On 3/16/2018 7:02 AM, rangerssuck wrote:
On Thursday, March 15, 2018 at 5:23:54 PM UTC-4, Bob La Londe wrote:
On 3/15/2018 10:37 AM, rangerssuck wrote:
As you may remember, I've been making prototype printed circuits on
my Taig micro mill. I was pretty well stuck, not being able to get
traces small enough. It didn't seem to matter much what my z depth was,
the conical bit wouldn't cut a line that was anywhere near as narrow as
its 0.1mm tip.
Then one day, it dawned on me that that the head was vibrating like a
sonofabitch. I mounted a dial indicator on the bed and positioned the
plunger on the head. There was ± .003" of vibration! pretty much a
killer when trying to mill lines less than 0.008" wide.
I have since replaced the whole headstock with a 1.5KW 400Hz spindle
motor. It runs up to 24,000 rpm and there's no detectable vibration.
This is great. Of course, now that that problem is fixed, it made others
more visible. I noticed that lines in the +Y direction were a different
width than those in -Y. This turned out to be loose gibs, so I took
apart the whole XY mechanism, tightened the nuts to remove all but
0.0005" of backlash in X&Y which I had previously compensated for in
mach3, and adjusted the gibs properly.
All I can say is that I wish I had taken the time to do these things
from the get-go. What a difference. Next (after I get some work out of
the way), I'll do the same with the Z axis.
Hmmmm... I think somebody here might have suggested the 24K spindle on
the Taig. Did you use Paul Heinreichs spindle mounts?
I can tell you from experience its better to adjust out backlash to the
best of your ability and ignore the Mach 3 backlash compensation. I've
watched it steadily work a profile cut across the part as it gets
confused before.
My experience showed me a Taig mill could be adjust reliably to under
.001 backlash and run ok, but it would limit the ability to rapid
without losing steps. With it loosened up to about .003 backlash
overall mine would cut upto about 60IPM and rapid up to 80ipm. Of
course that is also dependent on your motors and controller. Whether or
not that is "good enough" depends on the type of job you are doing and
how you approach the job.
Yes, you did suggest the spindle. No, I did not use his mounts - he only offered me a complete set including 220V motor & VFD, radiator, water pump etc for a bit too much money. So, I bought a kit with 110V motor & VFD (easier for me) and a mount. I made a bolt-on plate that replaces the dovetail plate, and it works fine.
That's too bad. He used to sell his mounts separately on EBAY. I have
two sets. One for a 65mm and one for an 80mm spindle.
I also have the material and tools to make a dovetail mount. I have tested the design in a plexiglass scrap and it fits fine. I still need to come up with a locking mechanism for it, and I'll get it done after I finish some "real" work.
I need to break out some dovetail cutters and make some specialty tool
holders for the QCTPs on my lathes.
I got the backlash down to 0.001" in X&y and I have never tried to run at anything greater than 30ipm.
20-30ipm is what most folks recommend for the stock Taig V-lead setup.
The newer ball screw version runs much faster. If you do not need to go
faster I wouldn't mess with it to much.
At some point soon, I will experiment with trading looseness for speed.
Also, the possibility of making delrin nuts is there. I posted something
about that years ago, it involves heating two pieces of delrin and the
leadscrew and squashing the delrin around the screw. Perfect fit and
zero backlash.
I remember that. I have not done it, but I have tapped a few delrin
nuts. On my little MaxNC5 when I converted it from crappy V-leads with
split tube nuts over to acme screws I used the anti backlash nuts from
Dumpster CNC. I used it pretty hard, and only greased the screws and
ways. Never set it up to auto oil.
With the cad drawings available at
http://medw.co.uk/fisheye/gallery/114, those should be pretty easy to
make. I used those drawings to make the spindle mount as well.
Also still to do is getting the VFD talking to mach3. I have the communications parts, just need to install the plugin.
I hope you mean hardware plugin, because there is no software plugin
needed that I am aware of for VFD control. (I guess if you are using
RS485 there might be.) You just need a board to convert a pulse stream
over to 0-10VDC and a couple relays you can trigger with outputs off of
your breakout board for forward and reverse switches. Some higher end
breakout boards come with that stuff built in. I used low current
trigger relays on my good breakout boards, and I used 48VDC relays from
Altronix when I was still using a G540 with a 48VDC power supply.
My C23 and and my C32 boards have voltage and spindle relays pre-setup
and built in. The C32 did something weird with the relays though that
resulted in unexpected spindle reversals. I didn't need reverse on that
machine so I hooked both relays to forward on the VFD and all was good.
Probably won't buy another C32 board.