HMC retrofit

I'm thinking about getting a 2001, out of service, Roku Roku Libero RX6 HMC. At this moment it is due for scrapping, pricing should be near scrap metal, and I contemplate about saving it.

I do have a 9-5 in the machining business, but this will be a hobby gig. Not sure were I will end up with it. If I won't fail the retrofit, I'll keep it as a white machine for black days. Who knows.

I do have my fare share of diy, and I do like wrenching so, it can't hurt a bit of play.
Main concern is my lack of knowledge via electrical/electronics part.

Either way, I do know what am I getting into.

Short machine description:
- very small footprint
- double pallet
- Nikken CNC-201 unit as the fourth axis
- 2.2KW 12 or 15000 rpm
- ISO 30 spindle
- 30 tools chain + shifter, index ATC
- Fanuc 18
- Small Fanuc aplha series servos (red caps)
- 40.000mm/min rapids
- roughly only 6 years in use
- though it was kept indoors, it wasn't prepaired for preservation. Few rusty dots on the guides, spindle was blocked etc
- all the amps, controls etc are missing. Though, I could recoup them from her sisters

For starters, I would wan't to avoid keeping her with Fanuc controller. I do like Fanuc, but for hobby use, I can't afford their service costs. A middle way it's impossible since it is way too proprietary.

My first priority is to make sure there is a solution for what I'm trying to do. At this point barriers that block my view are as follows:

- Peripheral stuff: pumps, brakes, pallet changer, doors, sensors, ATC, pneumatic valves etc. Should I govern them via macros in Mach4, go via a PLC or go the old school way of relays?

- I have Fanuc servos for the tool magazine and for the pallet changer. Do I need drives for these motors? or they can be driven directly via a PLC?

- Fanuc red caps - should I use a converter for the encoders? or just change the encoder?

- Controller would be CSMIO analog version. If I definitely need drives for the pallet changer/atc magazine servos, I'm thinking to replace these with steppers. This case would require the step/dir version of CSMIO

- HMI - heard the term Pokeys as a solution, or some other gadget from CS Labs that can do the trick. I'm looking for stable, 100% reliable analog controls. Example feed override.

- is Mach4 stable enough to keep things safe? 40m rapids can get dangerous

- anybody have experience with Nikken 4th axes? drives, motors etc

- not sure what holds the tight precision on these machines. Quality ball screws? or is there some serious mapping in the old controller?

Any advice is more than welcome,
Thanks,
M.

Dont really have any good advice, never heard of that machine brand, but it does sound like you have your head around this and have thought it thru. The small foot print in a horizontal and 12000 rpm spindle would be sweet, especially if it was like 20 hp and 40 taper. I think the electronics will be the hard (expensive) part for you (just based off your description of experience). This is were I end up spending money sending stuff out to get repaired because I just dont understand it or have the right equipment to test and repair it. I would seek some advice from someone that does retro fits so you can get a better idea of what your looking at budget wise. I would also decide what kind of parts and machining your going to do on the machine when done, that way you can focus on what aspects of the machine that are going to be the most beneficial to you.

As usual, there are more ways to skin a cat.
At this moment I would like to know what is the best option.

It used to do small prismatic parts in aluminum with very tight tolerances. On occasion mild steels. I'd like to keep it the same as it was.

M.

If you can get the hardware, lots of guys run Fanuc in their home shop. Including me. So much info can be had for the asking because of the large user base.

I think you are really underestimating the amount of work here. A simple 3 axis mill is not a bad retrofit. Add something like an umbrella tool changer and it gets much more complex. Add something like that side mount tool chain and it's going to be a serious endeavor.

No way I would take on a project like this. You would be much better off to buy a running 3 axis machine, even if it's much older.

We run mostly Fanucs at work. I might get a hold on what's missing for the control part of it. You know from the shelf with "yellow boxes" or as I mentioned before, from it's sisters.

Personally it's a love and hate kind of thing with Fanuc. The hate part is the lack of support here. Usually it's 2 weeks time until somebody arrives. Things can never be fixed, just replaced, huge hour rates, etc. 5000 for an amp seems steep considering other offerings. I know it's apples and oranges, but both do the job. As for reliability they will both fail at one point so.

