[Alex_Cole]

Ok. Here is what I have. I have these servo motors http://www.cadcamcadcam.com/index.as...PROD&ProdID=61

and i am using this drive http://www.geckodrive.com/product.cfm?pid=14

My problem is I keep loosing position in X. I have never been able to get this machine to cut right yet. I have the cnc4pc c1 breakout board and I also bought the cnc4pc c1g breakout board to see if that was the cause...no luck.

When I cut a 2" square...then run the same program again my X axis is off by about .045 That is a lot! if I run it again it is off again only a different ammount. It does not seem consistant with how much the error is. I have worked on tuning the motors all winter and have defenatly decided that it isnt going to fix it.

When I am running my mill in jog mode I can hear a crunching type sound that comes from the servo. Now this is not the bearing or my screw cause I can swap my servo encoder and power wires and the problem seems to follow the drive.

My Y axis was doing this also at one point and I ended up sending the drive in and they replaced it...which I appreciate a lot but the X axis is doing it just like the Y axis was.

I am wondering if the motors are not suited for these drives?
I am running a 24V power supply and am running my machine at about 12ipm when this happens. I can run my machine at 55ipm with these motors and drives so I am not going too fast.

I am about to throw this whole project into the scrap bin I am so frustrated.

Can anyone help as to what might be going on. I can and will provide any additional info if needed.

Thanks

AC

[Bubba]

AC,
Not sure I can offer much help, but what type machine is this on, what is the reduction ratio and lead screw pitch?
As it is, there is not much I or anybody can really offer.

Is this repeatable if your only "cutting air" or only when your actually cutting material?

[Alex_Cole]

The machine is a small benchtop mill. The make of the machine is Dyna

The error follows the axis that I connect to the drive I am trying to tune.

US digital encoder 500 count. so that is 2000 steps per rev.
Rounded off belt reduction is 2.8 to 1

Lead screw is .098 per rev.

I have a G320 drive and I can hook up this drive and not get these problems but I need the step pulse multiplier.

If i take and put 24v to the terminals on the servo (NOT through the drive) then the motor spins with no bad sound. It is quitet and smooth as it moves the axis from one end of travel to the other.

[Bubba]

AC,
OK, the last post indicates that using a 320 gecko allows smooth operation??
What multiplier are you using in the 340 and have you also tried setting it to 1 to see if that eliminates the problem (in other words make it act like a 320)?

Also, what makes you think you "require" the multiplier? For instance, on my mill/drill, I am also using 500 line encoders and it works quite well.

Another thought, can you and have you tried playing with pulse width etc? I'm just guessing.

[Alex_Cole]

I am using th multiplier due to the belt reduction and screw pitch I was un able to run at desired speeds.

I have not tried moving the switch and that is a good idea. I will try that and see if the problem goes away. I wonder if it is an issue with the mulitplier on the drives...that is with these small motors. I know these servos are on the small side for these drives.

When I started this project I was planning on using some other drives...it didn't pan out as they were not compatible with my breakout board and I had some issues so I scrapped the cheapo drive and bought the Geckos. I think these drives are great and I feel I am doing somthing wrong as the feedback and history of the Gecko drives I feel speaks for itself.

I will post the results after the my new test. If all else fails I will take the 340 off the z axis and move it to the X. Then put the 320 on the Z axis.

I had to replace the Y axis drive before for the same type issue. I sent it in and they sent me a new one. They never did say what the issue was. They just sent a new one. I am now getting the whole thing put back together.

The breakout board had opto isolators on the XYZ outputs and I was told that it would cause a problem so I bought a new breakout board made for the gecko drives to fix the issue. The thought was that the double opto isolators were not really ment to work together and they were droping the steps.

The axis seems to shift in one direction each pass.

Thanks and I will let you know after I change the step multiplier.

AC

[Alex_Cole]

Here are more details on my screw and the math I did to get to my desired speed and so on. My target speed was about 90ipm. The servos can not really do this and I fault going over 55 so now I am just happy if I can get 55 ipm. I was using the 5X multiplier but think now I should try the 2X and see what happens.

