Retrofitting a 1986 Maho MH400E

To J Green
I have noted what you have written. I owe you an answer. At the moment there is so much going on and I have to make some decisions. I was entirely ignorant on this subject but thanks to you and others on this forum , I am learning fast. Yes it is Spring time in New Zealand but the weather is still very unsettled. Thanks for your effort.
Cheers Chris

Hello Mark
The foot switch is a great idea. Thank you. The TNC has been partially dead since I have had the machine. One circuit board had a return trip to Germany. Still no full TNC. This no doubt sounds a dumb question. I like the idea of C.N.C Can I leave the existing controls and gear shift as they are and use a computer with Linux installed to run the D.R.O and CNC for the table motions only? This to save complication. We can add to the system at some future date.
Thanks for helping. Cheers Chris

Heidenhain LS-403 scale --> Heidenhain EXE (you can buy them from Ebay I have 602's) --> pretty much any DRO display will be able to work with the TLL output from an EXE.
Mark

Hello Chris,

i still have two ideas but to check this i have to power up my tnc first.
But this has to wait till the weekend, i am out of home for production at the moment.

i have a little reading/understanding problem:

You have got an exchange board from germany? But the eproms may be are still old?

Hauled in the rest of the Maho machines I picked up. 600P and 600E.

Also, I started dismantling one of the 400P machines. CNC conversion appears to an option without too much issue. Removed the X axis lead screw components and drilled out the bevel gear. Installing a ball screw will be straight forward and there is plenty of meat on the cast iron end cap for the table to firmly attach a servo motor mount. Y axis should be a similar operation. Not sure what I'll do for the Z axis. May mount a motor the the hand wheel location, a heavy table should eliminate most backlash using the screw that's there. Other option is to use the quill. Not a lot of travel but it wouldn't require a very big motor.

The 600E in the picture powers up but throws a code. I don't care for messing around with old electronics and find they're program size limits to be a pain. I plan to retrofit with easily available parts. I had a lathe dump an EPROM and the factory wanted $8k for a reprogrammed piece. No thanks. The Daewoo lathe now runs all day with a Centroid Acorn.

I haven't done much looking into the Indramat servos. I'm assuming they're brushed DC which would have them easy to drive with a Gecko 320X. Will see how it goes.

J wrote:

Glemigobles
My MH600E has table droop - sag . Was wondering where your machine is worn?
Have you considered inviting Mark over for a Three week MAHO scraping party ?

The most wear on my MH600E is on the vertical column. You can measure it with a micrometer to check how much thinner the side of the column is in the middle of travel when compared to the top and bottom. When milling with the vertical head, this translates to a tilt/skew in the Z axis that measures approximately 0,13 mm across the table. With horizontal milling this of course becomes the Y axis.

I face-milled the surface of an aluminum plate to create a tooling/sacrificial plate for mounting stock and minimize this error. It helped a lot and reduced the measured error down to 0,05 mm. Still, the actual wear on the ways is bad and not getting any better with use.

An alternative solution to scraping that a service technician suggested to me would be to shim the sliding member of the Z axis. But that would create problems in the ends of the travel. So, the machine should be torn down, the ways ground and scraped or just scraped (but there's turcite to think of keep all the elements in their original planes so some machining would be necessary anyway). I would also take that opportunity then to grind the table and both spindles' inner tapers, and just do a general rebuild (probably would also replace the servos with newer ones). But it's the advice of everyone I met who works in machine tool rebuilding/maintenance (and not just online posts, of which there are plenty) to scrape the ways, having them ground is not enough.

I have this sneaking suspicion that my encoder problems on the Z axis may combine with the wear on the column.

Anyway, back on topic, despite the work required on retrofitting a P series machine, they are almost guaranteed to be in better mechanical condition than an E or C of the same age just because it's far more laborious to use them so much. A CNC mill in a production environment will not be coddled and will wear much faster than a manual/mostly manual machine (just think about quill use vs programming drilling and tapping where the whole table moves up and down; it all adds up over 30+ years).

All you guys who got mint-mechanical-condition E and C series Mahos got great deals. A LinuxCNC retrofit is very doable and completely changes these machines, taking them to a new level. The learning curve is very steep, but the documentation coupled with online help is comprehensive and fantastic (you just have to get past the point of learning what they're trying to tell you ). Anyone who has the time should go for it.

eflor8234 wrote:

The 600E in the picture powers up but throws a code.

I haven't done much looking into the Indramat servos. I'm assuming they're brushed DC which would have them easy to drive with a Gecko 320X. Will see how it goes.

What is the code? It will always throw up a code on start up until you power on the hydraulics with the 3SH-1 button next to the E-Stop. You have to press it twice, and then on the second time you press it also press clear or enter (I don't remember which now).

The Indramat servos are brushed DC and both Mark and I connected the 3TRM-2 module to Mesa 7i77 cards.

EDIT: just in case you don't have all the manuals, I have a list of all the Philips 432 error codes available. I never had any problems with that control apart from the CRT monitor burning out. But it only did 2,5D milling (simultaneous 2 axes interpolation) and had 16 Kb of memory with no drip feed. Retrofitting is a great idea, especially since you seem to know your stuff and won't suffer like I did.

eflor8234, Indramats are very reliable. Do Gecko 320X even support resolvers? Why bother with the expense of changing to Geckos, when you only need two wires per axis from a MESA 7i77 to the indramat, and the encoder wiring from the EXE to the 7i77 to control an axis. The resolver closes the velocity feedback loop to the Indramat, and LinuxCNC closes the position loop through the Heidenhain encoders. (Or does yours have the Phillips encoders?).

If you send me your Email address (I have the same user name on other forums like Practical Machinst since PM is not activated here), I'll send you the document where I collected all the x00E retrofit information.

May I suggest to document the MHx00E retrofits in this thread, but make a separate thread for the x00P modifications. ThisOldTony is also doing a 400P, so it would be helpful for you both, but also for others searching for type specific info to have it in a clean thread rather than good information buried on page 130 of this one. I am sure that all the MAHO guys will keep engaged on a 400P specific thread as well.
Mark

RotarySMP wrote:

eflor8234, Indramats are very reliable. Do Gecko 320X even support resolvers? Why bother with the expense of changing to Geckos, when you only need two wires per axis from a MESA 7i77 to the indramat, and the encoder wiring from the EXE to the 7i77 to control an axis. The resolver closes the velocity feedback loop to the Indramat, and LinuxCNC closes the position loop through the Heidenhain encoders. (Or does yours have the Phillips encoders?).

I haven't looked at the servos to see what type of feedback they use. If it has encoders, I'll try to use them. If not, I'll put some encoders on the motors.

I want to retrofit with simple parts.My problem with machines this old is that I feel like I'm always waiting for the next problem to come up. I've rebuilt a couple few VMCs and and a lathe, I've been really happy with the setups. Most of my parts are done in CAD leaving me with pretty big files sizes. I don't like drip feeding controllers...too much seems to go wrong.

I'll post a new topic for the 400P build out.

I think what Mark is suggesting here is to ditch the old control but retain the servos, they have closed loop feedback with tachos and linear glass scale encoders. Both exist in the cabinet as separate modules which you can easily connect to a LinuxCNC control built with Mesa cards.

Unless you also don't trust the drives and motors, but those will be expensive to replace with new parts.