[cwalker1960]

I have purchased an old hurco bmc 30 VMC 1991 year model running on max32dsp . The controller is dead right now (possibly only the hard drive) and no telling what it would cost to get it back up. Seriously considering a retro and at the moment the 2 top candidates are linuxcnc with mesa boards or kflop with kanolog and kmotioncnc.
I do have a very limited knowledge of linuxcnc as I've used it with the para port on an old homemade router table with steppers but no experience with kflop at all.
The machine has yaskawa ac servos on all 3 axis with yaskawa analog drives, it uses a dual 5000 count encoder input on each axis(the encoder on the motor is wired straight to the relevant servo drive and the one on the axis screw is wired to the hurco controller). the encoder on the axis screws has 8 terminals,a+.a-,b+.b-.index+index-. vdc,and ground. The spindle motor and drive are also yaskawa 7.5 kw. and has all the windings brought forth on the motor for high and low speed via a relay (contactor) . It also has a 24 position ( I think, could be 20 , have to look again) tool changer.
My needs are basic but when it is done I do wish to have electronic gearing (for rigid tapping ,etc.) and at some point I plan to add a rotary axis (the "inturn 4th axis" by cubed comes to mind)( it should be able to function in lathe mode or index mode) that will allow me to make gears, do some turning and various other cuts with all (any needed) axis geared together.
I don't think my very limited experience in linuxcnc is enough to deter me from learning something new , I just want to get the best product for my application. I don't have the funds to "try it out" so when it's done it's done. Price is a deciding factor for me as of right now this is strictly hobby.
One thing that would be a huge plus for me would be conversational programming for one off parts, and it does seem that linuxcnc users have added quite a few functions in this regard.
What I'm looking for here is user input on the pluses and minuses of each system and why one would (or should) choose one over the other. The price and learning curve doesn't seem that far different for either all things considering. Although linuxcnc does have some very knowledgeable followers in their forum to assist with the growing pains.
I do have mach3 to use as the front end should I decide to go with the kflop , but that doesn't seem like the determining factor to me.
I would appreciate any and all responses and any other controllers to take into account that I have overlooked.
Thanks

[Tkamsker]

i have done similar to an bridgeport 412 i using linuxcnc it is up and running but it was a verry hard ride because it is like an old car if you touch something wrong it ll cost you a lot of time but i learned a lot so how do you intend to control the servos and motor ?


Gesendet von iPad mit Tapatalk

[cwalker1960]

Well that is what I'm asking... I am hopeful that I can re use the yaskawa servo drives at least for the time being. If linuxcnc,, I am considering the 5i25/7i77 boards for analog output,, if kflop then with the kanalog board and probably the kmotioncnc front end.. Any advice???
One thing I can add,, I do feel comfortable enough with linuxcnc that I am very confident that I can make all 3 axis move as intended,, the tool changer is my biggest fear if I go that route.. on the other hand I am quite intimidated with the kflop not knowing anything at all about it , but it sounds like a very capable solution to my end.

i have done similar to an bridgeport 412 i using linuxcnc it is up and running but it was a verry hard ride because it is like an old car if you touch something wrong it ll cost you a lot of time but i learned a lot so how do you intend to control the servos and motor ?


Gesendet von iPad mit Tapatalk

[Tkamsker]

in my project the owner wanted to keep as much of old machine so i had to work like a surgery verry little steps a lot of testing in my case i found granite device controller which where able to control the sem servos from 1990 ! it took me another month to find the perfect tuning from encoder to motor driver but now i can control that beast like an stepper machine. and because of atc beeing an developer i wrote an component for linuxcnc which controls the atc works like a charm only thing is that the z axis is so insane heavvy that i have to run it slow to no get in error state so the owner has to decide how much money it is worth to have faster z axis ,..


Gesendet von iPad mit Tapatalk

[bhurts]

I won't push you towards one solution or another. In the end you have to be comfortable with your choice. I am however getting ready to retrofit a very similar although newer hurco bmc 30 using k-flop and kanalog. Different servos and drives but the same idea. This Will be my 3rd k-flop setup. First time trying a retrofit using all original hardware. If you go that route you won't be alone.

Ben

[cwalker1960]

I won't push you towards one solution or another. In the end you have to be comfortable with your choice. I am however getting ready to retrofit a very similar although newer hurco bmc 30 using k-flop and kanalog. Different servos and drives but the same idea. This Will be my 3rd k-flop setup. First time trying a retrofit using all original hardware. If you go that route you won't be alone.

