PDA

View Full Version : Micro lathe Zero backlash* conversion



Jonathan
28-04-2012, 06:47 PM
I recently found a Clarke CL250M micro lathe for a good price, so I got it to convert to CNC. This lathe is very similar to the C1 micro lathe from Arc Euro.

Parts bought:


Lathe: http://www.machinemart.co.uk/shop/product/details/cl250m-metal-lathe?da=1&TC=SRC-lathe%20metal
4 E8 Magneto (angular contact) bearings
RM1204-600mm ballscrew
RM1204-300mm ballscrew
4-RM1204 ballnuts
Belville washers to preload nuts for zero backlash and to tension Z-ballscrew and preload X angular contact bearings.
Two 12T and two 30T HTD pulleys and belts for ballscrews, 36T,18T and 12T for 1:2 and 1:3 ratio for spindle.
Already had 6.5kW brushless motor, seemed a crime not to use it: http://www.hobbyking.com/hobbyking/store/__5142__Turnigy_80_100_B_130Kv_Brushless_Outrunner _eq_70_55_.html


Unfortunately I didn't take many pictures to start with, so this is going to be more like a 'look what I've made' in parts than a build.

Not a chance of getting an RM1204 ballnut in the cross slide, let alone two, so I've mounted them behind. To do this I milled a block of steel to extend the saddle for the ballnuts' mount and drilled it:

5877

Next I turned some 3.5" aluminium bar to accommodate a two ballnuts, one in each end:

58725873

One ballnut fixed in this side:

5874

The other goes in this side. On 'top' of that goes a steel spacer, then and end cap to accommodate a pair of beleville washers. This cap is bolted through the ballnut to apply a preload of about 300N and rotationally constrain the ballnut:

5875

Cut the aluminium motor mount on my CNC router. Just a few 5/8" thick aluminium plates with a cavity between to accommodate the pulleys so that swarf cannot get in the belt drive or bearings. I used a 12T pulley on the stepper motor and 30T on the ballscrew for 0.008mm resolution with a 200step/revolution stepper motor. If you count using microstepping it's a convenient 1um resolution. Once assembled I put a DTI on it and you can see the needle move for every microstep...so reasonable to assume somewhere between 8um and 1um, either way plenty good enough for this sort of machine. The angular contact bearings are held in a separate part and clamped with a 5mm steel plate also cut on my router. I machined the ballscrew with an M8 thread on the end used to apply preload to the belville washers / magneto bearings.

Z-assembled, but only with one ballnut:

58765887

After much deliberation I decided I couldn't get an RM1204 ballnut to fit under the lathe bed without milling some of the bed and raising it. I just milled a couple of mm off the webs in the bed to get the ballnuts to clear, so that shouldn't weaken it to much. Forgot to take a picture of that.

Next I milled the middle protrusion off the bed stand:

5878

Then made some 10mm spacers from ground flat stock to go between the bed stand.
I cut the ballnut mount from 2.5" aluminium bar. Bored, drilled and tapped both ends to accommodate the ballnuts in a similar fashion to the X ballnut. Then some milling and finishing with flycutter. Next I machined the ballscrew and made a quick tube to fit over the end and transferred it into the ballnut. This enables fitting the ballnut without all the balls dropping out...

5879

Belville washer pressure piece and one screw machined to closely fit the ballnut on the cylindrical portion. This constrains the ballnut rotationally.

58835882

Sliding the ballnut on:

5880

Assembled and preloaded:

588458855881

Assembled with Z-motor mount, not shown:

5886

The 6.5kW (allegedly) brushless motor I had was the same diameter of the somewhat limited 150W DC motor this lathe came with, so it bolted straight in with the mount it came with. I just had to make a shorter shaft (so the original covers fit) and machine the pulleys. Although this motor initially seems excessive, it's arguably well matched since it has plenty of torque over the full speed range, I probably wont need to use the 1:3 ratio.

58895888

I've not had chance to do anything with the lathe for a couple of weeks, so I've left it like this:

5890

I will soon add plates to the Z-motor mount to seal the belt, similar to Z.

Now I can cut hemispheres!
5891

Enough for now... more to follow shortly.

*Well near enough zero backlash..

