Any home built CNC'ers - Mach3 E-stop
 

I've got a lashed together 3 axis mill/router.

I replaced the Dremel tool with a 12V DC spindle.
Now the limit switches (wired to the E-stop input) false triggers when
the spindle is loaded rapidly.

I've bonded the spindle supply gnd the motor/controller ground.
I've placed a 0.1uF cap across the E-stop loop at the controller.
I've placed a fly-back Schotky diode across the spindle motor terminals.

By configuring the E-stop for an active low input and shorting out the
switch loop I've proved its the spindle causing the false E-stop signals.

I currently run the E-stop loop open circuit with the switches closing
when the axes travel too far.

Can I simply swap the E-stop pin configuration to active low and run the
axis limit switches normally closed (opening when they are hit?

Any gotchas?

Thanks

Any home built CNC'ers - Mach3 E-stop
 

snafu fired this volley in news:5NqdnfiSF-
:

Can I simply swap the E-stop pin configuration to active low and run
the
axis limit switches normally closed (opening when they are hit?

That's the safer configuration, anyway. That way, if a switch fails (bad
contacts, etc), it shows an unsafe condition.

But you really need to find the source of the noise. Feeding the spindle
AND returning the limit switch signals via STP (shielded twisted-pair)
AND routing them so they aren't in _tight_ physical parallel to one-
another is likely part of the solution.

How's the spindle varied in speed -- variable voltage or PWM? PWM is
quite noisy.

"Loading" the E-stop with some significant current will probably help.
If it took (say) a 50ma swing in loop current to cause a signal change,
rather than a couple of ma would likely help, too.

Lloyd

Any home built CNC'ers - Mach3 E-stop
 

On 18-May-15 6:49 PM, Lloyd E. Sponenburgh wrote:
snafu fired this volley in news:5NqdnfiSF-
:

Can I simply swap the E-stop pin configuration to active low and run
the
axis limit switches normally closed (opening when they are hit?

That's the safer configuration, anyway. That way, if a switch fails (bad
contacts, etc), it shows an unsafe condition.

But you really need to find the source of the noise. Feeding the spindle
AND returning the limit switch signals via STP (shielded twisted-pair)
AND routing them so they aren't in _tight_ physical parallel to one-
another is likely part of the solution.

How's the spindle varied in speed -- variable voltage or PWM? PWM is
quite noisy.

"Loading" the E-stop with some significant current will probably help.
If it took (say) a 50ma swing in loop current to cause a signal change,
rather than a couple of ma would likely help, too.

Lloyd



Yep, normally closed is definitely preferred.

The source of the motor seems to be the 12V DC spindle motor. It seems
to trigger the E-stop when it unloads. ie at the end of a vertical
plunge into the work - I'm thinking back emf? Various caps + snubber
diodes tried.

Wiring is well separated but not shielded. I may shield it when I rewire
it depending on what cable I can find....

It's got a 10k pullup at the moment - only 0.5mA. Reducing that is
another thing to try.

Thanks for the ideas.

Any home built CNC'ers - Mach3 E-stop
 

Guessing from a distance: seperate the 12v supply from the logic/estop
supply. As a check, use a completely different supply - a 12v wall plug
type with adequate power if possible; or a stray computer supply.
Out of curiosity, what is the 12v spindle?

Hul

snafu wrote:
I've got a lashed together 3 axis mill/router.

I replaced the Dremel tool with a 12V DC spindle.
Now the limit switches (wired to the E-stop input) false triggers when
the spindle is loaded rapidly.

I've bonded the spindle supply gnd the motor/controller ground.
I've placed a 0.1uF cap across the E-stop loop at the controller.
I've placed a fly-back Schotky diode across the spindle motor terminals.

By configuring the E-stop for an active low input and shorting out the
switch loop I've proved its the spindle causing the false E-stop signals.

I currently run the E-stop loop open circuit with the switches closing
when the axes travel too far.

Can I simply swap the E-stop pin configuration to active low and run the
axis limit switches normally closed (opening when they are hit?

Any gotchas?

Thanks

Any home built CNC'ers - Mach3 E-stop
 

On Monday, May 18, 2015 at 9:41:44 AM UTC-4, snafu wrote:
On 18-May-15 6:49 PM, Lloyd E. Sponenburgh wrote:
snafu fired this volley in news:5NqdnfiSF-
:

Can I simply swap the E-stop pin configuration to active low and run
the
axis limit switches normally closed (opening when they are hit?

That's the safer configuration, anyway. That way, if a switch fails (bad
contacts, etc), it shows an unsafe condition.

But you really need to find the source of the noise. Feeding the spindle
AND returning the limit switch signals via STP (shielded twisted-pair)
AND routing them so they aren't in _tight_ physical parallel to one-
another is likely part of the solution.

How's the spindle varied in speed -- variable voltage or PWM? PWM is
quite noisy.

"Loading" the E-stop with some significant current will probably help.
If it took (say) a 50ma swing in loop current to cause a signal change,
rather than a couple of ma would likely help, too.