Yes, it would be much wiser to retrofit a simple VMC, but wasn't fortunate enough

IMHO i would not touch mach 4 for something like this with a barge pole. Hell i would not use it for much at all. Even tormach have shied away from mach for a whole bunch of reasons. Put simply linux in a real time environment makes controlling anything reasonably easy, you have to do a mass to even begin to get real time in gates land!

Now LINUX CNC and one of the Mesa interface boards should be easily able to do this. Nothing more than a tower type PC monitor and keyboard - mouse. Does not need to be anything special, simple 32 bit and a gig of ram + as muich hard drive as you want (250mbs, is a lot of G code!!! so a typical small 30-60Gb ish hard drives plenty and a zillion times the capacity of a fanuc control!) If the drives are there, chances are you can interface to them through the appropriate mesa board combination. Tool changers not hard, the shear qty of IO lines the MESA boards support will easily do the function you need for tool change (90 something IO lines is a sub $80 board). Mesa board, pc and some wire, assuming you have all the working drives + limit switches and there currently working, your easily looking under $1000 to get it working. Probably a weeks solid wiring and a week or so of tweaking - code.

Your going to have to get a good basic understanding about electronics and there interfaceing + a basic functional idea of code, but its not hard.

LINUX CNC will do everything fanuc controls can, just at zero cost for all the features. Things like rigid tapping will need spindle encoders, but software wise its no cost! The Mesa stuff is very powerful - very capable industry grade interface equipment.

Lots of people retrofit manual equipment, retrofitting CNC's is easier because all your drives, sevo's, ballscrews, limit switches are all there to start with. There should be very little mechanical work actually required if the base machine works and you just have a partial stripped - sick control.

VMC or HMC makes little diffrence, get your basic 3 axis working, then the spindle, then the tool change, then pallet change if relivent, then add a 4th axis. Once you have the 3 main axis working, adding a 4th or 5th is easy. Linux will support 9 axis of sim motion if you can work out how to programme it :-) + run a extenisive and powerfull software based PLC function with a ungodly number of inputs - outputs.

My biggest concern if i was you would be in the machine moveing costs, that could be your worst problem!

Mantinance wise, once you have done the retrofit, your your own OEM tech - MFG help line etc, its all on you. Repairs are as fast as you can do them and theres no being tied to any one drive servo or anything.

Its the above and improved capabilities that i ditched the anilliam controller on my mill to go retrofit for in the first place!

adama said:

If the drives are there, chances are you can interface to them through the appropriate mesa board combination. Tool changers not hard, the shear qty of IO lines the MESA boards support will easily do the function you need for tool change (90 something IO lines is a sub $80 board). Mesa board, pc and some wire, assuming you have all the working drives + limit switches and there currently working, your easily looking under $1000 to get it working. Probably a weeks solid wiring and a week or so of tweaking - code.

Click to expand...

2 weeks to retrofit a machine that you have never seen run with a open source PC based control with sketchy documentation and no support?

I'd love to see that. I guess it depends on your expectations. I'm sure he could make the 3 axes and spindle run in 2 weeks. Where theses hobby control retrofits drop the ball is with the details.

Tougher aspects that few address:
Monitoring the pressure cycles of the lube system.
Getting the timing optimized on the tool changer.
Rigid tapping.
Homing from the encoder marker pulses.
Pallet changing.
Spindle gear box.
Probing (tool, work)

Then there is the safety stuff. A real CNC machine will have an E-stop string designed to some ASME spec. All the push buttons will be 24VDC.

Next is the alarm messages. A typical CNC might have 20 pages of alarm code messages. Most of these will need to be programmed into your retrofit so you have a clue what is going on when the machine screws up. Things like tool changer recovery will require macros that you will have to write.

If I recall with Linux CNC, that's going to be a lot of Python programming.

If you can do that in 2 weeks you're hired.

I have never touched Linux, it's windows forever.
That's pretty much the only drawback, but it's a huge one. Not sure how Linux deals with parameterizing codes, subs etc. I presume it should be able to, but it's another learning courve.

