Hello Guest it is December 10, 2019, 05:01:32 AM

Author Topic: Ethernet SmoothStepper first test  (Read 78372 times)

0 Members and 1 Guest are viewing this topic.

Offline Hood

*
  •  25,856 25,856
  • Carnoustie, Scotland
    • View Profile
Re: Ethernet SmoothStepper first test
« Reply #100 on: February 22, 2012, 06:05:56 AM »
Did you choose one of the spindle options in SS config page?
Hood
Re: Ethernet SmoothStepper first test
« Reply #101 on: February 27, 2012, 11:41:49 PM »
It seems to work now that I moved it to Step and Dir (4 pulse width) instead of Relay or PWM. Is 4 pulse width the correct number to put? I am using the Sieg KX3.

Offline Hood

*
  •  25,856 25,856
  • Carnoustie, Scotland
    • View Profile
Re: Ethernet SmoothStepper first test
« Reply #102 on: February 28, 2012, 04:03:56 AM »
I have a  lower pulse width but I am using a high resolution servo for a spindle motor, you will likely be fine at the 4.
Hood
Re: Ethernet SmoothStepper first test
« Reply #103 on: March 03, 2012, 01:51:01 PM »

Big trouble with touch probe input for a while, but it turned out the BoB was set up very marginal and not driving the ESS reliably. Wierd symptoms, but all gone now I fixed the setup.


I am having some issues with probing also. PDMX 132 BOB to ESS. Basically I cannot get the "Probe Active" LED to come on in Mach, unless I switch to "active high" on the ports and pins, at which point it goes green and stays green.

 I get 3.8V on the touch plates dropping to 0.3 when it touches the spindle. I get the same 3.8 voltage when I compare the plate to a ground plug on my extension cord, so the spindle seems well grounded, even though it is a PC890 and double insulated with a two prong plug. I do have a ground wire running to the spindle but it isn't doing anything at the moment.

Breakout board seems to be powering the plate well enough.
Spindle seems well enough grounded to trip a TTL
I have the touchplate hooked to pin 15 on port 1. Configured pin 15 in Mach to "active low" and assigned it to "Probe". Seems pretty straightforward.

I wonder if this is something to do with the Smoothstepper not seeing and relaying the input signal to Mach. In the ESS data monitoring screen there are two options with checkboxes for the input pins; I tried both ways for pin 15 (but didn't restart Mach or ESS). Is there some other config option I should be looking at? The PDMX BOB is opto-isolated and inputs are filtered. Maybe switching to a CMOS (Port 2 apparently) from a TTL input port would help, though the CMOS seems more particular about the voltages it needs to trip...

Any ideas? Kind of driving me batty. Thanks.

Re: Ethernet SmoothStepper first test
« Reply #104 on: March 03, 2012, 03:17:51 PM »
OK if I short the plate input to the ground on the BOB I can get the LED on the BOB to light up and guess what - the screen LED comes on also.

Apparently a 0.01 microFarad capacitor between the input and ground helps; I will see if I can find one today and report back.

It is odd because the input is not tripping low prematurely during the autozero routine; it just won't turn the BOB LED on at all unless I short the plate wire to the ground on the BOB itself. I tried touching the common ground wire on the cabinet, and the BOB LED would not light up.  The BOB ground is wired directly to the power supply for the control, which goes to a ground plug in the wall and a rod jammed into the earth outside my garage. So I don't understand why the ground on the board is any better than the ground elsewhere inside the cabinet. Is my ground faulty somewhere?

Nothing I was doing before was lighting up the input LED on the PDMX BOB; looks like the Smoothstepper had nothing to do with it.

Comments still welcome.

Karl
Re: Ethernet SmoothStepper first test
« Reply #105 on: March 03, 2012, 06:07:46 PM »
Thank goodness for Radio Shack. Unfortunately the capacitor makes no difference.

Given that the machine side of the BOB is at a different voltage than the computer side, I'm wondering how people get their cutting tool to the same voltage as the breakout board common pin. This seems to be what it takes to light the LED up.

Perhaps I will have to run a separate wire from the board all the way out to the spindle.

Offline Jeff_Birt

*
  •  1,107 1,107
    • View Profile
    • Soigeneris
Re: Ethernet SmoothStepper first test
« Reply #106 on: March 03, 2012, 06:55:10 PM »
Quote
I'm wondering how people get their cutting tool to the same voltage as the breakout board common pin. This seems to be what it takes to light the LED up.

Perhaps I will have to run a separate wire from the board all the way out to the spindle.

You don't want to try and use the machine itself as a current return path. Not only do all the bearings a sliding parts not make for a good conductor but if the machine is bonded to earth ground you are also running trying to run a low voltage return signal through earth ground (which is a bad idea all the way around).

