Help Request - EStop has a waveform on my mill...

I have been converting my SpectraLight 0200 which normally requires a proprietary ISA card (and equally proprietary software) to a generic parallel port interface. The goal of this project is to be able to operate the mill using EMC2.. It has been pretty simple, it only took a couple of evenings with the o-scope to sort everything out.

I have nearly everything working. However I have two signals from the light machines corp controller that I can't determine how to interface. The two signals are Z Limit and E-Stop. Both of these signals are 5v carriers with a pulse every 25us (See attached image). This pulse drops to near 0v, jumps up to near 10v and then returns to 5v. It seems that these two signals are doing double duty, providing a clock source when the mill should be running, and dropping to 0v when it shouldn't be running.



The problem is the drop to near 0v. Every so often (always less than a minute, usually 3-10 seconds) EMC notices the low voltage with predictable results.

Is there a way to tell EMC how long the signal should be "held" before accepting it?

Details of my conversion are available at build.spaceopera.org/

Thanks in advance,

Shannon

Are you sure that isn't just noise from a step motor driver, sure doesn't look like a valid digital signal
but rather looks like noise from a motor drive coupled into a high impedance input signal
I would just filter those signals with a ~0.1 uF capacitor to parallel port ground as a first try at getting it going

I forgot that the motors "hold" when not running. With that in mind, it could be noise. Thanks for the observation.

I tried a 10uf cap, in line, to filter it. It didn't work and I wasn't sure why. Most of my projects are digital circuits, I get in over my head with analog circuits pretty quickly.

I have .01uf, 10uf and 37uf Electrolytic caps on hand that I usually use to filter noise out of supply power. I'll substitute the 1uf in parallel as suggested. Otherwise I'll have to wait and get my hands on a .1uf cap. The parts drawer never has the right stuff.

Thanks for the input. I'll post the results when I have them.

1 uF should work (but perhaps be a little slow) I had suggested .1 uF (that "point" is just about one pixel)
Note the capacitor goes in parallel with the signal (from the signal to ground)
When you say inline that suggests a series connection to me (and series is not right)

PCW wrote:

1 uF should work (but perhaps be a little slow) I had suggested .1 uF (that "point" is just about one pixel)
Note the capacitor goes in parallel with the signal (from the signal to ground)
When you say inline that suggests a series connection to me (and series is not right)

Yea, I meant series, and yes I can prove empirically that it isn't right.

Thanks for the input, I'll warm up the iron in a few.

Shannon

Success! I dug around my parts drawer only to discover that my smallest cap is .01uf and the next largest is 10uf. However the 10uf cap worked just fine, there isn't even a perceived delay.

Thanks for the guidance.