[louieatienza]

I've been meaning to document this a bit more, but have been busy lately and just finding the time to work on this let alone document the build has been tough. Anyways...

Lately my machine time has consisted mainly of cutting parts out of Garolite for a local speed shop; i.e. intake gaskets, throttle body spacers, etc. These jobs require a few hours of machine time, and I have some dedicated fixtures for the repeat jobs. I also process aluminum plate for another client, which also requires a lot of machine time and dedicated fixturing. This makes it tough for me to do some of the smaller work and prototyping that I enjoy doing. I also want to be able to take some of my work "on the road," like on-site engraving for local sporting events, maybe even some craft fairs. Plus I still need to fabricate some small parts for my aluminum mill/router and get that done as well.

I was able to obtain a whole slew of THK rails from a fellow guitar forum member during a fundraising auction a while back. All SHS15s, but all of them had 400mm rails. I was able to find some 508mm and 200mm rails pretty inexpensively. Since the rails I received had only one block each, I cherrypicked the best rails and paired the blocks; now I have a bunch of 400mm rails that I need to rid of.

For mounting the rails, I used Misumi extrusions, that have factory machined-flat surfaces. The surfaces are really nicely machined and very carefully wrapped. I used my favorite material, phenolic, for the machine frame and other parts. I also used 1/2"-8, 4 start ACME screws from McMaster Carr and dumpsterCNC nuts and threaded collars. I used 1/2" x 1-1/8" x 5/16" plain radial bearings with a very light preload to support the screws. Since I don't expect a lot of cutting forces (and certainly less than the 25lb rating of the anti-backlash nuts) I think I'll be fine with them. The bearings, like in my current machine, are integrated into the frame of the machine.

For electronics I will be using a G540 box with 24V psu, and SmoothStepper. Motors will be Danaher 200in-oz NEMA23s.

[louieatienza]

Some pics I do have:

Pic 1: Parts from Misumi... Real nice work, wrapped excellently. THK rails in background.

Pic 2: A pic of the assembled base frame, with the gantry on top. I have to make brackets to mount the bantry to the uprights. Once I have everything tweaked I may make a spacer to fit between the two extrusions but my preliminary hunch says it might not be needed for what this machine is intended for.

Pic 3: A view of the business end of the leadscrew. Ruland spider coupling, 1/8" brass spacer, radial bearing. The other side of the gantry has a bearing, 1/8" spacer, and dumpsterCNC threaded collar that will be used to put a very light preload on the bearings.

Pic 4: Closeup of how the base and extrusion go together.

Pic 5: Side view. Overall footpront (minus stepper in the rear) will be about 19" x 19". This is a moving table fixed ganry design, and the work envelopw will be about 14" x 9" x 4". I should have 4" of Z travel with the table installed.

Pic 6: Closeup of Misumi extrusions, machined face and mounted THK rails.

Pic 7: A 3/4 view of the machine. This will mainly be in my office. I have an enclosure also designed but not yet made.

[DonFrambach]

Looks real nice! I think you'll enjoy having a small router than can do long runs making small detailed pieces without encroaching on the other work you do on your larger machines. Portability could be a plus too!

[louieatienza]

Thanks Don! I actually might even have a gig lined up already, and it's not done, so I got to step it up (no pun intended!)

I completed the stepper mounting brackets for the X and Y axis, as well as the leadnut blocks. Using phenolic for the leadnut blocks.

I had to open the holes on the leadnuts to allow me to use M4 bolts. The blocks are threaded.

The stepper mounts are similar to the ones on my current machine. Basically the four standoffs are 1-3/4" long threaded rod couplers. I used 1/2" socket cap screws to both bolt the couplers to the frame and the stepper plate. The stepper plate is 1/4" MIC6 cast aluminum plate. The steppers are also mounted to the plate with M4 socket head screws. These were a lot easier to make then to try and cut four pieces of rod exaxtly the same. Also, not having a bolt going through the rod means the stepper copuld be removed from the machine without unbolting it from the plate.

[roy_okc]

Louie,

Nice looking build. I'll definitely be watching your progress. What's your expected capacity?

Interesting use of your large machine (gaskets/spacers). Googling "garolite" has your Youtube video cutting it on the first page.

After a year off/on with mine, I can see how a smallish one could be handy for a lot of uses. If my middie ever gets busy, I'd probably build a small one of similar size.