Ben

Have you done one with a toolchanger yet? If so will we need any additional hardware to control the toolchanger ?
Will you use the kmotioncnc for your front end or mach3/4?
I've breifly looked at the kmotion front end , it's clean and looks like everything needed is there,, are there any or could one easily write some wizards to add to it for simple conversational?
Is electronic gearing possible with the kflop?
What are the options for mixed signals (some analog and some step/direction ? or can that even be done once it's incorporated on the kanalog board?
From what I've been reading seems people that are using the kflop are very satisfied and it seems to be rock solid. Oh and I've read a few of your posts bhurts,, you seem to be very knowledgeable about it too so that's a big plus that I could follow along with you ..
I would still like to hear the pros and cons of each system though.
is there anything that you know of that linuxcnc/mesa combo does better than the kflop or vice versa? Do you feel there are any features in linuxcnc that you wish you had?
Have you seen any of the videos of the cube studios "inturn 4th axis" , thats quite an impressive piece of hardware he has put together. Any immediate thoughts of controlling a setup like that with kflop?

[Tkamsker]

Hi,
i also don't want to push you in an direction
i recommend do you an excel list with pro , con , weight , points
what you want to achieve features and so on then you get an kind of almost feeling free decision,..

i don't know kflop in my case/project paramount was to keep existing hardware and as cheap as possible money wise.
seeing and reading all the old drawings and power stuff of that machine did lead to linuxcnc decision
on the beginning i did document a little on my old hp if you want to read http://www.tkamsker.at:8080/web/pub/analyse

and about inturn the designer will now integrate it in linuxcnc and probably the explanations why and how he does will help you as well.
thomas

[cwalker1960]

Hi,
i also don't want to push you in an direction
i recommend do you an excel list with pro , con , weight , points
what you want to achieve features and so on then you get an kind of almost feeling free decision,..

i don't know kflop in my case/project paramount was to keep existing hardware and as cheap as possible money wise.
seeing and reading all the old drawings and power stuff of that machine did lead to linuxcnc decision
on the beginning i did document a little on my old hp if you want to read http://www.tkamsker.at:8080/web/pub/analyse

and about inturn the designer will now integrate it in linuxcnc and probably the explanations why and how he does will help you as well.
thomas

Thomas,
First,, thanks for the link to your build,,, sorry i don't know the language,, German??
Which mesa boards did you choose for your conversion? the 5i20 seems to be very popular among retrofitters.. I am guessing because of the additional I/O available. I will probably communicate with Paul at Mesa before making my final board choice decision should I decide to go the linuxcnc route. From just looking though it would seem that I could probably get by with the 5i25/7i77 combo and possibly a 2nd I/O daughterboard. I don't know if speed is a deciding factor on Mesa board choice or not but I will find out before purchasing anything.

Ben,,
I'm still trying to wrap my mind around the kflop but i think i sort of got .. Is it that the kflop really controls all the machine functions, has it's own processor and could basically run the machine without an external computer aside from the fact that without a computer connected to it you would have no communication to the kflop? I may have worded that strange, what i'm asking is when people use mach3 for the front end on a kflop ,mach3 is just a portal to send the gcode to the kflop and have an interface to control various functions,, where as the kflop is actually doing all the trajectory planning and gcode interpreting ,,, is that correct??

If I got that right I sort of understand it as I built a little 3d printer a few years ago ,, it is controlled by an arduino mega, a modified version of grbl is loaded directly onto the arduino. it has a 4 line lcd screen, a few buttons and a slot for an sd card , gcode can be loaded on the sd card and can be ran directly from the printer not even attached to a computer,, second choice connect the arduino to a pc and use a front end such as printun and you can send the gcode to it from your computer.. and the arduino sends feedback to your computer interface to let you know where the axis are, nozzle and bed temp and a graphic display of the object as code is being executed. If i got that right then I would have a few more questions in regards to kflops control software..
can kflop do screw mapping? Does it have backlash comp? Tool offset comp? I think I read that kflop inherits it's g and m code interpreter from linuxcnc , and i also think I remember seeing that mach3 has a few not industry standard codes. Does that mean that kflop and mach3 may have a few communication problems ? even if that is so , it's probably not a deal breaker for me because I am more leaning towards using kmotioncnc anyhow if i go with the kflop , especially if there is some way to add some conversational programming to it. If not though,, I guess machs wizards could be a handy thing to have. One thing i have learned from using my plasma table , it would be extremely nice to have a way to edit some things at the controller screen instead of having to go all the way back to the cad program , make mods then back to cam to process, that can get tiring.
I'm sure i'm not finished asking questions,, it seems the more i learn the more I realize just how much i don't know..
I would like to thank both of you for your time and input and maybe we'll get a few more people chiming in.
Does anyone have any experience with centroid? that's most probably a way out of my budget option but i wouldn't mind hearing others thoughts on it..

