Retrofitting a 1986 Maho MH400E

And now for something a bit odd. I've fitted my Maho with a frog warmer! And no it doesn't have a problem with frogs. It's actually a humidity control system. As you may have seen the rear of my house is basically underground. The concrete roof has a garden on it. You can see from a previous post, on occasion it can look more like a submarine. In the lifting crane post you can just see the front of the house, which has window for its full length which faces the sun (north in the southern hemisphere). So i have a situation where the rear of my shed, through the glass work, faces a cool sandstone wall the is usually damp and stays around 18 degrees C all year round. I can pull damp air along between the glass and the wall to an air conditioner radiator yet to be fitted. Nice and environmentally friendly... eventually.
The awning on the front of the house shades the front rooms from sun in summer and allows it in in winter. Nice. So I currently balance the house temperature by having a set of small fans circulating house air though into the rear shed and back. There is also in water heating tubing throughout the concrete floor including the shed. But here is the downside of all underground houses. If you take warm air from the house and circulate it to that colder zone near the underground part, the absolute amount of moisture stays the same but the relative humidity goes up. And of course dew forms when that cooler air gets to the dew point. In Sydney that can happen even in summer on a high humidity day. So that glass wall sometimes gets wet with dew and puddles of water on the sills. Being right next to it, the radiant temperature of the cold glass can form dew in the contactor cabinet. Expletive deleted here!
So I've fitted this relative humidity controller. I've just set it to give 70% relative humidity and it turns on 40 Watts of the "frog heating" elements. It's currently runs the heating pads about 1/4 of the time, so not expensive. 70% set point gives about 5 degrees C distance from the saturation line on the psychometric chart. I've fitted some 15mm rigid foam to the inside of the cabinet doors and to the inside roof of it to reduce the heat loss which makes a noticeable difference to the on off ratio. When the machine runs this system simply gets overwhelmed by the cooling fans, but it's not really worth switching it off. The elements look like sheets of Mylar with resistive material fused between them. I've simply slipped them up the gap between the contactor mounting panel and the cabinet wall. The mounting panel standoffs give a 15mm gap for them. Perfect place for them. The humidity sensor is on the cabinet roof.
For a few bob extra I got the one with WiFi so now I can check if my "frogs" are doing okay from anywhere in the world!
Note on the side of the mill photo, the air exit mesh as mentioned in a previous post. It now has a Mylar flap over it that stops the humidity system losing warmth. It swings out a bit when the cooling fans run.

David

Here is the coolant pump after a lot of paint stripping and painting. Note: the plastic impeller just unscrews, but left hand thread.
I replaced the ball bearings in the motor with new double sealed ones.
The pump shaft I linished on the lathe.
There was a coolant thrower on the shaft but it was just a rubber grommet that someone put there. I replaced it with a Vee ring seal. I smoothed the face of the pump casing so the lip of the seal can run on it. I've dry run it mounted on the tank. It runs so smooth I had to put my finger on the shaft to tell if it was running. If anyone knows where I can buy a small 415 Volt plug and socket for a 300 Watt motor so I can shorten the very long cable to it I'd like to hear.
The tank I had sand blasted with a bunch of other stuff. it was a nasty mess inside. It's leak testing in the photo, with a bit of detergent to lower the surface tension for leaks. It has a new level indicator as the old one was smashed. The cover plates blanking the holes for it were my first CNC part off the Maho.
There is a new linear grommet around the perimeter for the lid to sit on.
That little bucket on the side of it by the sink tap goes into the foot of the mill and catches the coolant returning to the tank. It means the tank must be put in that position which is awkward. I've placed some big plastic slides underneath it so I can slide it out for cleaning. It needed raising anyway as the whole mill is sitting up on rubber mounting pads and the side bucket wouldn't fit otherwise.
David

