Default X and Y steps/mm are wrong

By default, the Robo 3D printer uses 80.69 steps/mm for the X and Y axes.

The printer uses GT2 belts and 20 tooth pulleys. The number for zero belt stretch is:

2mm belt pitch * 20 pulley teeth per revolution = 40 mm of belt per revolution

360 degrees per circle / 1.8 degree stepper motor = 200 full steps per revolution
200 full steps * 16 microstepping multiplier = 3200 steps per revolution.

3200 steps/revolution / 40 mm/revolution = 80 steps/mm

This Robo number might be intended to compensate for something but the mechanical parts (the motors and drive pulleys) will be exactly 80 steps/mm (on average per revolution. The tooth to tooth spec might have variation but there is no physical way for 20 teeth in 360 degrees to be anything other than 80 steps/mm). The only way you could have a number that is not 80 steps/mm is if the belt stretched. However, that would result in less than 80 steps per mm, not more.

To test my theory, I printed a 0.1 layer height 25mm calibration object with settings at 80 steps/mm this morning. The X direction came out as 25.02 mm in X and 25.05mm in Y. With the original settings, you would see approximately 25.2mm.

This isn't a huge thing for smaller parts but a part that fills the bed would be 2mm off at the extents.

 

So if this is the case why then did my print of several size cubes give the same deviation regardless of scale, +.2mm, eg 100.2 and 10.2?

 

That is a fantastic question. Perhaps the published firmware is not the same as the installed firmware? What I know for certain is that my Viki case is just a bit larger than it should be. Not enough to be noticeable until I went looking for it but quite obvious now.

At 100 mm, you should be seeing 0.8mm of "growth" in X/Y. The thing that makes all of this tricky is that the amount of plastic extruded, the plastic you use, the nozzle, etc. all play a part.

 

http://www.thingiverse.com/thing:93812 this was what i printed.

 

I printed that yesterday afternoon. With 80mm/step, I'm under 0.2 for all of them. Depending on where, how I measure, the variance is 0.0 to 0.2 off of the expected value for all parts.