What's new
What's new

The "fix your Maho" thread

akb1212

Aluminum
Joined
Jun 26, 2009
Location
Norway, Vestfold
I'd like to start a thread with tips and tricks to get a Maho mill up and running. The idea is to collect them in one single thread if possible since it's not easy to find what you are searching for here through searching. Most of them is already in this forum, but it's spread out so much that it's dificult to find.
It is solely because of this forum here I'm now finally able to run my mill. And I have come across a fair amount of the normal problems experienced when trying to get it up and running. So I wanted to start off, but I'd like everybody who feels they have a useful hint or tip to contribute.

My mill is of 1990-91 vintage, so my knowledge is from that age machine. But I think most of the problems I have applies for both newer and older machines as well.

I think it will be a good idea to separate the different issues, so I plan on making one post per issue. Here is a list of topics I'd like to cover:

1. Constants. This could possibly even be presented in a thread of its own. A list of different machines constants and such could be given in that thread. That way we would have a library of constants to look at. But a quick intro on the basics of them should be in this thread as well.

2. Communication. How does it work, and what do you do when it doesn’t. A check list of how to get it working is also a good idea. Tips on what hardware to replace is also to be presented.

3. Monitor. It's not straight forward to replace the monitors on these machines because of their low refresh rate. An easy way to fix this is to use a converter. Hints on where to get them and how to use them is next tip.

4. E-stop chain. Hints on how to get a machine out of e-stop. Including how to bypass the ****ing door safety system.

5. Drives. There are really good hints in this forum on how to get a drive working again, and what is their weak points. This is a common problem for these machines as I have learned. So it will be probably be useful for someone with drive problems stopping them from using their machine. Even if it's not a problem now it will probably become a problem in the future as the caps in these do fail quite reliably.....

6. How to get the table up if it is at the bottom. There is an easy way to do this by changing constants only. Doing it mechanically is a nightmare on the y-axis because of access. So it's much more efficient to do it by the constants. And it's not straight forward because of the software limits. A detailed description on how to do this will be presented.

7. Tool changer. An easy way to reset it after something happens during tool change.

8. Electrical schematics. I have made an overview of what component does what in an excel sheet. This has proven to be very helpful during fault searching. This is very machine specific. But there is a system to it, and it might be useful for other machines as well. And it could easily be updated for your own machine if you have the electrical schematics.

9. Hydraulics. Hints and tips on how to know if your hydraulic accumulator is working as it should. And some experiences from fixing the valves.

10. Retrofitting. This is something I haven't done yet. And it is probably also something that I will start a new thread for. The idea is to ditch the 432 controller only, and leave the rest as it is. The stone age interface and lack of hand pendant, proper override controls and limited memory (with the need for DNC) along with only RS232 communication and stuff like that is getting on my nerve. I'd rather not spend years to learn an aincent system like that. I'd rather spend the effort on replacing it, even with one with only similar capabilities as far as machining goes. Replacing it this way should be easy enough, the hard part being back engineering the ladder logic needed to replicate some of the logic functions. All hardware interfacing are easily accomplished with Mesa hardware in LinuxCNC. But more on this in a separate thread when I get around to start that work.


If anyone feels like they have a hint or tip that would be a good thing for someone trying to get a Maho mill up and running please contribute. But I'd like to keep this thread a quick and easy place for accessing info needed for this. So please try to sick to the topic and present hints and tips only in this thread to keep it clean and easy to search through.

Anders
 
Last edited:
3. Monitor. It's not straight forward to replace the monitors on these machines because of their low refresh rate. An easy way to fix this is to use a converter. Hints on where to get them and how to use them is next tip.
The horizontal frequency output is what matters. Typically a "converter" is a frequency doubler, which will only work on the really old controls where the horizontal rate is so low that doubling it works out about right for todays LCD monitors. As you get newer than say 1987, the horiz frequency gets a bit faster such that a doubler will get you too much horiz rate and the result will be a blank screen on new monitor. In that case the best solution is to buy a multi frequency LCD from Australia that is will work with various horiz rates. Cost about $425 ...plus UPS shipping another $100.

http://www.converters.tv/products/vga_to_vga/849.html

The Au monitors come with standard VGA connector but also have separate wires for Red, Blue, Green, Horiz, Vert, Ground so the only trick is to figure out which wires on your Maho are each of those wires. I figured it out years ago and have the info somewhere... but seems like someone posted that info here on PM in the past.

maho81.jpg


maho80.jpg


10. Retrofitting. This is something I haven't done yet. And it is probably also something that I will start a new thread for. But it might be useful to have some hints on what is needed in a thread like this. The plan is to replace only the 432 controller and use the rest of the components. This is what the limiting factor is for me anyway.
Did you get the TV show "Fantasy Island" in Norway ? If so...welcome to it.. ;)
 
Ok, you jump ahead to 3 already.

Milacron, first off all you are wrong about the capabilities of such a converter. Secondly you are wrong about the age at which Maho started to use "proper" frequencies on their monitors. From what I know all Philips 432 controllers use CGA frequencies (to low for any LCD not equipped with a scaler). At least mine does, and it's from 1990. This might have changed on the 532 controllers, I don't know. But for backwards compatibility I assume all 432 controllers use low frequency monitors.

