Messages

3651

General Mach Discussion / Re: Limiting 5th axis rotation, new setup

« on: September 15, 2018, 10:21:45 PM »

Hi,
my preference would be something like the attached.

Note that the home switches are independent of all limit switches. In an industrial machine if any of the limits switches activates the machine
de-powers, not just stop, but kills the electrical supply.

Thus when you home the machine, if you combined a home and a limit switch, then the machine would stop. Mach can handle this by ignoring limits
when the machine is homing. That means when homing the machine is in a potentially dangerous state, it could accidentally exceed it limits and Mach
would not shut down. For large equipment this would be contrary to safety law in many countries.

Note that in industrial machines each limit switch would have its own input as well, it would allow the controller to permit the operator jog the machine
in the direction that brings the machine back into bounds and disable th jog direction that makes it worse. This would require even more inputs and wiring
for which very few hobbyists bother.

While I recommend the advantages of having separate and multiple inputs for limit and home switches you may, and probably are, constrained
by the number of inputs on your controller or BoB. If you have only two inputs to spare then your previous diagram will work but Mach will not know
which limit switch has just operated, just that one has. Likewise should any of the home switches operate while Mach is expecting the Y axis home
switch to meet its pre-programmed homing order Mach will interpret it as Y Home, when it was in actual fact an false activation of the B axis home switch
say. Consequently the Y axis would be mis-referenced and a crash is now likely.

3652

General Mach Discussion / Re: Limiting 5th axis rotation, new setup

« on: September 15, 2018, 08:55:41 PM »

Hi,
no not really, but looking at what you have drawn you probably don't have a choice, you don't have enough inputs to do as
I suggested.

I would in preference, I must point out that I use an Ethernet SmoothStepper with two ports equipped with BoB's which give 18 inputs
and 16 outputs, have all your home switches on separate inputs. I would have your X, Y and Z axes limit switches in series on another input.
I would have your two fifth axis limit switches in series on yet another input. Given that you fifth axis is likely to suffer damage it makes sense to
have limit switches and an input dedicated to monitoring that axis.

My suggestion would require four inputs for the home switches an two inputs for the two limit switch circuits, for a total of six inputs. Do you have
six inputs to spare?

Craig

3653

General Mach Discussion / Re: Limiting 5th axis rotation, new setup

« on: September 15, 2018, 03:25:19 PM »

Hi,
the ideal is to have three switches, one home and two limit switches. Industrial machines are wired that way.
It means that there is no confusion between a limit or a home event. That does require that your controller or
breakout board has sufficient inputs, not always the case with hobby CNC gear.

I personally would fit a single home switch on a dedicated input and use Soft Limits. If you want the extra protection of limit switches then fit them
and wire them in series to a separate input.

Craig

3654

General Mach Discussion / Re: Limiting 5th axis rotation, new setup

« on: September 15, 2018, 10:33:28 AM »

Hi,
if you use two switches, one at each end of travel and wire them together how will you home the axis?

If you use the one wire then it would have contact at 100cw but also 100ccw, so which is home?

Craig

3655

General Mach Discussion / Re: Limiting 5th axis rotation, new setup

« on: September 15, 2018, 12:38:56 AM »

Hi,
it is common to use a limit switch as a home switch also on an axis.

Take say the X axis with two limit switches, one at each end. Lets say the leftmost one is X-- and X Home and the rightmost one is X++.
Your Soft Limits on the Homing and Limits page would be 0 (0 being home AND X--) and X++=400mm for an X axis of 400mm travel.

Imagine you have only one switch, a home switch on the X axis. Lets say that it is 100mm from the lefthand extreme of the axis. Your settings on the
Home and Limits page would be X-- = -100 and X++=300mm. Note the travel (400mm) is the same in both cases its just that the zero point is somewhere
in between the left and right extremes of the axis.

