twin headed hydro punch how to cnc

More
20 Aug 2009 21:38 #736 by jd896
the nature of the machine is to punch steel sectons for cantlever racking (the bolt holes for arm mounting) just i am having trouble thinking of firstly a way to move a 250kg beam and stop it accurately an how you would program the machines to punch independantly as they are off set i.e. they have to start and stop punching at different points, also how would the machine know where the start of the beam would be, as to set its start position

i will upload photos of the machine as i know they will help to understand the task

many thanks john

Please Log in or Create an account to join the conversation.

More
21 Aug 2009 15:44 #738 by BMG
Sounds like you will need a beefy setup for moving your parts in and out.

For initial placement, you could that the beam placed into a table with a backstop for zeroing the first axis and a second stop which the beam is snugged against for zeroing the second axis. Perhaps hydraulic dogs could then clamp the beam onto the table. Once clamped, the table and beam would move together under the control of a CNC positioning system. The stamping could be triggered by a Z axis move.

Once you can move the beam and table under the punch, the punch locations are trivial gcode.

Please Log in or Create an account to join the conversation.

More
21 Aug 2009 20:04 #739 by jd896
the problem with that is the beams need to be moved by the top flange as the section size changes depending on the weight the racking will have to carry. only the one direction of movment is needed as the punches are permenantly set to the same centers, ie the center of each side's flange and guide wheels keep the beam centered, the drive could as far as we can think of needs to be some sort of wheel gripping the top flange possibly a stepper motor and geared down if needed speed of the beam moving is not important it's just consistancy required as i said i will get pictures of the machine as this will definatly help with the understanding

thanks john

oh and g-code hard to understand and program ?

Please Log in or Create an account to join the conversation.

More
22 Aug 2009 12:28 - 22 Aug 2009 12:29 #741 by BigJohnT
The basic g code is pretty simple if your just doing straight line moves.

G1 X1.25 F25

The above g code means move the X axis in a straight line from where it is to 1.25 from the X0 position at a feed rate of 25.

I use the spindle on to turn on and off my plasma torch and suspect for a punch you might do the same. You would need to provide feedback when the punch cycle is complete before moving to the next location.

Moving a 550 pound beam around with a stepper is a pretty tall order... How do you move them now?

and pictures are a plus...

John
Last edit: 22 Aug 2009 12:29 by BigJohnT.

Please Log in or Create an account to join the conversation.

More
22 Aug 2009 13:28 #742 by jd896
well the progaming should be ok then

we move the beams manuly at the moment jus t using the already punched holes to drop a pin in thats on the end of a lever and that is used to pull the beam across to the next hole

also could it be programed to have two spindles as the punches need to be opeated seperatly at points
and the picture are comming

Please Log in or Create an account to join the conversation.

More
23 Aug 2009 13:20 #744 by BigJohnT
For that you would want to use the M1nn command.

www.linuxcnc.org/docview/html//gcode_main.html#sec:M100-to-M199:

For example your g code might look like this

G1 Y1.5 F25 (move to first location)
M100 (left punch cycle)
Y2.5 (move to second location)
M101 (right punch cycle)
Y4.5 (move to third location)
M102 (both punches cycle)
G0 Y0 (return to home position at the rapid rate)
M2 (end of program)

Some details here on how to wait for a stroke finished signal.
wiki.linuxcnc.org/cgi-bin/emcinfo.pl?CustomMcodesToHal

John

Please Log in or Create an account to join the conversation.

More
24 Oct 2011 20:03 #14187 by jd896
hi again sorry to drag this back up i have been thinking about this latly again and have a bit more of a clue asto hoe too do it i think

right the beams will be driven buy a pair of rollers pinching the top flange of the beam this drive will almost deff need to be a servo possibly ac motor and encoder type the question here is how do i interface to a servo drive/ inverter drive and does the servo/ inverter have the encoder feedback or does the pc have this back

next the machine would have to have an independant encoder to also run pn the flage of the beam maybe two as not too loose the beam through both punches (as they are 600mm appart) may also need twin drive but still single axis so no problem there ?

the next thing how could emc cope with getting a beam in i.e. loading as the one of the encoders wouldnt have a signal till 600mm down the beam and loading in general as it would also have to be the zeroing process i suspect ie find end of beam and set 0 position and start program

right lastly i think lol

the interface would it be possible to have the machine ask for the hole centers ie the starting patteren and the the pitch (centers) for the remainder of the beam and also if possible check the overall lenght of the beam , ideally this would be in the way of ask for the info and then generate the gcode for program and then prompt start cycle

sorry it a big one and thanks for the info so far good buch of users hear

many thanks john

Please Log in or Create an account to join the conversation.

