Correcting for spindle runout

[NuttyMonk]

Hi all,

new member here. Got myself a Sainsmart Genmitsu 3018 PROVer and am reasonably happy with it so far. Seems kinda rigid but does have some flex in it due to the thin guide rails on the X axis which are only 10mm diameter and the mount for the spindle being made of plastic.

I plan to upgrade the whole X and Z axis in one go and have sorted out a nice list of bits and pieces I need to buy. Surprisingly cheaper than I was expecting to be honest and using the Openbuilds components should make it easy to assemble and dial in.

I'm trying to mill PCBs and with some WD40 on the PCB and a German spiral engraving V-bit (0.1mm 30 degree) I'm getting nice clean cuts with a good Z accuracy (I'm running at 10k RPM and 30mm/min plunge and 60mm/min XY feed rate). The main problem I have with the machine is the runout I am getting on the spindle which is a cheap 44mm diameter 775 motor. I've measured the runout with a plunger gauge angled up into the bore of the collet chuck and I am getting 80 - 90 microns of runout in there. As I rotate the spindle the runout seems to be circular, not some nasty notchy runout. This means that if I want a 0.3mm track width in my PCBs I need to design the PCB with a 0.6mm/0.7mm track width or higher as the runout causes a lot more of the material to be milled away than should be. This is messing with my PCB designs and putting me in an existential funk.

Therefore, Is there any way to counteract runout on spindles? I'm thinking along the lines of shimming the motor in its mount. I'm having a hard time trying to visualise what I would need to do to correct for this problem. My thinking is that the inside of the ER11 collet extension isn't rotating on the vertical in the X and Y axis and therefore the end of the V-bit is rotating in a small circle. By shimming the motor with some shim stock or even just paper or aluminium foil can I correct for that?

Hopefully I will be rebuilding the X and Z axis in the near future to increase rigidity and therefore improve accuracy and allowing me to run at higher feed rates but it's gonna take me a couple of months to get all of the parts together and I'd like to sort out this problem sooner rather than later. Even if just for my own sanity. A new spindle will probably be high on my list of new parts to get. One with less than 0.01mm runout if possible and no ER11 collet extension. I'm beginning to dislike those cheap little poorly made things.

Thanks in advance for any advice or suggestions.

Cheers

NM

[Doddy]

Shimming the motor wouldn't work... you'd have to shim the collet or cutter.

I tried to bodge an altogether bigger machine with a half decent spindle... the additional weight cause all sorts of deflections across the axes. Be realistic.

[NuttyMonk]

Thanks for the reply. I wasn't sure if that was a good idea or not but i guess not.

Cheers

NM

[ZASto]

If you plan to rebuild your axes, invest in new motor, for example FME (Kress) FME 1050 - 1.
Also In my opinion, 10k rpm is not enough for milling traces with such small bit 20k+

29153

I usually only drill my boards on CNC, but prefer to chemically etch them. Quicker, easier, no extra unhealthy dust :)

[NuttyMonk]

The WD40 that i spray on the board helps with the cutting and also soaks up the dust. Haven't had to clean the mill since i started doing that and the cuts are so much cleaner than before i used it. Plus i think the spiral V-bit helps too with the cleanliness of the cuts.

I've been looking at spindles recently and think i have narrowed down the search to this one or very similar...

https://www.ebay.co.uk/itm/2-2KW-CNC-Spindle-Motor-ER20-2-2KW-VFD-Inverter-Drive-Water-Cooled-220V/254569831940?hash=item3b458b5e04:g:0CAAAOSweatelsZ i

Decent power and RPM plus it's water cooled so it will be quieter and maybe last longer too.

Cheers

NM

[Doddy]

Those spindles are good, but fairly heavy. You'll find the twin unsupported rails bowing under the weight mid-axis, resulting in the cutter gouging deeper into the PCB substrate. Same for the other axis. You can try to counter by Z-mapping the bed but it's a faff. For such a small machine I'd look at the 800W (smallest/lightest) spindle - but even that will have similar problems. As above, be realistic.

Edit: Measure your backlash as well - that could be similarly limiting.

[NuttyMonk]

I wouldn't be using that spindle on the current setup. I'm planning on building the X and Z axis out of C-Beam extrusions from Openbuilds with XL C-beam gantry plates joining them together.

I have looked at the 800W version of that spindle and it might be the one i buy. The 2.2kw one was just an example.

Also, removed most of the backlash on the X and Y axis using a couple of grub screws in the supposedly anti-backlash nut that comes with the machine. I don't know why it doesn't come like that as standard. It's such an easy fix and makes a big difference.

Cheers

NM