[gpraceman]

I just got a G540 and now have all 3 axis running. I've tried adjusting the trim pots, according to the Rev4 manual, but never come across a "null" or "sweet spot" point as I have seen mentioned. I have the trim pots adjusted all the way clockwise. Going counter clockwise just increases the drive's vibration. Any clue on why I can't get to that "sweet spot"?

Mach3 Setup
Kernel Speed: 25KHz
Steps/inch: 2780
Velocity: 172 in/min
Acceleration: 40 in/sec/sec
Step Pulse: 2us
Dir Pulse: 2us

This is on a Zenbot, which has a belt drive.

[resistor]

What are the motors?
The motors amp requirements?
The current set resistor value that you used?
res out

[gpraceman]

Motors: NEMA 23 3.6V 3A 285 oz/in 0.9deg
Resistor: 3KOhm
Power Supply: Keling 48V 7.5A

[resistor]

The reason i asked...
One late night had similar unsorted vibration.
Next morning metered resistors, had missed them by one decade.
Very little chance of the same problem with your three amp requirement.
Mine were being over driven.
Sure that you have rechecked your cabling.
Might try backing the velocity and acceleration way down to see the effect.

[gpraceman]

I used the DB9 breakouts from Soigeneris. They have a trim pot to set the resistance. I made sure to double check the setting with my DMM before putting the back shells on. I put a dab of fingernail polish on the tops of the trim pots, so they would not get out of adjustment over time.

I've rechecked the wiring many times. The motors are wired in the same way that I had them for another driver board that I had first tried (had to return it as it kept dropping the Z head all the way down).

I tried following the instructions in the Rev4 manual on adjusting the trim pots, which mentions tuning at 2 revs/sec. If I calculated correctly, that would be about 172 in/min (0.695 turns/inch). Nothing was indicated about what acceleration to use. I just used the recommended starting point setting from the Zenbot manual (30-40 in/sec/sec). Of the 3 axis, Z is the smoothest, but still a little rough.

EDIT: Not sure what is up, this forum replaces "Soigeneris" with "cnczone" in the URL.

[resistor]

Would be happy to work through the numbers with you, if you can bear with my typing on this phone.

Not very familar with the zenbot.
Just looked at their web.
"set the axis being adjusted to move at 2 revolutions per second"
Did not find the tooth count for drive pulley.
See that it is 3/8" wide belt.
What is the tooth count or pitch diameter of the drive pulley?
3/8" belt is normally .1 per tooth, so tooth count times .1" would be inches per revolution.
Usually travel speed is expressed in inches per minute so 2 rps is the same as 120 rpm.
If drive pulley is 10 tooth, 120 rpm would require 120 ipm (inches per minute)
If drive pulley is 20 tooth, 120 rpm would require 240 ipm (inches per minute)

[ger21]

It would appear to me that your machine travels from one end to the other in about 3 seconds? I think that in order to tune the Geckos properly, they need to be running at a continuous 120rpm while you make the adjustment. Your machine is so small and fast that it's nearly impossible to do.

[gpraceman]

According to the manual that I got with the machine (not the same one they have posted on their site), the pulley diameter is 0.458" and circumference is 1.439". They state that this means it rotates 0.695 turns/in. They do not list a tooth count.

At a 10 micro-step setting for the G540, the Zenbot manual lists a steps per inch of 2780, which is what I have set in Mach3.

[gpraceman]

It would appear to me that your machine travels from one end to the other in about 3 seconds? I think that in order to tune the Geckos properly, they need to be running at a continuous 120rpm while you make the adjustment. Your machine is so small and fast that it's nearly impossible to do.

The model that I have (Zenbot 1216) has a travel of 12" x 16" x 4". It definitely takes more than 3 seconds to travel from end to end on X and Y.

[ger21]

I'd program it to run continuously back and forth while you adjust it.

[gpraceman]

I'd program it to run continuously back and forth while you adjust it.

Well, been having my better half jog it back and forth while I was trying to adjust the trim pots. The best results, at the current settings, is with the pots all the way turned clockwise. Haven't reached any "null" or "sweet spot" point that people have described.

[resistor]

Sorry was in holding pattern, my number finally came up.
Looks like Gerry has been taking good care of you though.

"pulley diameter is 0.458" and circumference is 1.439"
In mxl belt would be 18 tooth pulley.
If it is a 18 tooth pulley, the math seems about right.
18*.08=1.44 inch travel per rev.
1/1.44=0.69444 rev per inch of travel

You can do a basic check of your Mach linear factor settings.
Measure the actual travel without regard for acceleration or velocity.
Start with axis close to one end.
Measure with a tape measure or yardstick between axis fixed point and point on moving gantry.
Zero axis counter in your controller software.
Jog exactly ten inches on that axis per the counter.
Measure again between the previous points.
Should be real close to ten inches.
If not, determine which setting is wrong.

