Z axis drive issues - a couple

My printer is generally running good, but I noticed today in trying to set my Z0 position that there is some backlash in my Z positioning and maybe an inconsistency in the way Marlin handles the Z offset. The Z mechanism is 3 lead screws driven with one stepper with one idler for tensioning. The belt was loose when I started this but is now tight. The results did not change. BTW, my Z endstop is a nozzle activated piezo with an M48 std dev of 0.0229.

question #1: How much deflection should I get on a span of 47mm GT2 6mm belt?

I have set my Z offset to 0.05 in Marlin. When I home Z, Repetier shows Z = -0.05. When I step Z by 0.01, Repetier shows it going from -0.05 to 0.00. The next step goes to 0.01 etc. If I step Z by 0.1 initially, Repetier shows Z = 0.05
This seems like an inconsistency with Marlin. I would expect it to either take 5x 0.01 steps to get to 0.00, or any single step to go to that position. IE, a 0.01 step goes to 0.01, a 0.1 step goes to 0.1, a 1.0 step goes to 1.0, etc.

question #2: Does anyone else see this inconsistency or is there a reasonable explanation?

If I use a 0.08mm "shim" (paper), 9x 0.01 steps from home yields Z = 0.08. The same actual height as 1x 0.1 step + 3x 0.01 steps (Z = 0.08) based on the shim test. Does anyone else see this result, or do I have something set wrong?

In playing with this I notice that my shim tests were inconsistent. What I found was that if I go to Z = 0.08 then do -0.01 steps it takes 4 steps to get the screws to move. The stepper appears to move and the belt on one side of the stepper feels like it is moving, but the screws do not move until step 4 or 5. Therefore Z doesn't change and the shim gets confused.

question #3: Is this to be expected or do I have a bad pulley, belt, or idler and if so how do I figure out which is the problem?

Edited 3 time(s). Last edit at 02/23/2019 01:44PM by ruggb.

What type of couplings are using on your steppers?

I've had similar issues on two of my printers when I was running flexible couplings - the type with a helical slot between the screw and the stepper. It drove me nuts for a while trying to work out what was going on. I switched to solid couplings just to eliminate that as a possibility and my issues were solved.

thanks, I have no flex couplings, just 1 stepper, 4 pulleys, 1 idler, 1 belt, 3 screws, and 3 nuts.
It only affects the retraction function so the effect on the print is pretty minimal.
I have another belt coming, I guess I may as well order more pulleys too. I don't know how the idler can create that but I guess I will pull that apart and have a closer look. It is pretty difficult to determine what is/is not moving for a 0.01 step.