This is near the end of the saga. I thought I'd add a little on the lube oil mess these machines make. I'm using oil lamp wicks to help direct the bulk of the lost lube from the various slides etc.
Starting with the little moat at the base of the Z jacking screw: Much of the knee's lost oil ends up in the moat. On mine is unevenly cast and ends up with puddles of lube oil in it. To drain that oil, I've simply put a wick of the type used by old oil lamps in it. The end of it goes down the drain hole in the casting where it is caught in an old sandwich tray. I threaded that drain hole and put a bit of brass tubing down below for fear that the wick might get caught in the Z axis drive belt as you see. The tray is actually a bit full. I'd prefer that the wick wasn't touching the oil, not that it's important. The thing that makes it work is the head drop down the wick to that point. It will actually cause concave menisci in the wick laying in the moat from that suction head. So the puddles get sucked empty, not dry unfortunately.
Along the X axis on mine is a rather hopeless piece of sheet metal that is pretending to catch oil running down the face from the slide way above. I'll probably improve that at some point, but currently I've simply used gasket cement to seal against the casting face. It has drain hole at each end back into the casting but the oil from them drops on the shed floor out of other casting holes. It's not clear in the photo but I've  threaded the cross drillings in them and attached plastic tubing. The tubing arcs around back into the casting and drips onto the shield spring cover of the Z axis jack screw. But again, I've put a wick in them. In this case there is a bootlace that runs a bit along the tough formed by the bend in the sheet metal oil catcher, into it's drain port and some distance down the plastic tube. The same capillary action as above sucks the oil from the trough and deposits it down the tube. I can say that there is a lot of oiling on that spring shield. Cheap and cheerful and easy to do. Noticeably saves mopping the oil off the floor, but it's not a magician.
Also,unless you machine is very stable and level, that sheet metal trough will flood at one end. My whole machine is on rubber mounts so it can flood.

David

I had another look at my gear change system just to be sure I had things in a row.

First thing I should add is the relay board on mine is hardly fit for purpose. The 3rd PCB, that I removed as redundant, is clearly marked "LUBRATRONICS 4 ACHES". So it's for lubricating 4 axes, not gear changing in particular. On who knows what. I can see why the later Mahos have a much nicer looking board.

My tracing drawing didn't make much sense to me. Some while back, after a while thinking about it, it seems this is how mine works:

The Izumi PLC on mine never actuates 11d9 relay. I've just checked every gear change to make sure 11d9 never actuates. That means the forth cam on each change motor is bypassed as is the voltage dropping resistor.

The result is my gear change motors run full speed at 24 volts and full power. Note my gear change motors seem to be Bosch windscreen motors repurposed. So the later machines may have more suitable ones.

The redundant track notes on my drawing are incorrect. Sorry, I forgot to remove that. What happens is that the normally closed contacts on 11d5, 11d6, and 11d7 short circuit the motor winding when they aren't energized. That also means the motors are dynamically braked by the motor back EMF short circuiting through the motor winding when the the relays turn off. So the motors stop faster.

I guess the previous Philips controller had some strategy to use 11d9. But these boards were like that everywhere. I took me ages to figure out what was redundant and how whats left actually worked. A nightmare. The wiring diagram of the later board looks like Maho cleaned it up and made their own.

David

 

I've been doing a bit more on the gear changing and some of it may help those updating older machines to LinuxCNC.

For these older machines the spindle motor is above the main gearbox making the gearbox hard to inspect. Following the explanation below may mean you don't have to.

The gearbox isn't as complicated as it seems at first glance. Without going inside the gearbox here are a few things to know:

There are 3 selectors each with it's own motor: 11M1, 11M2, and 11M3.

Each selector has 3 gear positions. That should give 3 x 3 x 3 =  27 gear ratios.
But the selector on the output end of the gearbox, on the 11M1 selector motor, has neutral at it's middle position.
So you get 2 x 3 x 3 = 18 gears.
Neutral is on the output end of the gearbox, presumably to eliminate one gear set at the high torque end where the gears need to be bigger, heavier and more expensive rather than elsewhere.

