[GLCarlson]

"Can you post a picture show in your setup? If if I understood correctly you are saying that your router spindle is coaxial with the main spindle? How do you accomplish this?"

Here's a brief overview and a couple pictures. This design was a response to frustration with the common design which loses X or Y travel because the auxillary spindle is mounted to one side of the main spindle.

Mount: Fortal, about 0.8" thick, triangle dimensions are about 5" on a side. Top and bottom machined together in one set-up. Two rough sawed pieces of rough stock were faced, joined with cyanoacrylate, then drilled and tapped for bolts. After bolting together, the spindle nose and router body mounting holes were bored. Side clamping holes drilled/tapped, and finally the two pieces of each triangular mount were cut and then cleaned up. 1" 8020 was cut/faced to length and tapped 1/4-20. Total vertical dimension is 9", when mounted adds about 7.5" to spindle length- that is, Z travel is now around 10". Concerns for the mount were making sure the mounting holes were concentric and square. Obviously the top hole fits the Tormach spindle nose (3.375"), and the bottom hole is sized for the router body (2.75" or so). Have your router body in hand before starting.

Router: Dewalt DWP611, variable speed, bought a spare body on E-bay for 70 bucks. Collets and collet nut from PreciseBits.com. I bought a high precision collet and 1/8 and 1/4 high precision collets as a kit. About 75 bucks. See that website for a lot of detail about which router bodies work and a range of precision collets and cutters- do not use the collet/nut that comes with the router. Spindle runout on both the Dewalt and an earlier version (using a Bosch) was unmeasurable with a Last Word (<0.0002 or so), which amazed me but seemed typical for mill-run router bodies these days. The DWP611 is variable speed, 15-27K. Other router bodies vary, but 10-30K is common. My 1100 tops out at 5800rpm, so I've got a hole in rpm coverage, but that has been of no practical concern. DWP611 claims 1.25 hp, speed adjusts with what must be a triac or similar control, don't know whether it's variable voltage or PWM. If I had infinite time, I'd build a PWM controller and vary speed with a digital input, but in practice I just set an approximate speed and run, and that works. Power on/off is just the router switch.

Mounting on 1100: as seen, just a simple clamp to the spindle nose. Unfortunately, it doesn't play well with the spindle nose LED; one or the other but not both. I generally make sure I've got a bit of clearance under the 3/4" collet to avoid accidental main spindle damage. The LED lighting in the router body is minimal (that is, inadequate).

Same approach should work for a laser cutting head or a 3D printer head. In the latter case, the decreased Z axis might become an issue.

Happy to answer further questions.

[jcposada]

Thanks for posting these. Very clever!

[Stigoe]

If you want better variable speed, you could use a Super-PID.

[polar8]

Very interesting. Thanks for posting pics. Any thoughts on using a spindle motor like this instead of a router?

http://www.ebay.com/itm/New-CNC-80MM-24000rpm-220V-1-5KW-Water-Cooled-Spindle-Motor-ER11-CNC-Router-US-/272225937363?hash=item3f61ee27d3:gMQAAOSwCypWnOUF

[Zetopan]

"If if [sic] I understood correctly you are saying that your router spindle is coaxial with the main spindle? How do you accomplish this?""

Note that there are other solutions that are inherently concentric with an R8 taper spindle. I got this BIG Daishowa Seiki UG-125B Universal Air Grinder in mint condition for a low price from eBay. A precision dovetail with adjustable offset allows for grinding internal or external radiuses, and when it is set to zero it is terrific for engraving with an appropriate cutter. There are fairly often other less elaborate air driven turbines or electric motor spindles on eBay, just keep a look out.

[GLCarlson]

Replies to various posts.

Looked at the water cooled spindle motor. Surely, it would work. Reasons I didn't choose that: the Dewalt (and other) router bodies had both power switch and variable speed already in the body, power was more than adequate (roughly 1kw vs "more than that" is irrelevant for small cutters- a horse should be plenty), and 50% higher price. I also had some -minor- concern that the water cooled spindles looked as if they needed more than a simple clamp mount to incorporate power switch and speed control, where the router body was obviously an easy clamp and go. All that said, if I'd had one of the water cooled spindles sitting around, I'd'a used it.

Air turbines (and implicitly, an R8 mount). Definitely an option. I didn't have an air turbine in the scrapbox or anything else that was already R8. I will note that using the R8 taper for mounting implies that the spindle is locked somehow, and the spindle drive disabled. Mounting on the spindle nose, not through the spindle, eliminates a potential accident-in-waiting and is the specific reason I didn't make an R8 mount for the Dewalt. The other reason was that by using the spindle nose, I don't have to demount my TTS collet. Swapping out the 3/4" TTS collet for another R8 collet or mount is a bit of a pain with a PDB installed. Clamping onto the spindle nose is simply easier.

PID for speed control. The router speed control works. I'm sure it's not the best answer, but it's an adequate answer.

My solution is one of a set of possibles. It's cheap, easy, and -so far- works very well. I hope I've encouraged others to consider keeping the X/Y envelope intact using the Z axis as a trade-off. I've found that to be a better choice than losing XY work envelope (my first attempt at this was a Tormach/Kress style side mount. Drove me mad.)

[luv2ride]

I'm running a Proxxon with side mount as I type and yes, it will drive you mad at times. Like earlier I had my part zeroed and moved the proxxon over to verify.... 0.1 shy of being able to clear my part... reclamp part, rezero and away we go. I really like the idea of having something inline with the spindle.

