[KickFarter]

As the title states, I have, for better or worse practically inherited a Seiko D-Tran XM5000, an older cantilever, cartesian assembly robot.
Its working area is 600*400*200, and its resolution and repeatability for movements are boasted quite high in the manuals that I found.
I am almost dead set on starting to build it up as a router, unless you all see any red flags in the details below.

I have been exploring CNC builds and entry level machining for months, so this might get lengthy, but bear with me. Disclaimer this will be my first build!!!!!!

Here are the nitty grittys on this particular idea:

1. Deflection - Cantilever and Router certainly don't go in the same sentence together. However! The machine is industrial grade, cast iron constructed and designed for a 15KG payload. I won't be hogging steel, but 30 lbs seems to be enough to get away with aluminum. I was playing with an indicator this morning and it holds up well to vertical forces. (Within 0.01" when I'm really leaning on it) Thats really alright for me, but I really don't know how much force is transmitted during lateral operations, and how the thing will take it. The rails are beefy though, and the ballscrews are ground/high pitch. I'm impressed overall with the construction of this thing.

2. Tooling - Are those 4million RPM chinese spindle motors really any good? The collets look tiny, and have never seen someone doing any real damage with one. I read an interesting thread on AC Servo motors for a lathe spindle, why not for milling? Shooting for aluminum in terms of power torque speeds etc. Since the machine isn't as rigid as a VMC, maybe higher spindle speed and a little slower feed would help? This is where my lack of actual machining experience shines!:withstupi

3. Controls - The proprietary controller looks ancient. BUT I can get a working one if I need it. I'm tempted to steer away from this option for a few reasons. Its huge. Its old. And I have no idea how to mess with it. The PC interface software needs to be run on Win2000 or earlier, and is geared towards pick and place assembly work. How I will get it to accept commands from Mach3 (or similar), internally, I have no idea. I have a mechanical background and software is not my forte. The other extreme is replace both the motors and the controls with new, or like new, matching pairs for each axis. The plug and play aspect of this idea bodes well to me, BUT that looking at around 2-3 grand. How do I get these brushed DC servos running on a modern controller?? Seems to be a 600ct encoder with 6 wires coming from it. In all my research I find that mitsubishi motors like mitsubishi controls etc. This company, Tamagawa no longer supplies DC servo controls, atleast not on their website. Before I call someone up over there what are your immediate thoughts? Keep the motors or swap them. I would love to keep it under 1000 for electronics.


This last one is what I've really been thinking about lately and could use some guidance here for sure. Like I said I am a greenie so go easy on me. I've been doing my HW as best I can!!
Any feedback or comments would be greatly appreciated. You can find a picture of the machine if you google "Seiko XM5000" What do you think??:idea:

[ger21]

Do you have any pictures?

[KickFarter]

Attachment 277788
This is one of the robots I have available to work with. Not the exact one, but this size will fit my garage well.

Attachment 277790
These are the guts of the lower axis. Nice couplers, ground ball screws, and 2 inch wide THK rails. I believe the perpendicular axis has 4 carriages.

Attachment 277792
Attachment 277794
The old school controller. It has serial port connections, so I can plug it up to a PC. But I would have to learn the proprietary software as well as teach myself to write programs/pointers for windows, which would involve messing with the front end software (mach3 or the like). That sounds pretty steep to me. Id rather go the route of breakout board control.

[awerby]

We built a router with similar actuators a while back. The motors were even more incomprehensible than yours, so we replaced them with brushed servos, running them with Gecko G320 drives at 78 volts. If your actuators came with brushed servos already, you might see if that would work for you. We mounted the actuators on a welded steel frame (I don't believe in cantelievers either), and got a big old rotary table to use as a 4th axis. It was all pretty straightforward, although we had to machine our own adapters for the shaft ends in order to be able to use standard couplers. I don't know about those Chinese spindles either; some people seem to have good things to say about them, but they don't seem as beefy as the Perske spindle I ended up using on this build. (They're a lot cheaper, though...)