Robin Hewitt
28-04-2012, 07:33 PM
Bravo ! :beer:

Nowt like Belleville washers. After years of trying to reduce the backlash on my mill I finally did a proper job and sprung it. When you can't turn the handle at all without the slide moving, not the tiniest nudge, you will never want to go back.

Jonathan
28-04-2012, 08:01 PM
Bravo ! :beer:

Thank you... but spoke to soon I'm afraid.

The backlash is nowhere near zero - 0.04mm on Z and 0.024mm on X. I've tried various preload settings on the Z ballnuts and different amounts of tension on the ballscrew to make sure there's no end-float. I'm therefore not sure where the backlash is coming from. The preloaded ballnut is possibly a little tight in the mount on Z, so that might be it ... I'll have to investigate. Not sure about X :(

To do:
Replace 80mm chuck with 100mm.
Z-motor mount cover.
Get 50V ESC to replace current 24V ESC so I can run up to full speed (about 3000rpm)
Program PIC to read signal from LinuxCNC and output PWM signal for ESC
Currently using 0-50V 20A Lab PSU with a 3000mAh LiPo battery in parallel. The battery allows the motor to be used to it's full power with an underrated PSU so long as the average power is below the rated power of the PSU. I need to buy a more permanent PSU.
Make little 'control panel' to replace old one.

Robin Hewitt
28-04-2012, 08:57 PM
Should not be possible to fail.

Quick methodology test...

One nut is bolted to the slide, the other nut is merely prevented from turning.

John S
28-04-2012, 09:28 PM
Leave the chuck at 80mm, the spindle can't handle the weight of a 100m chuck.
Ring Ketan if you don't believe me.

Jonathan
28-04-2012, 09:41 PM
Should not be possible to fail.

Quick methodology test...

One nut is bolted to the slide, the other nut is merely prevented from turning.

Yep. Although when I bored the ballnut mount on Z I didn't leave much clearance, so maybe the ballnut is skewing and binding somehow. Will have to investigate when I've got time.


Leave the chuck at 80mm, the spindle can't handle the weight of a 100m chuck.
Ring Ketan if you don't believe me.

That's rather annoying, especially since I started making the backplate yesterday. Have you tried it (not that I'm disputing your experience, just wondering)?
Looks like I need to get some external jaws for the 80mm chuck as they're missing.

John S
28-04-2012, 10:35 PM
No not tried it, bought a C1 to play with but the little TOS got hungry and ate it.

JAZZCNC
29-04-2012, 11:39 AM
Jonathan,

I've been given a Myford Super 7, really don't see why folks get so excited about these things but anyway no prizes for what going to happen to it, much to the relief of my lovely little boxford who was under serious threat. So a few Q's if you don't mind.

Did you get the Screws from Chai.? and was that the lowest pitch he does.? (I know I'm being lazy and could ask him direct but hey oh he can't answer th rest so here ya go.!!)

Where did you get the Bellville washers.?
What strength did you get and how many per screw.
How did you stack them series or parallel IE: )))) or ()()

Nice job by the way, what MAx cutting dia can you do and when do we see the video.????

Jonathan
29-04-2012, 12:15 PM
Did you get the Screws from Chai.? and was that the lowest pitch he does.? (I know I'm being lazy and could ask him direct but hey oh he can't answer the rest so here ya go.!!)

As far as I know RM1204 is the smallest pitch (and diameter) he does, although I didn't ask directly. He doesn't state the ballnut dimensions, but they are a lot smaller than RM1605. RM1204: overall length=28mm, diameter=24mm,flange length=6mm, flange diameter=40mm, between flats=25mm.


Where did you get the Bellville washers.?

http://www.leespring.com/


What strength did you get and how many per screw. How did you stack them series or parallel IE: )))) or ()()

For preloading the nuts I got:
http://www.leespring.com/uk_product_spec.asp?partnum=5000351000&springType=W&subType=

For preloading the bearings I got:
http://www.leespring.com/uk_product_spec.asp?partnum=323028709&springType=W&subType=

I chose the springs based on:
Preload force=10-20% of ballscrew static load rating
Belleville washer compressed no more than 75%

I stacked them like () to get greater adjustment whilst maintaining the same spring constant. The Z-ballscrew has two Belleville washers at one end, the other end fixed and one angular contact bearing on each end, so the screw is tensioned. On X the end of the screw is floating, so clearly both washers and bearings at one end.


what MAx cutting dia can you do and when do we see the video.????

