New Machine Build - Router for aluminum

Hello everybody,

At work we're running after a decent contract for aluminum molds. At this moment we don't have a machine to deliver goods, so we are on a thin string of buying one or building one.
There are few decent products out there, but the price to pay for them is pretty hard to come by. Some of them go over 3 times or more the budget of a build. For just a little over the build, commercially, sizing is limited.
We'd require something that can work in the area of 2500 by 3500 by 350 mm.

If, there will be something out of it, here it goes, attached a basic idea.
This will be my eighth CNC so far, and I would mention, that all the others are junk. Good for hobby, but not professional.

So, here we go:
1. Aluminum: - I am interested in high speed machining with light cuts. If I'll be able to go to 5 mm depth with 4-6 m per minute I'm happy. I'm not into the hard deep milling, because for this kind of job is worthless. Yeah, great for roughing, but when I have 2 sqm of finishing, I don't want to sit 2 months for it to finish. So, light cuts, and high speed. My actual machine, which is like an old tractor, will do 2 - 3 mm depth at 2 m per min. But the tool will wear out in 2hrs due to vibration. 0.5 ~ 1mm is the safe with it.
I usually mill with single flute cutter from Datron 4, 6, 8mm at 24.000 rpm. I am considering to go for 40.000 - 60.000 rpm in this case.
2. Construction: - 2500 x 3500 total dimensions. Frame will most probably be made out of steel. Leveling will be made with epoxy. One bucket, hoses, and pour on both X axis supports. This way I believe, I will get it planar. For the main table, I will fill the base with epoxy granite, inserts, and an extra coat of epoxy which I will mill afterwords. Work holding, most probably a sacrificial ally plate with venturi labyrinth. Gantry, will be 5 or 10mm steel sheet, bent, ribbed inside, and enforced with a 100mm pipe. Not sure yet if I'll fill the gantry with EG or not. On one hand is good for dampening, but it's bad for inertia. Considering that I'll go for the fast and light cuts option, I tend to go for the empty idea.
3. Drives - helical good rack and pinion are hard to come buy, so I am considering a 32 dia ballscrew with 32 mm pitch. Fixed screw and rotating ballnut. 2 per X, one per Y, and one per Z.
4. Motors - I have seen some closed loop drivers for steppers available. If that will be the case I will probably opt for something in the range of 12, 20 or 32 NM steppers. It all depends on weather I'll fill up the gantry with EG, or not.
Ideal encoder placing would've been straight on the ballscrew, but in my case I'll have to settle with on the stepper axle. Servos are the other option, but never worked with them. 400W, 700W doesn't translate much in my head at this moment. Also, I'm clueless about their torque, gearing necesities, rpm and so on.
4. ATC - this would be a nice option to have. Had worked with ATC before, but never tought about implementing one. I'm quite fond of these chinese spindles out there, or I might have been lucky with them. I have seen a version with 24.000 rpm and ATC option for ISO 20. Not sure how does it come by, with locking and releasing. Via air, or via solenoid. Also, haven't seen one for 60.000 rpm with ATC.
5. Mach3 will be the way to go, tough people that I spoke with, machine sellers, constantly tell me it's worthless in the professional environment. Same machine will triple it's speed only because of their dedicated drivers and software. Don't know what to believe, is it hype or plain facts. True, theirs run with 1.5 m / sec!!!

The reason I went for this layout instead of a classical one, is because it looks, at least in theory to be easier to level/square and all. Also, the gantry supports are taken by the table. And table or lower frame, I have to do it either way. So that's another plus. Jobs will probably be fed via the forklift so it will be easier this way. Probably it'll come with back aches when doing a small job, but it's a win/lose situation.

If we'll decide to go with it, instead of buying, I'm hoping at 4 months with it.

I might be very wrong about the logic of it, so if you guys have walked this path before, I'd appreciate your thoughts.

Thanks,
P.

I'll watch this with interest. 5mm cut at 200ipm is not what most would call light, it's in the territory of a serious VMC

My 5 Cents ,.. If you find an Stepper which is delivering the Force and Speed you Need let me know ill also use it in my next product.
I think you will Need think like Mesa Card because of frequency of steps ,..
But i think it is possible what you try to achieve ,
My Solution would bev(because i use it happyli in Projects ) generalmechatronics Card which Requires linuxcnc a Touch Screen and servos ,..


Gesendet von meinem iPad mit Tapatalk

Hi guys,

Well, 5mm ADOC HSM isn't easy to achieve I'm sure, but that would be the top target.
Option are: https://www.youtube.com/watch?v=_PC9ezN6cno - triple the expense over the diy, and less than half the size of my requirements. This is top notch quality.
This is more available to us mortals: https://www.youtube.com/watch?v=oqqtMK89B2M
Nevermind here about the material milled, check the speed: Najszybsza frezarka do modeli 1m/s! -Profesjonalne obrabiarki CNC

I'm still waiting on quotations, and, to be honest, I would prefer the buying option anytime over building. CNC building tends to get frustrating at times. Especially when you're running for precision. Been there so many times. Hard work, and the end result is not satisfying. I always tend to say "ok, this is my last build ever. It's just impossible make 'em like you buy 'em." Though, I have seen quite a few expensive machines that have huge shortcomings. I won't even mention chineese option. I believe they stay in woodworking, with woodworking tolerances.
On the other hand, I haven't seen anywhere a decent DIY that fits into my requirements. Maybe I'm aiming high, maybe not.
This build won't be cheap either. Start calculations for it allready reached 10 thousand, so I'll have to be very very carefull with it. It's pretty easy to flush it all down the drain.

Thanks

To buy a machine, a large format router, to do what you want won't be cheap at all. Especially if you want it to be more rigid than what you have or have made in the past. Frankly you would have a major project ahead of you to DIY a machine this large and get acceptable results.

