Questions Wiring Up Servo/Gecko

[richlandrick]

Hello all,

having spent much more time than I expected on the mechanical opart of an IH mill conversion I'm finally ready to wire up my controller. This is my first servo system.

I have three each:
KL34-170-90 850 Oz servo motors
Gecko G340 drives
US Digital: E5S-500-250-EG-PKG2
CA-3131 Connector

I'm using a KL- 7220 power supply.

1. The KL34-170-90 has a black, red, and green wire. Green is Ground, White is +, Black is Common. I assume the white wire is Arm + (terminal 3)and black is Arm -(terminal 4). Where does the green ground wire go?

2. Are all 5 wires from the encoder used? Where does the green index wire go?


Thanks in advance for any help rendered.
Rick in Washington

[123CNC]

1. The KL34-170-90 has a black, red, and green wire. Green is Ground, White is +, Black is Common. I assume the white wire is Arm + (terminal 3)and black is Arm -(terminal 4). Where does the green ground wire go?

Green ground should run back to your control cabinet's ground bus/point.

2. Are all 5 wires from the encoder used? Where does the green index wire go?

You'll just need 4 wires for your Gecko servo loop, +/- power and a/b single ended signal. The index is not used for your Gecko servo loop. It can be used for a highly repeatable homing signal, but don't worry about that now.

[richlandrick]

Thanks 123cnc

I hooked up my parallel cabel turned on the power and tried to get the motor to move. I hadn't changed my Mach 3 motor tuning from the previous stepper settings. I assumed the only difference would be the amount of movement for a given signal.

The Gecko was in fault until I reduced the "limit." Until reducing it the motor would spin (pulse) on its own first one way then the other. It wouldn't respond to jogging either before or after adjusting the "limit" down.

I'm just at the beginning of the wiring and tuning process, but it sure would have been nice if it just worked the first time.

Any ideas?

[123CNC]

First, I'm assuming you are bench testing, or at a minimum do not have the servo physically engaged with their respective axis/load, e.g. drive belt removed.

Generally speaking when dealing with servo loops you'll want to verify your encoders are working, live signal and proper direction. Next you would want to tune the amplifier and then the controller. Then check proper operation of limits, physical and software, and check e-stop.

Then your ready to test the system with drives engaged.

Follow the 320 manual for tuning the drive. They describe both a multimeter and an oscilloscope approach to tuning.

Servo's require a little more caution then steppers. Although the 320 should fault out quickly if you have something wrong. If the encoder is opposite the step/dir signal the servo will start to move and stop when it hits the 128 count (IIRC, actual # may be different) off of target. In other words if you send a step/dir to move clockwise and the encoder is reading counterclockwise, the servo will turn briefly until the following error is hit. You would need to either swap the encoder a/b wires, the servo's white+/black-, or change the active low/high for dir in Mach. Any one of the three. This is why bench testing is recommended, even when you think you know what will happen.

[richlandrick]

Thanks,

I am bench testing the motor. I'll follow your guidance and test again this morning.

Does it matter if the other two drives are in fault mode while I test a single motor?

Rick

[TOTALLYRC]

Originally Posted by richlandrick Thanks, I am bench testing the motor. I'll follow your guidance and test again this morning. Does it matter if the other two drives are in fault mode while I test a single motor? Rick

It shouldn't make a difference as long as the error signal is not wired to the estop circuit. I would as a matter of course disconnect the 2 unused drives/motors and have on connected at a time. This way it will prevent any accidents. I have only bench tested my g320s and it is not that hard to do. Print out the directions from gecko and take your time.

Mike

[123CNC]

I agree with Mike. The other 2 drives connected in fault shouldn't effect you unless youv'e tied the err/res (terminal 5) into your control scheme. And, disconnecting the other 2 and working one loop at a time is a prudent move as well. This method may help you cut your teeth with 1/3 the risk (assuming we learn from are mistakes and don't just say, wow those sparks, flames, and smoke were cool...I wonder if I could do that again.)

