Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: GytarrMann19 on April 06, 2012, 08:29:48 PM
-
Hello all,
I'm working on retrofitting a Hardinge CHNC. It's running pretty well now, but I'm having troubles with my original solution to turret indexing. I'm using a stepper to index the turret as if it were a rotary axis. I have one home switch which corresponds with tool 1/0deg. From there its just a matter of indexing to each tool position by degrees. The turret locks in place via a raise/lower situation.
Unfortunately, it's proven unreliable. There's a lot of backlash in the gear train that drives the turret, and the steps per degree don't work out to an even number causing further frustration.
I'd like to set it up with 8 sensors, one for each position, and I have enough inputs left for them.
So my question is, could this type of setup be easily controlled by a macro? Just run the stepper till it hits the right sensor? I'd rather go this route as I have no experience with PLCs or ModBus.
Thanks for any ideas,
Kevin
-
I am not sure you need EIGHT sensor inputs. Put a good RC filter on each sensor output and then combine them all together, and bring them in on one pin. (How you combine them depends of what sort of sensors you have.)
Then when you want to index to another tool, watch the input pin. It will start giving an 'active' signal (meaning the turret is at one tool), then it will go inactive while the turret advances to the next tool. When it gets there the pin will go active again. Assuming good filtering so you don't get noise, this should be fairly reliable. It can be done with a macro.
Cheers
-
Suggest you do a search and have a look at what Hood has recently done. My only comment is that you want the turret to lock up tight.
RICH
-
My preferred method would be to have triggers at each position and use multiple switches for each tool, that way it should be impossible for it to be at the wrong tool and at the same time only use 4 switches for the 8 positions.
Hood
-
Hood, can you suggest a possible arrangement for the switches as you describe them? My plan is to use stationary Hall sensors and a rotating disc with magnets.
Kevin
-
Well for 8 positions you would need 4 switches and they could be something like this
Tool 1 = 0 0 0 1
Tool 2 = 0 0 1 0
Tool 3 = 0 0 1 1
Tool 4 = 0 1 0 0
Tool 5 = 0 1 0 1
Tool 6 = 0 1 1 0
Tool 7 = 0 1 1 1
Tool 8 = 1 0 0 0
Hood
-
you can do it with three switches.
000
001
010
011
111
110
101
100
-
Not really as when it was in between positions it would think it was at a tool position if you had 0 0 0 as one tool position.
Hood
-
that's only true if you allow them to go to all zeros between changes
-
How would you not?
Hood
-
reflective disc and sensors.
make before break rotary switch
-
The OP said
My plan is to use stationary Hall sensors and a rotating disc with magnets.
Hood
-
have the magnets set up on a disc so they dont have a gap, using that flexible magnetic strip
-
Yes a possibility but then you would have to be accurate and with the backlash in the gears it may not be so good, but possibly worth a try to save on a switch if you needed to.
Hood
-
4 sounds good to me too
-
there will be a fair bit of work to get them to fire at the same time anyway (a read delay in the macro possibly). An 8 position rotary switch would be the easiest to work with.
-
Possibly if it would stand up. I have just written a ladder for a friend who has a duplomatic 8 pos turret and it uses a rotary switch with 8 positions so I suppose they can last, then again I dont think its a standard off the shelf switch they use.
I use a servos with indexing drive on mine and makes it a breeze :)
Hood
-
Barrafuldi use one in there 8 station toolchanger, and it's been known to fail.
-
My original turret used normal limit switches and it worked well, it was 6 pos and just had 3 different cams (one for each switch) and it worked well but when I first got the lathe they were a bit hit and miss. I put it down to the switches internal springs being weak, the lathe hadnt been used much in the 20 or more years before I got it, probably only about 50hours so some of the switches had likely been sitting in a compressed sate for years at a time. I replaced the limits and it never missed a beat up until a month ago when I hauled it off to replace with the new turret.
I think if I was going to use switches again I would use optos and slotted discs or possibly if space wasnt a problem some of the honeywell limit switches that I use on the axis for limits.
Hood
-
Gytarr,
You mention avoiding PLC's but my take on this is that a PLC could really simplify the design. I say that having 12 years in factory automation that almost eveything ran on PLC. My concept on this would be to use a combined mechanical and electrical/electronic solution much like the orginal turret. If you could index by turning the turret with a stepper or other motor to get you into the near vacinity, turn the motor off, then use a simple air cylinder to clamp the tool table secure with air pressure, you could get a secure repeatable position on every tool change. This is a bit oversimplified but you get the concept.
Regardless of my opinion I am fascinated by the concept of a tool turret and would like to keep tabs with you on how you work this out. Gang tooling is killing me!
Robert
-
I'm using the existing turret. It was originally turned by an air motor, and upon triggering a sensor, an air cylinder shot a catch out into a ratchet wheel type of deal to stop the turret at the right position, at which time another air cylinder clamped it down on to a set of tapered lock pins. Trouble is, that design was endlessly problematic. So I'm using a stepper to rotate it into position, at which time it locks onto the pins via the original air clamping system. Trouble is, if the stepper didn't get it exactly where it had to be, it rotates slightly upon locking. Which means it starts slighlty out of position for the next tool change. The problem keeps compounding, and after 10 or 20 tool changes, its out of position enough that it can't lock. Why doesn't the stepper always get it there? The steps per degree don't work out to an even number. Stir in a little backlash, and the whole setup fails to be reliable. So, I'd like to use sensors at each position to stop it, rather than rely on the stepper alone, because it really has no way of knowing that it's gotten out of position slightly.
I know this can be implemented via PLC, but I don't know jack about them, or how to utilize modbus to talk to them. I have enough inputs left to just use mach, whether using 8 inputs, or using some sort of switch matrix that could reduce it to four inputs.
Kevin
-
Hood,
Space is very tight for my turret sensor, so I planned on using small Honeywell halls, in a TO-92 package. They would end up being epoxied into a plastic block, and then the arm/wheel with magnets could just ride over the top. The original 8 position sensor was made this way, and I modified it with one of these honeywell switches to make the turret home sensor I'm using now, which has worked well.
That being said, I'm still kind of unclear on how to arrange the switches and magnets mechanically, to utilize your suggestion about only needing 4 inputs instead of eight.
Thanks again for everyone's replies!
Kevin
-
I use PLCs with my machines and it makes things like turrets so easy to automate and really they are quite easy to use or at least the Direct Logic ones are, I have written quite a few ladders for toolchanger, 3 for myself, one for a guy in the USA with a 12 pos turret, one for a guy with a toolchanger on his mill and only a few days ago I did one for a friend that is about to put an 8 pos turret on his lathe.
Having said that it can be done with direct inputs/outputs to/from Mach and VB.
For the sensors you would just need differing magnets at each position. If you look at the example I gave earlier
Pos 1 would have 1 magnet to trigger switch 1,
Pos 2 would have one to trigger switch 2,
Pos 3 would have two to trigger both switches 1 and 2
Etc
Hood
-
you could still use mach, use a second parallel port with pins 2-9 as inputs and have a common feed to the turret and eight switches (electronic or otherwise). Use the inputs and write the code to look at each position/input, it would be pretty easy code to write.