[Buoyen]

I have an idea for a mini mill. To machine small parts in aluminium and steel. I see LMS sell the Hitorque mini mill castings.

Base:
https://littlemachineshop.com/produc...4308&category=

Column:
https://littlemachineshop.com/produc...4307&category=

How would they be for a base of a small diy CNC machine. Remove some of the dovetail to make space for rails and bearing blocks and use the base surfaces which should be flat on these as mounting surfaces for linear rails.
I read the ways on them are not perfect, but I guess they are better than what I can do myself. I do have a router which I can use to remove som material where needed, but I do not thrust it to make things flat and parallell.

Like shown in the picture. This is with 20mm rails. 1 long bearing block per rail.

240mm rails on Y to get about 140mm of travel. The drawing has 250mm rails, but they stick outside the base.
300mm rails on X to get about 200mm of travel.
350mm rails on Z to get about 250mm of travel.

Maybe want to have longer rails on X also, and possible 2 bearing blocks on each rail on X and Z.

But the question is, the castings. Are they good and worth the money?

[skrubol]

I had an earlier version of that mill, wasn't bad.
The single carriage per rail doesn't look too stiff to me though.

[asuratman]

I bought 2 ea base casting, but for next built I will use a block steel for the linear rail.Better to use linear rail than dovetail. You can get size as you like and can get from market very easy. You can get BK/BF and ballscrew nut easy, no need to modify the casting if you want bigger screw.


Attachment 408216

[Buoyen]

Nice. Sounds good. Agree single rail don't look good. Attach a photo of an improved version. Bigger table also. Longer rails and stickout on both sides, but I see the bearing trucks is not supportet to the end either.

Have thought about starting from scratch built steel base. Then I'd need a machine shop to prepare them for me as I can't do it myself.

Material properties is different with materials, but generally the steel base would be stiffer but offer less damping than cast iron? Maybe using high carbon steel makes the differenses less. Or maybe all of that is insignificant on this level?

These castings look fairly well priced. That's why they caught my attention. Buying steel and have it machined at a shop would be more expencive I guess as I live in a country with fairly high rates.

[asuratman]

I have not made one from scratch with steel base, but in cnczone here some people have made one. They got good result, looks easy and cheap to build one. They used 1605 on xy on my casting base mill. In the future with steel base, I will use 2510 or even 2520 on xy. Not much price different between 2510 and 2520 right now. BK/BF and nut adapter are cheap. What to do just machine flat the steel. For my machine above, I just made z axis with 2 blocks steel and machine flat on the face for linear bearing.

Maybe it will be too expensive to machine steel in your country as you said, I know your country. I have been to some scandinavian countries.

[Pardonmebuthelp]

The column is the most valuable part of the diy mill world. This base and column without added support has flex(i have a cnc 3990mill). Think about the long run as far as what you can do with a larger solid base. If you go with solid base hopefully out of cast iron if you have router you can take off a few thousands off the top of what ever block of metal for the base and then scrape it flat with a dremel, surface plate and some dye

[Buoyen]

True, good tips!
I'm gonna check out availability of some cast iron blocks before I make any decision!

[wizard]

Nice. Sounds good. Agree single rail don't look good. Attach a photo of an improved version. Bigger table also. Longer rails and stickout on both sides, but I see the bearing trucks is not supportet to the end either.

Generally you want two trucks but for some applications you might be able to use longer trucks. you don't go into detail as to what the machine will be doing so if the goal is a general purpose mill use two trucks.

Have thought about starting from scratch built steel base. Then I'd need a machine shop to prepare them for me as I can't do it myself.

Well I don't know about local conditions but employing a machine shop may well be worth the money. The reality is you would need to machine the Chinese iron if you wanted to install linear rails so it is a bit of a wash. Around here we have specialized shops like Home | Nifty Bar that are focused on processing plate stock and even stocks some machined materials. If you went with an A36 plate about 3" thick you would have a nice heavy and stable base to build the rest of the machine on.

Solid plate is just one possibility here. If you have access to steel tubing in your location you could go with a heavy 8 x 4 or 10 x 6 tube that you then fill with an epoxy granite fill. Such a base would be as good as cast iron with respect to vibration.

