Retrofitting a 1986 Maho MH400E

Hi Andreas,
I saw that the deckel gearbox was similar, but with no end stops to the shifter shafts.  Cool that it was useful for your Deckel. Remember the rules... photos or it didnt happen! Or better, a link to a video of the gearbox under LinixCNC control.

That behaviour of the gearbox.comp putting the gearbox in neutral each time the spindle stops is an unfortunate side effect of LinuxCNC behavior of interpreting an M5 as both spindle stop and spindle speed zero. I didn't know that when I spec'd the comp. db1981's input is a good one, to stop it reacting to a neutral and off command. Jin wrote a minor mod to the comp to stop it doing that neutral thing, which I think is also up there and is the one I am using, but it has the side effect of never going into neutral, so If I want neutral for swinging an indicator, I have to hunt for the big allen key, and manually shift the gearbox.

I am no coder (I cant read what Jin did), and also a weak Github user. I passed your  request on to Jin, to ask how he wants the Github done.
I also have the issue that I could not identify a free pin to net to to turn the spindle on or off from the physical control panel. Jin and I talked about making a big change. Currently the Comp requests spindle on / off from LinuxCNC and linuxCNC implements that. We were going to switch it around to give the comp control of the hardware. and have linuxCNC request spinlde on/off to the Comp. But we never got around to it. On the Maho the brake is hard wired through contactors, so I didn't need to do anything. Switching on the spindle contactor on switches off the brake and vice versa.

The CW /CWW on the maho is also hard wired contactors, which I am controlling from the Mesa 7i84 through the MAHO 24V relay board. The Comp does use them to rock the gearbox back and forward during changes to ensure meshing, but through LinuxCNC.

Hi Andreas,

to be honest I did not look into the code for a long time (although - yes, I did promise Mark to rework it - sorry man, still did not get to it!); the issue we had was that LinuxCNC seemed to be more oriented towards VFD driven systems which did not match with my initial design of treating speed 0 as "neutral gear".

My idea to work around this problem was to put more logic around speed control and manage speed changes and shifting in some smart way where stop would not automatically mean "neutral". From the top of my head this also has influence on reporting the current speed, I think right now that is still bound to "which gear am I in?". So those changes - that is the part I never got to...

> Also I want to provide the finished comp and .c .h files with a wiki as soon as I am done. Do you want to include it in your github or should I create my own?

Are there many changes or can the comp be adapted to handle both machines via some #ifdefs at compile time?

If we can share code, then it'd make sense to fork the original repo, add your changes there and then submit a PR against upstream to get your changes merged.

Kind regards,
Jin

Wonderful. So a
net spindle.enable to an in pin of the comp and a simple if in the comp
if (spindle.enable==0){
return
}
That will work. I will also handle the relays for ccw/cw and slow/fast in the comp.
Thanks

Hi Jin thanks for your reply.
A combined repo / comp would work.
I use less inputs and haven't changed the program that much apart from some True/False statements. Everything could be handled by #ifdefs.

For now everything works with the gearbox. I will try to finish the PID setup / Homing and the wiring by Thursday. Friday I leave for Spain and will work on the rest and the combined comp after I am back.

Thanks again.
Stay healthy
Andreas

I look forward to seeing your machine up and running.

Also like to be able to use the code on my FP5 which has only 2 motors and 3 cams for each one. I assume the switches are outputting LH, center and RH shift fork position. I have posted PLC logic etc. in my FP5NC thread.

