Nice machine, needs upgrade, any advice?

[BikeCourse]

X-post from LinuxCNC.org forum

Thanks in advance for any help.
We’re a non-profit arts/makerspace that inherited a homemade 5-axis CNC router that was made in the late 90’s. It’s in nice shape and doesn’t have many hours on it. Back-story is apparently that the original owner had it made for cutting tire molds but that it didn’t meet their needs. Second owner picked it up as a hobby but never had the resource/time to get it going. After 10 years or so in his garage he donated it to us on the condition that we never call or ask for any assistance from him on it. Ha! Deal!

Here are a couple pictures
Machine
Z-Axis

It’s a robust 5’ x 5’ fixed gantry, movable bed machine that uses all Kollmorgen H-series servos. Motion control is handled by a discontinued, unsupported, proprietary Galil DMC-2100 controller that uses a Galil PCI card to communicate with the Windows XP computer.

We’d like to upgrade the software and motion control on it to get rid of the Galil and use LinuxCNC instead. Since we’re a non-profit makerspace we don’t have a lot of room for expensive mistakes. I think we’re got the resources to pull it off but we don’t have anyone experienced in building CNC from scratch. We’ve got Linux admin, electrical engineer, old-school machinist, CNC operators and skilled CAD/3D modeling folks. Between us I’m confident we can do it.

I’d love to start with any 30,000 ft view issues we should consider before we take a wrong turn and waste resources on a dead end. Ideally, we’d also convert the bed to handle 8’ sheet goods and add vacuum hold down at some point in the future when the software/control is all squared away.

If you wanted to turn this machine into a relatively simple and flexible sheet goods cutting router, how would you approach it?
Cheers.

Similar Threads:

• looking for upgrade advice
• Nice DRO for mill machine
• New Grizzly Machine! Looks nice!
• Nice machine

[awerby]

From my view at 30,000 feet, it seems like you're looking for trouble. If this machine works as-is, I'd say leave it alone and use it for 5-axis work on light materials. If it doesn't, then fix it until you can. If you want a larger 3-axis router for cut-out work on sheet goods, that's an entirely different machine, with a lower Z axis. Unless you scrap most of this thing, you won't get that out of it, and you'll probably find that what you saved from it wasn't optimal for the task. It sounds like you've got enough talent around there to build the router you want from scratch, and it would probably be a better machine than whatever you could manage to make out of this one.

[Jim Dawson]

If I were doing this project, I would install my software on it and run it. There is nothing wrong with the Galil DMC-2100, but what I don't understand is what PCI card you are talking about. The DMC-21xx is a serial or Ethernet stand alone motion controller, no PCI card required. The DMC-18xx cards are PCI motion controllers, but would not normally be used with the DMC-2100. But they might be connected to an ICM-1900 inter-connect module.

You actually have a really nice system and unless you are going to upgrade it with more modern equivalent hardware (read that as really expensive), you will be taking a step backwards.

I am actually serious about installing my software. Since you are a non-profit, I would give you my software along with the source code for educational purposes. It is fully compatible with the Galil hardware and the XP computer. If you're interested, I'll PM a link to an installer so you can look at it.

[wizard]

I'm a tipping a bit towards Andrews view, this is an oddly constructed machine and might be more trouble to rebuild than it is worth. The whole machine just doesn't look robust but that may be an issue of not seeing the scale of the device, you mentioned a 5 foot gantry but that isn't obvious in the pictures. Looking at the pictures again the grease gun does give some sense of proportions and frankly those proportions are odd. The Z axis looks like a mechanical disaster and frankly I have to wonder if it is sturdy enough even for light machining which might indicate where the machine went wrong for the original owner.

Now given that the machine is rather large I might suggest converting it to a three axis machine as being the low cost way to success here. Throw out that part of the controls causing you grief and implement a new controller with any of the many options we now have on the market. The reason here is fairly clear, old electronics are maintenance nightmares when something finally does fail. I don't follow Galil but you are likely talking 20 year old electronics, you already note that the electronics is not supported and make no mention of the ability to self support. I only bring this up due to the experiences I'm having supporting 20 years and older CNC and automation hardware in the factory I work in. Some of this stuff is very difficult to maintain due to the zero support and the fact that so much has been phased out. Transistors are a pain to source, capacitors are all smaller these days thus not fitting on bus bars correctly, processors are in many cases unavailable even on the used market. Then you need to realize that anything bought new is light years ahead of what it is replacing.

I might also suggest lowering the Y axis gantry to a more reasonable level for 3 or 4 axis work. This will blow out the cost though but could lead to a far more functional & useful machine.

To put it another way I wouldn't bang my head against the wall trying to get a controller to work that hasn't worked in years. Just about anything you buy new will be more reliable (well hopefully) and certainly easier to use and more friendly. There are so many options two that you will be able to fit something into your budget.

There is also a more practical reason to consider going back to 3 axis. That is the fact that there is a big jump in costs for 5 axis CAD/CAM solutions. 3 axis and even 4 axis have plenty of CAM solutions out there, good 5 axis is far more difficult and costly.

I don't want to sound like a party pooper but I don't see much value in this machine as a 5 axis CNC.