I would even go so far as to suggest swapping a/b and verify results, swap white+/black- and verify, change software high/low and verify...etc.
Confirm for yourself and grasp a better understanding on the first motor of cause and effect. Some people prefer to have the wiring consistent across all three axis and make all the 'difference' settings in the software. Others may prefer software all alike, hardware changes for differences (think +/- direction of z on a knee mill and ballscrew direction of motion cw/ccw as compared to +/- on x/y).

[richlandrick]

I think I have experienced a $360 failure.

Drive 1: The problem I was having earlier was that the motor leads needed to be reversed. I checked all the wiring many times for proper polarity, connection points, etc. I finally switched the polarity of the red and black motor wires to red ARM- and black ARM+ and the motor seemed to work fine. It was pretty cool watching the motor respond back to its proper position when I tried to turn the shaft.

Drive 2: I moved on to the next controller and experienced some arcing at the power ground (terminal 1). Still have 5 volts at the test point and big blue cap, but the motor doesn't do anything witht he encoder pluged in or not.

Drive 3: Appears dead. No voltage at the test point and the fault light does not turn on.

Drive 1 Again: Fault light works, but doesn't go out. Motor doesn't do anything at all.

It occurred to me that with $360 for Gecko drivers, close to $200 for a power supply, $70 for a breakout board, that I might have been nearer to the cost of a package servo controller than I previously thought. At least I was hoping so because this exercise was getting old fast.

I plan to call Gecko Monday to check my options. Meanwhile anyone want to provide suggestions?

[TOTALLYRC]

Originally Posted by richlandrick I think I have experienced a $360 failure. Drive 1: The problem I was having earlier was that the motor leads needed to be reversed. I checked all the wiring many times for proper polarity, connection points, etc. I finally switched the polarity of the red and black motor wires to red ARM- and black ARM+ and the motor seemed to work fine. It was pretty cool watching the motor respond back to its proper position when I tried to turn the shaft. Drive 2: I moved on to the next controller and experienced some arcing at the power ground (terminal 1). Still have 5 volts at the test point and big blue cap, but the motor doesn't do anything witht he encoder pluged in or not. Drive 3: Appears dead. No voltage at the test point and the fault light does not turn on. Drive 1 Again: Fault light works, but doesn't go out. Motor doesn't do anything at all. It occurred to me that with $360 for Gecko drivers, close to $200 for a power supply, $70 for a breakout board, that I might have been nearer to the cost of a package servo controller than I previously thought. At least I was hoping so because this exercise was getting old fast. I plan to call Gecko Monday to check my options. Meanwhile anyone want to provide suggestions?

From what you are describing, you were using the power connections on the gecko's to go from one to the others and back again. This is the same thing as switching the dc power on and off. A big no-no.
The drives are repairable if I remeber correctly, but the call to gecko will tell you for sure.

Mike

[richlandrick]

No. Nothing that reckless. The common on drive 2 came loose while I was testing. I had all three drives hooked uop at the time.

The lesson learned - test one drive at a time.

[mtk]

Is there someone that has a servo package out there. I called Gecko but just drives. They do not support differential encoders wich I have. Can someone help with some input here.

[TOTALLYRC]

Originally Posted by mtk Is there someone that has a servo package out there. I called Gecko but just drives. They do not support differential encoders wich I have. Can someone help with some input here.

If you want to actually use the differential encoder you can do a couple of things. One is to get a chip that decodes the differential signals at the drive and converts them to single ended which the Gecko accepts. This way you get the benifits of the differential encoders and can still use the Gecko drives.

You could also use just the a and b signals and ignore the a prime and b prime signals and use it as a single ended encoder.

The third option is to go with something like the control box that I am using on my retro fit. That is of course if you are using +-10v analog drives.

http://www.cnczone.com/forums/showthread.php?t=66994

The DSPMC/IP directly takes the analog signal and closes the loop and reads the encoder position back in Mach3.

If you don't have servo drives at all then you could consider step and direction drives from AMC or Copley that handle differential encoders.


There are more options but it depends on exactly what you have and what you want to do.

Mike