Another in Need of Expert Help with Gecko 320s

[bill south]

Good Morning Guys and Gals;
I'm have some issues with my servo drive setup and need some guidance. I've completed two

other cnc projects and this is the first time with these issues. I may be asking too much of

my machine but here goes!!!!

Machine is a home built router 34"X34" with .2" lead ball screws. The axis are all on bearing

trollys and all mechanical movement is bind free. My drive system is as follows:
3 each Gecko 320
2 each ElectroCraft E643 (X&Z) Nema 34, 90V Max 4800 RPM motors.
1 each Electrocraft E712 (Y) Nema 34, 60V Max 2100 RPM motor.
BOB is a PMDX-126 wired as the manual describes for Gecko Monitoring (stop on fault).
Power supply is a home built high amperage torroidal design with filter cap and diode. Max

voltage is 52V DC.
The motors are belt coupled with a 1:1 ratio.
Control computer is a dell 2Ghz with 1 gig of memory and dedicated to the machine.
Running mach 3 at 45000 Hz
Motors are tuned at an acceleration of 8 (z) and 10 (x&Y).
System was calibrated using a digital caliper and the built in Mach calibration feature.
Motor Speed!!!
Thats the issue. I can't get the sytem to run past 70 ipm without faulting. I was targeting

150 ipm on rapids. The fault occurs when the drives reverse.
I've attempted to tune the geckos with a scope but can't figure how to hook set up the scope

and have it resemble anything in the documentation. I've tuned the system by ear many many

times. I'm almost afraid I'm wearing out the pots!
One test that I conducted this weekend kind had strange results.
What I was attempting was to test the PMDX 126 gecko monitoring feature. I loaded some code

and started the machine cutting air. I then removed the HV fuse on the motor side of the gecko

for the y axis. The position error was noted and the machine faulted and stopped mach. Just

like it is suppose to work.
I then replaced the fuse, restarted the program and removed the x fuse. The gecko did not

fault until I tried to replace the fuse. Strange. I replaced the fuse and started the system

again. I then removed the z fuse and the system just kept on running. I replaced the fuse and

the axis started working again but of course it was out of position. Almost like the gecko

position error monitoring feature was not working.
I then tried to test the position error circuit (encoder) as is suggested in the 320 manual. I

connected a vom to the test pins and 5 volts was noted. Of course to move the motor/encoder

out of position, I had to remove the hv from the sytsem, otherwise the motor cannot be turned

due to motor torque. Well, the 5 volts is no longer present when the hv is removed, even

without turning the motor. I guess the 5V is supplied from the hV circuit.
After reading all the above, I kind of realize that the motors are badly mismatched with a 60 v

motor on the y and a 90v motor on the x and z. Will this make a difference?
One other bit of information. I checked the voltate on the motor output from the geckos when

doing a rapid movement at 100 ipm and the voltage was only 7 - 8 volts. I haven't done the

math but the power supply is most definately capable of that. Will the low voltage demand

impact the error monitoring feature in the gecko?
Any thoughts or help will be very appreciated.
I'd like to get 150 ipm out of this system but reliability is the most important!
HELP.
Bill
Oh yeah, sorry for the long post but I've been working on this crazy system since Christmas!

[morestuff]

Hi Bill,

The tuning of servo systems is a process that can cause
plenty of frustration.

The G320X instructions describe the P, I, and D adjustments,
but there are other factors that can be used to affect
the servo response.

First is the encoder resolution. If you are using the
AMT-102 style encoders, you can set dip switches internal
to the encoder to control the resolution. You should set
only as much resolution as you need. For example you will
want the encoder feedback to offer between 4 and 10 times
as much resolution as the minimum commanded movement that
you need to achieve. Excess resolution can offer tighter
performance, but may also make it more difficult to tune
your system. There is also the possibility that high resolution
settings could require pulses faster than you computer can
generate them.

Remember that the G320X will use quadrature decoding, so
if your encoder offers 500 pulses per revolution, then
the G320X will be able to see 2000 counts per revolution.

The G320X following error detection circuitry has an
adjustable limit. If your servos struggle to maintain
tight following when moving fast, and you really only
need tight precision when moving slow, then you can
set the DIP switches (number 4 and 5) to allow greater
following error before signaling a fault.

The G320X also has adjustable torque gain. For most
users the default setting works, but sometimes you may
benefit from changing the setting of DIP switches
8, 9, and 10. This will affect the feedback stability.

Lastly, the G320X has a current limit adjustment. If
this is set lower than required by the motor specifications,
it can limit the amount of torque available unnecessarily.

Simulating a fault by removing the fuse is not a valid
approach. The G320X cannot monitor position and provide
fault detection when it does not have power.

The best way to simulate faults is to mechanically
overpower the servo motor so that it is out of position.
You can also connect a momentary switch between the
ERR/RES (terminal 5) and Encoder +5VDC (terminal 7).
This will cause all the G320X drivers that have their
ERR/RES terminals connected together to enter the fault
state.

This simulated fault will cause the G320X drivers to stop
driving the motors, and cause the PMDX-126 to signal an
E-Stop to Mach3. When you then click on the flashing RESET
on the Mach3 screen, the PMDX-126 will reset the G320X
drivers and they will again begin driving the motors.

