Hi, planning next big cnc project, a diy-vmc !

what do you think of this idea, structure is built from square steel tubing, base is 400x200x8mm, column 300x300x8mm, X-carriage 300x100x8
travel 500/300/400mm (X/Y/Z)
table 800x300mm

the idea is that we want a big accurate machine but during assembly each component should be possible to handle by two persons (<100kg).

If anyone has experience with the rigidity/damping qualities of this kind of steel tubing i'd be interested to hear about it ! especially the effects of filling the tubes with sand/lead-pellets or similar damping material.

any other ideas on how to build the main structure of a diy-vmc ? plasma-cut steel plate is another option but will require serious welding...

I've heard that steel often need to be heated to a high temp to relieve internal stress before it is any good for a precision machine like this. Does it apply to this kind of steel-tubes ? Heating would then have to be done after assembly of the main structure but before machining of the surfaces...

If you fill it with concrete, make sure to cast a big lift hook in the top. It would then have a second use.......for a boat anchor :D

But in the way of being a bit more helpful, I would say to you to get an older machine and retrofit it. Use those old iron castings, there is nothing wrong with them. Chances are good that there are a few other components on a used machine that could be reused as you do a thorough overhaul of it. It makes no sense to build one from scratch, if you are mimicking an existing design.

What accuracy and precision do you want to get? If you are satisfied with +/-0.025mm it could be possible with a fabricated steel tube structure. You would probably get some benefit from filling the tubes with loose aggregate to give some damping. But I think you are going to need access to a large milling machine for machining some surfaces on your parts. Steel tube is not accurate enough for you to just drill and tap holes to bolt your linear guides direct to the surface of the tube. You will need to weld on lengths of hot rolled steels something like 15mm by 100mm and machine the surface of these to get them straight and parallel for attaching the guides. Also I think the bed and column should be bolted together rather than welded as is implied in your drawing and again you would need the welded and machined strips. The reason I suggest this is that with this approach you can install and true up your Z and Y guides then to get these perpendicular in two planes you would shim the bolted joint between the base and column. You will need to do some stress relieving and I suggest doing it in two steps; stress relieve then rough machine to within 0.25mm of the finished dimensions, drill and tap then stress relieve again and finish the machining. With fabricated tube structures sometimes the stress relieving is done just by letting them age outside with changing temperatures. A Finnish summer and winter would certainly give you a good temperature change but two of them are a long wait so you will need to build yourself a big fire somewhere or find a friend with a big oven. Although I just realised ... Finland ... saunas; just take all the pieces with you to the sauna a few times but leave them behind when you do the jumping through a hole in the ice. For stress relieving they should cool very slowly. Also jumping through a hole in the ice carrying 75kg of metal might reduce your chances of finishing the project. And in a tone almost as flippant as my fellow Canadian I think I have reached the end of my suggestions.

If you fill it with concrete, make sure to cast a big lift hook in the top. It would then have a second use.......for a boat anchor :D

If I ever get a boat that needs a 500kg anchor then I'll be sure to remember your tip :cheers:

But in the way of being a bit more helpful, I would say to you to get an older machine and retrofit it. Use those old iron castings, there is nothing wrong with them. Chances are good that there are a few other components on a used machine that could be reused as you do a thorough overhaul of it. It makes no sense to build one from scratch, if you are mimicking an existing design.
Been trying to find a machine to retrofit for some time... however there's the problem of transporting a 3000+kg machine, unkown wear and tear on all parts (not least the dovetail ways which most manual mills seem to have), most manual mills being designed for totally different Z-movement (quill+knee) etc. etc.

Thanks for your comments!
What accuracy and precision do you want to get? If you are satisfied with +/-0.025mm it could be possible with a fabricated steel tube structure. You would probably get some benefit from filling the tubes with loose aggregate to give some damping. But I think you are going to need access to a large milling machine for machining some surfaces on your parts.

The linear slides and ballscrews are good to about +/-0.01mm (over short distances) so I would like the rest of the machine to not limit accuracy greatly. I do have the opportunity to machine the surfaces of the tubes on a big vmc.

Steel tube is not accurate enough for you to just drill and tap holes to bolt your linear guides direct to the surface of the tube. You will need to weld on lengths of hot rolled steels something like 15mm by 100mm and machine the surface of these to get them straight and parallel for attaching the guides. Also I think the bed and column should be bolted together rather than welded as is implied in your drawing and again you would need the welded and machined strips. The reason I suggest this is that with this approach you can install and true up your Z and Y guides then to get these perpendicular in two planes you would shim the bolted joint between the base and column.

that was the advice I got from a friend also, the linear guides won't be fastened securely enough by only bolting them to the 8mm thick tube... probably need steel plates under the linears to fasten them securely.

