X3 build - Kflop/Snapamp, Keling servos, AC servo spindle drive, cncfusion kit...

I started this build 2 years ago and had to abandon it after just a few weeks. It's now officially resurrected and I'm ready to start posting now that I'm making some real progress and have the time to dedicate to it. If interested, here's a link to the original build thread:

http://www.cnczone.com/forums/showthread.php?t=40861

I've made some changes to the build that are worth mentioning, mainly switching over from the Grex G100 controller to Dynomotion's Kflop/Snapamp combo. I decided to go with the Kflop controller because it looked like a promising USB controller w/ closed-loop servo capability and the Grex felt like a dead-end technology since the development and support had all but vanished. I'm very glad I did because (as most of you here probably know) the Grex development was dumped shortly after. :|

After only 15 minutes on the phone w/ Tom from Dynomotion we had one of the keling servo's running in closed-loop mode on the kflop. I then decided to pull off of the x3 build and focus on retrofitting a Denford MiracPC lathe that I picked up from Reliable Tools using the kflop controller and a couple of used IMS drives from ebay. I'm actually using the same kflop controller to run the lathe and the mill since it supports up to 6 axis.

I had zero electronics experience and Tom practically walked me through the entire process so without him it would have never became a reality. Besides being extremely knowledgeable his support was way beyond anything I have experienced from any other vendor I've dealt with so far. In fact he helped me with other vendor's products when I could hardly get a response from them! (Thanks Tom)

Moving forward... the first agenda was to bore out the gearhead couplings that were too small to fit the keling motor shafts. I had to bring them up to 1/4" from 3/16" and I ended up using 3 bits for this job - the first two were standard drill bits and the last one was a 1/4" milling bit. I had to break out the fogbuster coolant for this or it would have never worked! Whatever treatment those couplings had gave it some serious hardness!

The first coupling didn't come out so well since I used 3 standard drill bits (before switching to the 1/4" milling bit) and the diameter was too large. The fit wasn't even close to snug but I managed to get it torqued onto the shaft at least. The second one came out better since I switched to the milling bit for the final pass but it still had plenty of slop when fitted to the shaft. The last one came out very well and had such a tight fit that there was a slight pop from the air suction when trying to slide it back off the motor shaft. Very happy with that one. I guess I got a better feel for aligning the drill bits on the lathe and also the speed and feed rates.

I then turned some motor mounts from some 1/2" Al bar stock which were the first parts I made on the lathe (or on anything for that matter!). I had to drill and tap some 4mm metric threads on each end. These all came out pretty well and I learned a lot about speed, feed, and d.o.c. since I just jogged them out by hand using the kmotion software with on-screen pendant keys. I actually started coding some gcode after getting bored a few pieces into it which worked pretty well and saved a ton of time.

For the most part all of the mounting of the motors and gearheads to the cncfusion kit went pretty smooth. I used the drill press and a milling bit to mill out some pockets into the backside of the Y axis mount that sits flush on the side of the table and drill out the threads for the top two holes to get the Y axis mounted. I had to do the same to all 4 of the mounting holes on the Z axis plate.

For the X axis plate I ended up tapping and force-threading some 4mm metric screws into the motor mounts and then cut the heads of the screws off with the lathe (see pics). This gave me a threaded end on the mounts that I could thread into the top two holes on the x-axis mounting plate.

All 3 axis are mounted and aligned and working well. I used a cordless drill attached to the encoder shafts to drive the axis' and they appear to be functioning in good order.

Next up are the gas struts to support the head weight...

Keep up the good work.

Unfortunately, I have similar problems with my mill. I either brake something or work/life gets in the way and I have to put it on hold.

The servos that you have appear to be very similar to mine. What was your reason for using the gear heads?

The servos and the g320s are probably the one thing that I got mostly right in this advetnure...So far....

For counterbalancing the head weight I decided to go with two 40lb gas struts from McMaster-Carr mounted on either side of the column. After scratching my head over it for a little while I decided to mount them at an angle to help counter any overhang due to the weight of the head extending out over the column. This seemed to make the most sense at the time since mounting it at the rear of the head didn't feel right to me. Mounting it at the rear or closer to the column would have helped relieve the strain on the Z-axis motor but wouldn't do anything for offsetting the the overhanging stress due to the weight of cantilevered head.

Choosing the path of least resistance I borrowed the hole from the old spindle motor tension bolt since it appeared to be in a near perfect spot for this design. I just drilled out the threads and ran a bolt through the inside of the head to tie down the mount/spacer that I turned on the lathe.

