Advanced Search

If you are writing an interpreter to plug into LinuxCNC than it should probably be written in the same language as the existing one. That isn't to say that Python couldn't be made to work (I don't actually know) but there would be more to copy and adapt in C++

RotarySMP wrote:

... Starting over I would probably have just injected a little fresh grease. It is easy enough. 

 

OK, so I can just do that and tell myself I’m not being lazy — it’s science. We’ll make an A/B test and see how the machines behave over the years.
Yes, that sounds like it could actually be a thing.

 

It is not difficult to pull the spindle. I just welded up a U frame to pull against and used all thread to draw it out. 
Schaublin recommended disolving Kluber in some solvent like trichchlorethane which we cant get. I don't yet know whether I screwed up good bearings or saved them. Time will tell. Starting over I would probably have just injected a little fresh grease. It is easy enough. 

If the dovetail fits, you could also just drill and tap the height adjuster holes on the other side.You can never have too many tool holders

I wonder why the replaced the X servo, but left the Z one?

 

I've added U and W to the G7x cycles, but those letters cannot be used outside the subroutine interpreted by those G7x codes.

Change the belt and leave it for now.

Focus on what does not work, then make little improvements.

Thanks ... that confirms what I suspected. I really hope the auto-tuning on the Delta B3 drives works better once they have the proper inertia of the machine. Running them completely without load probably doesn’t give the controller anything meaningful to work with.

I’ll revisit the tuning once everything is mounted mechanically. Fingers crossed that the B3’s auto-tune behaves more reasonably under real conditions.

The question about the grease in my head is that the spindle seems properly lubricated right now — no strange feeling, no signs of anything bad. So if I pull the spindle and put it back after installing the replacement belt, I’m hoping I won’t make it worse than it is. When it comes to regreasing these high-precision bearings, I’m hesitant — I’d rather learn on less costly spindles first.

Regarding Mach: the seller provided a copy of his Mach3 folder. I didn’t know this before, but you can just put that folder on the PC and it works—no installation, no configuration, no compiling anything custom.

If the Mach 3 install is new, is it even configured?

If you can add grease to the bearings without disassembly, do that.

I recommend synthetic grease.

You don't want ultra-thick grease, but you also need it to be thick enough to prevent wear.

I'm back home again, having been away at work during the week. I've created an initial config using pncconfig, then edited up the contents the best I can, based on my previous working instance in 2.8.4 Of course, those old files no longer work with the current version of gmoccapy and LinuxCNC, presumably due to changes that have occurred along the way, so I have no working system.

I'm in a bit of a Catch-22 situation, as I can't start from a working copy and simply(?) make a few changes. The SIM versions have a range of extra files to fool LinuxCNC to work without a real machine and I don't know which those are. My "starter file" created in pncconfig never got as far as being fully operational, not least because the bitfile in my 5i25 for isn't listed in the pncconfig options.

When I try to run LinuxCNC with my new embryonic config, it seems to bomb out when it tries to start gmoccapy. The debug info is very curt and doesn't seem to give me much of a clue what to look for.

I'm pretty certain I'm missing some of the files required for operation but I've no idea what those are. Without a functioning example, I can't work from a "known good" example. Ideally I'd have an example of a complete file / folder set that I can compare with - is there something I can refer to?

The various gmoccapy sims work fine but as I said, they seem to come with a variety of additional files that I won't need with a real machine. Can somebody help to identify which files I need and which I don't? Many thanks!

 
 
 

The aft-mounted one, I didn’t even think of that. I’ll try out tomorrow if it fits. It’s not just the wrong way around; there are no threads on the other side.

The servo is really much larger than needed, but buying a more suitable one when I already have this one on hand seems silly.
The controller is a CSMIO/IP-S
(link: de.cs-lab.eu/sklep/6-achsen-cnc-steuerun...und-simcnc-software/)
They seem to be sold on eBay for a decent price sometimes, so this might be the first retrofit in history that doesn’t cost more than expected (…it won’t).

To really run the thing I “only” have to replace the encoder belt on the spindle. If there’s one part I didn’t want to mess with, it’s the spindle — but I don’t think I have a choice here. I’ve seen some really bad belts, but never one that still worked despite having sections with no plastic left.

I’m considering just pulling the spindle and putting it right back in. What do you think? You replaced the grease without disassembling the bearings, right? Would you do it that way again?

Welcome to the Schaublin retrofit club.

Could it be that that corner tool holder just has the height stop mounted on the wrong side, and flipping it fits? Or could it be that it fits on the aft mounted single tool post?

That AC servo installation on the cross slide looks nice. Your Z axis AC servo is a beast, and still smaller than the 1970's drive.

Could you please post some close up photos of that MACH interface board at top RH in the control cabinet. Is that something like a smart servo for Mach 3?

If it works, I would probably run it for a while and make some parts as is, just to see if there are any other issues you need to know about.
Cheers,
Mark

The reasons for upgrading the operating system are:
1 Security. There are no longer security fixes for Wheezy.
2 When my previous system was taken out by lightning, I could simply move the hard drive over to another machine and had it up and running the same day. Moving old software over to modern hardware won't work - a risk.
3 Support a current version of LinuxCNC.

The reasons to be on a current version of LinuxCNC are:
1 A lot of work goes into the product with enhancements on functionality, usability, compatibility, easy of use, hardware support, simplified configuration and more.
2 I might benefit directly from such enhancements.
3 If I run into an issue and ask for help, the first question is usually about the version - generally speaking. If old, the first advice is to upgrade, for people to be able to help. The problem might well be solved in later versions. Nobody wants to, or is even able to support old stuff.

Z3n wrote:

Thankfully I have a nice bench power supply from my bench flashing for ECUs / binary reversing, so stable power from 5-24v DC isn't an issue

Sweet!

Your servos need manual tuning.
10 mm, one mm, or even one tenth of a mm is too much following error.
Standard servo tuning techniques applies.
Maybe don't spend too much on it now, you need to redo it in the machine with its inertia.