You want to program limits into your fifth axis. On Machs General Setup Page check that the fifth axis, B, is in angular units and also that rotational
soft limits are engaged. I suspect you will want Rotational Rollover unchecked.

If you have one B axis switch at the center point of rotation (home) then your softlimits would be -100 and +100. If you had two limit switches at each end of travel
with the home at the ccw end then the limits would be 0 and +200.

Provided you reference (home) your machine when you turn it on prior to use and your soft limits are set correctly they work pretty well. Mach will refuse,
in fact throw an error, if the Gcode or jog inputs attempt to cause an axis to go out of bounds. If you forget to reference (home) your machine then its
entirely possible for your machine to go out of bounds because you forgot to tell it where it was at the start, so even having the boundaries correct doesn't
help because it didn't know where it is to start with.

The safest method is to have limit switches, one at each end of travel. Most high speed industrial machines have separate limit switches like that, it means that
if any one of the switches is activated the machine is shut down.....no ifs....no buts.....completely de-powered, its a safety measure. Such a machine will have
soft limits as well set to go off before the machine touches the physical limit switch. The idea is that 99.9% of the time the machine will work within the soft limit
boundaries and only as a result of some dangerous failure will the limit switch activate, and to comply with safety laws shut the power off.

Craig

3656

General Mach Discussion / Re: cutting a spiral

« on: September 15, 2018, 12:07:35 AM »

Hi,
codes G2 and G3 are circular arc/helical interpolation moves.
See for instance 'CNC Programming Handbook' by Peter Smid, its pretty much the Gcode bible.

They are fixed radius, you cannot to my knowledge use them as a spiral.

The way to do it is use CAD to draw the part including the spiral toolpath then use CAM to generate the Gcode.

Craig

3657

General Mach Discussion / Re: CS Labs CSMIO/P-A

« on: September 14, 2018, 03:40:18 PM »

Hi,
3 rpm represents an effective input voltage of:
V_in=3 /300 (rpm/V)
   =0.01V or 10mV

That could be the input bias voltage of the input amplifier alone. You can chase your tail for hours and then the temperature changes and its out
again. Close the position loop already!

Craig

3658

General Mach Discussion / Re: CS Labs CSMIO/P-A

« on: September 14, 2018, 03:25:23 AM »

Hi Mick,
your servos are in velocity mode. An analogue voltage at the input will cause the servo to rotate. It is tunable, something like
300 rpm per volt. Even when there is no input voltage there is still electrical noise, thermal noise if nothing else, that's a fact of physics
and you could end up with some net small voltage which will cause rotation.

When you hook up the CSMIO it will close the position loop. Thus if Mach commands 3184 steps, say, or 114.624⁰ past the index point and
should the servo drift a few tenths of a degree due to noise then the CISMO will detect that the servo is not at its commanded location and make a voltage
correction such that the servo moves back towards 114.624⁰ again.

If your servo were in position mode it would not rotate at idle, the position loop is now in the servo drive but it has the same result as the CSMIO, namely any drift
will be detected and a restorative action will be taken by the drive.

Craig

3659

General Mach Discussion / Re: Z losing zero position in rapid move up

« on: September 13, 2018, 02:37:31 AM »

Hi,

I was thinking more along the lines of a backlash problem

You may very well be right, I hadn't considered backlash. If backlash is an issue why is it that it occurs when you are jogging
not during a Gcode job?

A reduction in acceleration or velocity or even both may prove effective and therefore point you in the right direction. If the problem remains unchanged
then you can all but rule it out, either way its still a useful experiment.

Craig

3660

General Mach Discussion / Re: Z losing zero position in rapid move up

« on: September 13, 2018, 12:48:10 AM »

Hi,
sounds like its losing steps, try reducing the acceleration of the z axis a bit.

When running a job the machine tends to run at a percentage of max speed/acceleration whereas
your jogging will be at max.

If that fixes it you might have to look further as to why its doing it now but not before, maybe some
mechanical binding or similar.

Craig