I used the mount as a spacer in order to align the strut parallel to the column or the direction of travel. I did this cause I didn't want to introduce any additional force that might cause the head to twist on the ways.

I chose to mount the bottom of the strut at the base of the column and so had to hand-drill through the cast iron and thread the holes for the strut mounts. This took about 8 drill bits starting with about 5/32" and working my way up. Took some time and some grunt force but I managed it with a cordless 12V drill.

After getting both of the struts mounted and turning the ballscrew by hand using the timing belt it feels like the upward travel is now slightly easier than the downward travel. After adding the AC servo motor to the top of the head it brought it closer to the middle. I figure that after I get a power drawbar of some sort rigged up I should be really close to the perfect balance point. It may be off slightly but most likely just a couple of pounds at which point I could always add some weight to the inside of the head since there is plenty of room for it. Either way it's a TON better than it was without the struts.

I lost about 3-4 inches of travel at the top of the column but will worry about that later on if it becomes an issue. I could always design a better mount that is adjustable - something that will have two settings so I can switch it over for more height if needed. The bottom clearance should be perfect since I plan to add a tooling plate from a large mic6 plate I picked up a while back.

Next up is the electronics. I started on an enclosure and will post pics of the progress next...

Hey Deviant, yeah for me it was working and moving. I did an absurd amount of overtime the last couple of years and I moved a couple of times to boot. :|

I went with the 3:1 gearheads to both slow down the high rpm's of the servos and to triple the torque power since they are only rated at 230oz/in. This should be a much better fit for this mill and will allow me to run the servos at their optimal rating without driving the table off the mill. ;) I was able to pick up a box of 5 for a good price so I went for it. I ended up using 3 but could have gotten by with only 2 by going with a smaller timing pulley on the Z axis since it was already using a timing belt to drive the ballscrew.

Some very nice work there! Really nice job on all the servo mounts!

But won't those gas springs be getting in the way? Looks to me like when using short tools (e.g. - small endmills) your Y- travel may be limited by the table, workpiece or vise hitting the springs.

Regards,
Ray L.

Hey thanks, I appreciate that! I'm hoping the struts will not be in the way. In this pic the struts are all the way compressed and with the collet holder mounted in the spindle there's only 1/2 clearance to the top of the table. So, at the lowest possible position you could hardly mount a bit in the holder w/o hitting the top of the table.

When I add the tooling plate (1" thick mic6) and a vise then there is a lot more clearance at that height compared to the fully compressed position since the angle decreases pretty quickly once it travels upward. I will find out soon enough though!

I started with a large enclosure and used a 17" drill press to drill the 100+ holes needed to hold all of the I/O cables, switches, fuses, etc... This was a lot of fun to say the least since there was no real way to clamp the box down to achieve any form of rigidity so you can see by the pics how ridiculous it got!

I started with a small criterion boring head that I picked up on ebay in used condition. Of course the threads for the screw fasteners were stripped so I had to re-thread them to a larger size, but even so, it was definitely the wrong tool for the job. The box just swirled around in circles with the bit as it circled around and it took forever to cut a single hole. One good thing about that experience was that I had to bear hug the enclosure to try to keep the bit cutting and this helped to give me an idea of what it felt like to scrape some metal. After about an hour, what initially seemed to be a super hard metal felt more like a soft malleable substance that could be peeled away in layers.

I then tried one of those funky drill bits from Harbor Freight that has all the sizes in one bit and it proved to be ideal for the job. I was able to drill all of the larger hole sizes using this one bit. And it's still sharp! Since I had no way to clamp down the enclosure in any ideal fashion I resorted to straps and pretty much anything I could find to shimmy it in place. This method took forever because I had to repeatedly unstrap and re-strap and loosen and re-tighten about 5 things in order to line up each hole! Even then it required the bear hug to get the hole bored out in the right place. I just drew circles for the hole diameters and used them as guidelines when drilling. It worked.

For damn near every large hole I drilled/bored I had to make 2-4 smaller holes for mounting screws. This required 3 bits so needless to say it took a while.

I used a jigsaw to cut the two large fan holes in the top and bottom of the box after drilling out the corners first with that HF drill bit.

I then sanded all the high spots by hand with a block and then I bought a cheap HF palm sander and pretty much went over the entire surface sanding it down to bare metal. I then primed it with a spray can of primer before applying some rustoleum acrylic paint with my HF spray gun and compressor. Using the acrylic was a terrible idea and I wouldn't recommend it to anyone. Definitely not a good choice for this sort of application! The paint never seemed to setup and would rub off on your hand even after days of drying! I ended up trying to rub a coat of sealing glaze over it to try to stabilize it some and I had to be careful to not remove the paint entirely from the surface. This totally changed the look of the paint but it pretty much did its job with sealing it for the most part so I am satisfied.

