difference in scale between the stl file & print

[Polygonhell]

Repetier - the one that SeemeCNC posted..... only changes were to flip direction on motors.

Do you have the 15 or 22 tooth pulleys and is your RAMBO an earlier or later one?
If the pulleys are plastic they are 20 tooth, and unless your kit is one of the very early ones you'll have a later RAMBO board

[Polygonhell]

ok i have a problem with scale. My X/Y cartesian coordinates seem to be scaled ok size wise on a 20 mm print, but my Z axis travel seems way off. when i move through setup i only get a height of 265mm or so on Z in firmware, but my actual length is 350-355 mm from the top of my glass build plate.

So my question is if i move the Z axis steps per mm does that effect Z on just the Z motor, or the cartesian z axis?

Have you checked the actual position with M114 or are you looking at the value in Repetier Host, the latter is only accurate after homing if you explicitly set the height of the printer correctly in Repetier Host.
If you move up or down 10mm does it move 10mm ?

[ROBRMI]

plastic pulleys, and rambo board is newer one i'm pretty sure. kit is like 2 mo old. did M114 when setting bed level and all that jazz.

I think a 10 mm move up in Z direction is approximately 12 mm measured, but x/y moves seem to be accurate....

[Polygonhell]

plastic pulleys, and rambo board is newer one i'm pretty sure. kit is like 2 mo old. did M114 when setting bed level and all that jazz.

I think a 10 mm move up in Z direction is approximately 12 mm measured, but x/y moves seem to be accurate....

What I'm trying to clarify here is how are you measuring the 265mm?
If you home and type M114 what does it report?

It really ought to be impossible for X and Y to be correct and Z wrong unless the arm length is set incorrectly in the firmware, since you haven't change that, I'm trying to establish how you are measuring the error.

It does occur to me that you might see something like this if the steps per in the firmware didn't match the steps per in the EEPROM, there is currently a bug in repetier that requires they match, or the geometry is incorrectly calculated.
Is the firmware set for the correct number of pulley teeth?
What steps per mm values are set in the EEPROM menu?

[ROBRMI]

Gotcha, I home and type M114 and the Z on both the LCD and on the Repetier host show 265.8 (which is the max length setting in Repetier firmware for:

#define X_MAX_LENGTH 265.8
#define Y_MAX_LENGTH 265.8
#define Z_MAX_LENGTH 265.8

this is the position where it touched the printbed when i did the bed height calibration measurements.

I printed a big block basically 3.5" x 1.5" x 10.5". when i physically measured the block, i had 3.6" x 1.5" x 12.9"

the block was drawin in Autocad inventor, and exported to STL.

I have the EEprom disabled currently just working from downloads.

I'm not sure about the pulley teeth. I thought Repetier took pulley diameter and calculated that:
#define PULLEY_DIAMETER 10 // pulley diameter in milimeters on stepper motors - THIS IS FOR STRING DRIVEN SETUPS LIKE KOSSEL, CEREBRUS ETC....
#define PULLEY_CIRCUMFERENCE (PULLEY_DIAMETER * 3.1415927)
#endif



#define STEPS_PER_ROTATION 200 // 200 is common, which are 1.8 degree per step motors, .9 degree motors would be 400 steps per rotation
#define MICRO_STEPS 16 // RAMBo 1.0 and earlier use 8 (1/8 stepping), 1.1 and on uses 16 (1/16 stepping) RAMPS and Pololu drivers are normally 16

/*Number of delta moves in each line. Moves that exceed this figure will be split into multiple lines.
Increasing this figure can use a lot of memory since 7 bytes * size of line buffer * MAX_SELTA_SEGMENTS_PER_LINE
will be allocated for the delta buffer. With defaults 7 * 16 * 30 = 3360 bytes. This leaves ~1K free RAM on an Arduino
Mega.
*/
#define MAX_DELTA_SEGMENTS_PER_LINE 30

// Calculations
#define AXIS_STEPS_PER_MM ((float)(MICRO_STEPS * STEPS_PER_ROTATION) / PULLEY_CIRCUMFERENCE)
#define XAXIS_STEPS_PER_MM AXIS_STEPS_PER_MM
#define YAXIS_STEPS_PER_MM AXIS_STEPS_PER_MM
#define ZAXIS_STEPS_PER_MM AXIS_STEPS_PER_MM

#else
/** Drive settings for printers with cartesian drive systems */
/** \brief Number of steps for a 1mm move in x direction.
For xy gantry use 2*belt moved!
Overridden if EEPROM activated. */
#define XAXIS_STEPS_PER_MM 106.666
/** \brief Number of steps for a 1mm move in y direction.
For xy gantry use 2*belt moved!
Overridden if EEPROM activated.*/
#define YAXIS_STEPS_PER_MM 106.666
/** \brief Number of steps for a 1mm move in z direction Overridden if EEPROM activated.*/
#define ZAXIS_STEPS_PER_MM 106.666

It's very possible i have the firmware wrong, I swapped to Repetier a week or two ago....

[ROBRMI]

plastic pulleys, and rambo board is newer one i'm pretty sure. kit is like 2 mo old. did M114 when setting bed level and all that jazz.

I think a 10 mm move up in Z direction is approximately 12 mm measured, but x/y moves seem to be accurate....

What I'm trying to clarify here is how are you measuring the 265mm?
If you home and type M114 what does it report?

It really ought to be impossible for X and Y to be correct and Z wrong unless the arm length is set incorrectly in the firmware, since you haven't change that, I'm trying to establish how you are measuring the error.

