Retrofitting a 1986 Maho MH400E

That's a good suggestion Andy.

Now I need to start the main spindle with the hal component to salvage what I can from the busted HF one (mill out the broken screws).

To accomplish that aim, I must remap the IOs from the 7i84 pins to the 7i77 pins (they have been wired all along, but I wasn't planning to do any of this this soon).

I've skimmed through the files from the github repository and it appears to me that the pin connections for IO to the Mesa card (hal net command) that I need to change are contained in the machine.hal file, and not in the component itself.

The mentions of the Mesa 7i84 in the gearbox.comp file appear to be comments. Is this correct? I am unfamiliar with the programming language in which the gearbox module was written.

Before I break anything else, once the gearbox_comp.hal is compiled, it doesn't matter which directory it's in right? Because the install command creates a real time compnent, like trivkins, that can be loaded by any config in its main.hal?

I got some warnings when compiling but the hal seems to have loaded since linuxcnc starts with it, and none of the nets have generated any errors.

I'm going to double check all my wires before commanding any gear changes, my encoder problems have taught me to distrust all my >0,25 mm2 crimped wiring.

Ran my first test program from Fusion360 and immediately crashed the spindle into the table

<vulture>Damn that really sucks, but we still want to see photos!</vulture>

Since I am only getting started with CNC milling, I run all gcode programs I generate first through NCViewer.com , and then also use the excellent backplotting of linuxCNC, switching to look along X to see what the minimum Z is in the file. I picked up a Z-254 instead of Z-2.54 in the last program I ran.

Correct, the gearbox.comp provides pins which you connect through the machine.hal. We tried to capture all our knowledge in the gearbox in the wiki in Github
github.com/jin-eld/mh400e-linuxcnc/wiki

Mark

It is also my understanding that it doesn't matter where the comp file is located.

I believe that MH500E and 600E's have the input stage physically reversed, so I think you might need to reverse the feedback from that stage. The only datapoint was some guys in Russia who used our comp. Their feedback was that running a gearshift on block in the wrong direction did no harm. Your mileage may vary.

Once you get it worked out, please let me know so I can add that knowledge to the WIki. My friend is going to revise the gearbox.comp this summer, to give it control of the spindle, and thus compensate for the current behaviour of going into neutral all the time.
Mark

Mark, I think this type of mistake will not be visible in a simulation environment unless you also create a virtual configuration listing all of its axis settings.

I will trade you the source of my misery for feedback regarding the compilation of the gearbox component. I copied the output from the terminal to a text file. It lists some warnings but does create a usable real time component because the machine starts with it. I'm asking about this because using the source rip environment thing yielded a "no such thing exists" type of response. I couldn't find anything like it via Synaptic package manager either.

I'm worried that the file didn't compile properly.

I have to edit the pictures from my phone before I can post them, I'll do that in a separate post.

Here are three pictures: 1) the spindle face with a broken endmill inside 2) broken screws inside tapped holes that were holding the spindle to the mounting plate 3) inside the holes of the mounting plate that connected it to the spindle you can see threadmarks imprinted in the aluminum by the steel screws.

The spindle hit the plastic, the mounting screws broke in half and the still-running spindle bounced off into a large pile of rubbish in the chip tray while I smashed the machine off button. It's good that I didn't empty the chip tray in the past few months because otherwise it would have hit hard against the sheet metal. It's also good that the move was fast enough for some of the energy to go into the mounting plate before it annihilated the spindle. If the move was slow enough, I imagine the bearings would be crushed. The way it happened, they still move without a noticeable increase in friction, but they make a louder noise, which indicates that the ring in which the balls are set might have cracked a little (at least that's what an old mechanic who used to work for a local machine tool building company told me, and I trust his opinions).

Lucky that those mount screws sheared, before something worse happened, like cracking the Y-Axis dovetail.

I really didn't understand how to compile the comp. Sergey gave me a list of instructions and I followed them. The recipe is here:
github.com/jin-eld/mh400e-linuxcnc/wiki/...-Component---Testing

That complies the cop, and also the simulator, which is very useful for the initial testing.

Frustratingly, the manual for the spindle doesn't list the type of bearings that are inside. It's a type of one-size-fits-all manual which contains general info on the available variants of the product.

I have no idea how expensive replacement bearings would be before I disassemble the spindle. It's difficult to make a good decision regarding mending/replacing this. Since it's cutting plastic, I'm going to try running it before doing anything else. I can see how much of a difference the damage makes. I don't have a press in my workshop so it might be impossible to disassemble/assemble the spindle anyway.

I'm thinking that if I don't go under during this retrofit and I manage to cut some parts without killing myself or destroying everything I've worked so hard to achieve, it might be time to get a loan and buy an inline spindle mounted on an SK40 tool holder. But that's just a fantasy that helps me keep on keeping on today.