Yes it does have resolution more than 1 per degree, when it spits out a value in java it is scaled to degrees, but it ends up being more like 360.125 I am not sure if they are all significant digits because I have no idea of the precision of the encoder. What I am saying is we should know how many clicks per rotation of the wheels (the swerve action not the wheels that accualy move the thing) because the joystick gives out a number in degrees (from -180 to 180) and if we have the wheels be on a 360 degree scale the pid will work. On Tue, Jul 27, 2010 at 1:13 PM, Adam Czerwonka <Adam.Czerwonka@xxxxxxxxxxxx > wrote: > That information is in the spec, I believe that it should also be written > into last year’s software. Worst case we’ll look up the sensor info > online. I’m sure it’s got resolution better than 1 per degree > > > > *From:* team2039-bounce@xxxxxxxxxxxxx [mailto: > team2039-bounce@xxxxxxxxxxxxx] *On Behalf Of *Mark Amber > *Sent:* Tuesday, July 27, 2010 1:05 PM > *To:* team2039@xxxxxxxxxxxxx > *Subject:* [team2039] for tonights meeting > > > > Does anyone know how to find out exactly how many encoder clicks are in one > rotation of the wheel, I think that is an incredibly important part of the > program, I want it to be once we have one full rotation we can divide that > into 360 steps, and it needs to be extremely accurate, we cannot go by > looks, because every time the wheels spin it would just get that much more > off, so what ever error we have, say it is the length of this --> u on your > computer monitor it will be uuuuuuuuuuuuuuu off after 15 rotations. > > > > I was thinking we could put an encoder on the swerve action and spin it > around 20 or so times, so the a few millimeters of error would be 20 times > less significant, then we could divide by 360x20, instead of 360. > > > > Is my thinking correct, does anyone have better or more proven ways to do > this? > > -- > »»Mark Amber«« > -- »»Mark Amber««