Press brake control

[Bigtoy302]

This is kind of like the other thread but a little different. I am going to be building a cnc press brake soon and I am trying to figure out the control on it. I am going to use a Parker proportional valve with on board electronics. All it needs is +/-10V signal. I would also like a stepper controlled 2 axis backgauge so I can do angles. I will have a quadrature linear encoder on the die holder for positioning feedback. I would like to use mach3 for this if I could. Can mach handle encoder feedback from Galil? I am assuming I will will need a Galil card to handle the valve and encoder and use the parallel ports for the backgauge? Is there a simpler way to do this? I checked out Delta's motion controllers but the $5225 price tag won't make that happen.

[Al_The_Man]

Referring to the last thread on this
press brake cnc conversion
I have the software and manuals, all you need is a DMC-1020 (2 axis) Galil card.
These are legacy cards due to them being ISA bus, but for these applications they are more than powerful enough.
I have got them off ebay for as little as $50.00 each.
I have several program examples, back-gauges etc, using the OPINT software.
It is extremely easy to use and screens can be set up in minutes once you gain a little proficiency.
Steppers do not use encoder/scale feedback, I would defiantly look at servo's if using a scale etc.
You can also use a scale with a proportional valve controlled by Galil.
The DMC-1000 manual can be had off the Galil site and you also just need to learn the Galil two letter commands.
Al.

[Bigtoy302]

Referring to the last thread on this
press brake cnc conversion
I have the software and manuals, all you need is a DMC-1020 (2 axis) Galil card.
These are legacy cards due to them being ISA bus, but for these applications they are more than powerful enough.
I have got them off ebay for as little as $50.00 each.
I have several program examples, back-gauges etc, using the OPINT software.
It is extremely easy to use and screens can be set up in minutes once you gain a little proficiency.
Steppers do not use encoder/scale feedback, I would defiantly look at servo's if using a scale etc.
You can also use a scale with a proportional valve controlled by Galil.
The DMC-1000 manual can be had off the Galil site and you also just need to learn the Galil two letter commands.
Al.

I don't need feeback on the backgauge. I wanted to use steppers on the backgauge to keep it simple. The scale would be for the hydraulic part. So I would need a 3 axis system? Two for the backgauge and one for the hydraulic die control? I would not even need mach at all?

[Al_The_Man]

The Galil card can be configured stepper or servo by jumper for each axis.
The program I run does not even need windows.
I do not remember ever seeing a 2 axis back gauge?
The machines I have done so far with this method are 2 axis Tube cut-off lathe, 40ft 3 axis Angle Iron punch, 3 axis rolling machine, Cincinnati point to point Horiz. mills.
Back gauges.
Al.

[Bigtoy302]

The Galil card can be configured stepper or servo by jumper for each axis.
The program I run does not even need windows.
I do not remember ever seeing a 2 axis back gauge?
The machines I have done so far with this method are 2 axis Tube cut-off lathe, 40ft 3 axis Angle Iron punch, 3 axis rolling machine, Cincinnati point to point Horiz. mills.
Back gauges.
Al.

Many of the new cnc press brakes have 4-5 axis back gauges. Each side is independent from each other and they both can go side to side and up and down. All I want to do is move each side independently so I can bend angles. It would not be hard to put a motor and screw on each side of the back gauge. Programing might get a little bit fun.

[One of Many]

Cincinnati Press Brakes had what they advertised as 6 axis back gages. Basically twin 3 axis gages that can move side to side, independent, up and down and fore and back of course. These were servo driven ball screws with encoder feedback. The interface cards were stand alone and addressed in such a manner that the control would broadcast data streams to all, but each would only accept its command as address;mnemonic;position data. There were also means of communicating to setup home position, speed and envelope limits as parameters. The HMI was all done at the screen to accept the desired program as the software calculate by conversion, then transmitted, to what the gages required behind the scenes.

Some of the servo valves were setup as a master/slave relationship with 10'nths resolution encoders on each side. There were quite a few parameters that helped control the ram position tracking, which relied heavily on counter balance valve pressure settings. No easy task recreating that piece of software.....with high frequency feedback for +/- .0004 repeatability and tonnage protection. Which was critical on thinner gage materials and narrow dies.

The project sound like quite the challenge. Designing the backgage shouldn't be that tough, but controlling the ram position and tonnage and structural strength? Yikes!

DC

[Bigtoy302]

Cincinnati Press Brakes had what they advertised as 6 axis back gages. Basically twin 3 axis gages that can move side to side, independent, up and down and fore and back of course. These were servo driven ball screws with encoder feedback. The interface cards were stand alone and addressed in such a manner that the control would broadcast data streams to all, but each would only accept its command as address;mnemonic;position data. There were also means of communicating to setup home position, speed and envelope limits as parameters. The HMI was all done at the screen to accept the desired program as the software calculate by conversion, then transmitted, to what the gages required behind the scenes.

