Fadal Spindle Addition to a CNC Router

[Scott.Damman]

Looking for a little help on an ongoing project. I am adding a low speed spindle to my CNC router, specifically adding a Fadal spindle. Currently the machine has a 10hp 24,000 rpm BT30 spindle on it. I would like to also be able to cut steel with this machine. So the plan is to make the spindle swappable from the current spindle to lower speed spindle. I know that the machine will be very limited in capability on steel and I will need to take lite cuts and baby it. But I only have space for one machine so making it capable of both is key.

Some info on the machine it is a pretty hefty/robust machine as far as CNC routers go...
5ftx10ft machining area with 22inch on Z. All steel construction, stress relieved and all critical surfaces machined.
Ballscrew and servo motor driven
The bed is designed to be removable, in the photo below it has the vacuum table for wood hold down on it. This will be removed and a T slot table added along with coolant catching tray and side shields. The plan will be to start out with a small area (20"x40") of the actual table and build a partial enclosure around it in order to use coolant.

A couple of photos for reference and scale





So I have chosen to use Fadal parts because they are readily available and pretty inexpensive. I found a great deal on a 15hp spindle motor (way bigger than what I will ever be able to put to use but it was scrap price and came with pulleys, belt guide, and tensioner). I ran the motor in WYE configuration with a VFD I had seems to run very smooth. However it did pull nameplate current with no load so thinking I will have it checked at a motor shop. I will run this with a AMC vector drive and setup a WYE/Delta relay system. I purchased a used spindle which the seller says will need to be rebuilt at some point soon but my thought was it will get me going and I can rebuild it or have it rebuilt in the future.

Here's where my initial couple of questions come in.
Does anyone know the bore diameter in a Fadal Z axis casting for the spindle?
I have watched a few videos of people replacing the spindle and it seems to be a loose fit. Also the process for tramming the spindle is to add shims under the spindle bolt flange so assuming there is some clearance in the bore.

For the 2 speed system I assume the AMC drive does not "know" what speed range the pulleys are in. The control just asks for a motor speed (+ - 0-10v as well as the forward reverse inputs)and the AMC drive outputs to achieve that speed. Can anyone confirm this? I plan to set it up this way. I will likely have the speed range change done manually or make an additional Mach4 setup for the 2nd speed. Likely I will be running in low speed most of the time. Initially I may just manually tension the belt. I plan to make my motor mount setup slotted, since even if I want to use the idler pulley setup I may need some adjustment as I dont know Fadal spindle center line to motor center line. Should allow adjustment of the loose belt to allow proper slip.

Unless someone has the spindle center to motor center measurement?

Does anyone have any good thoughts on how to easily align these spindle with minimal setup adjustment time after swapping? I plan to rebuild the Z axis eventually to make things a little more robust (larger section). The current spindle has a 12mm slot milled on centerline to the spindle. So I could use something similar on the Fadal spindle mount bracket but wondering if there is something better.

Thoughts on drawbar knockout cylinder? I could buy the Fadal one but that is pretty pricey and machine scrappers seem to just throw that part away. I have a couple local scrappers that I have asked to keep an eye out for one. I will have the Fadal belt guide around the pulley so will have the same mounts. Other thoughts are welcome.

Thanks in advance for the help

 

[mmurray70]

Sounds like what you need is the whole head out of a scrapped fadal. Seems crazy to redesign and make all this from scratch when there is a bunch of these melted down for scrap everyday. If your trying to make it more compact or less weight start with a Fadal casting and just chop out what you need.

Your right about the spindle drive, it receives 0-10v and outputs motor speed accordingly without any idea of what range your in. You will need to have your control adjust signal accordingly based on what range your using.

If your not in too deep you should consider ditching mach for a more reliable controller like Dynomotion or Linuxcnc or maybe even centroid or something. Anything but machcnc...

 

[Vancbiker]

.......
Does anyone know the bore diameter in a Fadal Z axis casting for the spindle?
I have watched a few videos of people replacing the spindle and it seems to be a loose fit. ......

IIRC, the bore was not machined, just left as cast. Fadal only flange mounted the spindle and used a sort of floating clamp ring at the top of the spindle to provide some support. Most spindles had to be shimmed between the cartridge flange and headstock to get decent geometry.