I use a two lead cable with alligator clips. Clip one lead (the common) to the bit, clip the other to the part/tool touch plate.
Happy machining , Jeff Birt
 
Re: Ethernet SmoothStepper first test
« Reply #107 on: March 04, 2012, 04:37:03 AM »

You don't want to try and use the machine itself as a current return path. Not only do all the bearings a sliding parts not make for a good conductor but if the machine is bonded to earth ground you are also running trying to run a low voltage return signal through earth ground (which is a bad idea all the way around).

I use a two lead cable with alligator clips. Clip one lead (the common) to the bit, clip the other to the part/tool touch plate.

I finally got it to work. Jeff it was your comment on some other thread about the breakout common voltage being different than the earth potential that put me on the right track. There is a lot of discussion about how you don't need a second wire if your machine is properly grounded, and I am here to tell everybody that is a load of poppycock.

The machine has a three wire power cord running to the spindle, which was connected to earth ground at the control box. Given that the spindle itself (PC 890) is double insulated and has no ground prong on its two conductor power cord, that ground wire from the control to the spindle was just sitting idle. There was no way I was going to run a third wire out there when I already had an unused one routed nicely through the cable carriers. Plus I could not find an alligator clip large enough to fit on a half inch end mill. So I took a few liberties.

1. I put a three pronged cord on the router and tapped the ground wire into the top bearing block with a #6 machine screw. This allowed me to shorten the cord also, which was overdue.
2. I took the other end of the ground wire (in the control box) off the earth ground and tied it to the breakout board common ground. There is still plenty of earth ground connection going on in the control box; the only thing I moved wasn't being used anyway. I am a bit nervous about some sort of stray voltage flying around inside the router frying my breakout board, but so far so good.

Now my LED lights flash on reliably when the bit contacts both plates.

There is much I still do not understand about what is getting pulled to what voltage; for instance the voltage at pin 15 stayed reliably at about 3.9V relative to earth whether it was shorted to common or not. Isn't it supposed to get pulled low? And now the plates register zero volts to earth at baseline, whereas I thought they should have some voltage on them until they are shorted to the tool, at which point they get pulled low.

But this is all academic as the autozero routine works fine now with Tempest, the 2010 screenset and ESS. I did have one issue where the Z homing limit switch was not working; this seemed to go away after I disabled "persistent DROs" and now it is working fine again. Some scary sounds came out of that little stepper but I think it is OK.

I am really happy to have this feature functioning. However I am also amazed at how difficult it was to find the two or three threads that deal most fully with the issue, and how no one really seems to have discussed the fundamental aspects of "pulled high", "pulled low" etc., how resistors play into that, and why the breakout board common ground voltage is not the same as earth ground. I suppose it is all on Wikipedia.

Back to making dust!
« Last Edit: March 04, 2012, 04:39:31 AM by Karl W »

Offline Jeff_Birt

*
  •  1,107 1,107
    • View Profile
    • Soigeneris
Re: Ethernet SmoothStepper first test
« Reply #108 on: March 04, 2012, 10:35:38 AM »
Danger Will Robinson, Danger!

#1: A double insulated tool is designed so that no matter what goes wrong the operator will not be shocked. Both the housing is insulated and the armature (spindle shaft) are insulated. You just undid one of the built in isolations that were designed to keep you safe.

#2: Earth ground is not the same things as DC common. Earth ground (the green wire in an AC cord) is primarily a safety device. It provides a low impedance path to the Earth. It is not used to carry electrical current, that is what the hot and neutral wires are for. Earth ground is also used by various filters to 'bleed off' electrical noise. You do not want to use Earth ground as a return path for your DC signals, doing so just invites noise into your system and can create ground loops (current running through the Earth ground causes voltage drops so different devices will 'see' a different potential for Earth ground.)

#3: Since your using Earth ground as a return path for a DC signal and that wire happens to be in the same cord as your AC power to the router that means your return path for a low voltage signal is running a long distance in parallel with a high voltage noisy signal. This will lead to noise being induced on the low voltage signal.

Now, this probably won't cause the world to end but it is not a good idea nonetheless. I don't mean to belabor the point but circumventing a safety feature to avoid running one wire seems like false economy to me. I want to your machine to operate properly and I want you to be safe so please know that my warnings are meant in a friendly manner.
Happy machining , Jeff Birt
 
Re: Ethernet SmoothStepper first test
« Reply #109 on: March 04, 2012, 11:34:40 AM »
FYI for anyone trying to sort it out I think this Erniebro thread is pretty good on autozero, except the grounding stuff:

http://www.cnczone.com/forums/mach_wizards_macros_addons/56079-xyz_probe_modification.html