It's a little complicated, but the answer is 1 of them is offset, the other is stock. Changing from fine contact to full contact affects arcing, but HV is always there. The best current, has been with less brush surface area contact and when that offset brush is near 80 degrees - see pic.
Don't forget though, nothing is scientific in my setup yet...it's all plug/change/note/change. Not much scope work yet or multiple multimeters.
This motor is from a car vacuum cleaner, Dirt Devil.
The motor was encased in plastic, which supported the brushes and bearings. I've removed one side and 1 brush, but that means 1 brush is still stock.
The other brush sits on a piece of thick copper (house gauge) which is nailed to the wood piece that the motor is also affixed to. So one side is stock, the other can be moved around.
To complicate further, the stock side brush is filed to have a thinner contact area, probably a mistake. The other side is using a square brush from another motor. Mainly, because the arcing will likely eat a brush in short order and these were short to begin with.
I've got it running on 12V again, but, is only a temporary measure.
Am using a 3rd brush
The third brush connects to the same input (Positive supply) as the one on the end of the copper wire.
It then is placed where the original stock brush would go LOL
Upshot of all this.
If your motor won't spin up on 12V after modifying the rotor, attach a 3rd brush to a clip lead and put it near TDC (Top Dead Center) and move it slightly toward the brush side that you're taking the power feed from.
The only point of which, up to now, is to try out various other measures while running at the original 12V.
I do however see a link to thoughts on 4 brushes. In which case, a 4th brush at the other side (BDC) at a slight offset would also likely improve running.
Pic attached.
On left is me holding the 3rd brush in place, where there used to be one anyway.
A few degrees off from the top is the mid brush, held with the thick copper wire. On the right is the other brush.
Oh and the coil and LED on top is there because of the wireless noisy field, the LED lights up