[bhurts]

I am far from an expert especially in electronics but I'll try to answer as best I can.

I have never done a tool changer but plan on doing so on this machine. The only thing I'm not sure about is the air motor driving the tool changer. I may replace it with a servo and gearbox but haven't decided for sure. I will try to keep the rest of the pneumatic elements even if I replace the motor. Either way I will probably replace the tool changer encoder since its absolute and k-flop does not support absolute encoders as far as I know.

I am using k-motioncnc and will continue to do so.

Yes k-flop is a motion controller and offloads all the realtime work from the pc. So I think in general you have the right idea.

I believe it supports most of the g-codes in mach 3 and is based off of emc2 which is the same thing Linux cnc is based on. It should support all or almost all the g-codes as Linux cnc.

Backlash comp, screw mapping, and tool offsets are all a yes. As for non standard g-codes I believe that is the case to a certain degree with mach3 , k-motion cnc, and Linux cnc. For the most part they are still capable of the same things as any other control but some codes are a bit different.

Hurco nc also branched from industry standard although they offer a industry standard interpreter option.

I have no experience writing software so the viability of adding wizards is unknown to me. Probably not easy unless you have a programming background. The good thing about k-flop is you can have your cam program open on the same pc as you use for the control. If you have a g-code issue you can just maximize your cam change the appropriate fields and repost. I wouldn't make drawing and cam a habit while g-code is being run but I have done so without any issues. Depending on how good you are with reading g-code you can fix minor issues such as cut depths and feeds/speeds right in k-motion cnc since it has a built-in text editor.

As for Linux/mesa I have no experience with either so any comments I make on them would be misleading. I'm sure there fine from what i've seen but not the choice I would make.

Ben

[Tkamsker]

hi i dont use mesa i use general mechatronics because of de bus add on boards but speaking of axis and io they are verry compareable (i almost went the mesa route and used 2 mesa examples for my build which i simply reconfigured for gm ) but in my case there is an carousel mecanic with 12 tools in it 2 contactless pnp switches and around 8 110 volt relais which control direction of carousel start stop pneumatics of the arm with carousel. and then i wrote a component which controls the carousel whenever an t1 m6 (or whatever tool ) command is issued i also do rigid tapping and spindle orientation ,..


Gesendet von iPad mit Tapatalk

[Tkamsker]

hi cwalker if go to an language i don't know i use https://translate.google.com the original reason was for the others in the project to document the ideas ,.. if i do public documentation i of course do it in english even if the autocorrection often spoil some of the words -

[m_c]

My first suggestion would be to go over the mill, and come up with a list of what input and outputs you'll need. Once you have that, then look at what hardware you'll need.

The toolchanger isn't too big a thing to implement. It's a case of ensuring you have suitable inputs/outputs, know the required sequence to get it to work, then program that sequence into the controller. It's similar for most CNC controllers, it's just the actual programming method that will vary.

Regarding how the KFlop works, you almost have it right. The KFlop receives the trajectory data from either KMotionCNC or Mach3, then uses that to command motion i.e. the KFlop is doing all the real time processing. The actual trajectory planner remains on the PC.

[Tkamsker]

Hi I agree with your First part but i learned the Hard way that toolchanger in machines from 1990 are a beast. Starting with the Trafo which is so heavy that melting the copper you could found the whole retrofit - which have me a lot of interferences doing ghost triggerings at the Pins i ended up building an atc Simulator (using my 3d printer ) to get the timing and error Handling right. I guess if the electronics are more modern your headaces might be smaller or i just was unlucky ..

Gesendet von meinem SM-N9005 mit Tapatalk

[bhurts]

I'm not to worried about wiring since I have all the electrical and mechanical schematics for the machine. As far as I can see everything is wired into the pc and I have identified everything I need to get it going. The atc has a board next to it that I believe controls everything. Will need more research be sure.

Ben

[Tkamsker]

is this atc board documented like pin x on gnd is arm forward pin y on gnd gives turn left. pin z on gnd initiate the turn pin w on high means arm is really in forward position etc etc if this is the case it is simply


Gesendet von iPad mit Tapatalk

[cwalker1960]

I have all the wiring schemes for my machine as well.. I think i'm more worried about getting everything programmed right , Spindle orient, tool release, positioning for z and carousel , air blast for spindle cleaning, etc.
A lot of details in the correct order and safety checks to insure it happened has to take place during an auto toolchange,, it concerns me.

[skullworks]

While I feel lucky that I'm not facing this problem myself - my KM3 knee mill has no tool changer other than the operator handcuffed to the machine...