There are several manufactures that might make a suitable machine for you but none of them will be cheap. Haas has a machine that comes close to your needs but you need to be willing to lay out $150,000.00.

Yep, you're right about the Haas, but in Europe you could find for less than half the price. Ok not the same product, but the same result in what it offers as end result. Milled parts.

Nevermind it for now, I had few ideas about the building version of it. I'm not sure if you can call it a D.I.Y. for real, because it won't be made in the basement like all the hobby geeks of us. At work, this is pretty much what we do, and the workshop is pretty able- machines wise. So, if a client would come to us to build this machine, we would hop in for it. But either way, I'm sure it won't be easy.

I am contemplating now, to skip the big welded steel frame, an just pour a big block of reinforced concrete as the table. The price would be less than half, and we have the concrete plant just across the road. From here onwards, I stick to the plan, with a 100-150mm tall EG table. So, it'll be just a big cube for support, nothing more. I don't really see any other minus in this option than moving it. And if, we'll get to moving it, we'll just crush it all, and make another one in the new facility. If, it'll ever be the case.

As far as motors, closed loop steppers for my needs go for the same price as servos. So, leap ahead on this one.
I might be wrong, but I'm looking for 750W servos. The gantry would be somewhere in 200-250kg, and acceleration and deceleration would represent huge forces. But, if servos like to spin fast, I'm thinking to go for a 16mm dia, 5mm pitch ballscrew. So fast rpm to small pitch, this is ok. And I also don't know if I need to gear down the servo, considering it will be a rotating ballnut. Spinning that ballnut won't require much effort. The effort required comes after the thread of the ballscrew, where again, keeping it in 5mm gives me bigger pushing force. It should be the same result as with geared down servo, and 32mm screw with 32mm pitch. Mechanically is simpler, and I'll also get cheaper screws. Yes they will bend on this span, but I don't think it'll be a problem for the ballnut.

Next still unknown, is weather to fill the gantry with E.G. or not. I'm not skilled with F.E.A. to be able to predict how much improvement will those gussets and the crossbar add. They will add rigidity for sure, but damping? Maybe expanding polyurethane foam?
In my case, I'm not that sure if weight is on my side.
As for aluminum extruded profile, it is out of the question. The machine will work in 7-10C in winter and 30C in summer. Alu, steel, concrete and EG are not compatible on this one...

P.

16mm may be a bit thin. Not just because of whipping, but when you change directions with a large load you may induce twisting into a screw that small.

Spinning a nut on a 5mm pitch screw may be difficult at higher speeds. Imo, you'd be better of with a 2525 or 3232 screw, with a belt reduction to the screw. Spin the servos fast, but not the screws and nuts.

I would do the math First do you think 750 w servo i sufficient ? And i would also recommend thicker Drives probably 8 mm pitch is then Fine ,.. Look at Granite devices if you are in eu ,.


Gesendet von meinem iPad mit Tapatalk

Small updates,

Work hasn't gone great. We're bouncing back and forth with what it is exactly that we need, regarding work envelope, MRR's etc. So, design stage is bouncing also. Attached the current state.

Attachment 237842


Frame: RHS Steel, 60x40mm structure with 150x150mm. Base will be all filled with epoxy granite.
Gantry is made out of 2 opposite C channel beams (200x75), with 120x80mm elements welded perpendicular inside. For the Z axis, probably the same profile 200x75mm. Z will be quite high 200~250mm, so rigidity is a must there. Gantry final weight, roughly 180kg.

Attachment 237844
Attachment 237846
Travels 1600 by 1400mm

From here onwards:


Option 1:
R and P, 12nm Steppers - small, slow and cheap
Option 2:
R and P or rotating ballnuts, servos - bigger, fast and expensive

I would have a question for option 2. Does a linear encoder solve mechanical problems? backlash or repetability for example?
If so, I attached a picture of a diy idea. Probably it's dumb, but I need confirmation on it. It's a roller pressed on a track, with a rotary encoder mounted to it, instead of leaving it mounted on the servo shaft. This way I would have a confirmation of the position. Track would be some regular aluminum plate, not grounded. Probably the whole system would be spring assisted. But as far as the revolutions go, it should do the trick.

Attachment 237848

Now, IF, this would work, how do encoders mesh with servos as far as calibration goes? How and where should I imput data like mm per revolution etc?


Another question, servos are rated for example at 3000 rpm. That is max value? or I get max power, at those RPM's only? I'm thinking to go with a 10:1 ratio, perhaps even more on my case.
My math goes like this:
DHT series Servo 400w with 1.27 nm torque and peak at 3.6nm running at 3000rpm.
With a 10 to 1 ratio and a 12mm M1 pinion I should be getting 211,6 kg of force/servo. With 423.2kg on x axis (one servo each side), acceleration should be: 2.35m/s^2.
5~10m/minute feeds sound resonable for now.
Did I calculate things right? Numbers look too good to be true...

Last question:
Controllers;
I mentionned somewhere in the begining that I would be interested in HSM. Well, not the basic meaning of HSM. I mean HSM as light cuts and high feeds. But, I've read about limitations from Windows/Mach3/Controller etc.
I tried a fast simulation in Mach, and went nowhere near what I asked it to go. Ok, it was a dumb setup, for the test purpouse only - 10.000mm/min feeds, max acc/decceleration setup etc. And it was still contouring between 300 and 2000 mm/min.
Then I tried the demo software of this product:
Eding CNC - PC based CNC control
Got pretty much the same result as Mach. So, I haven't even reach to the controller, and I allready see limitations.
Both this and Dynomotion reading was fine, but is it even possible in real life to get these high speeds? assuming that one has the required power?

Thanks