The whole side ?
Seems like the collett and such would get in the way unless you had a very long milling bit.
You’d want shank to extend up as little as possible out of the collet and the collet to be close to the surafce. If flatsurafce .250" would be plenty. How thick a piece are we taking about? Anything under .500" - a .500" dia cutter would be fine.
Roughing cutter to get all but last .050-.100" of profile, then do final pass.
@artg_dms @nicksilva you guys work on CNC’s - how far back do you offset before final cut? How thick can you do with say a .750" cutter?
I was specifically thinking about how to finish the stack of plates for gantry supports.
So 2-6".
From your guideline, the 6" surface would require a 2" diameter mill at least 6" long, plus shank.
Sounds expensive…
If you are using the HAAS to cut these plates, why are you trying to cut the entire stack of plates at the same time. Suggest you just cut 1 or 2 at a time. This will allow you to use our existing tooling and all the parts will still be identical. It will just take more time to cut them.
The concept was to rough cut on the plasma, then finish on the HAAS.
I was wondering if the assembly could be finished at one go, or if it would need to be done one plate at a time.
If using a CNC why stack? The repeatability is very high. You’ll be forfeiting a LOT of repeatability between parts due to cutter flexibility and likely chatter marks.
I’m totally not getting this desire to stack parts. There is a reason they don’t do this in real world manufacturing. Also, thinner pieces travel at a higher IPM.
After you machine the individual parts, make a locating fixture so there is a common datum plane - then drill holes. If there is a high tolerance for concentricity (I’d guess high), positional (again high) and parallelism (high but of the three maybe shim adjustable) to a datum or a feature requirement and positional location, then maybe stack. What kind of tolerances do you require? If you don’t know, then you don’t know what process to use.
I know I drive people crazy always talking about tolerances required. But talk to a Manufacturing Engineer - without this, where to you start? Oh, and how are you going to measure all this? We have very limited capability - fortunately you could run a few test dimensional parts, say a 1-2-3 solid blocks with a couple holes. Run three of those, and see what the repeatability is and variance between parts. You can compare these against our gage blocks which are good to at least 0.XXXX5" for comparison. A plasma cam is probably ±.005" positionally, then add in the variable of the cut-out material (on thin stock where kerfing isn’t relevant.). I’m not sure anyone has ever run test parts on the Haas for accuracy or repeatability. This is probably something that should be done when we reopen and have it sent out for professional CMM measurement. Then we’ll know.
Yes, it will be expensive a "2 finish cutter isn’t cheap.
If any of y’all know how to do GD&T dimensioning - I strongly suggest you do so, it will force you to think what tolerances and how they relate to other parts in the assembly are used then that will drive the process to produce it. This is important as you are trying to make a machine that is accurate and repeatable.
There are methods to get around it needing to be that accurate, you can use this stuff to fill the gap, it sets to 30% of the modulus of steel in compression, and gives 100% contact area. So, you can just machine each plate individually, and bolt it together–voila, super rigid connection. Also, you can align it before injecting the adhesive, so you can attain geometric accuracy without precise machining of the mating surfaces (which normally is extremely difficult to do and have an actual good mechanical joint with significant contact).
https://www.diamant-polymer.de/en/products/dwh/ you can get it from Devitt Machinery Co here in the US
The other thing is, for a small machine especially, you don’t need that thick of plates. On my design I have 4 1" plates stacked up per column. Of course they are almost 10 inches long in the other direction, but with a 25lb sideload (reasonable for a conservative small endmill cut in steel using an HSM toolpath) they only deflect .00002 inch perpendicular to their thin direction. Which is more than stiff enough, as you will have more deflection from compression of your X ballscrew or from the deflection of the Z axis hanging down/rotating about the linear guides because of a lack of stiffness in the guides and their mounting surfaces. This is the beauty of the fixed gantry design, and why it is used in so many high end machines.
If you had a 10" column versus the 16" column on mine, doing some very approximate math you could probably use 2.5 inches or less on either side and have a similar stiffness. Realistically you would need even less because the machine wouldn’t have the torque in the spindle drive to generate enough force to cause deflection issues during machining, and the linear guides/ballscrews would have enough clearance in them to effectively eliminate any benefits of extra stiff frame parts.
