Waves NLS and SSL Channel - where to place?

As a virtual summing box or summing console - The NLS channel is inserted on the last insert of the DAW audio tracks, like a direct out routed to an analog console. The NLS Buss is placed on the first of the the master track or any other buss, just as the stereo return would be routed from the analog console back to the DAW.
To simulate console preamp coloration - The NLS Channel is placed on the first insert of your DAW audio tracks, and the NLS Buss can then be inserted at any point on the Buss or Master track.
 

Matron Landslide
NLS certainly makes a difference, on tracks/channels the idea is to place it last in the chain, on buss's you place it first. From the Manual

BenMMusTech
I believe that I may have the best understanding of this question, as I spent some time fathing about in the beginning with this very same question. And even though it may appear obvious to some - you should still really think about it and then decide how to employ these console emulators.

As two of the posters above have suggested, there are two ways to use these plugs. The first at the beginning of the signal chain, which I don't use very often, and would only use if I was going to create an almost exclusively analog track. Think acoustic instruments, and as the above post suggests this is for colouring of the input of the audio signal. Of course all processing is done non-destructivly. If the OPs uses the plugs this way, then the SSL channel strip would come second. If you want to put a DT effect or something along that line as well - you can put it after the SSL...just leave enough headroom for the DT effect - mostly I use these types of effects before any compression and EQ processing.

The other way to use the NLS plug is as above suggested as a summing box, which I recomend if you're like me and only a few instruments are acoustic. This means the NLS is the last effect in the chain and you would put the SSL channel strip before.

I would also switch on the 64bitfp audio file format and up sampling facility to get the best out of the Wave's plugs. To really bring out the idea of analog emulation you need to be able to replicate analog nuances like higher gain stage, because in theory, and especially with Wave's plugs, each of these plugs operate best in the sweet spot. This requires a higher gain level in most cases, and sometimes a lesser one to achieve this. Only 64bitfp audio files gives you the dynamic range to achieve this. 64bitfp audio, I believe returns digital to a similar state to analog. Think as I've mentioned, a larger dynamic range and indeed less glitches and mid harshness when doing a lot of processing to said audio file, which adding in second and third level harmonic distortion is...NLS.

Finally also be aware, that each emulated console is useful for different types of production - aesthetics. So for me, the SSL is ultra clean and really good for punch and indeed really big mixes. I've just complicated a 100 track piece and it would have sounded really bunched up in the low-mids if I'd used the TG12345...the neve might have been able to handle it...but it still would have been probably too dense in those low-mids. The SSL was designed and brought onto the market for these reasons I suspect. You really have to know some history to make these plugs work optimally.