Updated - Conversion of milling machine to CNC

[MeJasonT]

Hi all,

Previously i have posted information regarding the upgrade of my miller to CNC, about using GRBL as a controller and my Plasma table build.
Im now starting to get quit involved in converting and assisting others in various conversion projects so thought it was about time i wrote a serious thread.

During Covid i have been busy converting my mill yet again, It has gone down the linuxcnc/pathpilot route, I was starting to realise the limitations of using GRBL as a machine controller. I was looking for canned cycles, conversational programming and a better tool change method. Unfortunately GRBL does not handle these very well, By switching i have also gained a 4th axis. I have not got to make chips yet but as far as i hve read this is true 4 axis and not 3+1. Happy days.

Over the last 140 odd days (of the end of the world) I have helped a guy convert and old Denford mill to a linuxcnc styled machine and dumped a whole load of design information on another chap building a plasma table and torch height controller. There are also plans afoot to design and build a wire EDM machine.

The miller has gone through a number of development changes but is now starting to take on a life of its own, shortly before lockdown i was making parts and had managed to get the accuracy and repeatability nailed. The major issue was the worm screws, one of the conversions involved ripping them out and replacing them for precision ball screws. I went from 0.7mm of backlash (yep you read right) to 0.05mm of backlash. To be honest i hadnt even tried getting that improved on before i started to get peeved with GRBL. I am now in the unfortunate position of being a Pathpilot shadow user. If i had the choice i would buy a Tormach but unfortunately the UK market is one they are not very interested in. So clone it is.

Having purchased the software for the token admin charge, I went about upgrading my machine, I purchased 2x Mesa anything I/O interface cards, the type used in Tormachs machines then started the build. The first stage was to build a new controll panel. the old GRBL electronics had served me well but sadly were very limited and had injested so much grinding dust looked 50 years older than they were. wiring had turned into a snakes wedding after various ripup retries and repairs so i was focused on a new panel anyways.



The reasoning behind the 2 mesa interface cards was the i had an older version of Pathpilot which used a PCI interface card and i was intending to build the machine to match the new upgraded Pathpilot version which runs on its own network connection. The panel was constructed offline/off the machine so that the machine could be left operational until the last minuet. I etched and built a breakout board/pcb so that i could make wiring connections to the Mesa interface card and then made the appropriate wiring connections within the panel.

The build was really simple and straight forward, the advantage was that the machine was already wired with limit switches and stepper motors so the dirty stuff on the whole was already done.

The idology behind conversion to CNC and the alternative the retention of manual milling comes down to on simple question. do you love standing infront of a mill and get excited by the wonders of producing a single part every now and then, or is your ambition to have the ability to make parts. The debate for manual milling over CNC will go on for years. For me its about making a part to suit a requirement, yes i enjoy watch a tool all day ripping steel and making things shiney but the back ache and length of time it takes is non productive. Also im not retired.

The machines appearance has not changed much

[MeJasonT]

Conversion of milling machine cont.

The brains of the machine is a Gigabit PC motherboard, all the nerd gubings were attached to make it a pc ie. HDD and the handy move thingy and anger thumping buttons. The pathpilot USB drive was installed and the PC operating system installed. If only Tormach wrote windows, the whole task went like clockwork, hats off to the tormach guys for a fantastic product. As a shadow user i find it only fair to endorse their fine work. I am not affiliated with them in any way but feel they should get full recognition for their part in this project. I would certainly not discourage anyone from purchasing their product quite the opposite. - just to be clear we have strong rules on selling on this site.

Another point that should also be stressed is that pathpilot is a registered trademark of Tormach, they will sell the software as its core is linuxcnc which is open source software and therefore cant be sold or owned. But, the graphical layout and some of the functionality of pathpilot does belong to Tormach. hence the purchase. I went with PP as linuxcnc is typical linux people, i breed of individuals who all think they are in mensa and the rest of us are beneath them, asking for help is like selling your soul. mach3 - in my opinion the graphics look like they were drawn by a 3yr old. Its also expensive compared to Pathpilot which is really only intended for Tormach machines. I have heard Haas users say if only the hass Gui was laid out like Pathpilot it would make using their machine really easy. I think Tormach missed a golden opportunity, they could have had a side business selling controller and interface packages like Fanuc and Heidenheim. I guess its down to the linuxcnc open source thing. Enough politics already.

I decided to try building the Pc into the panel as it was going to reduce the footprint of the space required for the machine, having used GRBL and having experienced numerous issues with emf/rfi noise causing system crashes i was putting alot of faith in this build. by laying out and getting my head around location and spacing of various items such as the motherboard/transformers and stepper drives etc, i was able to figure out a way of including segregation between power components and control electronics by means of a steel plate divider in the middle of the electrical/control cabinet. whilst building up the panel a number of test were carried out to see if any item would be particularly noisy and affect the machine operation. Good news was not a single issue. with the wiring of the cabinet completed it was time to mount it on the machine, this stage took ages. Procrastination is a very serious part of engineering and takes up so much valuable time, I have come to the conclusion im more of a professional procrastinator than i am an engineer. This should be a new discipline/career path in engineering.