Material properties is different with materials, but generally the steel base would be stiffer but offer less damping than cast iron? Maybe using high carbon steel makes the differenses less. Or maybe all of that is insignificant on this level?

Honestly if there is enough mass in the base I don't see it making much of a difference in these small machines. You need to remember the plates will be solid materials and you can buy plate in just about any reasonable thickness. You just need to remember that you need mass, especially in the base to build up a solid machine

These castings look fairly well priced. That's why they caught my attention. Buying steel and have it machined at a shop would be more expencive I guess as I live in a country with fairly high rates.

Yes even here the shop rates can be high. You might get lucky to find a model engineering hobbiest in your area that would like to help. However I think you are missing a point here, if you want to put linear rails on these casting you will need to have them machined anyways, so machining costs will likely be a wash. You might be surprised at how cheap a suitable steel tube may be. if you can avoid paying "prime" pricing by finding drops or something in a junk yard you might be paying less that $75 collars to replace one of those pieces. That base might be good quality but it is only about 22 pounds. I can buy drops steel for $1 a pound or much less depending upon the direction of the wind. So if I spent $75 on steel I would have 3 times the mass right off the bat. Spend 75 dollars to have the beam faced and you you are still ahead. It will likely cost more than $75 at a machine shop but you might find an automotive shop with a machine to grind heads that might be more cost effective. For the true gluttons there is also hand scrapping a beam or plate flat. The point is you can find alternatives.

[Buoyen]

Very nice wizard. Thank you.

I have access to steel tubing loaclly. I have bought some small steel plates from the same company earlier and paid about US$ 3.50/kg incl cut to size.

I have one question though. How to mount the column to the base. If I can buy thick plates or solid stock I think it would be best if I can use the same material/size stock for both column and base.

Then I can think of 3 ways to mount the column to the base, what would you recommend:
- Through holes in the column and mount it to the back side of the base. Tramming front/back would need shims and I can tilt left/right.
- Plates on the sides which bolts to both the base and the column. This would also work with steel tubes and not solid stock. The plates could be larger than the column stock to increase stiffness. I can then tilt the column to tram fron/back and shim to tram left/right.
- Mill slots in the column to achieve a "lip" with holes at the bottom of the column to bolt it to the base which looks like the most common method of mounting the column. Tramming would need shims either way.

Added another picture from behind with the through holes in the column.

[Stigoe]

Another option would be to get a steel plate slightly wider than the column, drill 4 holes, weld it to the bottom of the column and bolt it to the top of the base (which would need to be equally wider than the column), similar to the new type BF25 or PM25. This would be a lot easier than trying to tighten bolts in milled slots.

[hackish]

I think the PM25 design is generally considered better. I do find it is more rigid than a grizzly, but maybe not only for that reason. It's especially noticeable using a boring bar or taking a more aggressive cut on something.

[Buoyen]

Good to know.
I'm sourcing some materials right now. It might not be ideal, but I'm trying to design something around a 180mm, several 160mm, 100mm and 80mm wide steel u-profiles I got for free and a base plate 33x33cm that is 25mm thick with 25mm thick braces that is welded on one side - unfortunately there are several big holes in the plate so the linear rails can't be mounted directly on it.

[Buoyen]

I have been running som simulations of stiffness with the free parts. I just need to buy some plates for closing the end of the column, XY-plate, spindle plate etc.. Trying to hit a stiffness value of 5N/um when the z is at it's highest (weakest column) and X/Y in the middle. Fixed table and applying the force at the spindle nose 45 degrees between X and Y. The linear bearings is not modelled as springs, but is fixed like a solid piece of steel, so actual stiffness is of course lower.

Using a casting similar to the X2 and the cj18 mini lathe (I have a casting like this already for a 4th axis project - just have to buy another). Was not able to measure any runout inside the 4" baseplate with my 0.01mm dial indicator so it looks okey?