Roy

[louieatienza]

I went to look for the wiring diagram for my steppers and they are gone. I have Danaher (Kollmorgen) CTP 2-phase steppers CTP21NLE23SAA11. While I have their latest literature, it seems their wiring color scheme was either changed, or this may be custom wiring?

Anyway, the wire colors I have do not match any I've seen in the Kollmoregen brochure. Using an Ohmmeter, I was able to determine which leads are paired: black/green, red/yellow, brown/grey, orange/white. The question is, how do I determine which direction each pair goes, and how do I know which pairs go to which phase?

I was thinking initially of wiring these motors bipolar parallel, but at 2.3A I believe I need 4.6A if I wire this bipolar parallel? It might not be worth it since I'm using a g540?

Thanks in advance...

[louieatienza]

Louie,

Nice looking build. I'll definitely be watching your progress. What's your expected capacity?

Interesting use of your large machine (gaskets/spacers). Googling "garolite" has your Youtube video cutting it on the first page.

After a year off/on with mine, I can see how a smallish one could be handy for a lot of uses. If my middie ever gets busy, I'd probably build a small one of similar size.

Roy

Thanks Roy. The Garolite (and most phenolics) can be tough to cut, and I use special tooling to lessen the cutting force and noise. Even then, the sound and smell are rough (even with dust collection) and I keep the basement closed and wait an hour or two for any excess dust to settle before handling the stuff, and even then with a mask or respirator. The resins can be a carcinogen, and there is always an issue with fiberglass dust so I take extra care. Which is another reason for the smaller machine. I can have it in the corner of my office, away from the dungeon as well.

As for work envelope, It's about 14.5 x 8-9 x 4.5-5.5, I don't plan to cut anything over 1-1/2" but I plan to mount a mill vise or drill press vise when needed. While these steppers have some pretty impressive specs, I don't know how they'll preform with the G540 and SmoothStepper. Apparantly these steppers can work with pretty large supply voltage and have a very high max operating temperature, though I don't plan on pushing it that far.

[jasoneule]

is that 1 inch phenolic? Do you source that locally or online? If you to not mind, what is the cost if that phenolic. I think you are on to something with that. Its an excellent material.

[louieatienza]

Jason,

Both. I have a local distributor that sells Arborite (paper based) for about $580 for 1" x 4 x 8. You can buy remnants on eBay for a good price.

On this machine, I used 7/8" Tufnol CE which is a canvas-based phenolic. This would be the third machine I made using phenolic. The extra mass and stiffness seemed to really help. On this machine I've made a point to isolate metal components with the stuff in the design to help minimize the transfer of vibrations. I'm actually surprised that I have not seen any other builds using it, though many commercial CNC routers and machining centers have it as an option for the table surface. But it does surprise me that with all the interest in epoxy granite, carbon fiber (which is also a type of phenolic), and such, that I havevn't seen more people using this stuff. It's a lot cheaper than aluminum, and can be as strong. The paper and cloth based stuff can be cut relatively easily with regular carbide bits. It can also be tapped for bolts that require little removal. You can even use sheetrock screws by making a slightly smaller pilot. Comes in mny colors, wood grains, and patterns if you desire as well (the paper based stuff of course)!

Them stuff nowadays is used widely for bathroom partitions, as it has some moisture resistance, and is very strong. The cloth based is used for everything, including gears, since it has a little "give" the gears can be made to mesh tigher against itself or metal gears.

The Garolite G10/FR4 grade is fire resistant and fiberglass-based, which I cut the car parts out of. They even have a graphite impregnated grade for making bearings and bushings. The earliest Fender Broadcaster and Telecaster guitars used Bakelite for the pickguards and such, which is really a paper-based phenolic.

[ger21]

The question is, how do I determine which direction each pair goes, and how do I know which pairs go to which phase?

Trial and error? Just pair them up and measure the resistance of the paired coils until they match. Direction isn't critical, as you can just reverse one pair to change motor direction.

I was thinking initially of wiring these motors bipolar parallel, but at 2.3A I believe I need 4.6A if I wire this bipolar parallel?

Wired parallel, with the G540, you'll lose 25% of your holding torque, but you may end up with more torque at higher speeds.