In addition to that my intended audience of this thread is people buying a mill of vintage of maybe early to mid 1980 to maybe mid 1990. This audience will most probably be on a budget. If not then why buy an old mill like that in the first place? Let alone one with issues....
So shelling out way more than $500 for a replacement monitor is not what I recommend. At least not when I know there are solutions that works just as well for less than $50. The one I ended up buying on eBay cost me less than $50 with shipping to Norway. So a tenth of your solution.....

Anyway, the converters I'm referring to are quite capable of converting whatever you feed them with and put out a normal VGA signal. The intended use for them is to replace the old monitors of old arcade games and such. And they are made so they are capable of converting most standards used way back then. And they are also easy to use. On eBay search for "CGA to VGA".

The one I bought gave me some problems. I found out that the h-sync input didn't work. It might have been because I fried that input or something. But I solved the problem by using an AND gate on the h-sync and v-sync to add them together and feeding that signal to the working sync input. That did the trick, and I have now perfect picture on my standard LCD screen. And because it's TTL level on the signals this is very easy. I just took the chip itself and used 5V from the converter to power it. I then soldered the wires directly on to the legs of the chip and that was it.
But due to my initial problems I'm not too keen on recommending one model in particular. But in general there are plenty models to choose from that should do the trick. The frequencies used by Maho are well within what all the available converters can handle. So I wouldn’t expect any problem no matter what model you use. I think it is CGA standard with TTL level. There are higher priced ones with casing that are made to industrial standard if that is preferred.

LCD screens are quite cheep these days, so I'm sure most of you are able to get one for almost noting. I ended up using a 17" LCD that I got for free from a friend. The ones with low native resolution (old ones) is probably best for this application as well. I removed the casing and it ended up fitting perfectly in the console with some modification and elbow grease. I had to modify the frame to remove parts of it that was made to accept the curved CRT screen. But I ended up with a much cleaner looking console, and a screen that filled the whole frame perfectly. I'm not sure its possible to integrated the monitor mentioned by Milacron. And I don't think I would risk doing experimenting like that with it either, it being so expensive.

So with regards to getting the monitor to work I'm not in favor of doing the expensive but safe solution. Save the money and use it in other places where it's more needed. The cheap solution works just as well here!

Anders
 
Did you get the TV show "Fantasy Island" in Norway ? If so...welcome to it.. ;)

you and your fanastasies.

:P

i think theres a split here. group 1 wants to restore the machine to its former glory... mach and emc just dont cut it. its a serious downgrade asside from the memory usage, and it is a big task to hook in even the simplest maho.

but then theres group 2, who wants to take a bargain machine and make it work. not work like a maho with .0002" absolute precision and all its features, but just work. 3 axes, a spindle, and not much else is needed.

i think this shouldnt be discouraged as much as it is. we may end up learing something. :)

i mean, its just 3/4 axes, and a spindle..... and a few HUNDRED relays that need to be engaged with the right timing, all controlled by the PC control directly.

my strategy (when i have time to get back to the machine and fix the electrical bugs) is to "upgrade" the control, not retrofit it. ive found memory chips that will bring it to 384k, which should make it very useful. ill do the LCD thing as well, BUT with a mini pc inside the control with a kvm and serial connection to the 432. this will let me sit in front of the machine with a propper UI to edit programs and do other tasks. ive also seen processor upgrades that more than double the speed, but i have suspiscions these dont work plug an play.

of course, my 432 control "works", and my setups is vastly more complex than some other mahos, so keeping the 432 is my best choice if i can. people with an outright broken control have a more complicated decision.

:)
 
mach and emc just dont cut it. its a serious downgrade asside from the memory usage,

Won't comment on Mach. But LinuxCNC is an upgrade, not a downgrade. Simultaneous handling of all axes. Quite good extensions for the G-code interpreter. Just a few cycles are missing (but said to come). Spindle orientation and pendant: No problem.

but then theres group 2, who wants to take a bargain machine and make it work. not work like a maho with .0002" absolute precision and all its features, but just work.

My glass scales do have a resolution of 0.5 µm. And LinuxCNC does read them without skipping any graduation and does position there (well, I do have configured a deadband of 1 *µm).

i mean, its just 3/4 axes, and a spindle..... and a few HUNDRED relays that need to be engaged with the right timing, all controlled by the PC control directly.
Hundred? I just counted them, and they were 28. I guess half of them are for the E-stop chain, and I wouldn't replace them with a PLC.

Nick
 
Milacron, first off all you are wrong about the capabilities of such a converter. Secondly you are wrong about the age at which Maho started to use "proper" frequencies on their monitors. From what I know all Philips 432 controllers use CGA frequencies (to low for any LCD not equipped with a scaler). At least mine does, and it's from 1990. This might have changed on the 532 controllers, I don't know. But for backwards compatibility I assume all 432 controllers use low frequency monitors.
It is possible I am wrong about converter capabilities NOW but I was quite right four years ago when I looked into this extensively and in fact bought a few converters myself. But perhaps now there are some on the market with conversion possiblities other than 2X. Someone buy one for a machine CNC control outputing 30kHz horizontal and prove me wrong. You speak of "CGA frequencies" but there is a range there. I need to know the exact number. If you have a CNC control that outputs 28kHz horizontal and the monitor goes down to 29kHz all you will see it a blank screen. But if the monitor goes to 28kHz it will work perfectly. The numbers are that critical. You sound like someone touting theory, whereas I have extensive real world experience in getting these conversions done.

