incorrect height

Hi, I recently bought a delta kossel pro BIQU, I had to make several changes because it was a real crap. I solved the problems of the concave plane and now I miss solving the last problem ... if I mold a 20mm cube the two sides are perfect while in height I have 20.30mm .. what should I fix? use firmware marlin ... thanks in advance and sorry if the question is trivial but I'm a neophyte with the delta, it's a new world

---

Printer 3D : Prusa i3 clone Zonestar P802NR (dual extruder)
Printer 3D : Delta Kossel PRO BIQU modified
Laser Cutter : Laser AS-4 3W
CNC : PCB Cyclone 2.0 (DIY)

First fine tune your extruder calibration, print objects to get the exact extrusion amount, a slightly rough top layer can add 0.3mm easily. Overextruding by a tiny amount can add 0.3mm.

Once you've done that, you need to print 2 cubes but different heights.

Print 1 cube 20x20x20
Print another next to it 20x20x40

Now if they are both 0.3 too high so 20.3 and 40.3 then the issue is with first layer height/squash change settings there, check your first layer nozzle to bed gap and adjust. Could just be rough top layers...

If one is 20.3 and the other 40.6 then you have an issue with your steps/mm. Unfortunately, on a delta it's not just a case of just changing them from say 80 to 78.8 ((40/40.6)*80) although you should do this and it will help but you then need to recalibrate your machine. Use least squares method its very easy even without a probe, go to [www.escher3d.com] put in your parameters, probe the height at the points suggested, then calculate and change your values. Easy.

EDIT - the only thing that can catch you with least squares is deciding when you enter the heights at each probe point if they should be negative or positive (depending on too high/too low) I can never get this right. So enter them calculate, try the new values and if its worse, change the signs around try again.

Also make sure your firmware allows you to go below Z=0 for areas where the nozzle is too high, or you have to measure the gap with feeler gauges instead which is tedious. Once properly calibrated you can set the firmware to allow Z=-0.5mm as min (in case you want to use grid levelling or something similar).

Edited 1 time(s). Last edit at 02/05/2018 06:48AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
DjDemonD
First fine tune your extruder calibration, print objects to get the exact extrusion amount, a slightly rough top layer can add 0.3mm easily. Overextruding by a tiny amount can add 0.3mm.

The extrusion is ok

Quote
DjDemonD

Once you've done that, you need to print 2 cubes but different heights.

Print 1 cube 20x20x20
Print another next to it 20x20x40

Now if they are both 0.3 too high so 20.3 and 40.3 then the issue is with first layer height/squash change settings there, check your first layer nozzle to bed gap and adjust. Could just be rough top layers...

If one is 20.3 and the other 40.6 then you have an issue with your steps/mm. Unfortunately, on a delta it's not just a case of just changing them from say 80 to 78.8 ((40/40.6)*80) although you should do this and it will help but you then need to recalibrate your machine. Use least squares method its very easy even without a probe, go to [www.escher3d.com] put in your parameters, probe the height at the points suggested, then calculate and change your values. Easy.

ok , I'll try ! thank you for now

Quote
DjDemonD

If one is 20.3 and the other 40.6 then you have an issue with your steps/mm. Unfortunately, on a delta it's not just a case of just changing them from say 80 to 78.8 ((40/40.6)*80) although you should do this and it will help but you then need to recalibrate your machine.

I do this on the prusa..but on the delta the steps/mm are equal not individual I would also shave the others

the firmware is

#define DEFAULT_AXIS_STEPS_PER_UNIT { XYZ_STEPS, XYZ_STEPS, XYZ_STEPS, 95 }

where

#define XYZ_STEPS ((XYZ_FULL_STEPS_PER_ROTATION) * (XYZ_MICROSTEPS) / double(XYZ_BELT_PITCH) / double(XYZ_PULLEY_TEETH))

Oh yes, I don't mean reduce the steps per mm for Z only. Reduce all the steps/mm for x,y and z by the same amount.

If you think about it when the effector moves up or down purely Z only, all that happens is all three motors turn by the same amount. So to position your nozzle at z=40 all 3 motors must move the carriages by 40mm exactly. However when you move in X and Y then the effect of lowering the steps/mm or increasing it will mean the movements will not be correct (not remaining within the XY plane, you'll get concave/convex motion), so you have to recalibrate once you get the steps/mm correct.