More
25 Oct 2011 14:31 #14228 by andypugh
jd896 wrote:

right the beams will be driven buy a pair of rollers pinching the top flange of the beam

Why not pinch the centre flange? I think this might make it easier to handle different section sizes.
Or does the beam hang below the rollers? In that case I guess top-flange rollers keep the beam under the punch at a consistent height.

this drive will almost deff need to be a servo possibly ac motor and encoder type the question here is how do i interface to a servo drive/ inverter drive and does the servo/ inverter have the encoder feedback or does the pc have this back

It depends... Typically with EMC2 you use a very dumb drive, and the encoder feedback goes to EMC2, and all the calculations are performed there.
Try to get a matched motor and drive, and make sure the drive takes +/-10V input, to keep things simple.
If you get a real bargain price on a brushed or brushless motor, though, we can advise on drives that work well with EMC2.

I would be a bit cautious of relying on friction drive for accurate positioning. I would be tempted to use a system with a hydraulic/pneumatic gripper on a loop of chain.

However, I understand that roller encoders work better than you would expect, so that might not be an issue. You probably know better than me.

the next thing how could emc cope with getting a beam in i.e. loading as the one of the encoders wouldnt have a signal till 600mm down the beam and loading in general as it would also have to be the zeroing process i suspect ie find end of beam and set 0 position and start program


You could probably work round the offset problem by connecting the encoders by a belt. If that is difficult, then there ought to be a way to achieve the same effect in software. It might be as simple as using HAL to choose the input from the fastest-moving encoder.

I am not sure that G-code is the ideal interface for this machine. You almost just want to load a file with two columns of hole positions. This could probably be handled by a much simpler bit of software than Axis.

I think you would need someone to write you a simple input filter that took the two columns of data (from Excel, for example) and translated into G-code.
(Thinks) Actually, if you have a WIndows PC and a copy of Excel I can write an Excel macro to spit out the G-code with very little effort.

Please Log in or Create an account to join the conversation.

More
25 Oct 2011 20:40 #14255 by jd896
The top flange drive system I'm currently drawing up and will post a render of this when done by it will be having a floting top driven wheel and a lever and ram assem to really grip the flange ether hydro or air not shure yet will have a look when have something together


Thanks for the drive output take it the dc voltage would be produced by a servo driver say geko type thing see I ment a 3 phase angular driven motor and corresponding drive I will find some needed torque figures and see about this when have settled on a roller dimamiter

And the seperate encoder is because I have seen this on flat bar processing machines so if it's Good for geka then it's fine with me kinda thing

Also the loading thing was I ment could we not have a button on the interface to go intoo say a loading. Mode where buy the machine would run the drive at a low preset rate untill a home switch or the encode reaches the beam and then the machine would know its start point and zero it's self ready to run the program

Lastly the my understanding of the axis interface is that it is coustomisible ( cannot spell buy the way ) so that I could as said have it have some boxes to put the hole centers intoo and then just spit the gcode into emc and run the program as it would only be as simple as punch x amount of mm for this many holes then this for the remainder there would be a brake inbettween. But that's pretty much it

Please Log in or Create an account to join the conversation.

More
25 Oct 2011 21:16 #14257 by andypugh
jd896 wrote:

Thanks for the drive output take it the dc voltage would be produced by a servo driver say geko type thing


I was referring to the command voltage to the drive. The high voltage / high current to the servos would be handled by a commercial drive, and the low mA commmand voltage would come from EMC2.

Also the loading thing was I ment could we not have a button on the interface to go intoo say a loading. Mode where buy the machine would run the drive at a low preset rate untill a home switch or the encode reaches the beam and then the machine would know its start point and zero it's self ready to run the program


Yes, quite possibly the "Home" button. (You could make Axis unhome at the end of each workpiece)

Please Log in or Create an account to join the conversation.

Moderators: cncbasher
Time to create page: 0.085 seconds
Powered by Kunena Forum