Regarding the age at which Maho started to up their horizontal frequencies beyond 15kHz I can't say exactly. All I know is I have worked on probably a hundred different CNC machines since year 2000 and all of the machines where I checked horizontal frequency there was a tendency for that frequency to increase a bit each year up until about 1996 or so. Same with the desktop monitors....sometimes I could match an older LCD with an older CNC and not need the Australian monitor. Such was the case with a 1989 Deckel FP2A with Dialog 11 where there was one standard monitor left in the marketplace that would go low enough horiz to work with no conversion at all....but only one monitor would do it. (which I now forget the make and model of, but it is mentioned somewhere in past PM chatter on the subject)

On the subject of CNC control conversions on Deckel or Maho's.. this subject has come up over and over many times since PM started in December 1999. It can be done with Heidenhain or Fagor industrial controls. Deckel Doctor and Wesley have done it with Heidenhain, plus an obscure fellow whose name escapes me has done numerous Maho Fagor conversions. But doing so with a PC control I don't know a single example...not one. Many dream about it, some start it, but no one actually does it to finality.

deckel79.jpg
 
Won't comment on Mach. But LinuxCNC is an upgrade, not a downgrade. Simultaneous handling of all axes. Quite good extensions for the G-code interpreter. Just a few cycles are missing (but said to come). Spindle orientation and pendant: No problem.

My glass scales do have a resolution of 0.5 µm. And LinuxCNC does read them without skipping any graduation and does position there (well, I do have configured a deadband of 1 *µm).

Hundred? I just counted them, and they were 28. I guess half of them are for the E-stop chain, and I wouldn't replace them with a PLC.

Nick

on the scales... i wasnt sure there was an easy way for emc to read them. the ones i have are sine wave, and few of the low cost servo boards will read them at all. what i meant for group 2 was that if the scales dont work... no big deal. any basic function is acceptable. i know YOU did manage to get alot of the features working on your machine, but i expect thats not a simple task for most people with a beat up bargain maho.

on the control, i suppose its going to depend on what maho you have, and what year it was from. in my case, i do indeed have hundreds of relays, and over 100 IO off the control. given the features i have in the 432, and its performance, emc doesnt seem to offer anything to me other than more meory for programs. on an older simpler maho, perhaps emc would add something substantial.

but i think you understand my point... alot of people will be quite content with even the most basic functions and performance and shouldnt be discouraged from retrofitting to pc controls.
 
My thoughts on how to retrofit my 600E to LinuxCNC

Ohhhhhh....

So it ended up in another "there is no point in "upgrading" a mill that old and complicated" thread after all....

I was shooting for a clean "this is the list of issues you might need to address and this is how to do it" thread. I will start that over again in an attempt of making it a clean and easy to use guide I wanted to make.

And I'm not going to even attempt convincing Milacron and you guys who think installing LinuxCNC is a downgrade that will be a major pain in the *** to undertake.

But for you guys who appreciate where I come from I can say this is my plan;

This is an excel sheet I made with ALL components involved in the logic of the mill. It states what they do, where they are connected and what signal is controlling each of the 42 relays/contactors (not hundreds then....). It also has all switches with what they indicate and how they are hooked up.
It also has all input and output of the controller fully mapped with function and where it comes from or goes. The first reason for making it was because I was getting tired of flipping through the electrical schematics. Then I also needed to map all the IO on the controller.

I had intended to post in here in this thread as a help to other trying to fix their machines as well as those looking in to retrofitting like I'd like to do. I intended to clean it up a bit first before I posted it here, but that is a lot of work so I present it here so you can see for yourselves. This will also give you an indication on what is actually needed for a retrofit, and hopefully see that it's not that may things in there that can’t be done with some work.

View attachment 83799

A quick explanation on the spreadsheet:
Everything is color coded. Each color is from the numbering of the components. Maho isn't quite consistent with this so there are some inconsistencies in the list. But this is because of Maho not me.
Sheet 1: Mapping IO of the controller and the KDA
Sheet 4: Overview off all components I have found involved in the operation of the machine. Drawing page is only relevant for my schematic. First number in column C is the page where the component is controlled, the rest of the numbers is where the component is controlling other components (or is not connected to anything but shown in the drawing anyway).
Sheet 3: Redundant, KDA mapping also on sheet 4.
Sheet 2: Some translated words in the schematics. Got tired of looking it up.
Ark 1: Colors of the encoder cables I had to change around on when switching controller. There are different ways to hook up the encoders. I was planning on making a separate post on this in this guide thread.

I have spent at least several hundred hours poking around inside the machine while getting it up and running now, possibly well over thousand. And this spreadsheet has been invaluable in that work. It also mean I have done a lot of quality checking of the spreadsheet. I can’t guaranty it, but I'm quite confident it's quite correct now.