 

[Scott.Damman]

I tend to agree with this statement I have had decent luck with short programs but anything longer than around 1 hour of continuous run time Mach tends to crash. The CNC controller is CSMIO IP/A they have their own control software that I should try just havent had the time to swap over.

As far as the head goes, I had the same thought about using the Z axis but I would like to swap the motor and spindle orientation keeping the spindle closer to the Y axis. I could possibly still accomplish this with the casting I just have not come across one.

Thanks for the drive info thats what I thought but helpful for conformation.

Ok so the bore was just big and open. I am surprised that it only needs support at the flange and that small bolted in ring at the top. But it does make it easier for me to machine not having to hold tight tolerances on that bore.

The spindle I was planning to use came in the mail today (was stuck at a reginal facility for over a month, seller shipped USPS...). It is missing major components specifically the lower bearings, spacer between the lower inner races, and the upper spacer that the threaded split ring sets bearing preload. Could have these made but I would need some dimensions and it sounds as though Fadal changed the spindle design alot and I dont know the year of this spindle. Seller is giving me partial refund and I can at least use it to take some detailed measurements off of for this build. But back to searching for a used spindle.

 

[Comatose]

I don't think you're going to be happy trying to mount a Fadal head on a CNC router. Just the head and the counterweight for it weighs more than your entire turret. There's a huge difference in torque between 10hp at 24,000 rpm and 15hp at 3000 rpm. To get the same deflection your structure would need to be 12x stiffer.

It's not actually impossible to machine steel with a router spindle. You'd need a small AlTiN hot mill type endmill, run it at about 8000rpm, use ONLY high speed toolpaths and keep your width of cut down to maybe .02" per pass. If you don't have the rigidity to do that without chattering to death, you definitely don't have the rigidity to run larger.

If you're wedded to the idea of a proper metal cutting head, I'd start by finding an old Bridgeport J or M head. You already have a Z axis, so just fix the quill in place and go to town. Or, fix the quill in the up position when using the router spindle and in the down position when using the metal cutting head. If it doesn't work, all you're out is whatever mount you fabbed up to attach it.

 

[mmurray70]

I tend to agree with this statement I have had decent luck with short programs but anything longer than around 1 hour of continuous run time Mach tends to crash. The CNC controller is CSMIO IP/A they have their own control software that I should try just havent had the time to swap over.

Can you do a proper feed hold with Mach now? My understanding was mach didnt work well without a second or 2 buffer, and feed hold wouldnt start until this buffer was ran through.

That seems crazy to me. If I had to wait 2 seconds for feed hold I would have crashed dozens and dozens of times over my career, destroying thousands of dollars worth of gear. I cant imagine running like that. Would never consider Mach just for that reason alone. Alright on a little benchtop mill I suppose, but not safe for a good size machine.

 

[Hardplates]

IIRC, the bore was not machined, just left as cast. Fadal only flange mounted the spindle and used a sort of floating clamp ring at the top of the spindle to provide some support. Most spindles had to be shimmed between the cartridge flange and headstock to get decent geometry.

Oh my, that might be worse than lapping the ways with sandpaper.

I don't think you're going to be happy trying to mount a Fadal head on a CNC router. Just the head and the counterweight for it weighs more than your entire turret. There's a huge difference in torque between 10hp at 24,000 rpm and 15hp at 3000 rpm. To get the same deflection your structure would need to be 12x stiffer.

It's not actually impossible to machine steel with a router spindle. You'd need a small AlTiN hot mill type endmill, run it at about 8000rpm, use ONLY high speed toolpaths and keep your width of cut down to maybe .02" per pass. If you don't have the rigidity to do that without chattering to death, you definitely don't have the rigidity to run larger.

I totally agree with this. I cut steel with a 3/8" AlCrN tool at 10,000 to 12,000 RPM all the time.

 

[Scott.Damman]

I don't think you're going to be happy trying to mount a Fadal head on a CNC router. Just the head and the counterweight for it weighs more than your entire turret. There's a huge difference in torque between 10hp at 24,000 rpm and 15hp at 3000 rpm. To get the same deflection your structure would need to be 12x stiffer.

It's not actually impossible to machine steel with a router spindle. You'd need a small AlTiN hot mill type endmill, run it at about 8000rpm, use ONLY high speed toolpaths and keep your width of cut down to maybe .02" per pass. If you don't have the rigidity to do that without chattering to death, you definitely don't have the rigidity to run larger.