Finally mounted the panel and started to hook things up, all went without a hitch. the only issue was that the spindle speed was not calibrated and the motor homing and directions had to be altered, obviously the software was not intended to run my machine.

My plan is to share the build notes and method of the conversion and to make a couple of youtube videos to support the build process, it is a little late as the machine is 90% there but at least it wont be a narcissistic thread of someone smiling say look what i can do, well jolly d for you. i hate such things with a passion.

The guys i have been following on youtube have all now evolved, they are posting less and less content and steering you to their Patreon site where they feel that $30 per month of $199 a year for a higher tier is a reasonable charge. This infuriates me as i think the money they as although reasonable is out of my pocket money limits, hell i barely pay that on broadband. They were all enthusiastic and got us all on board following their on line career to then isolate the few followers just to service the moneyed ones, yet again capitalism in your face. Isn’t that the reason behind the skills shortage and their ambition to share content in the first place to encourage their followers into engineering. So my content will be aimed at taking the viewer on my journey and into the production of parts. It will be like a Dummies book, might be to advanced for me but im sure our homemadetools site followers will inadvertently correct my sloppy machining practices by ripping the mick. intellectual and constructive abuse is always welcome.

[MeJasonT]

This is a list of the main channels i subscribe to and the people who have been the most influential in my endeavours to become a machinist at the age of 49.
The UK is great at offering apprenticeships if you are under 25, an ex-criminal or from a minority group such as females. To change career in adulthood is next to impossible and after covid the future will be even more uncertain. if i can start a job shop using a hobby machine so be it.


blondie hacks

https://www.youtube.com/channel/UC7J...4e74O53dL6arSw

NYCCNC

https://www.youtube.com/user/saunixcomp

Titans of CNC

https://www.youtube.com/user/titanamericanbuilt

Lars Christian

https://www.youtube.com/user/cadcamstuff

John Grimsmo

https://www.youtube.com/user/JohnGrimsmo

Jason Nerdly

https://www.youtube.com/channel/UCq9...z8dC9XJjjmOQ5g

This old Tony

https://www.youtube.com/user/featony

But not this idiot

https://www.youtube.com/channel/UCga...junai_bUlS849w

[marksbug]

Ive wanted to convert one of my mills to cnc, but $$ and brains to do so seem to be a issue...I used to program&operate some herco cnc mills 20+ years ago.they had simple language programing.I doubt I could learn g code now. as far as electronics go's...Im lost. the cancer&way to much radiation has stolen my short term memory& attn span. hell I have to take a pic of a std wall reciptial wires before I take them off or I wont know how they go back on a new receptacle 3 min later. but it seem all my old learned use full and useless stuff is still all there trying to get out. if I were to get the controllers and step motors and drivers and what ever else is needed and mount it all up it would look like old lucas stuff....lotsa smoke...well it could also look like the japan electronic stuff.when you let the smoke out of it it's dead too. so I got as far as buying some ball screws for my small mill...but doubt they will ever get used. keep yup the great work guys!!!!I still love to read it even if it's something I cant doo.

[marksbug]

I would think starting a job shop with a "hobby " machine would be great.Ive had many offers to do lots of stuff in small batches and large batches for defence contractors that cant get their machine shop to make small batches for a small project that needs to be done now. if I had a cnc mill I would do them and probably be quite bussy and make lots of $$$ .but at this point in my life....I want to do stuff for ME. Ive done everything for everybody all my life.and my stuff always got put off...almost 30 years ago i sold my money making race car to build a new one....( drag racing) last year I sold off all my unfinished racecar stuff including the unfinished new professional built racecar. 85% of it was new unused. I got about 15 cents on the dollar for it all. ( I got $15000 usd) it was all high dollar parts. thus the reason Im done working for any body else. so you there should be lots of work out there for you if you get word around and do good work. Good luck in you new venture!!!

[MeJasonT]

Hi marksbug (i take it your Mark)
Thanks for your kind word and support its very much appreciated, especially when i read your background and story. Again thank you.
I have another instalment I’m working on but have a small hurdle to overcome (actually 2). I was struggling to upload pictures again (Thanks Jon) they are too big - they were taken using my phone with the same resolution as the picture inserted above (don’t panic Mr Manering, I’m on it). I also had my PC monitor die and having to wait for delivery of another, the current monitor is a 15" one from out of the shop - man old tech is such a hassle.