I have a local distributor that sells Arborite (paper based) for about $580 for 1" x 4 x 8

Wilsonart makes a similar product. You could probably special order it from Home Depot, if they have a knowledgable person ordering their laminates. It would probably be a lot cheaper if you could have a cabinet shop order it for you, though.
Wilsonart Contract - Wilsonart® Compact Laminate

There's a similar product called Trespa. Trespa® Virtuon®. | Trespa
I've used some small pieces left over from a job on my new router.

I'm actually surprised that I have not seen any other builds using it,

I think that's due to the price, and the difficulty in obtaining it. Most people have probably never seen large sizes or quantities of phenolic sheet.

About 2-3 years ago, someone local to me sent me a PM offering me all the 1" phenolic I wanted. He worked at a shop building auto show displays, and literally had tons of scraps, in large sizes. He had built a 4x4 or 4x8 sized machine entirely from the 1" phenolic.
Unfortunately, I had to pass on the offer, as I had no place to store it. It's hard to turn down an offer for thousands of dollars in free material.

[louieatienza]

Trial and error? Just pair them up and measure the resistance of the paired coils until they match. Direction isn't critical, as you can just reverse one pair to change motor direction.

Wired parallel, with the G540, you'll lose 25% of your holding torque, but you may end up with more torque at higher speeds.

Thanks for that. I was hoping that there was a more "scientific" method. I believe each pair I have now reads about 2-Ohms. So you're saying I should pair one set to each remaining set, and measue the resistance of each pair, then do the same with the others, to find which pairs have equal resistance?

Wilsonart makes a similar product. You could probably special order it from Home Depot, if they have a knowledgable person ordering their laminates. It would probably be a lot cheaper if you could have a cabinet shop order it for you, though.
Wilsonart Contract - Wilsonart® Compact Laminate

There's a similar product called Trespa. Trespa® Virtuon®. | Trespa
I've used some small pieces left over from a job on my new router.

I've used the Wilsonart stuff before for a couple commercial bathroom partition projects. Never used Trespa before.

I think that's due to the price, and the difficulty in obtaining it. Most people have probably never seen large sizes or quantities of phenolic sheet.

About 2-3 years ago, someone local to me sent me a PM offering me all the 1" phenolic I wanted. He worked at a shop building auto show displays, and literally had tons of scraps, in large sizes. He had built a 4x4 or 4x8 sized machine entirely from the 1" phenolic.
Unfortunately, I had to pass on the offer, as I had no place to store it. It's hard to turn down an offer for thousands of dollars in free material.

Wish I was there for that! I use it a lot for jigs and such in my shop. While large sizes can cost some, there are a lot of off-cuts and such on eBay, and they in my opinion would be far superior to some other wood-based materials for things like router mounts, carriage plates, etc. A 3/4" bahroom partition door pobably runs about $200. But that same piece of aluminum would cost about three to four times as much. For a first machine by hand it might be tough to use, but I think for a second machine where you already have CNC capability the stuff's hard to beat (literally too!).

[louieatienza]

As to the bipolar parallel witing, I was hoping for more top end torque since I am using the two-start screws. Then again, my goal is for at least 150-175ipm rapids and 75-100ipm max feedrate, so maybe bipolar series would suffice?

[ger21]

With 1/2-8 2 start, 250oz motors wired bipolar series at 2.3amps, and a Xylotex @ 24V, I have my Y axis with 30" of travel set at 190ipm, and I frequently do 3D carving with it set at 175ipm, so it should be fine with a G540.

So you're saying I should pair one set to each remaining set, and measue the resistance of each pair, then do the same with the others, to find which pairs have equal resistance?

Yes, pair them up as if you were wiring it to the drive, and measure the two pairs that would connect to the drive.

[CarveOne]

The LinuxCNC website has a detailed procedure that should help you figure it out when you have no wiring instructions.

One quick way to find out if two wires go to the same coil is to short the leads together and rotate the shaft. If there is no difference in how hard the shaft is to turn by hand, the wires are not connected to the same coil. If the shaft becomes noticeably harder to turn, the wires are connected to the same coil.

In the case of a center tapped coil, there are three related wires per coil. You can tell which wire is the center tap using an ohmmeter on the Rx1 scale

[jasoneule]

I am sold on this phenolic....so many applications. I remember when I go some for a table saw insert. I tried drilling in it and basically made a mess. Then I got a special bit nd counter sink and it cut great.

so what bits and speed are you using to cut this stuff.

I am actually going to be doing a small robust mill or 3d printer next after I wrap up my build.....definitely going to use phenolic.