Edited 3 time(s). Last edit at 02/05/2018 09:26AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

my configuration.h

[pastebin.com]

I haven't really time to check through it all, but if the 40mm cube is proportionally larger still than the 20mm cube change steps/mm for xyz as your have shown above:

#define DEFAULT_AXIS_STEPS_PER_UNIT { XYZ_STEPS, XYZ_STEPS, XYZ_STEPS, 95 }

then recalibrate. Seriously try the least squares you can calibrate your machine with one set of probe point heights. It amazes me that anyone would use any other method.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
DjDemonD
If you think about it when the effector moves up or down purely Z only, all that happens is all three motors turn by the same amount. So to position your nozzle at z=40 all 3 motors must move the carriages by 40mm exactly. However when you move in X and Y then the effect of lowering the steps/mm or increasing it will mean the movements will not be correct (not remaining within the XY plane, you'll get concave/convex motion), so you have to recalibrate once you get the steps/mm correct.

I had the problem of the concave plane but I solved it, now I have a uniform print all over the plate. anyway tonight I'll try the two cubes of 20 and 40 and update you. for now, thanks

Quote
DjDemonD
I haven't really time to check through it all, but if the 40mm cube is proportionally larger still than the 20mm cube change steps/mm for xyz as your have shown above:

#define DEFAULT_AXIS_STEPS_PER_UNIT { XYZ_STEPS, XYZ_STEPS, XYZ_STEPS, 95 }

then recalibrate. Seriously try the least squares you can calibrate your machine with one set of probe point heights. It amazes me that anyone would use any other method.

#define XYZ_FULL_STEPS_PER_ROTATION 200
#define XYZ_MICROSTEPS 16
#define XYZ_BELT_PITCH 2
#define XYZ_PULLEY_TEETH 16

// delta speeds must be the same on xyz
#define XYZ_STEPS ((XYZ_FULL_STEPS_PER_ROTATION) * (XYZ_MICROSTEPS) / double(XYZ_BELT_PITCH) / double(XYZ_PULLEY_TEETH))

The calibration is not about steps/mm, but diagonal rod-length, printable- and delta-radius and such.

This guide is intended for Duet users but read the first part about getting your delta working right.

[www.duet3d.com]

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

but, I can replace these values from

#define XYZ_FULL_STEPS_PER_ROTATION 200
#define XYZ_MICROSTEPS 16
#define XYZ_BELT_PITCH 2
#define XYZ_PULLEY_TEETH 16

// delta speeds must be the same on xyz
#define XYZ_STEPS ((XYZ_FULL_STEPS_PER_ROTATION) * (XYZ_MICROSTEPS) / double(XYZ_BELT_PITCH) / double(XYZ_PULLEY_TEETH))

to

//#define XYZ_FULL_STEPS_PER_ROTATION 200
//#define XYZ_MICROSTEPS 16
//#define XYZ_BELT_PITCH 2
//#define XYZ_PULLEY_TEETH 16

// delta speeds must be the same on xyz
#define XYZ_STEPS 98

??

Its just a way to help calculate the correct steps/mm value. You can adjust it so if its 98 now and the objects are consistently 40/40.6 in size then do 98x(40/40.6) = 96.55 and use that value for #define XYZ_Steps

Then recalibrate and try some 20mm and 40mm tall objects again.

Edited 2 time(s). Last edit at 02/07/2018 06:16AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
DjDemonD
Its just a way to help calculate the correct steps/mm value. You can adjust it so if its 98 now and the objects are consistently 40/40.6 in size then do 98x(40/40.6) = 96.55 and use that value for #define XYZ_Steps

Then recalibrate and try some 20mm and 40mm tall objects again.

now is 100 (200*16) / (2*16)

so, can I put the value directly excluding the calculation?

obviously x and y will change, for these I will have to modify only the value DELTA_DIAGONAL_ROD?

Edited 1 time(s). Last edit at 02/07/2018 08:12AM by kyoday.

So calculate the steps/mm (xyz all the same) you need to get the z right. Then start adjusting diagonal rod and delta radius.

Edited 1 time(s). Last edit at 02/07/2018 08:24AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
DjDemonD
So calculate the steps/mm (xyz all the same) you need to get the z right. Then start adjusting diagonal rod and delta radius.

Now it's OK ! modified with #define XYZ_STEPS 98

after I adjusted DELTA_DIAGONAL_ROD, perfect cube! thank you

No worries, deltas take a bit more thought than straight cartesian or even corexy, as they require 3 motors linked to work, whereas corexy has 2 motors linked and straight cartesian has none.

Edited 1 time(s). Last edit at 02/08/2018 07:25AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions