[Valen]

After suffering from the inadequates in our supplied HM45 mill head for too long we have had enough.
No more 70C gearboxes, enough of so much backlash in the gear chain that you get chatter in the finished product. No more to spindles where one of the bearing surfaces is the casting of the head.

(hit here for images of the guts of a spindle on one of these things)
So we took the plunge and ordered a "cheap" chinese spindle off ebay
The hope is to mainly work aluminium and any steel we want to look at ceramic cutters so as not to overload the bearings (too quickly anyway).
2.2Kw water cooled from linearmotionbearings2008


its bigger than it looks in this picture, feels like a serious bit of kit.

Then the question became how to mount it, we could hang it off the side of the existing head, but the problem there is the existing head weighs 100kg and is giving us trouble with the Z axis, putting what looks like backlash but I think is stick and slip into it.

So we figured this was a great time to try out the phenolic basalt we were looking at, its like epoxy granite but should be cheaper. Inital testing on a 50x50x800 bar indicates it has roughly the same stiffness as a solid steel bar of the same dimensions.
So the plan is to make this.





The side plates are 8mm steel, with threaded bar running through it clamping onto an aluminium adapter from the round spindle into the PB. Part of the reason for doing it this way is so we can tram the front to back angle of the finished product. We are keeping the existing slide and mounting onto it with the ring of 6 pipes with bars through them you can see at the top, this should be an improvement on the 3 bolt system (of which only 2 did anything) of the old head mount.
Keeping this also means we can still tilt or at least tram the head.

Personally I'm liking the look of the cads at least, it looks like a proper bit of kit with all the nuts and bar sticking out everywhere.

Russell is hoping to get working on the molding soon and hopefully cast one up over the xmas break. We are interested to hear any thoughts you guys have on it too so don't be afraid to chime in.
Many brains make for fewer bugs

[janbbeck]

This looks very interesting. Any updates?

Is there a good reference on phenolic basalt?

[Valen]

Got a whole bunch of updates in the pipe line, I'm going to make a few posts then fill them in over the next few days.

[Valen]

ok so starting from the beginning this is what russell (rusel) did

before pulling the old head apart we made the holder (called a cradle) for the spindle out of some beefy flat bar.
Roughing



finishing



you can see the poor finish in the end result of that (which is why we wanted the new spindle) its like a file, you couldn't make it that way if you tried without a very expensive mill and hundreds of hours of machine time

This is the holes for the threaded bar to run through




And the whole assembly together


With that done we were able to make the cradle blank, this goes into the mold to leave a space for the part. A stack of MDF sheets were glued togther


Then covered in resin and milled to final dimensions.


The resin was to give a hard smooth surface for the wax to stick to, so it would release easily, unfortunately it wasn't to be...

Then we stared on the mold proper.


the curved segment was a little complex but not too bad

after routing the shape (can you "route" mdf at 1600RPM? ) it was covered in flexible laminex sheeting




The sides were relatively straight forward which meant the overall look of the thing came together pretty quickly.


then we added a bunch of inserts so we can bolt other bits and bobs onto it.

They were originally really long nuts, that Russell roughed up on the lathe so they stick better. The plastic tubes are to keep the resin off the threaded bar running through the blank, they are not part of the final product, we use the play in there to adjust the alignment of the head.

Speaking of sticking it was time to pull the Z axis slide off the mill.

as we used this as a part of the mold.

Using this as a guide we made the long pipes and stops to hold the head onto the mill.



the ends were then tacked on and the pipes machined


we then put the new nuts into the slide through the little insertion hole



Then we bolted it all together and filled the T nut groove with wax, in case any of the other bits leaked.



To protect the surface itself a laminate template was made, this also lined up the bars.


we then filled in the center and put a bit of bog around it to make sure it didn't stick there, its still pretty thin though so it'll locate on it somewhat.


this is what the back end looks like



when its all together it gets pretty tight inside the mold.


This is the whole thing ready to go.

[Valen]

Turns out you cant edit posts after a while :-<
oh well

First Russell dired out the stones, this was done by spreading them out on some black plastic.


First we mixed the resin. (these photos etc are kind of blury because I put my phone inside a bag to keep the sticky stuff off it.)
http://www.vapourforge.com/cnczone/h...xing_resin.jpg

Pot life was around 40 minutes at 25-30C but we didn't seem to come close to needing all that so it worked well, I wouldn't want to go much shorter though.

