Limit/Home Switch Strategy - Comments?

[Clockwork]

I'm in the final stages of assembly on router table for wood projects. Using commercial ball screw actuators with vitrually zero backlash.

I recently acquired some expensive Euchner limit switches for cheap on Ebay. They are the type that uses a rotary actuator shaft. Attached to the shaft is a 1.3 inch arm with a steel bearing at the business end.

The switches have high precision and are specified with an 8-degree difference between the NO and NC contacts. My big idea is to use the first contacts actuated (NO) for home switches and the second contacts (NC) for the limit function. It translates to about 2/10 of an inch between the home and limit positions.

If it works, it will be slick to have both functions provided by a single switch body. Don't know if 2/10 of an inch is enough buffer between the two functions - seems like it should be, but I'm new at this stuff. Any Opinions?

[Bubba]

What controller are you using? In mine, I have the option to use the home as limits and as a result, only one set of wiring! This also eliminates the potential for having space problems such as you mention (and you know Murphy will come into play here:})

[Clockwork]

Bubba,
Thanks for the response. Murphy's Law is exactly what I'm worried about.

I'm using Mach 3 with a hybrid controller setup. The motors are driven with a StepmasterNC controller. This unit doesn't have an option to connect any inputs. Even if it did, a standard parallel port only gives you 5.

I'm using a Ultimarc IPAC keyboard emulator for the inputs. The unit I have provides for 28 inputs. This allowed me to have separate inputs for plus and minus limits on the X,Y and Z axis and 3 for individual homing on the axis.

It also allowed me to build a push button control box on the router with plus and minus jog buttons for each axis along with an emergency stop button and a limit override button to jog off the limits. It should be very convienient compared to using the mouse and keyboard. Especially if an emergency stop is required. All total, I have 17 inputs heading back to the computer. Had to add an additional inverter chip to reverse the home signals.

Haven't wired anything up yet, I'm still thinking about it and looking for comments.

Hope that clarifies what I'm working with.

[jerber]

You realy want to connect limit switches through the lpt port.
Mach directly contols the lpt port.
If you do it through the IPAC there can be slight delay wich is not desirable with homing/limits.

Can't you make a "breakout" connector using a male and female db25
connector and some wires ?

Another option is to use some sort of modbus device (plc).
But I don't have experience with it yet.

[Clockwork]

Jerber

Thanks for the response regarding parallel port inputs vs. keyboard inputs. It was certainly a catalyst for thought and forced me to spend some time surfing the Web to dig out the details of each interface. Here’s what I found.

The time to transmit a complete key press serially from the keyboard to the computer is 324 microseconds. I rounded up to 400 as an allowance for the IPAC microcontroller to do its thing.

If the maximum stepper speed is 1000 RPM, the revolutions per second is 16.7. Using a 1.8 degree/step motor, the time between steps is 300 microseconds. A maximum of 2 steps will continue while the keyboard serial transmission is taking place. If the motor is connected to a 5 TPI ballscrew, this translates to a maximum of 2 thousandths of an inch of additional movement. Even if the motor speed is cranked up to 4000 RPM, it would only overshoot by 8 thousandths (compared to the parallel input) which is less than the mechanical slop in most limit switches.

Once the respective signals arrive at the processor, the keyboard interface is potentially processed much quicker than the parallel port interface. The interrupt priority level (IRQ) of the keyboard is set at a level second only to the system timer, the highest of all perhiperal devices. The IRQ for the parallel port (LPT1) is way down the food chain for getting its interrupt serviced so it may have to wait for a period of time before the input is seen by the system.

[Willbird]

I have not mounted limits on my bport yet (no motors other, so you know I'm not running it without limits). I keep thinking about putting a dial indicator in the system, most of them the stem goes clear thru...so it could bump the table, and the other end of it would bump the limit swtich, this would give me a way to verify the accuracy of the operation each time. Zeroing the indicator could be as simple as making a small jig with a setscrew the place over the bracket. And if you wanted to go belt and suspendors I think I can find a way to get a momentary switch in there too to back up the limit.

[marcusb]

Clockwork thanks for the calculations i was thinking of not using my IPac but now i have no reason not to again thanks

Marc

[skiddaw]

I came searching for these very calculations as I'm presently installing limit switches to the bed and carriages on my ongoing CNC router. Should consideration be given to the momentum of the system. A motor running at 1000 rpm would normally be brought to a stop using a velocity gradient down to zero. Simply removing power does not necessarily stop the rotation dead. I have no idea however as to how many revolutions, or part thereof, a particular motor might take to stop.

Would this be worthy of an experiment for those willing to test the operation of their limit detection circuit. Data to capture might axis, include motor equivalent, amps, stopping distance.

A say this because I was originally working on a push button design using 'recycled' devices from the front panel on a VHS recorder. These would have been inline with the carriage. Since then I've switch to typical micro-switches which appear to have a movement of 2mm on and 4mm max which I'm driving with a 10x30mm wedge attached to the carriage. Fortunately I have six of these with a roller and both NC and NO contacts. These came from old copiers; 'recycled' switches from old microwave ovens are often only single throw NO.

[stevespo]

I wish I had the time to do the experiment, but probably not.

I am using Honeywell 914CE microswitches for home and limits (2 on X and 2 on Y). The home switch also acts as a limit-- and the other switch is only limit++.

You're correct in recognizing that momentum affects the homing switch outcome. I noticed when homing my machine that the speed I was moving altered the actual "home" position. Often just by a .001" (or less), but measurable with a dial indicator. So, homing at 30 IPM gave me a *slightly* different home than when moving at 45 IPM, 60 IPM, etc.

However, now that I have a fixed homing speed - the actual home position is dead on each time. I wonder if other types of switches have more variability when they engage and that could also be an issue.

Steve