Wrong cut

[cncnescu]

Hi all!

I have a "small" problem with my DIY CNC. For some reason, I'm not able to get the right dimensions when I cut a 40x40 mm square in wood. I get 40 mm when I cut along the wood fiber and 39.7 when perpendicular to the wood fiber.I have tried to reduce the cut deep from 2mm to 0.5 but the result is almoust the same.If I cut in mdf ,I will get 39.8 mm on one axis and 39.9 on the other(so here the differences between the axis is smaller).
The backslash on all axis is less than 0.03
The deflection is around max 0.04-0.05mm when 15 kg applied on each axis
I'm running Uccnc with ETH400 and on each axis, I have a DM856.
Also 2.2 China spindle+VDF "NowForever"(China)
The cause of my problem can be the high noise made by the VDF. Probably I'm losing steps (when high cutting force), but I'm not sure about it.

Any idea?

THX
Stefan

[m_c]

It's more likely to be spindle deflection.
You say you've measured deflection on each axis, but what about on the spindle nose/cutter with the spindle at cutting depth?

[cncnescu]

I already checked the spindle and has no runout(0.01-0.02) and also no deflection. The cutter is a 4mm x22mm (not the best quality) and for sure will deflect. Tomorrow I will check this too.
However, the cut was done first at 2000mm/min and second at 1000mm/mm and decreasing the feed rate /deep of the cut improve the result, but not dramatically(from 39.7 to 39.8).In case of cutter deflection changing the speed/deep should improve the result.

[magicniner]

I already checked the spindle and has no runout(0.01-0.02) and also no deflection..

Have you investigated the side-force generated by your cuts and applied that force via a bar in the chuck, with the same protrusion as your cutter, to measure deflection?
If not then your measurements are not valid,
Regards,
Nick

[driftspin]

The cause of my problem can be the high noise made by the VDF. Probably I'm losing steps (when high cutting force), but I'm not sure about it.

Any idea?

THX
Stefan


Hi Stefan,

If you think losing steps is the cause,

Have the spindle return to a known position after cutting.

Setup a dti there.

Run the program return to that position and check for zero on the dti.


Grtz. Bert.




Verstuurd vanaf mijn SM-A320FL met Tapatalk

[cncnescu]

Hi Stefan,

If you think losing steps is the cause,

Have the spindle return to a known position after cutting.

Setup a dti there.

Run the program return to that position and check for zero on the dti.


Grtz. Bert.




Verstuurd vanaf mijn SM-A320FL met Tapatalk

Yes, it will return to the position, so it's not the noise.

[cncnescu]

It was from the cutter. When I changed to a new one(3.175x9mm) the situation improved to 40mmx39.95mm.Now I have to find a 22 mm cuter that will not deflect so much because the 9mm that works it's too short for what I want to cut

[cncnescu]

Bad news!!... Yesterday I made a mistake. The cutter was not 3.175 but 3mm. After changing in the Estlcam the cutter diameter to 3mm (the real one) I get again wrong dimension.This cutter it's 9mm long so the deflection is minimum at 1000mm/min.
I really don't understand what is happening.

[driftspin]

Hmm probably you have checked this already.


Did you check for play on ballscrew angular bearings for the y-axis and x-axis.
The ballscrew locknut loosened on me earlier too... :-(

Does the problem move to the other axis when you turn the stock 90deg?

Grtz Bert



Verstuurd vanaf mijn SM-A320FL met Tapatalk

[cncnescu]

I already checked the backslash and is max 0.03mm on all axis. When I cut a normal wood plate(acacia for ex) along the fiber I will get the right dimension and this I will get regardless of the fiber orientation of the wood plate if x along the fiber or y along the fiber. Against the fiber is always the wrong dimension.
In MDF bought dimension are more or less equal but bought wrong.

[cropwell]

What are your steps per mm settings for x and y axis ?

[cncnescu]

I checked this at the beginning and it was fine...320 steps/ mm, 1600 micro steps on the stepper drive, 5mm / rev on the ballscrew.The same settings for all axis

[driftspin]

Yes, it will return to the position, so it's not the noise.So this also rules out reversed polarity problems... right?


If you are convinced the total combined rigidity/flex cant be the problem....


You can try with a 6mm cutter with a lot less flex..



Grtz Bert.



Verstuurd vanaf mijn SM-A320FL met Tapatalk

[cncnescu]

Today I discovered that my VDF was set on CCW rotation and the cutter used for MDF was a right hand one. After changing to CW on MDF I get the right dimension.
However on the wood plate on a 40x40 square it's all perfect, but when I switched to a bigger 80x80 square I get again wrong dimensions 80x79.8.
Normally it should work also on a bigger square but at least half of the issue was solved.

[Dave_stoke]

Today I discovered that my VDF was set on CCW rotation and the cutter used for MDF was a right hand one. After changing to CW on MDF I get the right dimension.
However on the wood plate on a 40x40 square it's all perfect, but when I switched to a bigger 80x80 square I get again wrong dimensions 80x79.8.
Normally it should work also on a bigger square but at least half of the issue was solved.It isn't hitting soft limits and closing down on that axis?

Sent from my F8331 using Tapatalk

[magicniner]

With such a basic error as running your spindle backwards I suggest you run through all your settings and calculations again and check X, Y and Z travel in several areas with a dial guage to ensure your calibration and and other assumptions are correct.