Here is my doc about screw and encoder info.
------------------------------------------------------------------------

SCREW INFO
2.5mm = 0.09842519685039370078740157480315 inches per revolution
1inch of travel is 10.16 revolutions.


SERVO INFO
500 Encoder counts per 1 revolution
4 quadrents to each encoder count = 2000 steps per revolution.

TIMING PULLEY GEAR REDUCTION INFO
2.8 to 1 gear reduction

2000 steps per rev on servo means 5600 steps per rev on gear


10.6 revs to an inch equals 59360 steps per inch

computer can pulse at 25000 steps per sec * 60 = ppm 1,500,000

this totals 25.2695 Inches per min.

add 5x multiplier on Gecko G340 Drive and this equals to 7,500,000
steps per minute on encoder wich totals 126.3477 possible inches per
minute.
-------------------------------------------------------------------------

Thanks
AC

[Bubba]

AC,
To quote that famous and brilliant MR GECKO, "No-load means just that; even a flea will slow down a motor at no-load RPM, should it land on the shaft. Specifically, your available torque is about 3% of stall at your .............." and this was for my system with much higher voltage etc.

Based on his formula, the max "reasonable" rpm you can expect is about 1790 rpm (80% of max) or about 60+/- IPM for your conditions both mechanically and electrically.

Your limit is primarily based on the electrical characteristics of your servo motor and power supply.

Hope this helps"

[Alex_Cole]

I agree with you on that one and I guess I realized that when I started. My servo motors are somewhere around 30v but I only have a 24v power supply so I am loosing some anyhow cause of that.

I am hopeing that when I test the condition clears up when I change the step pulse multiplier. If I can get it to work on 2x instead of 5x then I can still run at 50 or so ipm....in theory.

AC

[Crevice Reamer]

I can think of 4 ways you could increase IPM with this setup. The most expensive would be to go to 5 TPI lead screws. Next most expensive would be 200 - 250 CPM encoders. Least expensive would probably be a change from 2.8/1 to 2/1 pulleys. You could also upgrade to 30-36V power supply. 36V is 1 volt over the envelope for these motors, but it wouldn't take much resistance to drop a volt.

CR.

[Bubba]

I disagree that a reduction in encoder count will help in this instance due to the max speed allowable for these motors.
Decreasing the reduction will not help much (might increase speed no load). but there will be no real power to do any real machining. Might be ok if this were a plasma machine with light inertia and force requirements.

[Alex_Cole]

Right now I don't care if it runs at 25ipm...I just want the axes to go where they are supposed to instead of drifting like it is.

AC

[Alex_Cole]

Question at bottom of post! Sorry for the length.

ok.....so I did some tests. I ran the mill the way it was and I got error. I actually found it was on all 3 axes when testing. My Z axis kept creeping up on me....so this lead me to try somthing...maybe I should have known this or maybe it shouldn't be this way but I decided to go into Mach3's ports and pins config for the motor output and switch the step pin to the "active low" setting.

I then ran a couple of test 1"x1" squares....here is the program I ran

G0X0.Y0.
G0Z.150
G1Z0.F4.
G1Z-.5
G1X1.Y0.F12.
G1X1.Y1.
G1X0.Y1.
G1X0.Y0.
G0Z.150
G0X1.Y1.
G0X0.Y0.
G0X1.Y0.
G0X0.Y1.
G0X0.Y0.
M30

Now this program cuts a 1"x1" square. Then it rapids around a bit to give the drives some direction changes. I did this so I could find any error quicker.

Now when I ran this program with the Active Low turned off then I found my axes were slowly getting more and more out of position from the 1st time the program was ran(where it cut the wax)

Now when I ran it with Active Low turned on then it appeared to cut in the same location(keep in mind I am cutting in wax).