I will clean it up a bit and post an updated version when I get around doing it. I just wanted to present it here now so you guys can see what it will take to do a retrofit. Hopefully you will see that it's all just a matter of being systematic about it.

So even if a retrofit can seem like a daunting task to undertake it doesn't look that bad when you break it down. Each and every logical function is easy to follow and rewire for a retrofit. In fact my plan was to put the same connectors (35 pin D-sub) on the computer to interface the digital IO. That way it's just a matter of swapping back and forth between the controllers as needed. And it's always possible to fall back to the original controller if the retrofit doesn't work.

That was the digital IO, so then there are the encoders and linear scales:

I know there are hard to interface 10uA sin encoders on the scales. But luckily the controller already has an EXE card that handles that beautifully. All I have to do is to reuse that card and interface to that and I'm good to go. The encoder on the spindle is a normal TTL level encoder I can connect directly to the Mesa card. So no worries there.

And then there is the interfacing to the servo motors:

I know these have Hall Effect sensors, and a taco that could be a challenge to interface. But I'm not planning on replacing the servo drives. So all I will need to provide is a +-10V analog signal to the drive (along with the appropriate enable signals). And that is easily done by the Mesa hardware. So the system will have 2 servo loops. The inner loop will be kept original using the present servos and drivers. The outer loop will be moved from the 432 controller and taken over by the LinuxCNC controller. This will take it's feedback from the scales as previously, making it an easy and clean retrofit.
A future upgrade would possibly be to get 3 off 8i20 amps from Mesa as well, and use these to run the servos. This could either be done by replicating all the interfaces (hall sensor and taco) and reuse the original servo motors. An alternative would be to use constant speed mode instead of constant torque (or was it the opposite way around, anyway it’s possible to get away with it). In any case it would mean braking up the inner servo loop and have only one servo loop like modern controllers have. Another alternative would be to sell the servo motors on eBay (along with the drivers) and use this money to buy new and more powerful servo motors. The 8i20 can drive 2200W motors, so there is room for great improvements here. But the beauty is that this can be done at a later stage at an appropriate time. And it should be possible to do it one axis at a time if so desired.

The only interfaces on the controller not covered now are the monitor, keyboard and the RS-232. But I trust you understand THAT will not be a problem to replace.

Then there are the upsides: The possibility to attach a pendant (with a hand wheel!), analogue override knobs, network, touch screen (I have a RS-232 type that should work in Linux already, actually I have 3 of them) and other modern features. Another thing is the increased memory, meaning no more need for dodgy drip feed over RS-232.

So that concludes what is needed for the retrofit I'm doing, at least as a start. With this setup I can always expand. And that is probably one of the biggest reasons for wanting to do this. I'd much like to add a 4th axis. But there is no way I'm going the Indramat route! I would also need to add an interface card to the 432 controller if I were to add a 4th axis without doing a retrofit. All in all a route I'm not keen on taking. Let alone the reduced functionality I'd get compared to LinuxCNC.

Some of you might think this is a downgrade because one would put a lot of effort in to it and not improve on it by doing it this way. The fact is that this retrofit will be more like an update to get a decent user interface and enabling me to add features. As the mill is now there are no way I'd like to spend the time and money to add the features I'd like using that out of date expensive technology.
By doing a retrofit like I'm planning I'll be able to get a decent user interface and the possibility to add features Id very much like to have.

And for the record, it's not that I think the 432 controller isn't capable. It’s the fact that it is as capable as it is I don't see the need to do more than copying its functionality. But most importantly I do it to get a user interface that isn't (seemingly) pre DOS style.

It's also the fact that the 432 controller is as capable as it is that makes it as difficult and cumbersome to operate as well. Modern machines doesn't add that much more functionality. It’s mainly giving us a better and more efficient interface and an easier and more intuitive way to access the capabilities we have. And this is what I want. I don’t want to spend the next few years learning an ancient interface good enough to do some decent work. This knowledge will not be useable on any other machines anyway, so it's not time well spent learning it. And I have already done my homework and figured out what it takes to replace only the needed parts to achieve this goal. And I expect to have to do less work to do this retrofit than to become a decent operator of the 432 control system. And knowledge on how to do PLS programming is way more useable as a skill compared to being a skilled 432 operator. I will even have use for PLS programming skills at work. And LinuxCNC is a PLS, and is programmed like one.
Even another issue is the dialect of G-code the 432 is using. I started looking in to G-code wizard as g-code verifier. And I discovered that the g-codes used by the 432 controller are quite different from the ones used by most other machines of today. That is another job I’d like to avoid, having to customize whatever CAM software and post processors available. CAM software in my price range doesn’t have anything like Maho style posts.... LinuxCNC ones on the other hand isn’t a problem as it is quite popular in hobby style machines, and use a more normalized dialect of g-code.

