Rotate Your Mousepads!

You are all patricians, but I'm still seeing so many MLT04s on horizontally laid-out QCKs. There is conclusive evidence that the MLT04 sensor operates better on vertical QCKs: http://www.overclock.net/t/1561041/reverse-en...t_24477902

As you can see from the bottom pair of graphs, the malfunction behaviour on vertical QCK is more desirable than on horizontal. On the former it just stops tracking altogether; on the latter it reports negative counts (throwing your aim to the other side or up/down).

If you use an OMB, IME3.0, WMO or IMO on QCK, turn the mousepad to portrait orientation rather than landscape. For all other mousepads, you can use MouseTester to check which orientation if any produces the same, clean malfunction behaviour (on my Goliathus it's the same: good in portrait, bad in landscape).

When opening up a WMO, these bitches: http://i.imgur.com/HWfSe.jpg sometimes break off. Causing the lower part of the upper shell to shift up gradually when using the mouse and blocking the M1/2 buttons. It's annoying to have to push it down every 30 minutes and generally messes with the feel of the buttons.
Has anyone found a way to re-fixate the shell after they broke theirs? Or just ideas on how you would go about doing that.

People are keen to tweak everything PC-related, but mice strangely enough are not really subject to comparable modifications.

After reading more into various engineering papers and data sheets dealing with optical input devices, I'm left wondering why apparently nobody has even attempted to mess around with their hardware/firmware. Just looking at this basic instruction sheet (http://www.pixart.com.tw/upload/ADNS-3090%20D...144306.pdf) it would seem easy to, once you gain access, control various aspects of a mouse with basic recalibration of the sensor following the manufacturer's instructions. Access can be gained by directly connecting to the PCB, but manufacturers ship out firmware over USB as well. Forcing/locking higher framerates, controlling voltage and current, overclocking the microprocessor, overdriving the LED? I'm not saying any of this stuff would necessarily yield performance increases, but it's just baffling that nobody has at least attempted to try and see what changing some of this stuff does to the device.

Or have you? I'd be interested in finding out about anything related you have attempted yourself, read about or any insight beyond that you could give as to how gaining access and in practice re-calibrating a sensor could work. I mean, mice manufacturers must do this stuff themselves as well, or they get the hardware pre-calibrated and just specify their needs to the sensor manufacturer. In any case, post-production modifications should be possible.