So after discovering this I decided to cut a complex 3d shape (star) that is surfaced over a 2"x2" square. I cut this by doing parallel cuts at .005 stepover with a .005 tip engraving tool. I cut this at .1 deep in blue machineable wax.

Results:
When I cut the 3d part the square stayed square(good..and a 1st) I started the program at a feedrate of 12. IPM. This was slow so I bumped it slowly from 12 to 16...then 25...then 28...then 30...and then to 36 close to the end of the program.

Everything cut fine but I noticed when done that in the Z axis at the locations where I changed the feed.(I used feed override in Mach) the Z axis seemed to cut deeper for a pass and then the next would be ok...that is until I got to 30 or so ipm...then it started cutting about .015 deeper until the end of the program....

Now the areas where the feed was left alone was great until I would change the feed again...

I am uploading video to youtube of cutting at around 30ipm I will link to here when I can.


Question: Is this a Mach3 problem? Is it not ok to change the feedrate while cutting using feed override?

NOTE: I programmed the part at 12ipm so I was over 100% on the override.

[Alex_Cole]

Here is a link to the video. This was cutting at 36ipm.

http://www.youtube.com/v/lkIzyJapq84

I want to say thanks to you who have been helping me, and those who will help me in the future.

I am almost there with this servo conversion...again I say almost.

Thanks

AC

[MrWild]

A step/direction drive is the same whether it is a stepper motor or a servo motor. What I mean by this, is that a step signal says move one step. With a stepper motor, the drive moves this step with no feed back. With a servo motor/drive, the drive says move one step, and sends a current to the motor to move UNTIL THE ENCODER SAYS THE STEP HAS BEEN MADE.

When steps are lost with a stepper, the first things looked at are drag in the mechanics that will over power the stepper motor. With no feedbaCK, THE DRIVE IS CLUELESS IF THE MOVE HAS BEEN MADE.

When using a servo drive, if there is binding in the mechanics, the drive will FAULT OUT. and stop. This is important in regards to your issue. You aren't having a mechanuical proiblem because if you were, the drive would fault out when it wasn't seeing the posssition it thinks it should see. Where you may find your problem is the encoder wiring. Your sheilding may be faulty and grounded at both ends creating a ground loop. Your encoder may be bad and giving erratic pulses so the drive reads the possition wrong and never faults out as it "thinks" it is where it should be. There could be a loose connection at the wires or internally in the encoder and you're geting intermittant pulse breaks.

With all the work you've tried, I think I'd try switching motors (and encoders) and see if the problem follows the motor. I'd also run new encoder cables and make sure only one side of the sheilding was grounded. I honestly think chasing mechanicals is the wrong chase to go with. Switching motors will give the quickest view of what's happening. and where.

[Alex_Cole]

does anyone have any recomendations on servo motors that I should get?

Is it safe to say that I should stick with a 100Watt servo minimum for the
G340 drives?

Thanks

Alex

[keithorr]

my suggestion is to not use feed override at all. start the program from the beginning with a new feedrate. bumping up the feedrate during a program gives me trouble and I'm using a proven solid machine.

[Alex_Cole]

I wanted to give you all an update. I was able to get my machine up and running correctly with my motors and drives....at least it seems for now. I ended up playing with the config settings, motor tuning, and triple checked all all wireing by removing the drives and re-seating everything.

Now all seems well I have attached some test pics of a small engraving I did last night. This was with about a .004 dia tip engraving tool. .002 step over at 14.ipm Yes that is a dime in the pic. I put it there to show size of the engraving.

Thanks to all for your help it is very appreciated.

Alex

[kimfmx]

Hi,

Did you find a solution to your problem?

Im seeing the exact problem and its driving me crazy.

Best Regards
Kim Mortensen

[Alex_Cole]

Check your positive high setting for step and direction. If I recall this is what was at fault.

Sent from my Droid

[kimfmx]

I did look into this, and changed these settings, and had to rewire my motor to run correct way.

but the issue is still here...

Mest Regards
Kim Mortesnen