test cube 17,5mm instead of 20mm

Hi all,
in the last few days I have built my first 3d printer, a prusa I3 pro B.
All seems to work fine and prints are nice except for measure. My test cube is 17,5mm on each axis instead of 20mm.
I have 20 tooth timing pulley and gt2 timing belt with steps setted to 80.

Every tip is welcome

Measure with digital calipers and calibrate your steps accordingly

Thank you.
So, all I have to do is to calculate the right number of steps with this(?) formula

CorrectedStepsPerMMValue = (20 mm*CurrentStepsPerMMValue)/(MeasuredSide)

and corret the value in configuration.h, right?

More or less
I found it more easy to take these formulas as a guide only.
You can always get some measurement error, like how much filament is over extruded or a little bump on the surface.
For this reason I actually measure the travel of the axis / extruder and adjust until within reasonable accuracy.

Did you count the teeth on the pulleys? 18 teeth pulleys exist and some people have received them when they were expecting 20s. There's no good reason for a 20T pulley to give 13% under the expected movement, so if you really have 20T then I'd check for loose belts, slipping pulleys and skipped steps.

And by "each axis" does that include Z or just X & Y? If you're including Z then the model is most likely scaled compared to what you are expecting, or you have a major coincidence involving your threaded rods as well as the belt drives.

Edited 1 time(s). Last edit at 11/16/2015 01:15PM by JamesK.

Hi JamesK,
I'll count the tooth in the afternoon, after work.
When I read the list of components, It said "20 tooth" and I haven't checked.

For "each axis" I mean also Z.

Pulleys have 20 teeth.
Maybe some setting in Marlin is scaling my print?

I don't remember anything in Marlin that scales the print (other than the steps per mm values). There's a couple of simple things you could do to check what's going on:

a) home the print head to 0,0,0 and mark the position on the bed. Use "g1 x100 f1000" to move the print head to position x=100, mark the new position on the bed and measure the distance travelled. Do the same for x=200 (or whatever fits on your print area). Do the same with the y axis. If these measurements match what you expect then we know the printer firmware matches reality and the problem is in the g-code. If these measurements show the same underscale as the print then we know it's a firmware config problem (or a strangely reproducible mechanical error). (You could look for mechanical problems by increasing the 'f' value to command increasingly rapid movements, and going back and forth between say x=50 and x=150)

b) Did you make your own test cube? If not, it's probably worth doing. Just make a solid cube in the cad program of your choice, and then slice it with a single perimeter, no infill, no top or bottom. The resulting g-code should be very easy to read, so open it up in a text editor and check what the ranges of the x, y and z coordinates are. This will let you see what the slicer is commanding the printer to do and should make it obvious if something odd is going on at the slicer level.

I measured the distance between x10 and x110 and is 98, same for Y axis.

Edited 1 time(s). Last edit at 11/17/2015 11:02AM by Alticcio.

Interesting, that's almost the same delta as on the 20mm cube (ie, expected was 100, observed was 98, delta is 2mm). If you do x10 to x60, and x10 to x160 do you get a similar delta for every case? If the delta is independent of the distance travelled it might indicate that something is loose rather than a calibration error. Binding bearings might also produce that effect if it takes several steps before the head first moves. You could look for that by doing successive 0.2mm steps and seeing if the head moves on every one or takes several before there's any movement (after a move from the opposite direction, this is basically a backlash test).

Same delta for every case.
x(y)10 to x(y)160 it's 148
x(y)10 to x(y)160 it's 148

I tried the test you suggest, moving x by 0.1 repeteadly.
It seems like moving on +x , the axis moves for every step, but moving on -x it moves one yes and one not.

Edit. even moving from 10 to 20 same result, almost 8

Edited 1 time(s). Last edit at 11/17/2015 01:11PM by Alticcio.

The mystery deepens! The constant delta strongly suggests some sort of backlash problem, but if you can only see it skip one .1mm movement then that only accounts for 0.1 out of a roughly 2mm error.

If you do a g1 move and then, while the stepper motors are still powered up, try and physically move the nozzle from side to side or the print bed from side to side, is there any movement that might explain where the missing 2mm is going?

What sounds really strange is the identical error on the 2 axis. It's like some math gone bad.

Nozzle and bed are not moving after G1 move.

Edited 2 time(s). Last edit at 11/17/2015 02:36PM by Alticcio.

I was thinking, is it possible to be a driver current issue?

I suspect your maximum jerk setting is too high for the motor current you are using. Try reducing the jerk setting or increasing the motor current.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Hi all, increasing the motor current seems have resolved the issue.

If so: All good
But keep an eye on the motor temp for a while - you don't want to overheat a motor with too much current

That's great, congrats!

Hey guys, after a day printing, all seems to work good.
Thank you and, If you like, take a look at the photo of one of my first, good, print attached.