Then into the mixing bucket we gradually added the fine and coarse rocks.

The drill wasn't really quite powerful enough for the job at 15% resin, at 20% it wasn't too bad. But definitely put the resin in the bucket then gradually add the stones.

Then we added 6kg each of the fine and coarse stones as measured by our "weighing system"

AKA a cheap digital scale from deal extreme









Then using the little scoop we shoveled the mixture into the mold.


Russell also made an air powered vibrator to help pack the mix into the mold. Named the wife pleaser 9000 we found that it was best to fill a chunk of the mold then use it to basically push through the mix and with a somewhat suggestive back and forwards motion work the mix to make sure it filled all the little pockets. Also putting it directly onto the inserts in the mold seemed to ensure a resin rich coating on the surface of those parts.
IMG]http://www.vapourforge.com/cnczone/hm45/PB_Head/casting_and_releasing/casting/thumbs/wife_pleaser_9000.jpg[/IMG]

This is the end result

Its not pretty but it seemed to work.





After a while it had started to go hard, Russell wrapped it up and put a double boiler under it to get the heat up on it, the thermistors we put in read up to about 50C



The next day Russell released it from the mold. It had exothermed as the temperature had risen after he took the heater off.

Unfortunately the spindle blank stuck, it was smooth and waxed and should have come out, but the vinyl ester resin used is extra sticky and seemed to punch through it, next time we might need to use a PVA or something.
As such the blank needed to be ripped out.



Whilst generally good the top side had some air entrapment

We think this was due to using the vibrator and focusing on the front while it was tipped up, and the top side slumped away from the mold.

[Valen]

this is the one you have been waiting for :-D

(keep in mind you can click the pictures for a high res version)

this is an overall shot of the setup
you can see the wires running up to the air cylinder for counterweight, even though the new head only weighs around 30kg, it still works better with the counterweight. We actually have the pressure cranked up higher than needed to balance it as it seems to make the slide operate smoother, with the new head our Z axis has gone from being the worst performing axis to the best, P in the PID loop has gone from .6 to 7.3, max following error is now ~.01mm or so and it'll normally sit at around +- .001mm or so during a cut.


We used radiator fluid for the coolant

says on the tin it should be good for 100,000km or 2 years
eventually the radiator will probably go outside the shop, but as yet its all in flux and hasn't gotten warm enough to notice really.



This is the business end, the white drops are coolant that sprays around far more than it used to on the old head, the shower curtains might not be up to it for long, it really does make a mess.
in the collet is a 6mm carbide extended length 2 flute aluminium cutter.
(was the best thing we could find for Al)

hit this for a youtube video of it in action
(ok its a little short and mostly you can see spray, but theres no music or anything silly so you can actually hear it working)




the mirror finish passes in the back right hand quadrant are the new spindle, the other rough looking ones are the old spindle. The new one is even better than it looks in this photo.

:wee::banana:

this is an overall shot of the setup, needs some work for added "pretty"
but its functional

[andypugh]

I am impressed by how you handled the interface to the column. Does it rotate freely?

Any feel for how rigid and damped it is relative to the original?

[samco]

Nice work! could the casting be honeycombed maybe to make it even lighter?

sam

[Valen]

@samco possibly, but if you look at the photos of the mold before its filled, there is rather allot of stuff in it. Might be able to put something in between the pipes that hold it to the Z axis i spose.

@andypugh it still rotates ok, so we can tram it easily.
It seems to be more rigid than the old one, not too sure about the damping yet, because we changed the spindle as well its hard to compare. Tapping on it with a spanner it sounds more "live" than the iron but thats probably mainly a result of its weight.

We might need to put the old head on again in the next few weeks to do some low speed milling of titanium, and we'll try and do some back to back comparisons then. But in the meantime its 100x better than the old one in terms of end results.

Its ripping material off at the limit of the cutter, (400mm/minute, 24kRPM, 3mm depth of cut,6mm full width cut) and only pulling about 500W to do so.

[svenakela]

You use an Eheim outer filter as a coolant pump?! For sure, you do have circulation!

Reall nice work with the head.

[rusel]

Thank for all the complements
The pump guts must have been broken when i pick it up on the side of road (it run ok with no back pressure) but it only cost $30 to fix it. Now we have some good circulation happening.