The clue here is what that goal is. And it is first of all to avoid having to learn and then have to live with a totally outdated and seriously hampered user interface. I’d like to be able to utilize my machine to do what it is capable of using a modern and powerful interface, still having the same physical capabilities it already has. And at the same time opening up for adding more capabilities without having to struggle with old, expensive and hard to get controller and driver hardware.
Another factor would be that I’m an electronics engineer by trade, and do know my way around this. I’m sure this helps quite a bit as well. This and the fact that I have spent the time I have fixing this machine make me quite confident that I’m capable of doing this. And by making it a swap out replacement with the old controller means I’m able to switch back and forth as I need, never making my machine useless at any point. Thus also making everything totally reversible.
But I do challenge anybody who bothered to read through all of this to point out issues they see could prevent this task from being possible. I have put a lot of effort in to this, but welcome all relevant info that would prove me wrong. Most of the effort put in so far is to have the machine working as original. But I will soon go on to spend a lot of time in an effort to retrofit it. And if anyone can point out serious and clear to the point technical reasons to why my scenario won’t work I would welcome them at this time. It would then save me a lot of work.
 
on the scales... i wasnt sure there was an easy way for emc to read them.

The Philips scales? No. I thought about developing a decoder, but buying Sony Magnescales was way cheaper. For the original scales, I made a deal and got the bellows for the vertical axis, brand new from the factory.

what i meant for group 2 was that if the scales dont work... no big deal.

You suggest to replace the servos with steppers?
Or rip out the tach and replace it with an encoder. Only to discover that now you need three new servo amps? Only to discover that you won't find ones for that voltage.

Nick
 
But luckily the controller already has an EXE card that handles that beautifully. All I have to do is to reuse that card and interface to that and I'm good to go.

OH! Why didn't I get that idea? :(
I have been looking at the schematics of that part finding out how it works. But I didn't realize that I could recycle that part.
Too late!

Nick
 
Interfacing linear scales on old Maho mils (Heidenhain analogue sine scales)

The Philips scales? No. I thought about developing a decoder, but buying Sony Magnescales was way cheaper. For the original scales, I made a deal and got the bellows for the vertical axis, brand new from the factory.
Nick

Why spend effort on making something you can easily buy existing products? The EXE "pulse shaper" cards do exactly what you want, and they aren’t that expensive on eBay? Did a quick search but didn't find any now. But I know I have seen them for a few hundred $. These are well built units from Heidenhain, so no worries with regards to reliability or accuracy. I already have one for all 3 axes so for me it's a no brainer. Did the new scales you bought have sub micrometer resolution? If not then you missed out by replacing them.

Just to be a 100% sure I went down in my garage just now to verify that this is what they do, and also verify the translated resolution they give. And sure enough, they gave me a nice and clean 5V 1kHz square wave when I set it to run at 100mm/min. That means it is giving me a resolution of 0.001 mm for each of the two outputs. In quadrature this gives a resolution of 1/4 of that, 0.00025mm or 0.25um.
So if I divide this resolution by a factor of 10 I should be able to get a repeatability of 2.5um, or 0.0025mm. That is with great comfort a great deal better than I expect to be able to get in terms of accuracy on my finished parts.
Initial tests with the present system are indicating that this is well within reason. I was able to move the y-axis what seemed to be quite exactly 0.01mm when using my 3D Taster (one whole division on the analogue dial). And it was consistently repeatable to what I precive to be well within the expected 2.5um. I really love that tool BTW. Perfect for aligning work pieces and such.

So I don't have any concerns in regards to interfacing a computer controller to the linear scales.
I also happen to have 2 linear scales that are sitting in a Microset (DMG) tool setter that I got for free from work. I know the light bulbs in the encoders fail eventually. So at that point I’m willing to sacrifice a scale from that to get my mill to work again if needed. Or only the encoder part if the length of the scale is off. For now I'm planning on using it as a tool setter. I only need to get an ISO-50 to ISO-40 converter. Too bad they are so ridiculously expensive (for the simple part they are). I will probably try to make my own at some point. I have a circular grinder.

As far as replacing the drivers and servo motors, I'm not considering doing that until I have everything working with the old drivers and servo motors on the new controller. That way it's possible to avoid having to deal with multiple problems at once. But when I replace them I'm not even considering replacing only one of them (either only servo drive or only servo motor). I’m expecting it to be a lot of work involved in trying to make the servo motors in the machine work with a new driver. And the price the original motors and drives fetch on eBay dictates it's better to sell them as original spare parts and replace them both with something that is known to work together and is newer.
I’m also considering attaching a 4th axis first and use that servo system as development system to find a good system that integrates well in the computer controller. It will then be a simple matter of replacing one axis servo system at a time as a whole with the exact same driver and motor until all of them are done.

Anders
 
You suggest to replace the servos with steppers?
Or rip out the tach and replace it with an encoder. Only to discover that now you need three new servo amps? Only to discover that you won't find ones for that voltage.

Nick

i mean "anything that works", steppers or new servos and drives included. or using a rotary encoder to feed back to the new control, while keeping the tachos and old drives. many different ways with different levels of acceptability to different people.
 
OH! Why didn't I get that idea? :(
I have been looking at the schematics of that part finding out how it works. But I didn't realize that I could recycle that part.
Too late!

