New Machine Build- 3 spindle cnc

[tjskcnc]

Hi!
Time to start the newest project for the shop - a cnc with three routers running separate bits.

The reason for three spindles:
First - drill holes
Second - cut slots and other types of pockets
Third - cut out shape

In a production environment, that means handling the product three times, which is not very efficient or profitable.

Decisions already made:
Primarily an 8020/Aluminum/Steel build, using bearings and parts from CNCRouterParts.
Shape cutting router 3+HP
other routers between 1Hp and 2HP
K2CNC router mounts
Mach3 (of course)
Gecko Drives (the only drives I haven't fried)

First problem to tackle:
Ball screw vs Acme vs R&P
This is an old discussion, but I would like opinions on the subjects of torque and holding strength. I will be cutting skateboard decks, which requires cutting though 1/2 inch 7ply Maple plywood in one pass (thus the 3+HP router.) The accuracy of all three methods are within required specs, but which system would best be able to handle the force required to plow through the plywood? The long axis (gantry) is the weak link in this area. The image (not to scale, of course) shows the basic layout of the machine. The "X" shows where the operator will be standing. The three Z-axes will be mounted on the same plate, using the same Y axis, but operating separately.



Thanks in advance for your help!

[TSJ]

I am not a expert more of a novice, but facing a similar problem (my design will have two spindles). If you want to have a cutting width of 48" then the gantry must be largerer so that either the left or right spindle can cut to the edge.

[blue_luke]

Just a question here....

On a point of view regarding manufacturing efficiency, I guess you will not drive the three routers at the same time.

Would'nt it be more efficient to install an automatic tool changer (ATC)
Somebody on CNCzone has built one for a CNC mill, I guess it could be adapted for router spindle...
Considering the building cost of 3 Z axis, plus the added weight to move around (bigger drive motors etc...) to me an ATC seems more practical.

Nevertheless, your idea is interesting and I am curious to see the outcome.

By the way, the cost of a 3hp ATC ready spindle is probably beyond reach!!

Luc

[wcarrothers1]

My current build is currently slated to have 2 z axis. Companies making routers are making them with 3 spindle carriers as aposed to ATC's. Their is an argument for cost savings and the typical need is uasually only for 3 bits for a given operation cutting panels. But another argument is switching between 2 spindles already loaded with bits is quicker then running to a tool changer and swaping bits then going back to work..

Either way something to think about. I do a lot of engraving so most likely I'll have my VFD on 1 axis and my engraving spindle on the other axis..

Becides bigger motors and heavier gantry is all in good fun.. Hopefully mine will not be to slow..hehe

b.

[blue_luke]

Another argument for the three head machine, even if its against my own with three heads you can do 3 times as much work, especially in the engraving business!

Three head with the same engraving bit and our production is instantly 3 times.

In the electronic business ( my trade) printed circuit boards are drilled on machines called 'Excellon' those machines have routinely 3 spindles with ATC !

Should see these go! Quite amazing!!

Luc

[ahren]

The force your machine will be able to generate while cutting is really a function of the speed you are cutting (which places you on the torque curve for your motor), the gearing of your mechanical drive system, and the efficiency of that system. Each of the mechanical drives you have pointed out has different options for gearing, so it's not as straightforward as just saying one type of drive is better.

That being said, ACME and ballscrews are effectively the same on this issue, except that ballscrews are more efficient (typically 80-90%), whereas the best you can get with ACME is around 50% using multi-start. ACME's advantages are primarily in lower cost and easier setup (which is why we sell them). If you are driving with stepper motors, there is a sweet spot for screws where your motor torque at desired cutting speed multiplied by your mechanical advantage at that speed gives you the maximum amount of power. Many machines on this forum have had good success with 2 turns per inch or 1 turn per inch for screw-based machines.

Rack and pinion systems tend to be geared even faster, simply because there is a minimum size for the pinion gear, so without an intermediate reduction (like the belt and pulley system in our R&P units), you effectively travel pi*gear diameter for each revolution. Our Nema 23 and 34 systems have been designed to provide excellent cutting force in the 100-300 IPM range. R&P is also highly efficient, also in the 80-90% range.