What do you think about this:
X2 casting
ER32 with 25mm straight shank
25x62x17mm angular contact bearings (one at the bottom, one at the top).
Preloaded via a belleville spring washer and nut at the top.
Belt pulley
0,5kw 2-pole induction motor with vfd or 1kw brushless motor with esc

Gearing of the induction motor might be doable:
3:1 gearing for low speed operation (max 1000rpm spindle)
1:3 gearing for high speed operation (max 9000rpm spindle)

[Buoyen]

Got 6,6N/um :cheers:
Not stiff for a vmc, I guess, but for a mini desktop machine with 500-1000w spindle power it should be ok?

I have simulated with one block per rail, but the plan is two blocks per rail, just haven't updated the model. But the base and column is prepared for longer rail in the model.

The beams are steel and I might fill them with reinforced concrete. The fill is modeled with 15000 Mpa young modulus. But acourding to some numbers I have found reinforced concrete might be in the range of 30-50000Mpa. So that will give me a nice bonus in stiffness I guess. At least until I take the linear bearings into acount.

[cncuser1]

Buoyen: I can't speak to the numbers for stiffness. But when I look at the length of engagement of the 3 axis to each of their rails, it looks short.

It is hard to judge from a sketch, over the internet, but I would like to see the engagement doubled.

again I don't know about stiffness numbers but over time the rails will wear, and the minimal engagement will show up more.

( every time I look at my Sieg X2 I am bothered by how little the saddle engages the X and Y axis.)

[skrubol]

I'm not sure if using normal concrete will do much/anything for stiffness. I believe a major reason so many have gone with epoxy granite (aka polymer concrete) is that normal concrete shrinks as it cures, so will pull away from the steel it's supposed to be filling. Do you plan on tying the concrete reinforcements into the frame?

[Buoyen]

Buoyen: I can't speak to the numbers for stiffness. But when I look at the length of engagement of the 3 axis to each of their rails, it looks short.

It is hard to judge from a sketch, over the internet, but I would like to see the engagement doubled.

again I don't know about stiffness numbers but over time the rails will wear, and the minimal engagement will show up more.

( every time I look at my Sieg X2 I am bothered by how little the saddle engages the X and Y axis.)

Thank you. Yes I agree, I have now done some more work on the CAD and using 2 carriers per rail. Also increased the lengths of the rails and spacing a bit.

The travels I have in mind are X 310mm and Y 170mm. Z I don't really know. But as much as I can get to have room for changing of longer tools etc.

I'm mostly working with smaller pieces - like this 50mm diameter aluminium heat sink. This is one is done on my router which is not so rigid and don't have any spray/coolant system so the surface finish is not so good
Coolant system is a part of the reason I wan't a separate machine for metal work. I have dust extraction system in place on the router for wood work.

Adding a picture of a steel plate I also did on the router. Faced it, marked the holes and drilled on the drill press. Then made the pockets on the router. Takes forever in steel...

I'm not sure if using normal concrete will do much/anything for stiffness. I believe a major reason so many have gone with epoxy granite (aka polymer concrete) is that normal concrete shrinks as it cures, so will pull away from the steel it's supposed to be filling. Do you plan on tying the concrete reinforcements into the frame?

I agree there is some uncertainties here, I don't really know if this will work, but one option is to use non-schrink concrete. I see some people have used that with great success. Also I do plan to fasten the concrete to the steel with anchors in the concrete. So doing this should increase the overall stiffness.

[Buoyen]

The model as it is right now. X2 head with ER32 straight shank and 0,75kW 6000rpm AC servo.

Missing all the details so a lot of work left. But I feel I have decided on some things.

I have switched to a design with moving carriages on X and just a plate instead of t-slot table.

20mm rails, 350mm long and 170mm travel on Y. 500mm long and 310mm travel on X and Z.

All the plates is based on 15mm thick 160mm wide steel sheet that just has to be cut to length and faced.

Nema 23 stepper motors and 15mm ball screws.

Forgot: Using 15mm GT2 belts on the axis. 40 teeth pulleys.

[cncuser1]

Buoyen : metal chips ans coolant get into every space and crevice. You have a lot of sensitive components that should be better protected.
some people use simple leather or rubber flaps to cover the sensitive components. They are not sufficient but they are better than nothing.

[Buoyen]

Thank you. Agree. What I want to do is something like this, printed in PLA. Can be a problem with available space, but I'll look into it.