[KickFarter]

We built a router with similar actuators a while back. The motors were even more incomprehensible than yours, so we replaced them with brushed servos, running them with Gecko G320 drives at 78 volts. If your actuators came with brushed servos already, you might see if that would work for you. We mounted the actuators on a welded steel frame (I don't believe in cantelievers either), and got a big old rotary table to use as a 4th axis. It was all pretty straightforward, although we had to machine our own adapters for the shaft ends in order to be able to use standard couplers. I don't know about those Chinese spindles either; some people seem to have good things to say about them, but they don't seem as beefy as the Perske spindle I ended up using on this build. (They're a lot cheaper, though...)

Well first off I like your build a lot. If I was going to design my own mechanics I would have done it this way too. For stronger materials, you can mount it high, pull your z axis way up and have a very rigid set up. Or you can have a large space for softer materials. What do you use this thing for mostly?

From the pictures I didn't show that we have more robots to play with too. You have me considering the benefits of just going ahead and planning for attaching another actuator to the other side of the cantilever. The thing would be BULLETPROOF. Depending on the spindle I think a set up like that could handle steel no problem.

Now some questions:
I don't see any encoders in the picture, are you running them? Overall how does the drive perform? Did you have to tune your system in initially? Being able to talk to the encoders was a concern of mine and I need to do some more research there. Up until now I had really been considering servo+encoder as one part, which is not really accurate. Once again I need some more research here but thanks for pointing me in the right direction. I was looking at more industrial components here, but TBH I don't even know what the difference would be.

Another reason I don't want to change motors is the flange/shaft size like you mentioned!

On the spindle motor, I had seen some things like this, but far out of my price range. The perske brand are a good find though and will definitely consider! I agree they look sturdy compared to the soda can spindles. Something like that would also offer less custom machining time for me which means I can get cutting sooner! Do you run yours from a vfd?

Thanks for the reply mate!

[awerby]

I mostly use it for sculptural carvings. Yes, having the other actuator on the other side does make it much more rigid - but not rigid enough to cut steel. That would take a whole different machine (we retrofitted an old Leadwell knee mill for doing that).

That picture was taken before we switched to servos; we replaced the steppers shortly afterward. Those boxes on the top are for constant force springs, which broke - we ended up using gas struts to counter the weight of the Z axis, which have lasted much better. My servos have encoders mounted on the back; nobody uses linear encoders for machines like this. If you've got brushed servos, the G-320 drives will probably work; ours had brushless AC servos, so we had to switch them out. Yes, we run the Perske from a Hitachi VFD; tried a Teco, but it blew up...

[KickFarter]

I mostly use it for sculptural carvings. Yes, having the other actuator on the other side does make it much more rigid - but not rigid enough to cut steel. That would take a whole different machine (we retrofitted an old Leadwell knee mill for doing that).

That picture was taken before we switched to servos; we replaced the steppers shortly afterward. Those boxes on the top are for constant force springs, which broke - we ended up using gas struts to counter the weight of the Z axis, which have lasted much better. My servos have encoders mounted on the back; nobody uses linear encoders for machines like this. If you've got brushed servos, the G-320 drives will probably work; ours had brushless AC servos, so we had to switch them out. Yes, we run the Perske from a Hitachi VFD; tried a Teco, but it blew up...

Got any pictures of things you've cut? Would love to see the result. Also more images of your machine would be awesome.

Can your machine handle aluminum? If so how does it cut?

Its nice to think that G-320 drives will work, since they are a lot cheaper than the industry solution. Just need to research the encoders. Are all rotary encoders the same? I have a Sanyo a/c servo with 17-19 wires, but these brushed dc motors only have 6?

It's funny you say that. We run several SS grinding machines at my work, and the dust gets everywhere. I have 6 blown up Tecos on my desk right now! haha Still can't convince my boss to shell out for a proper enclosure and filter fan kit. :argue:

[awerby]

There are some pictures of my work on my art site: Juxtamorph.com . I haven't tried cutting aluminum on that machine; I use the Beast for that (it's the retrofitted knee mill I mentioned.) or for smaller parts the Taig or ACT DMCIII. All rotary encoders aren't the same, but the usual type people here generally use are based on a disk with radial markings on it and an optical reader, and they mount on the back of a DC motor, turning it into a servo. The motor itself should only have 2 wires; the rest would go to the encoder. US Digital makes an inexpensive version: US Digital | Products

It's good to know that I'm not the only person who had bad luck with those Teco VFDs...