Nick


i had that idea to reuse those boards after reading a philips whitepaper on the scales. obviously a computer with an ISA slot would connect right to it, but wirting a driver to talk to it is beyond me. i also had the idea that one could simply "port" the OS to run on a modern computer but again, id have no idea where to start.

akb1212, dont misunderstand me. im on YOUR side, not milcrons. ive just decided that on MY machine, im better off using it as is, with a memory upgrade and if possible, faster cpu. my machine is different than most mahos. it genuinely does have 100+ IO (106 i think from memory). its got an atc, pallet changer, tool break camera, an insane spindle orientation mechanism, renishaw probe, air powered doors, and alot more stuff. its also very fast. in fact, if it had a hard drive, it would not be any worse control-wise than a brand new haas HMC and be superior in every other way.

EMC though... no thanks. ive run it on my simpler cnc machines. its extremely poor contouring performance makes it a no go for me. ive been looking at solutions like the kflop instead if i were to do a control replacement. ive also looked at some industrial replacements. most are too restrictive for my specific machine, and others (like the itnc530) cost a fortune.
 
Why a scaler is best for converting monitor signals

It is possible I am wrong about converter capabilities NOW but I was quite right four years ago when I looked into this extensively and in fact bought a few converters myself. But perhaps now there are some on the market with conversion possiblities other than 2X. Someone buy one for a machine CNC control outputing 30kHz horizontal and prove me wrong. You speak of "CGA frequencies" but there is a range there. I need to know the exact number. If you have a CNC control that outputs 28kHz horizontal and the monitor goes down to 29kHz all you will see it a blank screen. But if the monitor goes to 28kHz it will work perfectly. The numbers are that critical. You sound like someone touting theory, whereas I have extensive real world experience in getting these conversions done.

This IS my field, I'm an electronics engineer since 1996. I have worked as a service engineer at a national TV provider here in Norway for 7 years and after that as responsible for Media center in Norway for Microsoft for 3.5 years after that. So this is a field I'm quite confident in.

The converter I'm talking about here isn't purely a frequency doubler as such. It’s a scaler. What that is doing is scanning in the picture on the input and digitizing it to make a bitmap of each frame. It's then using that bitmap to make a new picture in whatever resolution and refresh rate you like! So it's just a matter of deciding what frequency ranges you think you need on the input, and make the circuits on the inputs with the correct bandwidth. The output is a straight forward VGA card with the normal resolution and frequency ranges.

In fact this IS also how a normal LCD screen is working. After all most normal LCD screens are capable of showing resolutions and frequencies outside its native resolution. They are also digital by nature, so they need an A/D on the input to make bitmaps to be fed to the digital input of the LCD display. T
his is also why you should always try to use the native resolution on your LCD monitors. If you don't you will end up trying to fit an odd resolution picture on a grid that doesn't correspond. This makes the picture blurry and less sharp.

But the thing is that the designers of LCD screens don't see the need to be able to view the low refresh rates and those low resolutions we need. After all there are no one else except us trying to use stone age old equipment that uses that low refresh rates and resolutions. So they have long gone abandoned the frequency ranges we use.

It's was also a period where frequency increased higher and higher. It's quite simply because higher resolution pictures and higher refresh rates looks better. And as technology evolved the frequencies of monitors kept increasing. And since the lower frequency ranges gave a poorer picture the manufacturers of monitors started to abandon the lower frequency ranges as they were deemed unusable and undesired by users. A higher frequency range is harder (and cost more) to make, so as a consequence it was skipped.

But then came the retro age, and people started trying to get the old arcade games back in to working condition. They also faced the same problems we do. These old games use the same low refresh rates and resolution our machines do. And to our luck the Chinese saw this as an opportunity to earn money. This is a considerably bigger market, so we benefit from that.
They started to make cheap scalers that made it easy to connect the outputs of these old game consoles to a modern LCD screen with only higher frequency capabilities. You can see the result on eBay by searching for "CGA to VGA". You will find more than 100 offers for units costing next to nothing ($30 to $40 including shipping from china, which is amazingly cheap for what we get), fully capable of interfacing whatever odd monitor output you could think of. These units are made to be as useable as possible. So they are made with a really large frequency range, ranging lower than we would ever need, all up to a point where we no longer need it in order to interface it to a normal LCD monitor.

There are "professional" or "industrial" units amongst the offerings on eBay, promising better quality and reliability than the cheap ones. And the decision on whether to go for one of these instead of one of the cheap ones is up to each user. Personally I wouldn't even consider it, and if reliability just buy 1 or even 2 spares of the same unit in case it breaks down. It will then be a matter of replacing it in a matter of minutes (depending on accessibility). But on the industrial ones you get a proper casing which you don't get on the cheap ones.

So I hope this is enough to convince you that this is an advice I'm able to stand by and say you will absolutely benefit from listening to. There aren't many other issues here we can say we have a cheap, easy and quick fix for. But for this issue we do.

Anders
 
Ohhhhhh....

So it ended up in another "there is no point in "upgrading" a mill that old and complicated" thread after all....

I was shooting for a clean "this is the list of issues you might need to address and this is how to do it" thread. I will start that over again in an attempt of making it a clean and easy to use guide I wanted to make.

.

So you have posted your ideas on an international internet forum but dont expect anyone else to offer their opinion or experience? I am sure you are proud of your accomplishments and I know your experiences will be useful to others but there are other ideas and experiences that are just as valid as yours. Any compilation would be incomplete without a discerning opinion. It is up to each of us to decide for ourselves what information is pertinent.