So in the next couple of days there will be more inspiration and cnc fiction uploaded.

I have been hunting out all my webcams and chasing down software to use my canon EOS camera as a web cam to get some material launched on youtube.
The mill is still presenting me with gremlins, A second transformer was installed in parallel to boost the current for the axis motors, not successfully i may add. It would have worked better for heat shrinking cutters into tools, spent ages chasing down erratic axis moves until i realised the transformer was nice and toasty. back to one transformer now and the machine is back running like a dream, all be it close to the power limit of the transformer. 3x 4A axis motors from a 10A transformer, umm another purchase is on the cards. Its unlikely that the transformer will take a hit of all motors starting up at the same time, its the inrush current that dictates the necessity for matching the transformer to the load. The 4th axis is going to be a bitch.

Transtronic in Ireland will make a custom transformer so getting the secondary voltage increased and a higher rated amperage.
I’m currently running at 30v but taking it up to 46v ac before rectification is well in range of the stepper motors and drivers - getting the feedrate up without screaming the motors would be a real advantage.

pathpilot can feed at 2000mm/min the best i can get is 500mm/min which is acceptable but not fantastic. Stepper motors average 1000 rpm for both nema 23 and nema 34 shell sizes but i also have a 80:1 ratio gearbox on both X and Y axis, I’m considering that may be the cause of my restriction. However the torque is amplified so holding accuracy should be mechanically that bit easier.

[MeJasonT]

This was the GRBL controller; you can now see why going down the road of rebuilding the machine became necessary





GRBL (actually the firmware) ran on an Arduino Uno. The hardware interface was/is relatively simple and straight forward, the Arduino uno pins are 5v tolerant and easily connect to Stepper drivers. The limit pins are connected to micro switches, the only big issue is the Spindle it required a little bit of signal conditioning and pcb building to get the pwm output for speed to a 0-10v signal for the variable frequency drive (VFD). It turned out the drive was particularly noisy signal wise so the spindle data was carried over a plastic fibre optic. Its not as if technology has all the answers, the fibre could only transmit data at a max baud rate of 2400. Where the Arduino was at 115200 baud.

The basic layout of the Arduino based cnc controller





Thanks to the advent of the 3D printer revolution anyone with the desire or basic skills and ability can build a CNC machine. Due to the memory size of the Arduino Atmega 328 IC it is limited in the amount of functionality it can provide, for basic 3 axis machine work its fine. It has no further expansion to allow it to become a fully-fledged controller for a vertical milling centre. Control interface/graphical user displays have been sparse, which has really let the basic Arduino cnc concept down. At some point you make the decision to grow, the Arduino becomes irritating when you can’t handle the tool changes or peck drill holes. The global search will inevitably take the person to linuxcnc.

When you look deeper into the control systems involved in obtaining computer control of machines you very quickly begin to realise the simplicity of controlling the mechanical aspects of a machine. looking at the interface connection for Mach3 and linuxcnc machines you can notice that the wiring methods are not dissimilar.



I’ve just been working on an old Heidenhain controller and was smiling with pleasure at how much I could glean from its wiring without even having the manual. Software is the key, the mechanics will deliver the axis movements but the software is what will give you the full functionality of a machine. The windows/linux pc desktop styled operating system is only just starting to appear in professional machines.
The plan to upgrade,
When you get to the point of realising that you can mill a part but it’s not what you hoped it would be and accuracy is for other people then its time to do a teardown and start again. The first upgrade or is it the 3rd or 4th I’ve lost count was to swap out the worm drives for anti-backlash ball screws. Wow, the accuracy was out of this world compared to the old worm drives.



The control/software was still hindering, my views of linuxcnc ive already covered. I chose to go down the pathpilot clone route. It wasn’t a simple decision to make; it was a toss-up between buying a second hand machine or the upgrade. The second-hand machine is still on the cards but the upgrade was already a financial commitment.
Researching the Tormach hardware and interfacing, I came across details of the Mesa interface card they use, oddly enough its wiring connection looked quite similar to the one above, in fact it’s identical. It wouldn’t have been to hard to simply swap the GRBL board for the mesa interface wiring. I wanted to upgrade the stepper motors from 2Nm nema 23s to 4Nm nema 23s, the transformer needed to be upgraded to supply a higher current and voltage output.
Improving the machine and planning to use it with much greater anger means a proper coolant system was required (bye bye Fog buster) ok perhaps I could adopt both. The enclosure was built for the best practice and health and safety practicality, na - just keep the coolant inside. My apprentice and I decided that we should send a picture to Haas and beg/plead for a VF1 sticker for it, now I have to ask Tormach as I owe them a debt of gratitude.