So which should you use? It comes down to more than just force. If your travel in any axis is over 4', I would recommend R&P to avoid screw whip, or the larger screws (and hence larger motors) you would need to avoid it. In addition, you will achieve excellent accelerations and top speeds with this system, so moving between holes for drilling will be faster.

For your z, I would go ballscrew if you can afford it. Travels are shorter here, and vertical axes, while possible, don't make a lot of sense for R&P (the axis can just fall when power is lost), and the lifetime of a recirculating ball screw is a nice thing to have.

Hope this helps!

Ahren
CNCRouterParts

[ger21]

If one spindle is for drilling holes only, mount a drill, not a router. A drill is smaller, and will last a lot longer.

And what Ahren said. You can gear all 3 systems to have the same amount of force. And proper tooling can dramatically reduce cutting force.

Luc, the cheapest ATC router spindle you're going to find is about $3500. Then add $150-$200 per tool holder.
3 complete Z axis should be less than half that.

[tjskcnc]

Originally Posted by ger21 If one spindle is for drilling holes only, mount a drill, not a router. A drill is smaller, and will last a lot longer. . .

Are you saying a regular hand power drill, or some other special spindle? I am open to all ideas at this time.

I plan on having all three spindles mounted on the same backing plate, so there will be a permanent fixed distance between each spindle. This should make it easier to program, plus needing only one set of power transmission, i.e. one lead screw or R&P set up, one set of bearings, etc. Having the Y-axis as the longer axis allows easier access to the table and changing of the boards. Through trial and error I have found this the easiest way to work in this situation.

Currently I will need to use a router for the heavy cuts, namely, the shapes of the boards. I am exploring different options for spindles, even contacting a few suppliers for recommendations and costs, but that will be a future modification, once the machine starts paying for itself. I don't know enough about spindles at this time to risk thousands of dollars on the wrong equipment.

Using an ATC takes too much time. I'm not dealing with just a few boards here. I just finished 300 decks last week. Today I started work on 2 prototypes for 500 decks to begin this week. I manually edit the g-code whenever I can see ways to shave even a few seconds off the cut time. Saving 20 seconds on 500 cuts adds up to almost 3 hours. The company I subcontract from uses a 4-Headed Bolo cnc, 2 for cutting, 2 for drilling. They cut at 600+ ipm and drill at 200+ ipm. No ATC in sight. The drills are cutting before the router spindles have completely stopped. I'll be happy cutting at 100 to 150 ipm, but fast rapids while drilling holes will help reduce cutting time.

I am leaning towards Ahren's R&P for the long Y-axis, possibly also the X-axis. I'll start looking into ballscrews for the Z-axes, but I'm comfortable with the performance of Acme 1/2 x 10 single starts, using Ahren's 8020 parts.

Thanks for the input so far. This is going to be an interesting build log.

[ger21]

If you have air, then a pneumatic drill is very compact, although maybe a bit loud.

Amazon.com: MSI-PRO SM710 3/8-Inch Straight Pneumatic Drill with Keyed Jacobs Chuck: James Mask: Home Improvement

Amazon.com: MSI-PRO SM710 3/8-Inch Straight Pneumatic Drill with Keyed Jacobs Chuck: James Mask: Home Improvement

If all 3 spindles are mounted to one plate, don't they need to move independently to provide clearance? At my last job, two 10HP spindles were mounted to a single plate, and raised and lowered independently with pneumatic cylinders.

I don't know if you've seen it, but I have an on and off design thread for a 2 spindle router I hope to be building soon. Still making a lot of changes, although I have cut a few small parts.

http://www.cnczone.com/forums/diy-cn...ew_design.html

[SORCHEROR]

why not use the three spindles to make three decks at once,and use a quick change
tooling setup like tormach does,you can manually change tools fairly quickly,the time lost will be made up on cutting three at once,and saves cost on having three seperate movable z axis,they could all move on with one set of guides and one screw my way
tormachs quick change tooling is just a setscrew and slip the new one in until it hits the shoulder than tighten,fairly quick,just a thought to consider,good luck

[M250cnc]

Three spindles each doing a different operation doesn't seem a good way to approach this.