What Gerry is suggesting is to run a short program to move the axis.
Sorry only example found right now is from a Fanuc VMC post setup.
Trimmed most out, and set spindle to zero rpm.
All you should need is the specific axis commands.
Gerry is a guru, he can provide a much better program,
with a sub call or in G91 incremental.
He will also know what default modals you need in the header for Mach.

O0210
(cycle x axis 0 to 10 to 0)
G00 G49 G40 G17 G80 G50 G90
M03 S0
F172
G01 X10
G01 X0
G01 X10
G01 X0
G01 X10
G01 X0
G01 X10
G01 X0
M00
M30

Add as many G01 X10, G01 X0, as you need.
Or just rerun the program.
This would allow you to tune without stressing the boss.

[gpraceman]

You can do a basic check of your Mach linear factor settings.
Measure the actual travel without regard for acceleration or velocity.
Start with axis close to one end.
Measure with a tape measure or yardstick between axis fixed point and point on moving gantry.
Zero axis counter in your controller software.
Jog exactly ten inches on that axis per the counter.
Measure again between the previous points.
Should be real close to ten inches.
If not, determine which setting is wrong.

Each axis is traveling the expected amount. No problems there.

What Gerry is suggesting is to run a short program to move the axis.
Sorry only example found right now is from a Fanuc VMC post setup.
Trimmed most out, and set spindle to zero rpm.
All you should need is the specific axis commands.
Gerry is a guru, he can provide a much better program,
with a sub call or in G91 incremental.
He will also know what default modals you need in the header for Mach.

...

This would allow you to tune without stressing the boss.

I can give it a try. I don't know if it would help to decrease acceleration to the lower end of what Zenbot recommends for tuning, which is 30 in/sec^2, but I can see if that does anything.

[gpraceman]

OK, I think I've got this figured out. I'm a noob to CNC and it shows. I was trying to tune while jogging, without remembering that I had the jog rate throttled down because this belt drive moves each axis quite fast. I ran the little gcode program and the drives sound much better at the target feed rate.

I get a little bit of hiss from the motors at times. I believe that I read somewhere that that is a normal thing. Hopefully, so.

Thanks for your help guys. Have a happy 4th :cheers:

[resistor]

Ok, very interesting that the measurement is as expected.
The thin thread that lead to that tangent was...

"At a 10 micro-step setting for the G540,
the Zenbot manual lists a steps per inch of 2780,
which is what I have set in Mach3."

From the online Zenbot manual...

<begin c&p>
Micro-step setting Steps per inch
1/10 1394.02
<end c&p>

The math seems to support roughly that 1394 value.

mxl belt has a .08 pitch.
18*.08=1.44 inch travel per rev.
1/1.44=0.69444 rev per inch of travel

0.69444 rev per inch of travel
0.69444*200=138.888
138.888*10=1388.88

Back to the original scheduled quandary "drive vibration".
With the motor unplugged, are you able to manually move the axis?
If so, does it seem to move smoothly or seem to bind?
If it seems resistive, is there much slack or backlash in the drive belt?
Is a motor easily removed?
If so, remove one of the motors and see if you get a different sound when tuning.
This would allow you to run the motor at 120rpm for a long period of time.
Might also be a great time to manually move the axis to check for binding.
And if moving the axis at a speed similar to what you have seen,
if a similar resonance is heard.
Not familar with the Zenbot, it uses steel tubing for the rails.
What type of bearings run against the tubing?
That is, are they roller bearings or plain bearings?
Way too many questions but "drive vibration" can be subjective.
And even verify that an input of .695 is equal to one revolution of the pulley.
Before you pull the motor.
If you run an axis at a slower speed, say 30ipm, is the vibration the same?

[resistor]

Great.
Now see that you got it sorted.
Bottom line is you worked through this one quickly.
"noob to CNC" is a valid excuse.
Have to look further for an excuse when i end up chasing my tail.
And notable that as usual Gerry was consise in his suggestion...
"I'd program it to run continuously back and forth while you adjust it."

Happy 4th to all as well!
res out

[gpraceman]

From the online Zenbot manual...

I think the online manual is for their smallest machine. I went with what was in the printed manual that I received with my machine. Anyways, I think it is all sorted out now. Now to work towards using the machine to make the panel cutouts in the controller enclosure, so I can put it all together.