Unsupported rail flex

[yngndrw]

Hi,

I have a Marchant Dice MDL-2-TR-UN, pictured here:
http://yngndrw.hostilezone.net/uploa...3112009153.jpg

I have only recently gotten the spindle mounted. I'm using a 1.5KW air-cooled HF spindle so the spindle is probably a bit large for the machine.

Anyway now that I've got it mounted I've noticed some movement if I push down on the front of the mount. I can see this movement by eye and it seemed to be coming from the Y axis bearings.

I'm not great at measurement and I have no steel surfaces on the machine to attach my magnetic-based DTI to so the actual numbers here are pretty meaningless, however they seem large enough for that not to matter.

I measured the deflection from the spindle mount approximately in-line with the spindle and while pushing down on the end of the mount the DTI indicated a movement of 0.4mm in each direction. I'd judge the force that I was using to be "Tightening a medium sized Allen bolt", so it wasn't too much.

I then measured the deflection at the back of the Y axis block and saw a movement of about 0.05mm in each direction.

Now these machines are not built for heavy work and are listed as "entry level", however these numbers to my untrained eye seem rather large. The Y axis span is approximately 350mm and uses a pair of 16mm unsupported round rails.

The X axis doesn't seem to have anywhere near the same kind of deflection and is a span of approximately 420mm using 20mm unsupported round rails.

I'd like to know if these numbers seem right and if I will be fine using the machine for what I wanted to. I wanted to mill PCBs, plastic and some metal. (Aluminium and copper.) The PCB milling is probably the most important to get the correct depth. Am I expecting too much ?

I'd also like to know if there is anything I can do within reason to improve this.

Thanks.

[hemsworthlad]

Hi, Have you try'd cutting any thing with it yet? try some circles in a hardish material and measure to see how round they turnout.

I think with the design of high-ish gantry sides with not a lot of support from flex and unsupported rails you will struggle with accuracy but 0.4 does seem a tad high.
Was you pushing down on the mount towards base? if so this will be an accumalation of both X & Y rails flexing and really you will be cutting in a linear direction most of the time not down, only when you plunge will the force be in vertical direction.

I think the proof is in the eating, just try cutting something that needs to be accurate and see how it turns out.

The problem with the design is that it doesn't give you much scope for improvement without a major costly re-design and then you may as well just build your own with decent supported rails etc.

Good luck!

[yngndrw]

Thanks for the reply.

Originally Posted by hemsworthlad Have you try'd cutting any thing with it yet? try some circles in a hardish material and measure to see how round they turnout.

I haven't cut anything yet, only drawn stuff with a marker. How hard is hard ? Is MDF hard enough for this test or so you mean something more ? I didn't want to cut anything until I build an enclosure for the dust so it will be a short while before I can do that, but I will do so ASAP.

Originally Posted by hemsworthlad I think with the design of high-ish gantry sides with not a lot of support from flex and unsupported rails you will struggle with accuracy but 0.4 does seem a tad high. Was you pushing down on the mount towards base? if so this will be an accumalation of both X & Y rails flexing and really you will be cutting in a linear direction most of the time not down, only when you plunge will the force be in vertical direction.

Yes I was pushing the front of the mount, both up and down.
I did measure how much the top of the sides were moving and it was much lower than the amount that the Y axis plate was. I just tried pushing the top of the mount backwards and forwards and while you can see the Y axis plate twisting by eye, I could not see much twisting of the sides. The X axis shafts seem to flex much less, but then they are 4mm larger in diameter.
I don't really have any experience with machining so I don't know how much force to expect so I may have been applying a lot more than would be seen when cutting. However one of my applications (PCB milling) requires very good control of the cut depth as it is using a V bit to engrave a thin surface of copper off a fibreglass board.

Originally Posted by hemsworthlad I think the proof is in the eating, just try cutting something that needs to be accurate and see how it turns out.

Yes you're probably right there. As mentioned before I'll test it out ASAP - Just need to get the enclosure done first.

Originally Posted by hemsworthlad The problem with the design is that it doesn't give you much scope for improvement without a major costly re-design and then you may as well just build your own with decent supported rails etc.

Yea, I was thinking of making a much larger (8'x4'?) rack and pinion driven machine for heavier work and just keeping this one for smaller work such as the PCBs. Will be a little while before I get the time, space and money for that though and I'll still need this one to be able to produce metal parts for the bigger one. I'm hoping that even if the flexing is an issue that it will just mean I have to cut slower.

Cheers.

[rowbare]

It is really difficult to determine how much flex you will have in practice till you actually cut something. The machine is small and it is easy to manually apply more force than the machine will ever see in practice.

The cutting forces involved in cutting pcb traces are minimal. A lot of people use 100W and smaller spindles for pcb machining and do fine. These spindles will stall at cutting forces much less than required to flex your machine. If you have ever used a Dremel tool, you know how little pressure it takes to stall one while cutting. And that force is greater than what is involved in cutting pcb traces (at least at the speeds most hobbyist machines are capable of). I don't think flex will be a major issue when cutting pcbs.

Cutting aluminum is another story. However there are a lot of constraints imposed by the machine that actually work to your advantage. I am assuming that your spindle has an ER-11 collet so that limits the cutter size thereby limiting the cutting force that can be applied. As well your maximum cutting speed is probably in the 1000mm/min range so that further limits the cutting forces.