[marksbug]

hell now days I get lost trying to post a pic..... yes i would have another transformer on the way that will handle the job.as soon as your ready to do something that needs to be done NOW....all motors will start at the same time ...and stop at the same time..... good luck with it all!!!! it's never too late to start something new if thats what you want to do and are able to do so. 25 years ago i thought I was to old to learn to program cnc machines, but it was simple. Ive been a machinist for most my life. but I was machining and building performance racing engines.not a lot of cnc machines in that back then.I for sure didnt have one.but I wanted one for my personal usage..so ..I got a job at a friends proto type machine shop.( we did a lot of short run prototype stuff for aerospace,dept of defence,ford, chrysler and gm. as well as many other things including taking care of the local drinking village with a fishing problem. ( a massive huricane also gave me time to change jobs since my former place of employment was no longer there, i ran the performance machine shop for hudson marina building high hp reliable "*****"boat engines.( cigerette boats with bimbos&massive wallets always wanting to be the fastest of the water).. so I know it isat too late for you.just keep on the path and get it done. and dont be afraid of having to put the engineers hat on many times to solve issues.

[Saltfever]

Thanks for the post Jason. Especially all of your trials and tribulations leading up to your choice of PathPilot. You bring a lot of experience with other languages and knowing your integration history (or pain, ) kind of pointed me to go over to Tormach and look at PathPilot. I didn’t specifically see them selling software as a separate product. However, they sell a controller with PP ver.2 installed. So I was wondering; did you buy just PP software or did it come as part of some other purchase option?

Also, thanks for listing all the channel links in one place.

Looking forward to hearing more about your progress. All the best.

[MeJasonT]

Hi Saltfever
I acquired the software only and built a PC to run the machine.

I became very disappointed with Grbl and the forum members lack of imagination to keep pushing the project. When they moved from the atmega 8 bit processors to the 32 bit ones they had a huge untapped potential of adding the features machine builders were looking for. don’t get me wrong, Sonny Juan and the other developers did a fantastic job of building a fully functional machine controller around an 8 bit processor. The development then stalled as the momentum was then passed to other members to create and expand, it is they who let the project come to a dead end in my opinion. The open source sharing revolution looks like its beginning to peak, open source designers and developers are beginning to realise the financial potential of their new found ideas. There are now a couple of the grbl forum members who are building a commercial units with locked software which goes against the spirit of open source. The hardware to build an open source 32 bit Grbl controller has taken over six years to appear and still does not look much like happening.
The PC is the most logical progression; even the Pro machines have cottoned on. Gone are the days of a dedicated plc alone running the machine, the new windows desktop styled controllers are all the rage now.

The use of operating systems other than Microsoft Windows makes a lot of sense, im so pleased that they do not use Windows 10, if this is their attempt at the best OS ever they are in deep doo doo. File explorer is garbage, the whole machine is slowed and the memory is maxed out. Like the other budding engineers on this site im screaming at the PC for Microsoft to sort its crap out, windows 95 was a far superior product to this junk.
Linux is the obvious choice. The beauty of it is if you can’t get to install on your chosen PC due to bios complications you simply put your hard drive in another machine that will create the drive and swap it back into the machine - Windows, not a cat in hells chance with that eufi thingy locking your drive to that pc.

Pathpilot (on linux os) is a brilliant choice so long as Tormachs openness continues. A bit like Fusion 360 we are all hooked now, how long until its free "loaned” licence comes to an end. I must applaud the commercial companies for opening up to help educate future machinists and engineers; smarter thinking is starting to turn the page on capitalism.

"Jasons cnc revolution" is continuing at pace, i just received a call this morning from a company called Chester Machine Tools in the UK who are attempting to source a second hand milling centre for me. The poor guy is having to pull hens teeth with both hands tied behind his back. I also spent a few hours last night looking for used machinery sales companies; it really gets my goat why not one of these companies displays prices, what’s the secret. When i want to buy, is it going to be 3k more than it would be to someone else?

Ive always been told and understand that in order to do business you have to gain someone’s trust - no prices gets them off to a bad start. I have a modest budget and would like the ability to compare what’s available and what i can get for the budget available, to not provide me such simple information as a machines price (which may or may not be negotiable) has taken away my decision making ability. I might spend 10k but I might spend 12k it depends on how shiny it looks and what features it has. Building a fourth axis is not beyond anyone’s ability on this site, however interfacing it to an old machine without its ability to run one is an issue. Having such a feature as standard on a machine makes it desirable even if the machine looks cosmetically like its been in a car wreck. Age wise, do you want a machine on its last legs with some mileage left in it ot one the cuts circles whilst it travels in straight lines (like my current diy build. lol).

The saga continues …..

For reference, this is an older post I wrote on HMT regarding Grbl as a cnc controller
Crash Course on GRBL public licence CNC control.