I have not tried a AlTiN endmill, maybe I give that try before moving further. What I wasnt sure on was if the bearings in the router spindle were up to the task. Its a giordano colombo spindle but the bearings are not at all as large compared to a 40 taper spindle.

I have successfully cut steel with it with acceptable results and what I would say minimal chatter. But the tool did not last long, granted it was an uncoated carbide tool I had on hand running around 10k.

Here is a video it is taking very small depth since it is taking full width. Trying to baby the spindle.

Here is a photo of the face milling, this was with a 2 inch insert tool. I dont recall the speed/feed. Since this picture I have properly trammed the spindle as well it was slightly out.


I know my machine is not as beefy as a Haas GR510 but that is what I used to design it around. The structure is all 10x10x3/8" box with 3/4" plate welded where critical parts needed to be machined. The gantry legs are made from 1/2" inch laser cut plate with 1.5 inch bar for the bearing block locations. The gantry is 12x12x1/2" steel tube with 3/4" plate added for linear rails. It was designed around what material I could get for cheap, all the box tubing sections came from scrap yard, there is a crane manufacturer near me that these are 12+ foot drops from their operation. I was hoping I could get decent results from this with small cuts and high speed tool paths. I would like to better understand the 12x stiffer if you can direct me. I calculated loads on the tool based on what I could find online and ran FEA on the structure to understand deflection when cutting steel. Tried to design around what I thought was adequate deflection for the results I was after...



The Z axis I need to redesign to give it more section this is my 2nd machine and I pulled the Z off the old machine to get me running. But the Gantry and main frame were designed to cut steel.

Mach 4 does a feed hold but since I only know this machine (I have no other CNC experience) I do not know what a proper hold is. It does hold as soon as I hit it but I dont really use that function (likely because I havent understood the use for it)

I appreciate the help and comments, a big part of the fun for me is learning as I go!

 

[Scott.Damman]

I'd start by finding an old Bridgeport J or M head.

This really has me thinking, could be a pretty inexpensive way to see if the machine was even capable. Going to have to do some looking for something of this sort.

Thanks for the thought

 

[Scott.Damman]

I have been learning alot of the last few days.


- It has become very obvious that I have no idea what I am doing when it comes to steel...
- The AlTiN coated tools are amazing thank you for the recommendation

I don't think you're going to be happy trying to mount a Fadal head on a CNC router.

This statement seems more true now that I have tried a little more.
The machine did cut steel after some attempts. I did find some issues with the machine that were not at all evident with wood/plastic. First my Z has 0.0009" backlash, thinking the ballscrew might need attention. The software can take that up but it was very evident in the cut. The cuts definitely chattered but gave ok (just barely ok) surface finish. My servos in X (my long axis) were overshooting around 0.0005 which again was not at all noticeable in wood but showed up in metal. I should be able to correct this with better servo tuning. The machine showed around 1.5 thou flex when cutting steel. In wood locally I can easily hold 1 thou with this (I say locally because its got rolled ballscrews on X so they are only so good). This has always been way better than I ever needed in wood. In the steel each feature measured big by around 0.003" which would tell me the system is flexing some.

Here is a video of it "cutting" 15,000rpm 0.25"DOC 0.040" WOC 1/4" ALTIN coated endmill

With that said its now obvious to me that I was over confident and this machine is not going to be acceptable in steel. It is still very capable of everything I want it to do in wood/plastic but not metal. So thank you for the comments and advice.

So now I am looking for a VMC to get started with. Initial thought is Fadal but interested to here other peoples opinions for a starter machine.

Thanks

 

[Comatose]

Honestly, it doesn't sound/look so bad that I'd necessarily give up and buy a VMC yet if you steel cutting needs are quite limited. You may find some similar things in metal with a proper VMC:

First, deflection matters. Wood's really soft and doesn't take that much oomph to cut. Steel, not as much. So if you've selected a roughing operation (Adaptive clearing, trochoidal, whatever) and you're cutting big, you may not be dead in the water.

First, are you cutting climb or conventional? Climb pushes the endmill away from the cut, conventional pulls into the cut. Typically, even with a pretty nice VMC, you'll hog out almost everything leaving .005-.01 inches of stock, then come back around again with a contour operation. That'll clean it up a fair amount.