Machine Shop Committee–the strongest opinions in the land about stacking objects 
1 Like
Filling the slop with glue really rubs me the wrong way, no matter how spiffy the glue is.
( Perhaps another ill educated opinion, but such is life. )
Did you do any machining on your plates to lessen the weight ? It would seem that with some appropriate FEA, you might could find places to do that. ( I am WAY out of practice on that, and was never that good at it. But I do have skilled friends… )
For my personal requirements, I’d like to be able to have 12" under the spindle. Fixturing to the bed would seem to be an option, so no vice, and I don’t need to do deep operations, but I do need to do some on the end of an 8" tall block.
I’d also like to make it a portable machine. Hence the desire to lighten where possible. That may wind up being contradictory to other requirements. Some though will have to go into making it as convenient as possible to disassemble then re-tram with minimal effort. ( yeah yeah - I don’t want much… )
In case anyone is interested here are some builds I have taken inspiration from. All super interesting to look at. I think they all qualify as milling machines (except the granite section one which is more of a router but is still awesome).
This one is easily the most impressive. Interestingly it uses the bargain basement $600 Aliexpress BT30 spindle with success.
Nice one with lots of cutting examples. Same cheap Aliexpress spindle.
This is a bit more DIY:
Where I got my inspiration about using adhered and bolted granite sections:
A nice take with a larger weldment frame:
1 Like
So if we machine the side posts one plate at a time on the HAAS, using the same program on each, then get them properly aligned and stacked, how close will the seams between plates match ?
“close enough” I’d expect within a half of a thou, and most of the difference being in the way you fixture the part.
If you care about that much of a difference, then you might need to start specing the temperature at which the plate is rated 
Not that critical for sure.
Kind of a wood working notion. Once I run the router over it, the edges match…
Mostly just a curiosity thing, not a quest for ultimate precision. It would look great. Will not matter much though, other than the ends of the columns. And that likely can be addressed otherwise.
whether it’s on the floor or the walls, usually between .010 and .015 inch.
This actually depends on the endmill and material, but generally a 2 or 3 flute for roughing and a 4 flute for finish. The overall depth is going to be determined by the length of the cut (LOC or the length of the flutes). If the software is smart enough to interpolate the cut (i.e. a gradual change of depth until the final depth) you can basically go to the full LOC as long as the chip load is kept small, like about .003 - so high speed and low feed. Personally I’d go multiple cuts to the LOC so that the feed rate can be reasonably quick. The last thing you want to do is plunge a large amount. Endmills and not drills. All the above applies if you’re coming in from the side as well.
1 Like
I agree. but as you point out the fixturing has to be right on for repeatability. all of this is assuming the actual run-out on the machine is good - especially if flatness is a concern.
I should have been more clear. My question was not for a critical uber precise something. More just how nice could I make a multi layer side post look. From a practical standpoint, as long as the ends are flat and parallel to each other, the sides don’t matter much.
I was just curious what the Haas can do. Still filling in the blanks on that.
Thanks for the answers.
I’ve put this one on my potential list. It looks nice and prices out nicely though it’s more a beefy shapeoko maybe than a true Mill. @Kentamanos, do you have a plan to buy any time soon?
I’m currently in a home automation rabbit hole (based around Home Assistant), but I should probably start to realize how long stuff will take to ship and start to source some things.
If you get to a point where you are thinking about buying some stuff let me know. I’d be happy to help offset some shipping etc. We can take it to PM’s, but I’m leaning towards a NEMA-23 version.
I have seen some good examples on you tube.
I have a half formed idea to use honey comb material for rigidity .
Built like a briefcase .folds closed for transport or storage.
Also have wondered if a 1/2 inch variable speed drill could be used for the spindle motor. ( on a different project.build.)
I know where to get aluminum scrap for 1.00 dollar a pound.
I don’t always find what I want. Lake June Recycling in Mesquite
I would think no because of side-loads.
That’s handy to know.
Have you guys see openbuilds.com? The have a couple open source router type builds.