jealous of your multi-colored recording experience 
This whole conversation is just an echo of what everyone has been saying: do what works best for you
If I should provide some context to my madness, there’s the fact that I’ve found that kits in general but Blackbox in particular, just doesn’t sound all that great if you max it out. You can push it, but it behaves at its best when it’s coming on strong enough to activate the compressor, but still does it fairly subtle. Which is around 70-80% of its max output capacity. So that’s how I came to that conclusion.
Going into the SiX, I found that if I pushed all channels to the extreme, there wasn’t much left to experiment with once the g bus compressor kicked in. So I was somewhat conservative on the lower orange end, but still wanted to be in orange, because all kinds of stuff happen with sound when it flows through the SiX. And the more you can push it, the more details emerge. But I want to save some for the compression and master output, which is why I decide to pimp the signal at that end. The movement, push and pull of the g bus compressor is like an analogue LFO on your input, if you find that sweet spot.
And then bring the master to the max to make all these details come out as much as possible.
So this is why I did (still do) what I did, by ear and trying out stuff.
Now, having just added the Fusion … well, that’s not perhaps a different story, but certainly a second chapter to the ongoing one
I realise I know nothing about db:s and signal chains and stuff, so I’m not saying I got it figured out. I’m just providing some context, this is how I did it and this is generally how I do things. It doesn’t always turn out great. But sometimes, it does.
Actually it sounds to me like you have it perfectly figured out
I think I’m finally starting to experience this on recent projects. It’s taken some time, but the time spent is paying off. There have been many incarnations, but SiX is integral to my setup at this point. I had it posted for sale a couple months ago, thinking about really scaling down my setup, but I’m really glad I held on to it. Plus, there’s this great group of folks helping me out with these great tips and tricks.
I’m experiencing the same with a about 1db bias to the right 
Reply from SSL support saying a plus/minus 0.5 db deviation per channel can be expected.
Has anyone rack mounted they SiX? I can’t see how many “U” it requires?
If I’m not mistaken 1U = 1.75" and SSL is 12.2" so 7U.
Thank you! Everytime I look for the measurements It just seems to make no sense.
Yes, they should give the damn U size as part of the dimensions, seeing as it’s rack mountable.
Also, the width (excluding that extra trim on either side) is the exact size so you can fit two next to each other in a 19" rack (nudge nudge wink wink).
Hahaha. It did cross my mind a while back, and the joke is the cost of the ears for x1 rack mounted is £100?!? I think I’d sooner put that towards a used SiX haha
I’m using the stereo cues as stereo sends to some FX units - and running the FX out into the 2 ext inputs. It seems that this means the sound returned bypasses the bus comp
is there any way to avoid this - except by running the return signal into one of the stereo channels?
Diagram on Page 16 of the User Guide shows Ext 1 & 2 is added to main bus before the main bus goes to the compressor.
Yes this was something I read as well and was confusing - but I think I realize my mistake now.
I was “adding” the external input to the monitor mix in the “mon source” section - instead of adding it in the Main section 
but of course then I can’t control the incoming level - level control is only from the send part + whatever is on the FX device