ETA...
The part I'm running now wouldn't benefit much though as I am running a 30 deg engraver, .040 Ball mill and a .375 ball mill. I would have to dismount to run the .375.

[Zetopan]

I will note that using the R8 taper for mounting implies that the spindle is locked somehow, and the spindle drive disabled.

A locked spindle is unnecessary and actually undesirable for this air turbine. The compressed air feed tube section is actually fully capable of rotating around the rest of the body. Stated another way, the mill spindle can rotate just fine while the air feed tube remains stationary. Since the lower part of the air turbine is connected to the upper part through a dovetail, the lower part can be concentric or not, relative to the mills R8 taper spindle. In the case when the lower part is not concentric the tiny collet in the lower part will move in a circle with the radius equal to the offset between the lower and upper parts.

In other words, no circular interpolation is required for moving the tiny collet in a circle, and it does that movement at the mill spindle speed. Hence the mill spindle is usually run at a low speed and that defines the feed rate along the circumference of the described circle.

When engraving, the mill spindle is generally off and the offset is adjusted to zero. When grinding or other specialized milling the mill spindle is generally running at a low speed. It is overall quite a slick arrangement for achieving a very high tiny spindle speed for grinding precise IDs and ODs without requiring any circular interpolation. A side effect is that is also works very well for engraving and can be adjusted for concentricity with the R8 spindle. Although I didn't say so before, the unit is quite heavy for its size - I would guess that it weighs around 25 pounds. The original wooden box with all accessories (collets, wrenches and grind stones) weighed around 40 pounds.

[GLCarlson]

A locked spindle is unnecessary and actually undesirable for this air turbine. The compressed air feed tube section is actually fully capable of rotating around the rest of the body.

Aha. Sounds like a very neat attachment. Not worrying about wrapping a power (air, in this case) cord around it is very handy indeed. I would think inherently more accurate than circular interpolation when used as described. Or, at least, it ought to be.

[GLCarlson]

I...
The part I'm running now wouldn't benefit much though as I am running a 30 deg engraver, .040 Ball mill and a .375 ball mill. I would have to dismount to run the .375.

That is indeed the biggest problem I've seen with mine- the chuck capacity is 1/4", and the number of precision collets is limited. Demount is easy, but there's still the registration issue between the two spindles. I haven't checked, but it's not likely to be small enough for precision in the tenths. Repeatability in that range is another issue.

There ARE router bodies that can handle up to 1/2" tooling -Milwaukee makes one, for example and sells quarter and half inch maybe fairly decent collets with it. However, I don't recall seeing precision collets like those from Precise for the big spindle bore. IF the Milwaukee router was as well made as the others (that is, similar runout despite the bigger chuck), then the only issue remaining would be the availability of a range of precision collets. I might be tempted to do something like chucking up an ER20 collet with a half inch shaft permanently on the router base. That'd cost another couple inches of Z travel, but if one could get around the runout, it might be a way to increase the tool range beyond 1/4". Those are available, 20 bucks on Ebay. And I've got one of those routers- I'lll set it up and report back on runout.

Obvious downside is that it piles kludge upon kludge. Errors multiply.

Added: My Milwaukee router, using the original Milwaukee collets (which appear similar to an ER style- but not ER, sadly; shorter) has a total run-out of 1.5-2 thousandths. That appears to be in the spindle; the spindle cone measured 1.5-2 at several points when rotated by hand. Adding the Milwaukee collet holding a carbide rod of known roundness didn't change the reading. The router is several years old, on it's second set of bearings, and has been heavily used in the wood shop. It would be a bit more of a problem to mount (to avoid the acme screw height adjustment) but the same basic approach would work. No variable speed on that model.

I conclude that an ER20 collet chuck could be made to work, with reasonable runout and a corresponding increase in tool capability, buy choosing the right router body and (perhaps) tweaking the collet. The Milwaukee router collet chuck is clearly better than the collets supplied with other routers.

[MFchief]

This is a topic the interests me a lot. I have made one of the aux spindle types and it works, but co-axial is much better.

One of the options I have thought about is very similar to the OP's setup but using one of the new style RC brushless motors. I think if would have a smaller form factor and some of the larger ones would certainly have the power for engraving. I haven't persued it mainly because of the bewildering variety of motors, and it seems most of the more powerful ones are "outrunners" with the outside part of the motor rotating.

Maybe the are some RC enthusiast here that could recommend some motor/ESC setups.

Terry

[Zetopan]

Maybe the are some RC enthusiast here that could recommend some motor/ESC setups.

It is very easy to find a 3 phase "Outrunner" and ESC that is near or even greater than 1HP at upwards of 30K to 100K+ RPM. The likely problem is coupling that to a spindle or some type of collet holder while keeping the overall length low.

[MFchief]

What I had envisioned was some type of spindle clamping method, like the OPs, with the motor bolted to the bottom of that. I have very limited experience with these types of motors, but from what I've seen it's not straight forward to bolt an outrunner from the end plate, like the old style brushed motors could do. I guess also, I'm not thrilled about the exposed portion of the motor rotating at those speed around head high. LoL

Terry

[cismontguy]

Thanks for your post. I really appreciate the effort you have made. A spindle router is the next this I am going to be working on for my PCNC 1100.

Cismontguy