Another thing to remember is all the HSM paths are rouging paths, so they don't typically leave perfect geometry. They'll be scalloped where each pass meets the boundary layer.

Second, not every .250 tool ends up exactly .250. It's not unusual if you're going for something with tighter tolerance to have to take a test cut and then program the compensation for the .248 or .249 diameter that your tool is actually cutting. Then if you're running a series of them, periodically tweak the comp value as the tool wears.

Hog out your part with .01 of stock left, then take a finish pass, and see where you end up.

A Fadal isn't a bad choice for a first machine. There's not a lot to them and aftermarket parts support is good. They're not particularly fast and they're not particularly accurate, so don't expect to be making molds with them, but for anything you might have been imagining bashing out with a router, they'll be perfectly fine.

But, again, if this is for personal or infrequent use, you might be fine with the router. (I have one CNC router and... more than one CNC mill.)

 

[Scott.Damman]

Honestly, it doesn't sound/look so bad that I'd necessarily give up and buy a VMC yet if you steel cutting needs are quite limited. You may find some similar things in metal with a proper VMC:

This is very helpful, maybe I am throwing in the towel to early.

This was climb milling. Here is another video of it pocketing

Here are some pictures of the surface and wall finish, would like to hear comments on this.




These 2 pockets were accidentally 0.1 WOC it amazingly did it but the deflection was worse. I didnt have a video of it but sounded like chatter when it was cutting.


The outside of this part was cut with a contour pass after the roughing with the adaptive cuts. The walls dont look too bad but the surface has alot of marking that I can feel with my fingernail. Maybe due to the lash in the Z screw? Also this is a high pitch screw 20mm per rev so it might be harder for the servo to hold in steel. Thinking when I redo the Z to go with a much smaller pitch to get the servo more mechanical advantage.



Thanks again for the comments helps me to learn...

 

[mmurray70]

That doesnt seem too bad to me for a router! You need to do another finish pass on the faces after roughing. Try taking another 3-5 thou off with maybe 75% stepover to finish the bottom, maybe slower rpm to save the tool. I would also add more finish passes to the walls too, looks like you have some marks left over from roughing. If most your parts can be cut with small endmills you should be able to limp along with that machine. As soon as you need to drill some big holes, or use long tools you may find it challenging.

Endmills vary quite a bit too. Sometimes the cutting edge is honed very slightly for more durability. Tools designed specifically for stainless and superalloys will have sharper edges and may work better on your machine due to lower cutting forces.

 

[Scott.Damman]

Have been learning a lot more over the weekend here.

Started on part 2 a mirror of part one. Much better results as I get things dialed in. Still some z stuff going on, thinking that might be ballscrew related. Overall I am pretty happy with the results running the contours conventional rather than climb really helped. The 2nd part is within 1.5thou on all dimensions. (With a calipers so...) Let me know your thoughts.

What’s a good helix ramp angle? I have tried 2-7 on different operations and am always getting red hot chips off. I am running 35ipm for the helix and 70 for the adaptive cuts.

https://youtu.be/mprAvDQvp4I


Sent from my iPhone using Tapatalk

 

[Comatose]

If you have a tool changer, drilling your starting positions will probably give you better results than helixing them.

There's definitely some slop in Z and the surface finish isn't exactly mold grade, but for what it is that doesn't seem like a terrible result. Are you also doing a floor finish pass? Do you have a lookahead available in the control? It looks like you're doing an exact stop at the end of each arc and line segment, which is fine for wood but it leaves those dwell marks in metal. It might be something like g60/g64, but then again it might be something else in your control.

You're never going to get the kind of finishes off a router that you have in a mill, they're just inherently a LOT less rigid, but depending on what you're doing, it might be good enough. For weldments you're probably fine now. For stamping dies you'll never get there from here. Horses for courses.

I will say, though, none of what you're running into is coming from the spindle, so none of that would be improved by mounting a heavier head onto the same gantry.

 

[mhajicek]

This should go without saying, but since you're used to working in wood it may have escaped you; always use the shortest cutter and holder you can for each cut. Deflection goes up with the third power of length, so if you can use a cutter half as long it will act 8 times as rigid. When cutting deep pockets it's not uncommon to use a shorter cutter partway down and then switch to a longer one, running slower, for the bottom.