Now just finding the best speeds and cuts
Any suggestion welcome
In aluminium for
2 and 4 flute carbide 6mm em
4 flute carbide 12mm em
4 flute carbide 6mm ball mill
4 flute carbide 12mm * 3mm bull mill

We will need to get more aluminium cutters so if there is any good sites out there let us now. Thanks

Russell

[svenakela]

eBay is a good source for mills. I buy heaps of the from German eBay, not so difficult to shop even though our English.

[Jason3]

Nice Upgrade

400 mm/min seems very much too slow - did you mean 4,000 mm/min? I run a 6 mm 3 flute at 20,000 rpm and 4,500 mm/min.

I like MariTool's 3 flute high helix cutters.

Cheers,

Jason

Edit; sorry, watched the video. Nope, you did mean 400 mm/min all right. I would crank up the feed, a lot

[rusel]

Hi Jason
The price look good on the MariTool's site. Thanks for the lead.
Yes we are running it slow but we are learn with this new beast.
What is the finish like at those speeds? Also what Kw spindle do you run?


Regards


Russell

[Valen]

The cutters we are using are designed for steel, as near as I can tell from the specs 400mm/minute is on the limit for them.

the mari tool 6mm uncoated carbide
6mm metric 3 Flute Carbide End Mill SE 38 Deg Helix 19mm LOC MariTool

comes out at 3000mm/minute at 13,000 RPM give or take.

[Valen]

Jason3
looked at it closer and you were right, I must have been looking at the wrong column

[nomedia="http://www.youtube.com/watch?v=frIssODebKA"]YouTube - 3600mm/minute 9600RPM 4 flute 6mm cutter[/nomedia]



hit it for some much cooler looking footage.

Acceleration on the axies went from 100 x and y and 30 z to 3800 on x, 2000 on y and 800 on Z.
Top speed is now 7.5 meters a minute, though there is a bit of oscillation there might do with some more tuning.

the new accelerations make it feel like the mill is breaking every time it stops, its just screaming along at 7 meters a minute then its stopped dead.

[Jason3]

What is the finish like at those speeds? Also what Kw spindle do you run?

The finish is great. Spindle - from 350 watt to 2.4 Kw on the small machines, I drop the DOC for the low power spindles but leave the feed and chipload the same.

Jason3
looked at it closer and you were right, I must have been looking at the wrong column

YouTube - 3600mm/minute 9600RPM 4 flute 6mm cutter

Acceleration on the axies went from 100 x and y and 30 z to 3800 on x, 2000 on y and 800 on Z.
Top speed is now 7.5 meters a minute, though there is a bit of oscillation there might do with some more tuning.

the new accelerations make it feel like the mill is breaking every time it stops, its just screaming along at 7 meters a minute then its stopped dead.

Great, that's more like it! Good work. How much power is it drawing now - still 3 mm DOC?

I like to cut a small test part with lots of sharp angles repeatedly while adjusting accel. and CV feedrate to get the best result. Usually I can get to a point where the mill is fast and happy at a cycle time of (for example, last test) 6 min 53 sec, vs 6:47 and sounding a bit bangy and clunky. That'll make the machine last longer, and is nicer to be around for a worst case 2% time penalty.

Regards,

Jason

[rusel]

Gave the new head and spindle a work out today.
Wanted to make a clamp for the spindle lead as when I fitted the 4 core lead to the plug I had to cut off the part that clamps the lead to the plug.
As you can see in the pic I put a diamond pattern on the out side of the part one pass at 30 and the second at 150 degrees. This made the mill run the x,y,z axis simultaneously all with no problems. This was done with a 6mm ball mill at 2500mm/min with a RPM of 9500 and took about 20 minutes to make the first one (both sides) from scrap.
The ribbing on the outside curved ends is from the gcode not being done in g2-3 but small segments. We have now sorted the problem between rhino and emc2
The last pic I have just sat the one part in place so you get the idea.
If I had try to make this with the old head it would have take for ever and and with the heave head and constant z axes changes it would of come not so good.
Just a little more tuning and it done
Very happy

Russell

[CROSSHATCH]

Me like the sound of that spindle and the speed of the nm45.

[Valen]

The finish on those last ones looks ok in person with that diamond pattern but for some reason it came out rather poor in the photos.
So I had a go with a new aluminium specific ball mill cutter and a new set of G-code, .1mm step down and .1mm scallop on the horizontals.
also it was done parallel rather than with z profiling.
I think it looks pretty good, the radiused bits have a mirror finish which is pretty cool.