I ended up making a mounting plate for the electronics out of some 1/8 aluminum and used a pop-rivet gun to attach the wire ducting and DIN rails. All the electronics are mounted on the plate and now I have to solder and connect all of the wiring. I am about to start on that next.

With all the parts mounted to the enclosure I made up some brackets from some angle iron and mounted them to the table I made for the x3 using some skateboard bushings to help absorb some of the vibration from the mill.

Now it's time to wire everything up. I picked up a soldering station so I could solder a lot of the smaller connectors so I'll be learning how to solder next.

I should be back with more progress in the next day or so...

I can see that I need to come live at your house. Maybe all my toys would spend more time working and less time aggravating me. *lol*

Is that a slant bed lathe in the set of pictures? I need to steal your enclosure also.

What wire management system is that? I'm assuming the little covers pop off and the wire tray is a bunch of little fingers?

Which solder station did you buy. I love my hakko 936.

Yeah, it's a slant bed lathe by Denford. I picked it up off ebay for a really good deal about a week after I bought a little 7x10 HF lathe. I quickly realized how much time I could save by starting with a decent cnc lathe and upgrading the electronics vs trying to build one from scratch. I ended up selling the 7x10 unused and I'm really glad I did cause the Mirac PC is a way more capable and far superior machine in comparison. All the hardware included with the MiracPC is top of the line and I tried to copy it wherever possible for the x3 enclosure build. 90% of the hardware used on the x3 enclosure is high-end stuff that I got for dirt cheap on ebay. Patience is the key there. ;)

Since the lathe is a 1994 version that ran off DOS I ended up scrapping the controller altogether and retrofitting some new (used) drives controlled by the kflop after failing to get it working on an old DOS computer. It appeared to be hardly used when I picked it up so I was really happy to say the least.

In the pics you can see the original control box before I added the bottom plate and mounted all the new hardware to it. On the side of the box you can see the break-out board that connects to the kflop controller via a serial cable that is now mounted in the new x3 enclosure. It looks all nice and tidy now but check out the nightmare photo (2nd to last pic) of the work in progress! In the last photo you can see the current temp setup I'm using now to make the x3 parts.

I got the idea to use that cable ducting for the x3 enclosure after retrofitting the lathe and being so impressed with how well it worked for cable management in the lathe controller box. I picked up a long stick of it from Mouser. The design is exactly like you explained.

Hey Deviant, I actually bought a knock-off of your exact soldering station! It's a Madell QK936A and it was only $40 new off ebay. We'll see how well it works once I get to soldering with it. Hopefully it will do the job.

Very nice job on the E-box as well! I faced the same problem with the one I built recently - WAY too big to even think about doing on the drill press. So, I made several large rectangular cutouts, and machined connector panels. The nice thing is, if I need to change anything down the road, it takes just a few minutes to make a new panel, and I can completely reconfigure the connectors if I need to:

http://www.cnczone.com/forums/showthread.php?t=87644

I like that lathe!

Regards,
Ray L.

Very nice work on the enclosure Ray! The removable panel is a great idea and I would have definitely gone that route if I had the x3 in working condition. I plan to do that on the pc controller box that I'm going to build once I have the mill up and running. I have a touch screen and a mini ITX pc that I'm using and am working out the design now for the the enclosure to mount them in. I'd like it to be mounted on a swing-arm so I can swing it back and forth between the lathe and the mill whenever I need it.

Check out the pic below to see the touch-screen and mini itx I'm using. The pc is under the mouse pad. I got that stainless TI keyboard off ebay for about $20 and it was brand new! I need to find a better mouse solution though... maybe a trackball if I can get a deal on one. Since the kflop has an on-board processor to generate the pulses I was able to get away with this slow pc. It's running a stripped down version of Windows 2000 since it seemed to run best with that OS compared to XP (only has 256mb of ram!). Working great for the lathe so far.

The lathe is great. I'll definitely have to keep my eyes open for one.

No idea how the knockoff will do. Hopefully they cloned it well. When I got the station ,the 100 bucks hurt my feelings. After using it, I'd have no problem buying a replacement.

