Unable to find home. Intermittent marker pulse

[frankljs]

Hello, im trying to finish setting up m15 to m39 upgrade.
Atrump knee mill.
SEM dc servo
Have not been able to finish system test because Y axis is “ unable to find home”. It will only home intermittently. I’m certain everything is connected correctly switches are wired correctly axes are moving in the correct direction and limit switches are tripping good.
However, I did notice that when looking at the PID screen, the Astrik */marker pulse is very intermittent to display when revolving the motor pulley. Sometimes it sees the encoder pulse, sometimes it does not-it’s very inconsistent.
Do you have any suggestions on how to fix this problem?
I will post a report.zip Tomorrow or asap.
Thanks

[cncsnw]

Check the differential voltage between the Z (pin 6) and /Z (pin 3) wires at the encoder plug.
See http://www.cncsnw.com/EncoderTest.htm.
It could be a faulty cable or plug, or bad connection.
It could also be a faulty encoder.

Also be aware that CNC12 has more stringent requirements for home switches than the old CNC7 software had. After the switch trips (e.g. while moving Y+) you have at most 1/4" or one turn of the motor (whichever is greater) going back the other way, for the switch to clear. If it does not clear in that distance, you get an "unable to find home" condition and/or a "407 ... limit tripped" error. This prevents the machine from running all the way to the opposite end in case of a stuck switch, but it also causes problems with long-travel switches. What type of limit switches are installed on your machine?

[frankljs]

My switches are pretty short throw. I dont suppose thats the issue, but I will post a pic or something so you can determine. I did rotate the motor pulley half a dozen times to see if my encoder pulse was too far off to no avail.
I will look at the dif. voltage between the wires asap.
Hopefully its a cheap fix. In the mean time, here is my report.zip
sytem Id: 0008dc111213-0118222705
mpu12

[frankljs]

Check the differential voltage between the Z (pin 6) and /Z (pin 3) wires at the encoder plug.
See http://www.cncsnw.com/EncoderTest.htm.
It could be a faulty cable or plug, or bad connection.
It could also be a faulty encoder.

Also be aware that CNC12 has more stringent requirements for home switches than the old CNC7 software had. After the switch trips (e.g. while moving Y+) you have at most 1/4" or one turn of the motor (whichever is greater) going back the other way, for the switch to clear. If it does not clear in that distance, you get an "unable to find home" condition and/or a "407 ... limit tripped" error. This prevents the machine from running all the way to the opposite end in case of a stuck switch, but it also causes problems with long-travel switches. What type of limit switches are installed on your machine?

So, does centroid offer an encoder test kit or something or do I need to buy the items from A.D. and Jameco etc?
Thanks

[cncsnw]

You need to buy parts and make it yourself, or improvise with a DE9 male-to-female pigtail with exposed test points.

[frankljs]

I purchased and hooked up the test fixture. Running 5v dc with an adapter plug. All I get is green power light on far right power LED. When i plug in the encoder cables(tested all of them), no other lights come on. A, B, and Z diodes do not light up. What am I missing here? Thanks

[frankljs]

Update on the testing:
My diodes are wired across 6-3, 7-4, 8-5, and 5vdc 9-2.
Using my meter, i get
4.1v on 6-2
4.3 on 7-2
4-5v on 8-2
No other diferential voltage can be measured.
The encoder cables have not been rewired or changed from the original system.

[cncsnw]

Using my meter, i get
4.1v on 6-2
4.3 on 7-2
4-5v on 8-2
No other diferential voltage can be measured.

I am not sure what "no other differential voltage can be measured" means.

Use your meter to measure volts (or millivolts) from 3 to 2; from 4 to 2; and from 5 to 2.

Also, do any of the voltages 7-2, 8-2, 4-2 or 5-2 change as you turn the encoder?

If the A, B and Z channel LEDs on the test setup do not light up, then you would appear to have near-zero differential voltage across each pair. That would indicate a faulty encoder.

What results do you get when you connect your X or Z encoder to the test setup?

[frankljs]

Sorry miss type. *No voltage measured on the other differential channels.

These tests were performed on the X and Z axis encoders (those didn’t have failures during testing) with similar results, except voltages were lower on Z, like around 3 on the channels mentioned above.

I will test other channels , also while turning tomorrow.

Another strange thing, the power LED will start to blink after 30 seconds or so and my readings change. I then pause testing for a minute and carry on.

[cncsnw]

If your +Z, +A and +B wires (pins 6, 7 and 8) are all ca. 4VDC above common (pin 2), then with a normally-functioning encoder you would expect to see around 0.4 to 0.8 VDC between -Z, -A and -B (pins 3, 4 and 5) and common (pin 2). But then, the LEDs should light up green.

If you test your LED-resistor assemblies in isolation, e.g. by connecting 5VDC to them in one polarity and then the other, do they light up green and red?

If there were really no differential voltage coming from your encoder on the A and B channels -- e.g. if -A on pin 4 is not driven low when +A on pin 7 is driven high -- then CNC12 should have been reporting "412 _ axis encoder differential error". If you were not getting 412 differential errors, and the resultant fault conditions, then it is likely that your encoders really do have valid differential output, at least on the A and B channels. That suggests there is some sort of error in your measurement methods here.