From what I read, Mach4/CSMIO though still not realtime, one shouldn't feel the nanoseconds while on the Godknob.
Mach3 had it's fair share of bugs, and that's what scares me about this route.
Wouldn't want to use a crash in order to find a bug...

About the maintenance you are right. That is what I'm looking for!

@Elwsley

You're right.
If I'd want a quick hobby gig, I'd go the stepper route, and keep the logic in Mach. But that's not what I want.
Time wise, there is no rush. I give it at least one year, a hard one if I may say so.

Machine wise, it's not the over complicated type.
Things like drawbar and shifter sincro are long way ahead. There will be million things to go trough, and learn, I'm sure. But one step at the time.

At this point it seems I'll have to change the encoders on the redcaps. Converters won't do it with ease. I'd stay away from changing motors if possible. Drives, probably DMM-tech. What does remain, is the spindle amp. I'd have to keep the original to keep costs down. Unsure how it would cope with the CSMIO IP/A.

M.

LINUX need not be feared, if you have used a mac or andrino based phone, your there already.

IMHO difference between linux and windows is slight these days, just a different back ground and slightly different icons, linux is very much akin to a windows working environment. There is a lot of options to chose from. Download one of the linux .iso files, make a boot disk and give it a try on one of your old pc's. trust me, if your even a half competent windows user, you will find it easy to use.

LINUX is not a black screen running lines of white txt as code, unless you want too!!! Its a graphical windows like user interface and theres a lot of versions to pick from!!

LINUX CNC has more documentation and support than fanuc has ever thought off. Go look at the forums, every thing is there, everything is explained, everything is modifiable - compilable to your hearts content. Theres hundreds of thousand of uses out there and they all support one another. Its the fundamental idea behind software like LINUX published under the GUNU license arrangement, its software free for everyone and modifiable by everyone! Unlike fanuc every user can add modify and share thoes features. Oh yeah its written in both english, french and a bunch of other languages, translated by people that actually know how it works, not just technical translation, so the instructions make sense. Because its global, a question on the forum oftern will get a reply any hour day or night with in minutes just like here.

Adding anouther axis, is but 20 lines of code in a text flie, sounds scary but you just copy and paste say the x axis, then rename it and adjust its settings. You then add anouther few lines in a diffrent file were you set what the control ouputs and to where they go for the new axis simple as that.

Mach and the non real time issue is a major problem, you can't do things like high speed moves reliably and safely out of a real time environment, you either go slower, or risk bigger crashes - loss of position. If you play online games, you understand PING is everything, trust me, REAL TIME MATTERS, especially if you want to do accurate high speed movements controlled by software.

Its not just the movements, its the way you talk to drives + other interfaces, lots of drives need a direction change signal held for a certain time before commanding motion then in that direct. With out real time control you just can not reliably do this. Its why you can't control things outside of real time environments or use interfaces like USB that do there own take on timings. You have to control the timings and issue them accuratly, that needs real time. You need to go look up the problems this causes, nano secounds matter, especialy when you never know if its 1 nano secound behind or a thousand. Mach attempts the get around the real time by allowing a more constant lag factor amungst other approaches. It works just, but its a mess, and for what your doing you need the better option.

The only programming I've done, outside Gcode, was some java to make the post processors in house. I get the logic, but I lack the sintax.

If I would've been a programmer I would've gone via KFlop, but since I'm not, I don't really want to have all the trajectory planner and other stuff on my head. I'm sure there are plenty of users who did got it right, but you know tastes don't get discussed. Reinventing the wheel neither.

My understanding is that CSMIO makes Mach4 more of an interface, a view layout with DROs. Hence, it's the CSMIO that controls things. Maybe someone with enough knowledge would approve.

In the end, the PC is a bigger risk than both Mach or Linux. This is why I need to bullet proof things via a PLC.

I had a conversion in the past that used to move while a colleague was welding 15 meters away. Call it screening problem, pc whatever - it's still a risk.

M.

Seems to me if you have the talent to retrofit a machine you should easily be capable of repairing a Fanuc. Even board level repairs.

For all the reasons above I would stay Fanuc or find another. But I want to make chips not controls. Good luck with either direction.

Make the Fanuc work. There are a handful of people I know on this board that could make that HMC work like new with a retrofit, all of them are talented in many different areas, all of which are required to make it work.