To make skateboards i would have the ply supplied cut in strips wide enough to allow a bit of clean up maybe even cut to length, the cost should be minimal extra

I would then drill holes maybe stacking the pieces using some form of location jig with the amount stacked depending on accuracy required

Then i would have two jigs to locate the parts using the drilled holes

I would load one part set machine running while loading the second part in the other jig

The same jigs maybe could be used to profile the outside again while one is running the second could be loaded

This is using one spindle as unless you are doing the same operation on all three spindle at once it doesn't make much sense

Sometimes the time loading of the part is greater than the time to machine that part.

The machine could be smaller and therefore faster and simpler to make

[tjskcnc]

Originally Posted by M250cnc Three spindles each doing a different operation doesn't seem a good way to approach this. To make skateboards i would have the ply supplied cut in strips wide enough to allow a bit of clean up maybe even cut to length, the cost should be minimal extra

The company provides me with the blanks. They press them in stacks (boards are custom 3D shapes) and they are not exactly the same from board to board. I had to visit their production facilities to see how they handle the problem. Basically, they ignore the discrepancies and just cut. Frustrated the #$@#$##$! out of me because they require precision work from me.

I would then drill holes maybe stacking the pieces using some form of location jig with the amount stacked depending on accuracy required Then i would have two jigs to locate the parts using the drilled holes I would load one part set machine running while loading the second part in the other jig The same jigs maybe could be used to profile the outside again while one is running the second could be loaded This is using one spindle as unless you are doing the same operation on all three spindle at once it doesn't make much sense

This is how I currently do it. Okay if I'm doing 50 decks, not so good for 500. Also, inaccuracies can occur going from one machine/jig to another. Prior to contracting me, the company manually drilled their specialty boards and cut them as you suggest, but it's not as accurate as when cut from the same machine in the same setup at one time. I've watched their manufacturing process - I'm welcome to visit anytime to learn - and I have experimented with different processes, and cutting it all on one machine at one time is the only way to do it accurately. We're not talking about regular skateboards. These are specialized skateboards, with specially designed trucks. Some of these companies have spent a lot of time and money designing the shapes of the decks, with patented designs for the trucks. One client actually assembles the boards, and, if they don't roll perfectly straight, the board is rejected. This particular board retails for over $400.

Sometimes the time loading of the part is greater than the time to machine that part.

After a few hundred units, you get pretty fast. I have the board changed out and ready by the time the router hits home. My helper is even faster. Boards are held by vacuum - easy on and off.

The machine could be smaller and therefore faster and simpler to make

I don't know how much quicker I can build than using 80/20 and all the parts available, especially from companies like CNCRouterParts. The questions I am facing are the best way to move those parts around (acme, r&p, etc., which is where this build is at right now, ) motor size, electronics (I'm obviously not going to be able to just slap on a G540), and spindles, which has been touched on a bit already.

why not use the three spindles to make three decks at once,and use a quick change tooling setup like tormach does,you can manually change tools fairly quickly,the time lost will be made up on cutting three at once,and saves cost on having three seperate movable z axis,they could all move on with one set of guides and one screw my way tormachs quick change tooling is just a setscrew and slip the new one in until it hits the shoulder than tighten,fairly quick,just a thought to consider,good luck

Sounded interesting so I checked their website. Doesn't look like something I could use for this situation, but I booked mark their site for future reference. I currently use K2CNC's adapters for my 1/8" bits, so I understand what you are getting at. Even if it can work out faster, I can't see myself making 100+ tool changes in a day!

Ger21 _ love your ideas on the multi-spindle design. I went through the whole thread and checked out almost every link suggested there.

About the air drill - is there any possible way to control the on/off through Mach3? Air is already running to clean the table, so it would be available.

For the Z-axes, I am thinking of using Ahren's design 100% and just mounting 3 of them to an aluminum plate. Below is a picture of one taken from another build log so those who don't know what I'm talking about can see one. Any opinions? advice? praise? jeers?

Ahren - do you know of any builds using your design and parts with ballscrews?

Thanks everyone so far for your input, even if the ideas just make me rethink my goals for this machine and the manufacturing process.

I've been online since 6am working on this. It's now 11:30! time to get to work making those prototypes!