The (best) easiest way to address this is to get a PLC unit (Programmable Logic Controller) to handle the tool changer. In a basic install all tools would be loaded into the spindle and a tool change would then stow the tool. Most likely you would not be able to identify which tool is in which pot as all those details would be managed internal to the PLC.

In use you would send the requested tool # and a start command and the PLC will return a operation completed signal telling the machine it can resume G-code execution.

Using the soft PLC in LinuxCNC you could likely look up which tools are where (which magazine pot) and more importantly add functionality to allow a diagnostic mode in manual/jog mode which would allow you to single step through each phase of the tool change cycle forward or backward. Of course to me the best reason is that I am cheap and would rather not buy a PLC then have to learn how to do its ladder programming.

Using soft PLC will use up a lot of I/O lines, but adding a MESA 7i90HD will allow you to add 72 more I/O lines for $59. Which brings up another question. What are your expectations for your new control console?

[engnerdan]

While I feel lucky that I'm not facing this problem myself - my KM3 knee mill has no tool changer other than the operator handcuffed to the machine...

The (best) easiest way to address this is to get a PLC unit (Programmable Logic Controller) to handle the tool changer. In a basic install all tools would be loaded into the spindle and a tool change would then stow the tool. Most likely you would not be able to identify which tool is in which pot as all those details would be managed internal to the PLC.

In use you would send the requested tool # and a start command and the PLC will return a operation completed signal telling the machine it can resume G-code execution.

Using the soft PLC in LinuxCNC you could likely look up which tools are where (which magazine pot) and more importantly add functionality to allow a diagnostic mode in manual/jog mode which would allow you to single step through each phase of the tool change cycle forward or backward. Of course to me the best reason is that I am cheap and would rather not buy a PLC then have to learn how to do its ladder programming.

Using soft PLC will use up a lot of I/O lines, but adding a MESA 7i90HD will allow you to add 72 more I/O lines for $59. Which brings up another question. What are your expectations for your new control console?

I am sure you will be able to log tools to pots if you can write the code. Being Kmotion has both tool numbers and slot numbers, I would let Kmotion handle the tool/slot/length offsets and just make requests to the tool changer. Personally I think PLC's are easier to program control logic in. So I would use a PLC for as much as I can, then setup Modbus RTU over RS232 between the KANALOG and the PLC.

This is what I am currently doing with my mill. I have an Automation Direct Do-More that handles as many non critical I/O and communications functions. So anything that is critical or time sensitive like feedhold, stop, e-stop, limit switches are wired right to the KANALOG. But things the cycle start, indicator lights, coolant control, enclosure lights are handled by my PLC. Additionally I have added a HMI to display some data from the machine and allow control of many functions, this is connected to the PLC. I have also connected a couple of the drives to the PLC via RS485 to allow for things like error decoding. Now if my spindle drive alarms/errors out I don't have to open my electrical enclosure to get the error code. The PLC reads the appropriate mod-bus register on the drive and one of the screens on my HMI diplays the error code as actual text, so error code ER-010 might be displayed on the drive, which then requires looking it up in the factory manual to determine what it is. I have it setup that the HMI does a message lookup for code 10 which might be "Drive Over Temp".

The modbus code for the KFLOP to talk to the PLC took me a while to figure out, with the help of some others. But now that I have it figured out I have it setup to exchange several data registers back and forth.

-Dan
-Dan

[m_c]

I wouldn't say using a PLC is the easiest way, as you then introduce the added complexities of having to talk to the PLC, regardless of what controller you use.
Plus you still need to ensure it's programmed with suitable failsafes, which you can do just as well without a PLC.

As for quoting costs, given this is an industrial machine, you'll most likely need to add 24V buffers/optos, which will bump up the cost.


For programming the ATC, my suggestion is write down a flow chart of the entire process, and understand that process, before you even consider programming anything.
Also write down what failsafes are needed and at what point in the cycle. That way you should know how things need to be done, and converting that to a suitable program/logic will be clearer.

[samco]

Andy has created a tool changer simulator demo in linuxcnc. (latest release - 2.7) This could be a starting point for other configs.

https://www.youtube.com/watch?v=PfZwpjUs1xI

one my my conversions use only linuxcnc (mesa cards for i/o). Tool changing/pallets was done in linuxcnc's ladder. The spindle gear change was done in a 'comp' C like language you can use to develop realtime components. I certainly could have done the spindle in ladder but I could wrap my head around it programmatically...

https://www.youtube.com/watch?v=39q6kvrSBSk

every machine is different - so some elbow grease is required.

sam