Some of the servo valves were setup as a master/slave relationship with 10'nths resolution encoders on each side. There were quite a few parameters that helped control the ram position tracking, which relied heavily on counter balance valve pressure settings. No easy task recreating that piece of software.....with high frequency feedback for +/- .0004 repeatability and tonnage protection. Which was critical on thinner gage materials and narrow dies.

The project sound like quite the challenge. Designing the backgage shouldn't be that tough, but controlling the ram position and tonnage and structural strength? Yikes!

DC

I am going to use a single 8" ram in the center so I don't have any balance problems. The die holder will guided on linear rails like the Boschert Profi brake. I am only making the brake 36" wide.

[PhoenixMetal]

For any of the accuracy afforded by the linear scales (linear encoders) to matter, you're going to have to build a very very rigid machine. Its going to need to be seriously overbuilt with a whole lot of steel so that the machine doesn't deflect under pressure.

Also, I would be concerned that the linear rails may not be rigid enough to deal with angular loads that would be generated by off center loading (bending parts not in the center of the bed). I just took a look at that Boschert brake online, and that is not the usual way how this is done. Usually the ram is guided on "ways" which are much beefier than a linear rail.

Hey, so long as you are taking on a project this complex, and so long as you want to make a really accurate machine, you should take a look at toyokoki press brakes. They are the most accurate machines that exist - ram repeatable to 0.00004" (1 micron). On those machines the ram is driven by servo motors and ball screws - instead of hydraulic cylinders.

To answer your original question regarding Galil cards and Mach - if you have a Galil card running the system, you don't need Mach. You would program and control it using Galil's software. A Galil system consists of the card (in your computer), and an Interconnect module (external) which has screw terminals to connect to the wires from your encoders, limit switches, etc.

If you use a stepper motor for the backguage, be sure that it is a very powerful one, otherwise you will have to run it annoyingly slow to prevent it from missing steps. Also, if you go this route, put an encoder on the stepper. Galil's software has a built-in function to use encoders with stepper motors - in which it corrects for missed steps. This essentially makes the stepper as accurate as a servo system but it costs a lot less. However, if you can find suitable servo gear cheap, that would be better than using a stepper.

[Al_The_Man]

Galil's software has a built-in function to use encoders with stepper motors - in which it corrects for missed steps. This essentially makes the stepper as accurate as a servo system but it costs a lot less. However, if you can find suitable servo gear cheap, that would be better than using a stepper.

I have only used Servo's with Galil, but according to their documentation I have read if you set the card for stepper operation, you do not have encoder feedback loop.
All the manuals I have mention open loop with steppers.
Al.

[PhoenixMetal]

Hey Al - Galil's encoders-with-steppers feature is called "Stepper Position Maintainence Mode." It isn't the same sort of closed loop as what you have with a servo, but basically the end result is the same.

The way it works is: After the move is complete, it compares the number of steps the encoder moved with the number of steps that were commended. If the difference is above the tolerance that you set, it does an additional correction move equal to the difference.

Truth be told, I couldn't get the Stepper Position Maintainence mode to work suitably (I was still getting errors), so I simply wrote my own code that did exactly that. My correction code would loop through the correction process (correcting multiple times if needed) until the position was within my tolerance. My programs would invoke the correction code after each position move. The positioning was perfect with my code.

This sort of thing wouldn't be suitable with a milling machine for example, since it is technically possible for the thing to overshoot and then correct, but for a press brake backgauge, its just fine. Positioning as reliable as a servo, but at a stepper price.

[Al_The_Man]

I don't know whether you are familiar with the old Opint S/W I mention, that was/is an excellent piece of s/w for things like back guages etc, I have done point-to-point mills, rolling machines and tube cut off lathes etc.
I am just experimenting with extending it to some sort of standard CNC interface.
The trouble was, it is DOS based and they did not carry on support after the lost popularity with ISA & DOS, otherwise it does an bang up job, I can write operator interface screen in minutes.
Essentially when the operator enters commands or data or reads in a file through the HMI, it applies this data or commands to a pre written program in Galil native language and then runs it. It does not send one command at a time.
It can monitor position or the content of any variable and display and update it constantly while running.
If the owner so wishes and is able, he can easily modify or write screens himself, by turning on the editing function.

I am just sorry that things such as the advent of Windows squashed what could have made a nice application even better had it been allowed to evolve.
It has made me alot of money!
Al.

[PhoenixMetal]

Hey Al, well it sounds like its time for you to learn to program in one of the newer windows-friendly languages!

What you describe (a program that can create an operater interface, handle lots of things simultaneously and interfaces with the Galil hardware with prewritten chunks of Galil code) exists in a modern version! Its called Galil's ActiveX Toolkit (for writing in ActiveX), and they also have a C++ API.

But that's a whole endeavor learning a new (serious) programming language like C++. I took C++ for a few semesters in college, and its not something that you can take up lightly. But who knows, maybe the ActiveX language is easily approachable. I think it is.