Unless you have the following skill set, I don't recommend it:

* Electronics
* Machining
* Mechanical Engineering
* Programming
* Computer repair
* Patience

msm said:

In the end, the PC is a bigger risk than both Mach or Linux. This is why I need to bullet proof things via a PLC.

I had a conversion in the past that used to move while a colleague was welding 15 meters away. Call it screening problem, pc whatever - it's still a risk.

Click to expand...

The PC is not the problem, if you want it hardened against HF, you have to harden - shield it and filter the supplies, theres no free lunch. You can't run a machine control on a typical PLC either, that’s not what they do. The CSMIO looks interesting, very similar to what the Mesa boards allow you to do with Linux CNC

Other Key thing you can do is use some real simple logic in your favour. Gold std in signal control is you Have a + and a negative signal line for each event, you then look for the swap over. Ie say the A+ encoder line pulls up, the A- line pulls down, when both swap over you can be pretty certain with real simple logic you just witnessed a movement not a interference blip.

I still stand by a 2 weeks lead time, weather the machine has 2" or 30' travels to swap the control over takes the same amount of effort. End of the day how ever you do it, its about half the time wiring then half the time configuring things. Once you have one axis running its the same to get the rest going. hell if anything it gets easier. A HMC or VMC only has a few axis, most of the logics in the tool change cycle, things like coolant and such are dead simple relay outputs,

Tool changers are not hard to do, what makes then hard is most retrofits were never designed around tool changing, like say a manual drawbar on a bridgeport. Starting with a cnc with a tool changer gives you a massive leap forwards there. You already have a spindle designed to easily hold and release tools, based on the pics you have a magazine and you presumably have some kinda changing arm. After that its just work out the interface and controls - execution order to do it in, there’s no magic, just a series of simple steps.

PLC or PC there really not the risk, the risk is generally interference picked up on the power , signal - control wiring, a PLC won't eliminate that, correct - good wiring and grounding practices + interference suppression is what eliminates it or minimises it, fundamentally its akin to makeing something more water proof, yep you can make it so it does not get wet easily, but long term holes will develop - most things will leak. 100% interference free is not something you can guarantee!

Welding within 15mtrs of a CNC of any description is not a good idea, especially if its one of the older transformer based welding plants that come no were near modern emission limits. HF start weldings even worse as is a lot of the HF start plasma cutters, its were the hypertherm blow back - aka lift arc approach really shines.

Here is a link to a tread about retrofits.

http://www.practicalmachinist.com/vb/cnc-machining/retrofit-mazak-monarch-hardinge-224896/

Forum member skunkworks has done a few retrofits:
http://www.practicalmachinist.com/v...sa-heavy-iron/k-t-milwaukeematic-iiib-149268/

http://www.practicalmachinist.com/vb/general/matsuura-mc-500v2-yasnac-mx1-m5x-manuals-279376/

I belive there should be some useful info in those treads.

Local Siemens would do the servos, drivers, plc, logic and controller somewhere between 10-15.000 euro

In the meantime I see plenty of Siemens controls and everything you need on Ali for way less than local.
As normal, if I get it from China, there will be no local support of the product.

Does anybody have experience with Siemens from China? is it worth it to even think about it?

Can't reply about there stuff from china, i only use it from local sources, but its good stuff, i use a lot of there VFD's on equipment, never yet had one fail either. If i had the budget for it, i would go for there controls too, i don't hence i use the Linux cnc option.

My concern with sourcing it (or pretty much anything) from china is not if local support is available, its a question of if its the real thing. Theres a lot of knock of stuff on ali, proceed with caution, unless they say guaranteed genuine, expect it to be fake!!!! Especially if the price is too good to be true.

We are in the process of converting the matsuura to linuxcnc. Reusing the yasnic/yaskawa drives. Pretty close. Working the bugs out. Using mesa interface hardware <$500. Using the integrated PLC software in linuxcnc. The second video shows a shot of ladder. The delay on the carousel down is the cylinder needing to be overhauled.

Matsuura's first 'successful' tool change under linuxcnc - YouTube

Matsuura running simple program with rigid tapping - YouTube