If you continue in this arrogant and selfish path you will find few who will care what you have to say. If you truly want to share information on a forum and have useful discussion, then be prepared to accept the good, bad and occasionally ugly. If you seek self gratification then publish a book.

Charles
 
So I hope this is enough to convince you that this is an advice I'm able to stand by and say you will absolutely benefit from listening to. There aren't many other issues here we can say we have a cheap, easy and quick fix for. But for this issue we do.
I'll be convinced when I see a modern desktop LCD/TFT displaying text from a CNC control that outputs a horizontal frequency too low for that modern LCD but too high to work if at 2X the frequency. It probably can be done but until I see it happen there will be some doubt.

Also consider there is some PITA mechanical factor when stuffing a desktop LCD with separate menu button board into a control meant for CRT when using desktop versus the Australian versions which come with frames that have screw down points and integral menu buttons on the backside.

Using the cheap desktop LCD can be done, as I have obviously done it on the Deckel D11 control..but it does involve a bit more labor to remove the "guts" from the plastic case and figure out how to mount it, and the menu board.

Having said that, you do have some good ideas for a PC based CNC Maho or Deckel retrofit, but we are a bit jaded here on that subject as so many and tried and given up in the past. There have been some brilliant engineers on here, one of which even went so far as to attempt to duplicate an entire Deckel Dialog 11 module brand new. As per usual after his grand details were laid before us we never heard from him again. (the supreme irony of which is a year or so later I had the exact same problem as him, same D11, same module, and figured out what was wrong with his original board but he was no where to be found to tell !! )

But maybe you will be the first on the Maho PC retrofit... I sincerely hope so.
 
Is LinuxCNC a bad choice for retrofitting a Maho?

i had that idea to reuse those boards after reading a philips whitepaper on the scales. obviously a computer with an ISA slot would connect right to it, but wirting a driver to talk to it is beyond me. i also had the idea that one could simply "port" the OS to run on a modern computer but again, id have no idea where to start.

That is a really easy thing to do. It's giving out a standard TTL quadrature output, compatible with most controllers out there. There are many readymade solutions to do this available. It's just a matter of finding them.

akb1212, dont misunderstand me. im on YOUR side, not milcrons. ive just decided that on MY machine, im better off using it as is, with a memory upgrade and if possible, faster cpu. my machine is different than most mahos. it genuinely does have 100+ IO (106 i think from memory). its got an atc, pallet changer, tool break camera, an insane spindle orientation mechanism, renishaw probe, air powered doors, and alot more stuff. its also very fast. in fact, if it had a hard drive, it would not be any worse control-wise than a brand new haas HMC and be superior in every other way.

With that kind of machine I absolutely understand your reluctance towards retrofitting that machine. It's probably also more important to you as a working machine compared to what mine is to me. For now it's just a hobby machine not tied to any economical investments other than I have already written off everything I have put in to it already. And given the complexity and level of equipment you already have on it I have no problems understanding your decision here. I would probably also come to the same conclusion if I were in your position. It's the "retrofitting a Maho is per se impossible for any non professional no matter what" attitude I don't buy.
It's also the level of understanding of the machine and how it's actually working I have come to get from fixing it that I look in to retrofitting it. I had hoped not to have to put that much effort in to getting up up and running, but that is how it ended up being. And to my benefit that lead to me now thinking I have what knowledge it takes to get this controller replaced.


EMC though... no thanks. ive run it on my simpler cnc machines. its extremely poor contouring performance makes it a no go for me. ive been looking at solutions like the kflop instead if i were to do a control replacement. ive also looked at some industrial replacements. most are too restrictive for my specific machine, and others (like the itnc530) cost a fortune.

Now that comment was not something I like to hear. Not because it's offending me, more like worrying me since it's probably my only choice as far as system to use for retrofitting my mill. Much like you I'm not even considering the high-end stuff due to price. But I'm curious as to WHY LinuxCNC (EMC) was so bad. I'd like to hear more about this and why it had so poor performance.
First off I'd like to know if this was a stepper or servo system. Then I'd like to know if it was using hardware (such as Mesa like I intend to use) to do the interfacing, or if it was using parport for communication.
And in what way did it have poor performance? Did it slow down because it couldn't process enough lines per second, and slowed down due to lack of processing power? Or was it limited because of inferior processing powers of the software, and ended up performing badly because of this?
There could be many causes for this, so I'd be really interested in knowing if this is an issue with LinuxCNC (EMC) as such, or if it is a result of your implementation in hardware. From what I have read there should be no reason to expect a decrease in performance due to LinuxCNC compared to the 432 controller.
It could also be an issue with how your CAM software generates the code. If it is spitting out thousands of small 0.01 mm moves and expect the mill to process thousands of these instructions per second there is an explanation there. Modern CAM software might expect to be interfaced to a modern fast controller fully able to handle this. But I don't think a small mill with a simple EMC setup should be expected to handle that.