If you click on the RESET on the MACH3 screen when
the drives have not signaled a fault, the PMDX-126
will NOT disable the drives. This was a design decision
to allow the user to reset Mach without causing the drives
to freewheel if there had not been an E-Stop.

Measuring the voltage to the motors, assuming that
your meter is not confused by the PWM square wave
signal, will tell you how fast the motor should be
trying to turn. It is quite possible to develop full
torque with less that 10 volts applied if the motor is
moving slowly relative to its rated speed. For DC
brush type servo motors, speed will be related to
voltage, and torque will be related to current.

The following error detection circuit and the encoder
are powered by the V+ motor supply voltage on terminal 2,
but should operate correctly at any voltage above 18 volts.
The actual voltage should not influence the error detect
sensitivity.

Having a lower motor supply voltage will limit the top
speed of you motor, but my guess is that your supply
voltage is not your problem. Even running at half speed
your motors will probably drive your machine at much
faster speed than you are trying to achieve. Do the
math keeping in mind that motor speed is revolutions
per minute and machine travel speeds are also in inches
per minute. Don't get tripped up inserting a factor of
60 somewhere.

Hope this give you some ideas without confusing things
even more.

Regards,
Steve Stallings
www.PMDX.com

[bill south]

Thanks Steve;
This is all good information and I now understand how the PMDX board works much better. Unfortunately, the other information relative to the Geckos will not help very much since I have the 320's and not the 320x. There are no selectable attributes only the pots to set. Thus the reason I cannot understand why I can't find a "magic" sweetspot after all my tuning. Basically, there are only two adjustments required and I can't seem to find the correct combination. Where is my speed?
Thanks again for all of your information. I'm sure the PMDX-126 board will work flawlessly. This is the 3rd BOB I've had in the system. All were upgrades but this is getting expensive!
Are there any hobby priced servo drives that have a self tuning feature?
Best and thanks.
Bill

[bill south]

With the information presented above (I don't want to bore anyone with repeating it), can anyone give me an idea what speed I should expect out of this setup assuming ideal mechanical conditions?
Thanks
Bill

[Larken]

Hi
What encoders are you running ?

If your interested, I can sell you 3 Viper 95 drives (@125 each) and If your not happy, i will give you your full money back including shipping.

I'm sure they will solve your frustration. But i recommend differential encoders

Larry K
Larken Automation

[Larken]

Oops !!

That was supposed to be a private message.

[bill south]

Thanks for the reply Larry! I kind of thought your response was PM material but I sure do appreciate it anyway. I think I have found a solution to my issue (again, I think). The values I was using in Mach may have been very wrong for what I was trying to accompllish.
Thanks for all the help guys.
Bill

[Mariss Freimanis]

You are grossly miss-geared which makes your application very short on torque. 70 IPM is 350 RPM on the screw yet your motor can turn I'm guessing 3,500 RPM so you are using only 10% of the motor's available speed. If you geared the motor down 3:1, you'd need only 1/3 the torque from your motor and you'd reach much much higher speeds.

Think of a manual transmission sports car when the light turns green. Do you want to be in 1st gear or 5th gear? It's the same engine and horsepower in either gear but 1st gets you a more spectacular getaway than 5th does. The engine has low torque compared to what the wheels need but it has a 7,000 RPM red-line and the wheels would tear off your car at 7,000 RPM if it were on jacks. You trade in excess RPM for torque using a transmission between the engine and wheels. You wouldn't build a car with its engine coupled 1:1 to wheels yet that's what you have when you couple a servomotor 1:1 to the screw.

Mariss

[bill south]

Thanks for chiming in Mariss. As always, your input is valued and appreciated.
I guess I was under the assumption that the torque curve was highest at the low rpm of the servos. If that is a misconception on my part then what you say makes good sense.
I''ve ordered a different pulley set from mcmaster to get the 3 to 1 ratio which is the best I can do with my current mounts.
This is in no way meant to be criticism of Geckos!!!! I've used gecko's for years and really like them. I've just been kind of frustrated with this particular machine is not working as I expected.
Many thanks guys!
Bill

[Mariss Freimanis]

Servomotors have a continuous rated torque. This rating is the maximum torque load you may continuously apply to the motor without overheating it.

This limit is independent of speed, which means it's the same limit at 1 RPM or 6,000 RPM. Power is torque times RPM; the motor produces it's rated power when it's turning at its rated RPM (typically 80% of its no-load speed).

To get the rated power from the motor, the gearing must be such that the motor is turning at 80% of its no-load speed when your mechanism is moving at its maximum design speed.

You won't get optimal gearing with a 3:1 reduction but you will be much better off than you were at 1:1 gearing. The geared output will have 3 times the torque you had at 1:1 reduction. The reflected moment of inertia will be reduced 9-fold and the motor will be able to accelerate the load much better to the speed you want.

Mariss

[tono88]

Hi, good morning. I'm trying to use the gekodrive g320x I follow all the steps and all ok but in the step of connecting the servo motor I get the fault led. The fault led didn't go away for a while, then the fault led goes away I repeat all the steps and in the same step I get the fault led. I think is the current limit but my question is how to remove the fault led to try again