It does occur to me that you might see something like this if the steps per in the firmware didn't match the steps per in the EEPROM, there is currently a bug in repetier that requires they match, or the geometry is incorrectly calculated.
Is the firmware set for the correct number of pulley teeth?
What steps per mm values are set in the EEPROM menu?

**********************************************
Also I agree on the height thing, I may be proportianally off on all three axies, and it just shows up on Z when i'm doing something tall. The 20mm cube is 20mm x20mm x 13.13mm?

[Polygonhell]

This is the code that actually set the steps per, the code you quote above is for filament drives not belt drives. Ensure that PULLEY_TEETH is set to 20 if you have the plastic pulleys.

Code: Select all

#if DELTA_DRIVE_TYPE == 0
/** \brief Pitch in mm of drive belt. GT2 = 2mm
*/
#define BELT_PITCH 2

/** \brief Number of teeth on X, Y and Z tower pulleys
*/
#define PULLEY_TEETH 15
#define PULLEY_CIRCUMFERENCE (BELT_PITCH * PULLEY_TEETH)

[Broose]

I adjusted the X/Y scale slightly in my firmware because I noticed parts were undersized, it's controlled by the arm length number in the firmware.
But I'd suggest printing something simple to test with like a large rectangle I used a 100x50 piece, I think my current scale maybe slightly oversized, but it isn't by much.

Polygonhell - So does that mean there is a linear relationship between DELTA_DIAGONAL_ROD and XY scale?

[ROBRMI]

This is the code that actually set the steps per, the code you quote above is for filament drives not belt drives. Ensure that PULLEY_TEETH is set to 20 if you have the plastic pulleys.

Code: Select all

#if DELTA_DRIVE_TYPE == 0
/** \brief Pitch in mm of drive belt. GT2 = 2mm
*/
#define BELT_PITCH 2

/** \brief Number of teeth on X, Y and Z tower pulleys
*/
#define PULLEY_TEETH 15
#define PULLEY_CIRCUMFERENCE (BELT_PITCH * PULLEY_TEETH)

Thanks for the info, out of town again, but Thursday i can try the changes. It makes complete sense though, 265.9*20/15 = 354.53mm, which is probably where my height should be.

[Polygonhell]

I adjusted the X/Y scale slightly in my firmware because I noticed parts were undersized, it's controlled by the arm length number in the firmware.
But I'd suggest printing something simple to test with like a large rectangle I used a 100x50 piece, I think my current scale maybe slightly oversized, but it isn't by much.

Polygonhell - So does that mean there is a linear relationship between DELTA_DIAGONAL_ROD and XY scale?

Yes, but it ought to be pretty close, I think I ended up increasing it by ~1mm, but I think my firmware results in very slightly oversized prints in X/Y.
DELTA_RADIUS also plays a part in this, but in a none linear fashion.
Note it has no effect on Z motion, that entirely down to steps per mm for X/Y and Z.

[ROBRMI]

Update, The 15 tooth 20 tooth problem was the issue, and after a long night of calibration friday, by saturday afternoon i was in buisness printing to scale with better quality than ever before. Thanks for the help to all !

[iad]

I know this is an older thread, but I have a similar problem, my scale is off by a factor of 1.0098 in all dimensions. I've gotten used to simply adjusting the scale of the STLs in Repetier Host, but I'd prefer to fix this properly, using the firmware configuration; what parameters should I be looking for?

Thanks!

-Luke

[bdjohns1]

I know this is an older thread, but I have a similar problem, my scale is off by a factor of 1.0098 in all dimensions. I've gotten used to simply adjusting the scale of the STLs in Repetier Host, but I'd prefer to fix this properly, using the firmware configuration; what parameters should I be looking for?

Assuming your bed is really truly level, you want to adjust the Diagonal Arm Length in the EEPROM settings. Take whatever your current setting is, and multiply by 1.0098 if your parts are too small (scaling up) or divide by 1.0098 if your parts are too large.

[iad]

Thanks, I'll give that a go. I'd say the bed is level with 0.002" or so. (i.e. no dial indicator, but it passes the "uniform-grab-on-a-sheet-of-paper" test nicely.) I assume I'll have to recalibrate the bed height after changing the arm length parameter?

-Luke

[bdjohns1]

I don't think so, but it would be worth a sanity check - I think the way the math works out, the diagonal length is only supposed to affect the scaling in X and Y.

[Polygonhell]

If it's off in ALL dimensions, i.e. X/Y and Z you want to change the steps per mm, not the arm length.
But I'd make absolutely sure that the error scales with the print before messing with either, the steps per mm is derived from the tooth spacing on the belt and I would assume in general that ought to be pretty accurate.

[Polygonhell]

I don't think so, but it would be worth a sanity check - I think the way the math works out, the diagonal length is only supposed to affect the scaling in X and Y.

The value has to reflect the actual pivot to pivot size of the arms, the value not only affects X/Y scaling but unfortunately introduces doming in the effector motion, the doming is a different curve that that introduced by delta radius and attempting to dial that out one using the other will then leave you with a saddle shaped motion at the outside of the build plate.

[iad]

Hmmm... Well, when I print the little 25mm hollow cube, it comes out 24.75mm on a side (every side), so it's definitely X/Y/Z scaling, and I've noticed the same scaling factor on a variety of other parts I've designed. It's very consistent... Granted, it's less than 1% off, but when I'm design press-fit parts that attach to off-the-shelf components, it definitely comes into play.

Ah, maybe I'll just leave well enough alone, and keep scaling things in Repetier Host.

-Luke

[geneb]

Wouldn't that 1% account for the shrinkage of the plastic?

g.