Lloyd



Yep, normally closed is definitely preferred.

The source of the motor seems to be the 12V DC spindle motor. It seems
to trigger the E-stop when it unloads. ie at the end of a vertical
plunge into the work - I'm thinking back emf? Various caps + snubber
diodes tried.

Wiring is well separated but not shielded. I may shield it when I rewire
it depending on what cable I can find....

It's got a 10k pullup at the moment - only 0.5mA. Reducing that is
another thing to try.

Thanks for the ideas.

You _definitely_ want a lower value resistor here. It doesn't take much to induce a half a milliamp of noise into a cable of any significant length. Your pwm(?) motor drive is, as stated before, inherently very noisy.

Just make sure your e-stop switch can handle the load, and that you use an appropriately rated resistor (100 ohm half watt will give you 50mA).

Any home built CNC'ers - Mach3 E-stop
 

On 19-May-15 7:54 AM, rangerssuck wrote:
On Monday, May 18, 2015 at 9:41:44 AM UTC-4, snafu wrote:
On 18-May-15 6:49 PM, Lloyd E. Sponenburgh wrote:
snafu fired this volley in news:5NqdnfiSF-
:

Can I simply swap the E-stop pin configuration to active low and run
the
axis limit switches normally closed (opening when they are hit?

That's the safer configuration, anyway. That way, if a switch fails (bad
contacts, etc), it shows an unsafe condition.

But you really need to find the source of the noise. Feeding the spindle
AND returning the limit switch signals via STP (shielded twisted-pair)
AND routing them so they aren't in _tight_ physical parallel to one-
another is likely part of the solution.

How's the spindle varied in speed -- variable voltage or PWM? PWM is
quite noisy.

"Loading" the E-stop with some significant current will probably help.
If it took (say) a 50ma swing in loop current to cause a signal change,
rather than a couple of ma would likely help, too.

Lloyd



Yep, normally closed is definitely preferred.

The source of the motor seems to be the 12V DC spindle motor. It seems
to trigger the E-stop when it unloads. ie at the end of a vertical
plunge into the work - I'm thinking back emf? Various caps + snubber
diodes tried.

Wiring is well separated but not shielded. I may shield it when I rewire
it depending on what cable I can find....

It's got a 10k pullup at the moment - only 0.5mA. Reducing that is
another thing to try.

Thanks for the ideas.

You _definitely_ wat a lower value resistor here. It doesn't take much to induce a half a milliamp of noise into a cable of any significant length. Your pwm(?) motor drive is, as stated before, inherently very noisy.

Just make sure your e-stop switch can handle the load, and that you use an appropriately rated resistor (100 ohm half watt will give you 50mA).



Mmmm, maybe I should just drop the resistor value first before doing too
much other work. It may just fix it. Thanks.

Any home built CNC'ers - Mach3 E-stop
 

On 19-May-15 5:53 AM, Hul Tytus wrote:
Guessing from a distance: seperate the 12v supply from the logic/estop
supply. As a check, use a completely different supply - a 12v wall plug
type with adequate power if possible; or a stray computer supply.
Out of curiosity, what is the 12v spindle?

Hul

Yep, they are already separate and as it happens I'm using a wall plug
unit for the e-stop etc.


The 12V spindle is a home brew thing.

I think the motor may have originally been off a large centrifugal fan,
maybe automotive.The motor is about 4" diameter x 5" long with a 1/4'
shaft. Its coupled via a small tooth belt to a shaft. The shaft is
mounted in a couple of bearings with a rotary tool collet fitted to the
bottom.


I'm only routing plastic with a 2 flute engraving bit so I don't need
too much HP. The Dremel was too fast and was melting the ABS plastic so
I changed to the 12V motor to drop the speed.

Any home built CNC'ers - Mach3 E-stop
 

On 18-May-15 5:21 PM, snafu wrote:
I've got a lashed together 3 axis mill/router.

I replaced the Dremel tool with a 12V DC spindle.
Now the limit switches (wired to the E-stop input) false triggers when
the spindle is loaded rapidly.

I've bonded the spindle supply gnd the motor/controller ground.
I've placed a 0.1uF cap across the E-stop loop at the controller.
I've placed a fly-back Schotky diode across the spindle motor terminals.

By configuring the E-stop for an active low input and shorting out the
switch loop I've proved its the spindle causing the false E-stop signals.

I currently run the E-stop loop open circuit with the switches closing
when the axes travel too far.

Can I simply swap the E-stop pin configuration to active low and run the
axis limit switches normally closed (opening when they are hit?

Any gotchas?

Thanks

I figure a follow up post is in order..........

I tried lots of different things, filter caps, changing pull up resistor
values etc.

I could not get Mach3 to work with the switches set in a normally closed
manner & gave up on that route.


What did work was using s shielded 2 wire cable. I kept the shield
separate from the grounded switch wire, commoning them only at the
interface pcb. Now its all good (for now!).

Thanks for the motivating suggestions gents.