@nardcox how silly of me to calculate from 100% SoC.. had quite the hangover on Sunday so my head was still at the after party 😆 

But yes, I want to be able to go ~100km/h for an extended period of time, other than that the acceleration of the 20kW feels like it's calling me 😇
Regarding the dimensions of the battery, I'm not planning to construct it with the plastic square grid, at most a zigzag grid but most likely without any plastic grid to maximise space efficiency. 
Without any plastic grid you can fit a 20S30P with BMS inside the original plastic housing. 20S35P is getting tight for just the cells alone.
This method has it's downsides but since space is an issue, every cubic cm I can save is a win imo. 
Main focus is that I have to make absolutely sure that the border cells of each parallel group can never touch each others plastic sleeves.. No one should be comfortable with separating 2x 40 cells in parallel just by their thin shrink tubing.
1mm resin infused carbon fiber lining in between the parallel group will do the trick (I'm using CF because at my work I got a near infinite supply of it).

I have yet taken apart the complete bike so I don't know exactly how I'm going to do it but I'm thinking getting rid of the original plastic housing and fabricate a larger one out of 1-2mm sheet metal (these sheets will ideally be placed just behind the plastic side panels).
Then load in the battery from the side before bolting it all up. 
But you're right, plenty of modifications to make and I suspect the downtime to be ca. 2 months as a lot of planning will be done last moment.
Fortunately the place I work at will come in very handy as I have access to welding equipment, mills, lathes, CNC machines and just about every handheld power tool you can think off.. all that is left to do is to convince my boss to let me use it 🤣  
Thanks for the heads up about that space being limited underneath the bike. I will think of a plan B in case it doesn't fit in there.

Milling the axle down 1mm on either side sounds like a good plan but I'm not sure about bending the arms a total of 1cm apart. Like I said, metal doesn't has to be pushed to it's stress limits in order to get actual stress fractures, constant stress will take care of that. 
But a big positive is that you can use your rear brake with it's original setup.

How did you manage to test the motor under load when you haven't been able to get it into the swingarm yet? (not a rhetorical question, an honest and curious one instead). 
Do you have the data cable to hook up the motor via USB? 

Hi Bob,

Yeah the plastic holders do add a little bit of space so if that's important hotglueing it together is the most efficient. The holder I now have add 1mm to each cell, they are placed 19mm HoH.

Regarding the motor, I'm now using the stock motor for testing purposes, haven't driven the new motor yet unfortunately, only on a bench without load and there it was just fine.

Hi Nardcox,  

I see from your photos that you managed to mount the QS273 in the end,  do you buy a different swing arm or modify the motor shaft?

Ive been considering buying a QS773, 40H 4000W but the mounting issues are putting me off...