About 75mm would clear the saddle, so in theory that diameter once I have found external jaws for the chuck. That reminds me I need to get/make an MT1 live centre. In practice clearly the spindle isn't strong enough to be doing that on a regular basis. I'll convert my Colchester Student lathe once this one's finished for much bigger parts.
Video I kinda forgot...just got one rather boring shaky video but I guess I'll upload it anyway.

m.marino
29-04-2012, 09:17 PM
Jonathan well done. Very well done. Have you looked at the Lathe in this video:http://www.youtube.com/watch?v=Y1IDHIcyHLY and if so when are you going to be building a tool changer? Not to egg you on but I know that something like that would be a good bit of fun for you to build.

Michael

wiatroda
29-04-2012, 10:41 PM
Have you looked at the Lathe in this video:http://www.youtube.com/watch?v=Y1IDHIcyHLY
Michael
Actually, this video was made nearly 4 yrs ago and since then no other videos came out showing some metal turning but plastic.
If it was able to turn metal - builder would upload it for sure. Just makes me think is it not rigid enough???? not able to turn metal???
Will you Jonathan treat us with some action video??

JAZZCNC
29-04-2012, 11:16 PM
Actually, this video was made nearly 4 yrs ago and since then no other videos came out showing some metal turning but plastic.
If it was able to turn metal - builder would upload it for sure. Just makes me think is it not rigid enough???? not able to turn metal???
Will you Jonathan treat us with some action video??

Actually I think this lathe was purposely made to turn plastic.? I think the bloke is the same guy who makes Dumpster Anti backlash Delrin nuts but I could be wrong on that one thou I'm sure it was made just for plastic.!

wiatroda
29-04-2012, 11:27 PM
Actually I think this lathe was purposely made to turn plastic.? I think the bloke is the same guy who makes Dumpster Anti backlash Delrin nuts but I could be wrong on that one thou I'm sure it was made just for plastic.!
I bet you're right. never heard of those nuts:whistle:. But it would be nice to see some fancy shape metal turning anyway :)

Jonathan
22-09-2012, 06:49 PM
This lathe has sat on the side doing nothing since I posted here, just not had time for it. A few days ago a 12Nm stepper motor arrived, which just so happens to fit the lathe mount, so how could I resist using it to power the spindle?

6947

A few moments later:
69486949

I had to take the covers off since now I don't have the right length belt which means the motor is in a funny place. I chose a 36:22 (motor:spindle) ratio since if we take the speed/torque curve from Zapp's site and convert it to more meaningful units, the peak power is around 600rpm so around 1000rpm spindle speed with that ratio.

6950

I've only got the motor on 70V (DM860 drive), so that graph isn't ideal, but it seems to work well and runs the spindle up to 2000rpm. The motor was fine with drilling a 6mm hole in aluminium and general turning. I think I'll make a bigger pulley for the motor and try it anyway. The finish leaves a bit to be desired, but I think that's mainly the lathe itself being weak and having to have the material quite far from the chuck as currently the rear shaft on the motor interferes with the saddle. I don't think that will be a problem with the longer belt and if it is I can easily mill a little bit off the steel block I attached to the saddle to get clearance.

I configured LinuxCNC to run the stepper motor as the spindle by editing the .hal file - pretty simple. Next step is to get G33/G63 to work so I can do threading. I could just run it as an A-axis and program a ramp (e.g G1 Z-10 A10), but I'd prefer to do it properly...

Anyway, it seems this thread is useless without a video, so here goes:


http://www.youtube.com/watch?v=fHCu40gmqLc&feature=youtube_gdata

Edit: Almost forgot to mention I solved the backlash problem on Z. Turns out it was just the belt being a bit loose, tightened it up and now the backlash is essentially zero - the 0.0001" indicator returns, as far as I can see, precisely to the same mark if approached from either direction.
I just need to make one piece for the X-axis so I can put the second nut in to get the same result. I wonder where the Belleville washers went. I think I've broken the magneto bearings whilst working out what was wrong, so I'll buy some better ones (http://www.vxb.com/page/bearings/PROD/Kit8330). He's got 708A and 708C so presumably that's the different contact angles - it would be good to know which is which so I can get the ones that tolerate higher axial forces!

Fivetide
22-09-2012, 11:56 PM
Looks fantastic Great video mate