Hi John and Hood,

Thank you for your input.
I am not happy with the one pulse solution, as if there is a power glitch or when the system is powered up, there is no feedback or handshake to tell Mach where the tools are. This could lead to some interesting adrenalin rush bending and mangeling. Would it be possible for Mach to output one pule for tool 1 and two pulses for tool 2 etc. The toolpost logic could then send a pulse back to Mach to verify.

Even further, could the transmission be an eight bit serial word both ways to ensure no mistakes?

regards,

Max

Hi Max
   I would think you could do this a number of ways. Encoders, prox switches, etc. As for not feeling safe about the one pulse, that wouldn't worry me so much. I know what you are talking about with power glitches and such. I would wire in an e-stop contactor that is normally open, if the power flickers, it will kill everything. The plant I work at has a lot of huge, high end machines. They all wok this way. Most of the time it saves the day, every once in a while, bang, boom, run  . No matter how you prepare, a power glitch is never good on a CNC machine weather you are running Mach or a multi thousand dollar Fanuc controller. As far as trusting the PLC, I would have no problem with that either. They are very reliable. In fact, the machines we count on the most are custom made manual machines that are all PLC controlled. They are tied into a cell which also has some large CNC lathes. They are loaded and unloaded by large 6 axis robots, feed by automated conveyors, also PLC controll. I'm in maint. Guess where I spend the most of my time? It's not on the PLCs. I like the thought of one pulse. I'm no programer but I know you can put in timers and such to help eliminate false signals. I look at it like this. We use 8 bit encoders to verify positioning of our turrets. Now the controller does look to see that all bits are right. But in the end in still only sends one signal. Yes, we are where we need to be, go to the next work coord. I have seen false readings on encoders as well. This makes a mess. Another note, on our newest lathes, Fanuc controllers, they actually use PLCs for a lot of the IO. Just food for thought.

Hi Chaoticone,

Thanks for your thorough reply. And yes, PLCs are incredibly reliable and only fail due to faulty components which is extremely rare. not like computers (which go wrong minute by minute when on Windows!)

What I can't understand is how you identify the correct tool in the first place. If Mach assumes tool 1 and you were on, say, tool 5 from the last programme, how do you get back to tool one?

Max

Call tool 1 at the start of your program, if the PLC software is written right it will know where tool 1 is by the opto sensors.

Graham.

Hi Graham,

Thanks for that.
But, how do you tell the PLC that it is the start of a program for it to set tool one?

Is it possible to send one pulse for tool 1, two pulses for tool 2 and so on? By, say, calling for one or more than one "change tool" pulses within a defined time window and then the PLC would definitely know which tool to select.

Max

Using 2 outputs you can select 1 of 4 tools, then pulse a third output to activate the PLC.

Using more outputs you can select many more.

If you use a RS232 to 8 bit TTL convertor you could select 1 of 255 tools. Then swap the outputs to inputs and see what tool is in the cutting position.

Graham.
 

Hi Graham,

So, there is more than one output possibility. I need only three 0-7 and a pulse, making four lines. Is that possible? Are the outputs parallel or is it a one wire serial? I am getting confused.

Max

You could use a second parallel port or a Mod IO card form Peter Homann,  then its just a case of writing a bit of code to set the outputs and pulse the PLC.

This can be done in the M6Start macro.

The PLC software first reads the inputs and changes the tool. When its finished it can set the output.

Then Machs M6end macro waits for the finish command.

Graham.

Hi Max

So you're the guy with the Starturn 5 - we've just been discussing your lathe over on the denford forum (www.denfordata.com/bb). I've got a Starturn 4 converted to Mach at home, a mate of mine is converting an Orac, and I also use an Orac and Triac at work.

The basic working of the toolpost is quite simple. The toolpost should only rotate in one direction, and there is a ratchet that prevents it being pushed back. Basically the three optosensors shine light onto a BCD coded disc. So what you will get back is 3 lines giving you the tool number. When the correct tool number is detected, the toolpost motor is switched into reverse and the toolpost drives back against the ratchet.

So in theory only 4 wires are required. You will need to find the circuit that controls the toolpost motor and see how this is turned on and off. With regard to the signal lines these run at 12V which is a bit more inconvenient and will require you to make up some optoisolators. I started a project to do the electronics for my mates toolchanger - unfortunately a year later he still hasn't go his lathe going so i've only tested the electronics on the bench.

Basically I used an AVR microcontroller - I then wrote a program that read the an RS232 string - converted this to the toolnumber - activated the toolpost motor and then rotated the tool until the correct one was detected. It then sent an RS232 string back to Mach to tell it that the toolchange was complete. John Pearson has done something similar - see the VB code to drive the RS232 comms here -http://www.machsupport.com/forum/index.php?topic=1658.0

Hope thats some help

Cheers

Dave

Hi Max,

Did you complete the conversion on the Starturn5 using Mach3.

I need some help to do the same job.

Any actual details on wiring the steppers and encoders would be helpful.
I already built a 3 axis CNC flatbed router, and want to use the same electronics package to control the Starturn5.

Many thanks

George