Hi Guys,I have just signed into this conversation. I have a Maho MH-C 700 that I have been resurrecting from the dead almost concurrent with this saga. Whilst I'm not refitting it with Linux there are some things that may be useful to those interested in this string.I'm a retired mechanical engineer living in Sydney, with family in NZ, so not so far away from this project.My machine was supplied in 1982 with the Philips 6600 series controller, but then refitted with a Heidenhain TNC 155 by a local agent in 1988. At the same time the encoders were replaced with Heidenhain ones. It has a 4th axis which is the the main milling table: 600mm diameter.The upgrade was done roughly and I've spent a lot of hours tracing wiring and removing a lot of redundant stuff they left behind. Whilst I had some diagrams of the wiring schematics it looks to me like Maho made endless variations to these machine. I ended up pull all the wiring free off the cable trays and loosely placing a cable tie on each wire, then sliding it along from end to end of the wire, sometimes through several bulkheads etc. A long and tiresome process. Then a lot of time condensing mud maps into the actual schematics. Tracing the tracks on the 24 V relay board on the door took about 2 months alone. Difficult to be sure that more than half of it did nothing.When the Philips controller was removed they also removed what I assume was a 19 inch rack mount in one of the side cabinets. That seems to have had a lot of interfacing relays probably with relay logic that was common in these early machines where the older guys didn't trust new fangled computers.In it's place is an Izumi FA 1 junior PLC. Fortunately one I used a lot early engineering days.The machine sat in my garage for 10 years before Covid lock downs left it to be the only project I could do.In that time the TNC155 lost all it's machine parameters. Fortunately there was a paper tape with the machine that I read by hand, plus a later engineer commissioning engineer had written them down in the manual. Add to that the Heidenhain TNC 155 was used on Bridgeports for a while and there are files out there to download which save a lot of typos and only need minor editing.To my amazement the Izumi PLC program was still there. They had used the battery backed ram and the battery was still good after 15 years! As soon as I saw it I sat there and wrote the whole thing down late into the night for fear of losing it. I've since replaced the memory module with an E2prom so it should last for some long time. The Izumi program is in ladder logic and quite easy to read. It reads the status of outputs on the Heidenhain controller that ask for a certain speed when flagged, and then converts them to logic that drives the 3  24 volt gear change motors directly. Readers of this string will be familiar with it. The logic may still be of interest to you. I can put it up here as an Excel file with comments if anyone wishes.I have some videos of it changing gears but they are big to attach.The Izumi PLC also controls the nid-nod action of the main spindle motor. Because the spindle motor gearbox is a "crash" gearbox it's possible for the straight cut gear teeth to crash into each other and not mesh. The nod-nod action fixes that problem. The spindle motor normally runs 415 volt ans starts star delta. But the Izumi switches that off and switches the motor over to another set of contactors that are wired to the 150 volt secondary of a dedicated 3 phase transformer. So the nid-nod happens at a reduced voltage (and power) on the spindle motor, and only in delta mode.The Izumi program replies to the Heidenhain when the gear change is finished.The gearbox speeds are ISO standard geometric jumps (or close to it) which is typical of many European systems of sizing. A Renard series as I recall.I think I'll leave it there for a first effort. If anybody wants more then just ask. I have lots of photos and a lot more for the saga. I could email the videos to anyone interested.Stay well.David

Thanks Mark for all the fantastic videos and helpful posts. I'm only about half way into this thread.
But I notice queries on the lube system. It may have been covered, but the lube pump doesn't actually meter the amount of oil delivered.
I've modified my lube diagram per my Maho configuration and hopefully can attach it here. Note those final distributors before the slide-ways. Every one has a dose metering piston in it.
So on my machine there are 35 of them!
From a long period of standing some were lightly suck by gooey oil. Also one oil port on the rotary table was gooed.
I'd advise cleaning them all.
Inside you will find a rubber check valve the presses on a central spindle that is very limited in movement. Around that spindle is a metering piston which is spring loaded against the incoming flow.
When the lube pump turns on the check valve opens and pushes the spindle but but keeps its central spindle port blocked.
As pressure builds the metering piston moves back against it's spring until it runs out of travel. I'll call that "cocked".Thus the pressure rises.
When all hit there stops the pump presumably dumps.
My lube unit has a Vogel IGZ 36-2 which is fairly dumb.
When the power is turned on it immediately runs a pump cycle that can't be set otherwise.
When it runs it waits for it's pressure switch to come on. When it does it presumably assumes that all the metering pistons are fully cocked.It then runs on a timer for a further 5 seconds, I presume to be sure to be sure.
Then it stops.
So now the metering pistons don't see pump pressure and their springs push them back to un-cocked. In this mode the check valves re-seat and the oil is delivered through the port in the center spindle.
If you wish to check for blockage simply undo the exit tube from any one of the metering devices until it gets air in it. Then replace it.
Watch as the pump runs and you will see no movement of the air bubble until the pump switches off. Then it moves a measured amount.
All mine deliver the same amount.
If the pump runs for more than one minute without the pressure switch coming on the pump stops and the red fault light comes on.
The IGZ 36-2 only allows the setting of the period between lube cycles. It does have a mode where it will count instead that could be programed by pulses from the traverse movement. Basically it would make more sense for LinuxCNC to do all that instead of the Vogel.
I've put a delay on timer in front of my lube unit so it doesn't turn on for a while. I can change that but it was handy during commissioning so I didn't waste a lot of oil.
BTW. I'm using Gulf & Western Tacway 220 in the lube system. Equivalent to the recommendation. Available in Sydney. Nice tacky feel to it. Works nicely on my lathe bed-way also.