Quote:

---

Originally Posted by yantra3d

Very nice work on the enclosure Ray! The removable panel is a great idea and I would have definitely gone that route if I had the x3 in working condition. I plan to do that on the pc controller box that I'm going to build once I have the mill up and running. I have a touch screen and a mini ITX pc that I'm using and am working out the design now for the the enclosure to mount them in. I'd like it to be mounted on a swing-arm so I can swing it back and forth between the lathe and the mill whenever I need it.

Check out the pic below to see the touch-screen and mini itx I'm using. The pc is under the mouse pad. I got that stainless TI keyboard off ebay for about $20 and it was brand new! I need to find a better mouse solution though... maybe a trackball if I can get a deal on one. Since the kflop has an on-board processor to generate the pulses I was able to get away with this slow pc. It's running a stripped down version of Windows 2000 since it seemed to run best with that OS compared to XP (only has 256mb of ram!). Working great for the lathe so far.

---

Yeah, those mini and micro motherboards are nice, aren't they? Think about using a touchpad for a mouse - nothing to get fouled by chips, and you can buy fairly large USB ones for cheap. That's what I'm doing. Since my "real" computer is a laptop, I'm quite accustomed to working with a touchpad, so even a mouse seems "wierd" now.... My LCD/keyboard/control panel will be mounted in a rolling cart, with 5 drawers below for all my tooling. The LCD/touchscreen/keyboard/touchpad and dedicated control buttons (about a dozen dedicated controls and another dozen "soft" controls, used by a custom screenset, around the perimeter of the LCD) will be mounted in a sloping-front instrument enclosure. on top of the cart.

Regards,
Ray L.

Thanks for the tip Ray... I'll keep my eye out for one of those touchpads. That sounds like a good alternative I think.

It sounds like our pc controllers will be kinda similar. I really want to add some spindle speed and axis feed overrides to mine but haven't really gotten far enough along to start any real planning for that yet. So far I've set up the electronics w/ relays to handle the coolant and potentially a power drawbar of sorts and would like to add some switches on the panel for those, but that's about the extent of it.

Are you planning on setting up a jog pendant of some sort too? I'm definitely considering adding one once I get to that point.

Quote:

---

Originally Posted by yantra3d

Thanks for the tip Ray... I'll keep my eye out for one of those touchpads. That sounds like a good alternative I think.

It sounds like our pc controllers will be kinda similar. I really want to add some spindle speed and axis feed overrides to mine but haven't really gotten far enough along to start any real planning for that yet. So far I've set up the electronics w/ relays to handle the coolant and potentially a power drawbar of sorts and would like to add some switches on the panel for those, but that's about the extent of it.

Are you planning on setting up a jog pendant of some sort too? I'm definitely considering adding one once I get to that point.

---

I had always planned to build a pendant, but a few weeks ago, I broke down and bought one:

http://www.vistacnc.com/b02_pendant_...dant_3_pro.htm

Granted, it looks like it was styled by someone who used to work for Jim Hensen on Sesame Street, but the functionality is perfect for my needs, and I don't have to build anything! It should arrive any day.

For adding buttons, MPGs, pots to your control panel, look into Modbus - Peter Homann's ModIO is very nice:

http://homanndesigns.com/store/index...&products_id=4

and Pokeys:

http://www.poscope.com/product.php?pid=3

Both are VERY well supported in Mach3.

Regards,
Ray L.

Thanks for the links Ray... much appreciated!

Not much progress to report on mill... doing a lot of soldering and wiring on enclosure right now and have all power supplies, switches and fuses wired in. Starting on servo cables now...

Right now I'm trying to focus on getting the lathe spindle encoder working w/ Tom for the kflop controller so it can do threading. We're going to use the A and B channels (both pulsing at 60ppr) to get the threading working and also rigid tapping. We're bypassing the 1ppr Z channel for now but may work on that later when configuring Mach plug-in (if needed).

Will post more progress on mill soon. Will try to post some more pics of the other components I'm using for the mill when I get a chance.

Quote:

---

Originally Posted by yantra3d

Thanks for the links Ray... much appreciated!

Not much progress to report on mill... doing a lot of soldering and wiring on enclosure right now and have all power supplies, switches and fuses wired in. Starting on servo cables now...

Right now I'm trying to focus on getting the lathe spindle encoder working w/ Tom for the kflop controller so it can do threading. We're going to use the A and B channels (both pulsing at 60ppr) to get the threading working and also rigid tapping. We're bypassing the 1ppr Z channel for now but may work on that later when configuring Mach plug-in (if needed).

Will post more progress on mill soon. Will try to post some more pics of the other components I'm using for the mill when I get a chance.