Each selector has 5 cams on it's shaft to monitor where the selector is. I've numbered the cams from the gearbox outward.
Every cam has a single lobe on it that sets a micro-switch for the controller to decode.
The lobes on the first three cams cover a small angle and tell you if it's in gear.
So, if a micro-switch is on the lobe of cam 1, 2 or 3 then it's in gear... or neutral if it's on the lobe of cam 3 of 11M1.

For all three selectors:
Cam 1 lobe will be on the micro-switch when the selector is fully counter clockwise
Cam 2 lobe will be on the micro-switch when the selector is fully clockwise.
Cam 3 lobe will be on the micro-switch when the selector is somewhere in the middle. (Odd order, but maybe the German mind works differently to mine.)

Cam 4 lobe covers a wide angle from full CCW to just short of the middle gear position.
Cam 5 lobe covers a wide angle from full CC to just short of the middle gear position.

If the selector isn't in gear, cams 4 and 5 can tell you which side of the middle gear position it is. In my case the PLC doesn't use cam 4, as you can assume that is where the selector is if it isn't in a gear or on cam 5.

All 3 selectors work in this way, but the "middle" is somewhat off middle on some of them. Presumably the selectors inside the gearbox differ in geometry.

If you have the older type 28A1 board like mine then cam 4 is hard wired into the motor circuit. But on mine it is always bypassed. I don't know how the older Philips controllers used that. On mine I've simply removed it from the circuit along with the 3 big resistors that limit the gear change motor speeds and added a wire shorting the gear change motors to the buss.

If you are converting to LinuxCNC then I think you will need to lift the terminals off X3 209 to X3 220 and redirect them to the LinuxCNC interface board. On mine they have been shifted to the inputs on my Izumi PLC that does gear changes on my machine.

Similarly you will have to drive some relays on the 28A1 board to make the gear change happen. Note the relay names on these older boards aren't the same as the later Schrack boards, but the schematics are similar enough to figure out.

David

A slight error on the Izumi outputs sketch in the previous submission. The 11d2 relay energizes the spindle motor brake to release it , not de-energizes it. I've added the corrected schematic.
Here is a comparison of the various gear change schematics. First Marks from page 1, then my original schematic and then my modified one.
I've drawn the schematic with the common buss down below rather than at the top. But originally mine was very similar, but for the resistor near the gear change motor.
11d9 on mine never activates. The equivalent on Marks is 11K3 or maybe 11S14.
Anyway, the 11K4, 5 and 6 contacts on Marks are shown normally closed. In that state they short the motor terminals and dynamically brake motors if it hey are spinning.
On my machine 11d9 normally closed contacts always shorted the speed trimming resistors, equivalent to 11R1, 2 and 3 on Marks drawing. Mine has no resistors 11R4, 5 and 6. So you can see in the last drawing I have removed all those upper components and simply shorted the equivalent of Marks 302, 303 and 306 to 301. That allowed me to remove 6 cores out of the door cable tray which went to the switches on the 3 cam 4's on the gear change selectors and stow them. In addition I removed the wires energizing 11d9's relay coil, removed the 3 big resistors off the board and the 11d9 relay. You can see that on the photo in the previous post.
It may be that the heavier gearbox on this machine, plus the older style windscreen wiper motors on the selectors allowed me to remove the speed trimming resistors.
All up there are so many redundant components on these old boards it may be better to just buy one of the Schrack 28A1 boards made by Maho specifically for these machines rather than mess with these old ones.

David

in original,
the resistors had been switched in series of the motor, for speed reducing if the control wants to set the middle (center) position, to prevent for overshooting the middle (center)cam . At these days the plc cycle times had been around 50ms or more ...