Yes, my intention was to bolt on the column to make it easily adjustable to 90degrees (sorry, did not draw that in the pic).

You will need to do some stress relieving and I suggest doing it in two steps; stress relieve then rough machine to within 0.25mm of the finished dimensions, drill and tap then stress relieve again and finish the machining. With fabricated tube structures sometimes the stress relieving is done just by letting them age outside with changing temperatures. A Finnish summer and winter would certainly give you a good temperature change but two of them are a long wait so you will need to build yourself a big fire somewhere or find a friend with a big oven. Although I just realised ... Finland ... saunas; just take all the pieces with you to the sauna a few times but leave them behind when you do the jumping through a hole in the ice. For stress relieving they should cool very slowly.


I thought stress relieving required much higher temperatures ?
summer-winter otside would be something like +25C to -15C
sauna would be +100C or so and cool to room temp.
is that enough ?

If it is then it will save a bit of money since the commercial stress relief I heard about heats the parts until they glow and costs maybe 2eur/kg.

What is your portability requirement? Does this thing have to be knocked down, moved and set back up on a regular basis?

Lots of mass in the machine base is your friend.

What are you planning to cut on it? How large will the servos be? You do realize that you will be spending good money on ballscrews, rails, motors and drives and a toolchanger, right? When you are going to spend $20K, sometimes it makes more sense to spend 30K or $40k, rather than end up with an expensive experiment. There are lots of older VMC's out there.

What is your portability requirement? Does this thing have to be knocked down, moved and set back up on a regular basis?

not on a regular basis, maybe after 1-2years. also, the current shop is a little awkward to get to, no driving right in with a big truck...

Lots of mass in the machine base is your friend.
What are you planning to cut on it? How large will the servos be? You do

Mostly parts and molds in Aluminium. A bit of steel (soft) but speed is not a requirement then.
Right now it looks like 400W servos for the axes. with 5mm/rev ballscrews, depending on the gearing, I figure between 2 and 5 kN of push force with rapids between 2000 and 4000 mm/min (thats 80-160ipm for you inch guys). accuracy and force are more important than fast/cool/bragging-rights rapids.
realize that you will be spending good money on ballscrews, rails, motors and drives and a toolchanger, right? When you are going to spend $20K, sometimes it makes more sense to spend 30K or $40k, rather than end up with an expensive experiment. There are lots of older VMC's out there.
Its more fun to DIY :banana:

some kind of budget:
I have the servos and I'm building the drives myself, might have to buy powersupply (300-600 maybe).
The rails and ballscrews are around 2k
machine structure as above would be maybe 1k including machining etc
spindle would be machined by my friend at material cost
atc can wait

so, I'd like to spend around 5k during the next 18months as the parts slowly come together...

"I thought stress relieving required much higher temperatures ?.....sauna would be +100C or so and cool to room temp.....".

I thought Fins were much tougher than that! The literature we get from the Finland Tourist Bureau gives the impression you come out of a sauna at +212 degrees and then chop a hole in the ice at -40 before jumping into the water. Although perhaps they converted to Fahrenheit for us poor benighted, backward North Americans.

Seriously; yes stress relieving needs much higher temperatures to be done correctly. Just letting a welded fabrication age for several weeks with changes in temperature can relieve some of the stress. The optimum would be to stress relieve correctly, rough machine and then stress relieve again before final machining. A compromise would be to rough machine and then relieve.

I just googled stress relieving and found:

"stress relieving is done at about 650 ºC (1202 ºF) for about one hour or till the whole part reaches the temperature. This removes more than 90% of the internal stresses."

Depending on where you are doing the work (this suggestion is not suitable for an apartment in the middle of Helsinki) you might be able to do effective stress relieving just using a 45 gallon (200 litre?) steel drum and sand with a large propane torch.

Your parts are all hollow and you could fabricate them in such a way that you have access to the inside; the bolted joint between the base and column could be around a hole and the long part that runs on the Y axis and carries the X axis guides could have a small hole at each end. Put some sand in the bottom of the drum, stand the part on it and fill the drum with sand to surround the part. Hang a steel baffle inside the part so you can aim the torch flame down one side of the baffle and have it pass around and come up the other side. With a big enough torch you should be able to get to what is called a dark to medium cherry red heat; you do not have much mass and you have a large surface area available for heating. Then let it cool.

You are right it is mroe fun to DIY provided you pick a doable challenge. I think you are just on the doable side of the borderline.