---

FYI - I got the iMach pendant a couple of days ago, and LOVE it! I've written my own "driver" as a macropump, that has more functionality than the provided plug-in, and it works just great with the SmoothStepper. VERY nice product for the money!

Regards,
Ray L.

Glad to hear that pendant is working out for you Ray!

Just a quick update on the mill... I managed to get all 3 axis' tuned and am now getting AC servo for spindle hooked up to take analog signals from c6 board.

The head was shaking like mad and tightening the gib strip to the max did nothing. After further inspection I decided to hack about 3/8" off the screw and then I was able to drive the strip down far enough to eliminate the shaking. It's now snug.

The x and y gib strips are way bent and after hours of lapping on a surface plate I was only able to get about 50% of the high spots taken off. I just picked up some more sandpaper so I can work them some more tomorrow. I am pretty certain that I will need to replace them both w/ some new ones cause they are just too bent to lap them straight I think. Might have to place an RFQ since I have no way to make them myself.

I used mini-din connectors for the small usdigital E4P encoders and one of the receptacles' contacts is not making contact with the pins. Have to place an order w/ mouser for a replacement. Kind of regretting my choice to use them for this application cause they seem too delicate to me. Should have went w/ a more rugged connector I think.

Tried to post a video of the servos driving the axis' but wouldn't accept file format. Will try to load it to YouTube and post a link here. It seems to me that the servos are too loud but would like to get some other opinions so will post a video w/ sound and see what you guys think. They actually sound louder than the steppers on my Denford lathe!! I wasn't expecting that.

Enclosure is more than 90% wired and almost ready to mount back onto table along w/ mill.

After I get the AC servo running I will need to get it mounted on the head. I'm going to hack the original z axis nema34 motor mount plate that came w/ the cncfusion kit for this. Will need to chuck it on the lathe and expand the hole size so the motor mounts flush on the plate and will have to bore out the original metal pulley to fit the new servo shaft and also another pulley to fit on the spindle shaft so I can get it mounted and drivng the spindle.

After that I will mount and wire the limit switches and the home sensors and then I can start playing w/ some gcode and try some test cuts.

Then I can move onto auto coolant via c6 relay and will fabricate a pc controller to house mini-itx and touchscreen panel before moving onto ATC and 4th axis, etc...
***

I've been making some progress on the mill in between working on the lathe spindle encoder with Tom. We got the 1ppr signal reading accurately in the software and we ran some simple tests to get the threading codes setup. Tom is compiling the codes so we can make an actual thread test cut so it's going pretty well. We had to add a capacitor from the 1ppr signal to the ground in order to filter out the noise but now it's reading very accurately. Well, as accurate as a 1ppr signal can be anyway!

While tuning the servos I was running into all sorts of hiccups that I had to get sorted out. One problem I had was that the x axis ballnut was scraping the underside of the table which was bogging the servo down considerably. After disassembling it I noticed the ruts that the ballnut had cut into the table and also the swarf that was built up on the screw. I ended up pulling the x axis assembly off the mill and filing down the ballnut until it cleared the table.

Next up was another round at trying to lap the warped gib strips. And after sanding my finger tips raw (see pic) I ended up rigging up a small sanding wheel on the drill press as a poor man's alternative to a surface grinder. This actually worked pretty well and the end result was a perfectly flat gib strip (after some more lapping on the surface plate of course).

In the photo you can see how I set it up. I used a 6" angle and then bolted some skate bearings onto the 1-2-3 blocks to use as guides. From there I just slowly fed the gib strip through the opening.

You can see a before and after photo to see how warped it was - even after I lapped it for hours. It was just too warped to lap out on the surface plate. This little rig saved me the expense of having to pay for someone to make new strips for me (not to mention it saved my finger tips!).

The servos are now tuned really well and running very accurately. I can command a move of 30000 counts and back at 250ipm before getting an error of 1. They have a ton of torque and when I lean against the table with all my weight there is no noticeable change in the motor's pitch and no loss of accuracy. The 3:1 gearheads were definitely a good idea.

I also got the AC servo motor for the spindle running through the kflop and c6 board. Now I just need to figure out a way to switch the polarity so I can get the +/-10V signal needed to run it in reverse. I am going to try to rig up a DPST relay to switch the polarity of the analog signal coming from the C6 board. I should be able to use the relay on the c6 board to drive the DPST relay and make this happen (hopefully). If anyone has any suggestions or ideas then I would definitely like to hear them!!

I have to break off the mill for a week but I'll be back with more progress when I get back...