The Philips controller on my machine was refitted with a Heidenhain TNC 155 controller and an Izumi PLC FA1 Junior for gear changing  around 1990 so they are a little faster.
Yes. I agree. The middle lobe is very short and could easily be over run. I had a quick look at the Izumi execution times for it and it looks like it is roughly 15ms per scan. So that may be why I can get away with removing the resistors.I gear changed a lot before I got the confidence to move those bits off the 28A1 board.
 I think that is why the guys commissioning it again wired the 11d9 relay to the Izumi where it was, to allow them to use the resistor if needed. But it seems that they didn't need to use it on my particular machine. It's nowhere in the ladder program and the gear change works nicely. It may be on these bigger machines the gear ratio has a bigger reduction and doesn't reflect the inertia to the motor as much? The rotating inertia reflects as the square of the gear ratio. I'm not sure if the older Bosch wiper motor they used is lower power and slower also.
But I agree to proceed with caution when commissioning any other Maho. There are myriad versions, all a bit different.They may need that those resistors. But go for the later 29A1 board if you can. Mine has masses of other redundant stuff on it that was difficult to debug and prove so.
Thanks for the reply.

David

Hi David,

Sorry, I had my eye off this thread for a while.

All my MAHO Gearbox knowledge is stored in the github repository with the COMP my friend wrote.
github.com/jin-eld/mh400e-linuxcnc/wiki

I would agree with Tommy. It is better to separate this excellent information into a separate MH700C thread, rather than burying it in page 164+ of a MH400E thread. Someone looking for this information would really struggle to find it located here.  I would suggest you start a new thread and cut / paste each of your excellent posts into it David, as you have done a real service to anyone else with a 700C.

It seens that Maho's were pretty rare down under. I live in Austria, so mine is closer to it's home, not mine
Cheers,
Mark

Hi Mark,
Thanks for the reply.
I came across your site by accident. It would have been helpful if I had found it earlier, but that's how it goes.
I've never posted anything like this before so I didn't have a clue what I was doing. The LinuxCNC site doesn't seem to have a beginners help about getting started. I had a lot of trouble getting attachment to work to start with, like some others, even when they were much smaller than 1.6MB.
I started out thinking you were in Auckland because of some of the early comments about arriving there, and with the Kiwi accent I jumped to conclusions. Soon realized you were in Austria and fluent in German.Unlikely for a Kiwi.
I took me a couple of weeks to get through all the 164+ page and the many videos you've posted. So much that I started writing notes on what I could comment on. Then I just decided to post replies to things as I went so they were still fresh in my memory. In a way it was nice to get the whole lot out without interruption. Part of the inspiration was some of the guys attempting to upgrade older machines than mine an struggling with the wiring and whether to upgrade to LinuxCNC or stick with the Philips. For them I think it would be better to go with LinuxCNC, but I doubt they will have the good experience that you had. Mine needed a lot of wire tracing to understand. There were components dead and bad wiring. I disassembled almost everything mechanical and cleaned and fixed it. Most had something wrong. I can see why this machine changed hands often. I suspect it caused a fair amount of financial distress for it previous owners.
Having commissioned a lot of PLC's running machines during the 80's a key thing to know is that they are very dead until all the I/O is all correct and working properly. It can take a very long time before that happens, and then the machine bursts into life. it often takes a lot of persistence and patience to get to that point.
I was amazed to see a Maho identical to mine on your site, in Christchurch as I recall. I've never seen one the same before out of hundreds of MH-C 700's I've searched for online.
When I worked in Australia as an engineer I mostly came across Deckels. Plenty of them, but all non CNC. I hadn't found any Maho's until later when some of my mates introduced me to them. They are so similar I wouldn't know how to chose one or the other. Mine came up on eBay.
I think I've posted all I want to for now. but there is plenty more that I could help with, not necessarily to do with upgrading to LinuxCNC. But a lot of mechanical stuff that guys upgrading might need to know. That might be better done face to face individually on Skype so I pan a camera on the machine live. But I'm happy to answer questions from anyone.
My feeling is your site is like a magnet for guys messing with these machines, not just with LinuxCNC.
I'm a novice at setting up a separate site for this, But if you guys want to paste my stuff to where it would be more useful, go for it.

Cheers,
David