If you select appropriate chiploads and feed speeds for the cutter size and use reasonable depths of cut, there is no reason why you can't make the components for your larger machine on this one.

Good luck!
bob

[yngndrw]

Thanks bob, yes you're bang on with the ER11 and maximum speed guesses. I was unsure if I'd be able to work around it or not, but you've put my mind at rest.

Cheers.

[carvemylog]

My CNC has supported X rails 43" long and I used unsupported 20mm rails with ~28" span. The gantry is 1" aluminum plates. The Y rails are ~7.5" apart, which meant I lost some Z travel, but thought it would make the overall stiffer. I have to admit I was pretty surprised when I finally got it all together how much I could deflect the Y axis rods just by grabbing the router mount and pulling on it.

I finished milling a 3/4" aluminum block and it turned out very well. I don't think the flex was a huge issue here because I was cutting so slow. In hardwoods, I've noticed that if I cut smaller sized (less then 1/2") holes faster then 75ipm they will start to go out of round. The other situation where it can be a problem is when the machine hits a false limit switch and makes an imediate stop (or like hitting the E stop). Since the router is still turning, the flex will leave a tiny mark in the part where the machine stopped.

If I had to do it all over again, there's no question I'd buy supported rods, but it hasn't been a big enough problem for me to want to replace them.

To reduce the effect, space the rods apart as far as you can sacrifice travel, and do the same on the linear bearings.

[yngndrw]

I didn't think about the spacing of the rods, mine are only 60mm apart but with only 50mm of Z travel as it is, I can't afford to change that. I will keep all of this in mind for the one that I build in the future.

I'm currently limited to about 1200mm/min or 40ipm so from your numbers I'm guessing that I won't have a problem. I was going to try it today but as mentioned in another thread, the ER11 collets that I have don't fit the spindle. I might try with some router bits if I can find them.

Thanks for your reply.

[hemsworthlad]

Originally Posted by yngndrw Thanks for the reply. I haven't cut anything yet, only drawn stuff with a marker. How hard is hard ? Is MDF hard enough for this test or so you mean something more ? I didn't want to cut anything until I build an enclosure for the dust so it will be a short while before I can do that, but I will do so ASAP. Cheers.

No MDF is not what i had in mind, it's not very hard just abrasive. . . kills cutters.
I was thinking a hard wood oak,ash something like that.

You shouldnt have any probs at all with pcb's and this size of machine is very much designed with that kind of work in mind, but like Rowbare says Alu is a completely differant ball game.

You will be able to cut Ali but will have to take very light shallow cuts, just means it will take a lot longer, especialy if large or deep job, Although with the spindle that you have you will get away with it, a normal small router would die very quickly with the amount of hours you would have to leave it running.

Have you got your collets sorted yet.?

[yngndrw]

Ah okay I'll see if I have anything harder. Not sure if I'll be able to find any in small amounts from the shops around here.

I got the size of spindle that I did thinking that I'd need the torque at lower speeds. My machine is quite limited in feed-rate but I'll work on improving that.

Regarding the collets: Yes and no.
I have two bits for it currently, a 1/8" endmill and a 1/4" shank router bit. The collets that I have are the supplied 6mm, an ER11 3-3.5mm and an ER11 6-6.5mm. The sizes printed on the ER11 collets are just 3.5 and 6.5 respectively. I also have both the supplied nut and the ER11 nut. NB: By "ER11" I mean the one which I got with the collet extension which match what you'd normally see as ER11.

First of all, the collets don't fit in each other's nuts. Now I tried the 1/4" shank bit in the supplied 6mm collet and it doesn't fit at all. It doesn't seem to be the "one size fits all" kind of collet.

I then tried the 1/4" shank bit in the 6-6.5mm collet and it works great. In fact I cut some wood with it a couple of days ago. (It's a 10mm diameter bit with two cutters. I was running it at 12,000RPM and was jogging it at 500mm/min at first but then increased to 1000mm/min. Although from what I've seen around the forums I'd be better off running it at 6,000RPM and 1000mm/min for the chip size.

So that leaves the last combination, the 1/8" bit with the 3-3.5mm collet. Unfortunately this doesn't clamp the bit at all when fully tight. I think that the taper in the spindle does not allow the "adjustable" collets to clamp across the full range. I've emailed the seller of the spindle to find out more, as they do list the spindle as ER11 and it's not really ER11. Still waiting on a reply from them.

[ger21]

The size etched on the collet id usually the largest size that will fit, so that's why the 1/4" (6.35) won't fit in the 6mm. and 1/4" is only .15 smaller, which is why it'll work in the 6.5 collet.

The 1/8", on the other hand, is .325mm smaller than 3.5mm, which may be too small to get a really good grip. From my experience with ER32 collets, using properly sized collets gives much better results than collets that specify a "range". If you're going to use 1/8" tools, get some 1/8" collets.

[yngndrw]

Aye I may have to get the right collet, although when I put the 1/8" bit in the 3.5mm collet when it's on the collet extension, it tightens without an issue.

[hemsworthlad]

YN can i just ask why you need a collet extension.?
I have the same spindle but ER20 water cooled version and the spindle shaft can extend down quit a bit from the bottom of my Z if i want it too by just sliding the spindle up and down in the bracket.

Can't see why you would need a collet extension or have i got the wrong idea what a collet extension is.?

Can you show us a picture.?