But if you are correct and LinuxCNC won't be able to perform as well as my present 432 controller then I need to look for something else. And I do know a guy around here who is trying to sell a system that should do the trick that would possibly be within my price range. But to make that decision I'd really need to know WHY LinuxCNC will ever be able to perform on my mill.

My plan was to use Mesa hardware for everything in regards to IO. So as far as I know there should be no reason why LinuxCNC shouldn't be able to do contouring just like the 432 controller can. After all the processor in there is SERIOUSLY outperformed by ANY computer I'd even consider using. Mine has a 16 MHz 486 processor on it. How on earth can that be able to outperform a LinuxCNC system with all hardware controlled IO and a processor that is more than 100 TIMES as fast? I'm not considering LinuxCNC when the paraport is used as IO as that will seriously ruin performance. I'm talking about when all IO is handled with dedicated hardware (Mesa in my case) that will not steal performance by swamping down the processor while reading IO signals.
I know the servo loop is set as standard to about 1 kHz refresh rate. And on a mill like mine I can't ever imagine 1 kHz being too slow for it to be able to move around exactly as fast and accurate as it ever can regardless of controller. But if it is it should be possible to increase the frequency of this rate to solve problems caused by it.
My machine has a max feed of 4 meter/min. So I don't see how an increased servo loop update frequency could improve contouring rate in a serious way. On a machine capable of 10 or 20 times that speed it's something else.

Does anyone know what servo look refresh rate the 432 controller is using? That would be interesting to know as a reference.
Come to think of it I happen to know that the output to the servo motors from the drives is a 2.5 kHz square wave. This means that there is absolutely no way the controller will ever need to run the servo loop faster than twice that, 5 kHz. Even if it did run faster it wouldn't be able to affect the trajectory of the servo because there are no transitions on the square wave to be altered more often that about 5kHz. But given the CPU frequency of the system it's safe to assume the actual update frequency is probably lower that that. So in this regard I'm confident enough to say that LinuxCNC should be at least equal to the 432 controller as far as the rate at which the servo drivers are updated.

But I do hope (from my point of view) that it's due to bad hardware implementation you were experiencing performance issues. If not the LinuxCNC team should realize they have a serious problem.

So I hope you are able to share some more detail as of why you experienced the performance you did. It would be really valuable to anyone contemplating retrofitting a mill with LinuxCNC. After all it's a better option than MACH3 (given that it isn't capable of interfacing to the number of IO needed for one thing). Other than that there are no other viable options except expensive hardware based stuff that I haven't offered much attention to due to the price of such a system.

Anders
 
So you have posted your ideas on an international internet forum but dont expect anyone else to offer their opinion or experience? I am sure you are proud of your accomplishments and I know your experiences will be useful to others but there are other ideas and experiences that are just as valid as yours. Any compilation would be incomplete without a discerning opinion. It is up to each of us to decide for ourselves what information is pertinent.

If you continue in this arrogant and selfish path you will find few who will care what you have to say. If you truly want to share information on a forum and have useful discussion, then be prepared to accept the good, bad and occasionally ugly. If you seek self gratification then publish a book.

Charles

I wasn't trying to be arrogant when I posted this. It was more like I was weighing back and forth between putting in that point about retrofitting or not, fearing that it would take away the focus of what I intended to post, which was a guide others could benefit from my own experiences combined with experiences from others in this forum who also have made their experiences and were willing to share them.

I should also point out that my inspiration for starting this thread was form other forums (in this case a laser building forum) who had this great "how to do it yourself" guide as a sticky on the top. And I came to think of all the help I had received from this forum. So I figured I wanted to start a thread here that hopefully could evolve in to that same "this is you starter guide" thread.

And I did invite others to join in and apply their knowledge, and so they did.

But it was a small sigh from my side that it took off and evolved in to something completely different.

Then again, seeing all the involvement it has created it wasn't a bad thing at all! I have really enjoyed the response here, even if there are those who seem to think that I'm trying to bite over more than what I can chew. And then again that have started a discussion that have put forward even more details and knowledge.

So I'm not at all sad or in any way hurt that it evolved in to what it has, on the contrary I appreciate the responses that have come. And as you hopefully also have discovered I'm now also trying my best to answer and challenge the others who are participating in the discussion it now has evolved in to.

I was also trying to change the name of the thread since it's no longer correct. But I couldn't figure out how to do that. Is it possible? If so can any one tell me how?

Anders
 
Anders, I dont think you can change the title, but those of us who have an interest already know what is going on. Your earlier tone did seem to me to be going in the wrong direction, so I thought a poke in your ribs might be needed. I agree, there is some good information being discussed and I am sure we will all learn a bit. Sometimes when we write the "tone" of our writing can be misinterpreted. If you are going to try the retrofit idea you have bit off quite a bit, Maho didnt make life easy for you but of course it is always possible to make it work.

The end result will be does anyone else think the machine is worth the trouble and that will only tell if the outcome works.

Most of the problems that people have is the electronics so hopefully those who have had troubles will post as well and your idea for the posting will work out. I have stayed away from the CNC machines and have stuck with a manual one for now but even I can see a desire for a more reliable cnc version that doesnt cost an arm and a leg. But I also know enough to know that I dont think that unicorn exists...:)

Charles
 








 
Back
Top