[louieatienza]

I am sold on this phenolic....so many applications. I remember when I go some for a table saw insert. I tried drilling in it and basically made a mess. Then I got a special bit nd counter sink and it cut great.

so what bits and speed are you using to cut this stuff.

I am actually going to be doing a small robust mill or 3d printer next after I wrap up my build.....definitely going to use phenolic.

You can use carbide and carbide tipped tools for the paper based stuff. For the fiberglass-based phenolic I use PCD tooling, which is very expensive unless you hunt eBay.

I was cutting at over 50ipm, but since my machine is not as ridgid as a commercial machine I start to get chatter, and it can cause the edge of PCD bits to shatter. You also have to plan your toolpaths so that you leave no slugs that can jamb the bit in the kerf, since the bit won't cut that slug easily, and you'll snap a bit. I dropped down to about 36ipm with a 2-flute bit and about 10k-12krpm. I also leave about .02" for a finish pass at full depth, and increase the spindle speed slightly as well.

[louieatienza]

The LinuxCNC website has a detailed procedure that should help you figure it out when you have no wiring instructions.

One quick way to find out if two wires go to the same coil is to short the leads together and rotate the shaft. If there is no difference in how hard the shaft is to turn by hand, the wires are not connected to the same coil. If the shaft becomes noticeably harder to turn, the wires are connected to the same coil.

In the case of a center tapped coil, there are three related wires per coil. You can tell which wire is the center tap using an ohmmeter on the Rx1 scale

Thanks for that link, that made it easier to figure out.

[louieatienza]

I finally figured out the phases for the Danaher motors and hooked them up with the Soigeneris stepper covers and XLR connectors. It was a bit of a pain, as Danaher uses a 4-40 thread on the hole and the Soigeneris kit comes with M3 screws. I simply chicked an M3 tap on the drill and re-tapped the holes very gently. As these screws are not carrying weight I think they'll be fine, especially with a little LockTite. Also the solder lugs on the XLR connectors leave little room with the shaft shield inside the case. Drilling the holes for the screws, the part I actually worried a little about, went very easily with the drill press, by peck drilling to clear the bit every 1/4" or so.

I think they do make the wiring look more pro...

I also ended up wiring these bipolar parallel. Although I won't get the full torque potential, I think it will give me more headroom and potential upgrade in the future should I wish to upgrade the drive.

I included a picture of the diagram of the wires for these steppers in case someone else may need them.

Thanks guys for your help with the wiring!

[louieatienza]

I got the Y-Z saddle made up and installed. It's actually three parts, two 1/4" aluminum flanges and one 1/2" G10 plate. The Y axis bearing blocks bolt onto the 1/4" flanges, and the Z axis bearing bloxks bolt onto the G10 plate, and then the plate gets bolted onto the aluminum flanges. I did not have and SHS15C blocks, which would have eliminated one step here nad made things easier. The G10 I called "Swiss Cheese" do to the 48 holes on the plate; 16 to mount the G10 to the flanges, 16 holes to mount the Z blocks, and 16 clearance holes for the Y block bolts; that was done to eliminate counterboring the relatively thin aluminum.

The dumpsterCNC AB nuts were modified slightly by opning the bolt holes to allow M4 bolts to pass through; the nut is bolted to a nut carrier I made of phenolic, that fits in a recess on the G10 plate. The slightly elongated holes on the aluminum allow for small adjustments to be made, though it did line up pretty square when test-fit.

I also made and installed the connector plates for the gantry assembly. It's just two 1/4" plates with countersinks for two M8 flat head socket screws, that bolt into tapped holes in the uprights.

I have the table plate and other parts made but will post pics later on.

[louieatienza]

Starting to come together; did some assembly and everything fits pretty well. This is a relief, as everything was jsut modeled in 2D.

Pic 1: The machine so far...

Pic 2: Table mounted, a chunk of 7/8" Tufnol

Pic 3: The router clamps are 1" phenolic. The cuts were done on the tablesaw.

Pic 4: The back of the carriage.

Pic 5: A view of the carriage dropped down. I will have a spoilboard and other changeable tooling plates that will bolt down.

Pic 6: Another view of the machine...

I need to make the parts for the Z axis stepper, and build the wires. Hopefully soon (if Hiurricane Sandy doesn't ruin things) I'll be able to get this thing moving...