David

In my first reply I mentioned that my Maho has been retro fitted with a Heidenhain TNC 155. To do this they had to put an Izumi FA1 Junior to do the gear change. I noticed comments on gear change strategies. I thought if I added the relay logic code and an image of it you might see a strategy that works (when on the Heidenhain). It's reasonably similar in I/O to that you are familiar with. The numbering of relays etc should be very familiar.I don't have a way of adding comment to the real code, so I simply pasted it into and Excel column (zip file) and added comments beside it. I have an FA 1 Junior manual so you can see what the code means in ladder logic, but it is too big to attach. I imagine many of you will be able to read it anyway.The first part is simply how the inputs on the Izumi read the bits on certain outputs of the Heidenhain. The comments explain that so you wont really need to look much at that part.The Izumi does the switch over of the spindle motor to 150 volts and uses the nid nod of that motor so the gears don't crash. You might get some ideas from how that works as it has yet to jamb a gear change. Note that it does each cam motor in turn and it always runs them in the correct direction to find the gear. It switches 11d8 to do that. I think my gearbox can go right around to the start like some mention here, but this strategy doesn't do that.At the end of the program it goes into a lot of timers. I didn't figure that out as by then it was all working nicely and my brain had gone to mush reverse engineering this whole machine. I have some videos of it changing gears and the spindle motor seems to nid nod more than some I see in the videos here. If anyone knows how to decrease the size of an MP4 I'll post some.
David

Some more information and words of warning:

Here is and attachment of the Z axis knee being put back on the tower. You can see the aforementioned problem with the lifter  feet you have been discussing, of the engine hoist only just clearing the base. But note where that black extension is on the lifter. It is actually forward of the locating pin to get the reach to mate the two assemblies. Not something I would recommend but it shows the limits of these lifters. I was very careful the the load was inside the castor footprint. And the knee is close to the 500kg for that pin load limit also.
The Lifter is from Supa Cheap Auto here in Sydney and is a little heavier than most.

The knee assembly is going on to the original setting of the Z axis slideway wedges, something I wasn't sure it would do. Just a few wiggles and it went on, much to my relief. The screw is fully up and the dunage is in case the knee stuck then suddenly went free if the sling wasn't taking the load.

Note that the Z axis brake motor is missing. So is the drive belt. That means the broken belt ratchet is actually locking the Z axis toothed belt pulley. Be careful that is working as the recirculating ball overruns easily and the the knee will accelerate as it falls and I suspect will hit the base hard enough to smash the castings.

Since then I've found I can hold the knee in place by hands on the screw pulley to inch it up and down a bit. But don't let go! it will get away on you. I had it propped while I fitted the oil supply to the recirculating nut as the banjo is difficult otherwise. But be careful.

A passing note: The original Philips controller had the Y and Z axis swapped from what is now more normal. It confused me a bit as the manual for the Maho had it described that way, but the Heidenhain refit has them with Z as the vertical. It means that the my machine has the servo closing relays in an odd order: X, Z, Y, B. Too much rewiring to change them.