Servo is weak when using a gecko 320...

[bitviper]

I went ahead and purchased the servo motor and driver card combo that can be found here: http://www.cadcamcadcam.com/index.as...PROD&ProdID=61

This system worked great but I needed about 25% more torque from it then I was getting. I'm running the drive at about 100RPM and the encoder is a 500 line HEDS unit.

Anyhow, I got myself a Gecko 320 and hooked it up. I think the wiring is correct because the motor shows a tremendous amount of holding torque when I'm not giving it pulses.

Trouble is, as soon as I give it pulses the torque drops to near zero. This is odd as the old driver which could go as high as 5 amps had tons of torque at the slow speed I was using (though I still ended up needing that 25% more).

My power supply is a 24volt transformer with a 20amp limit. It goes through a rectifier which then is filtered by a couple of caps that total around 15000 uF.

What is odd is that I'm measuring the voltage across the power supply terminals of the gecko when it's in holding mode (no pulses). I don't see a drop in voltage no matter how hard I pull on the carriage that's attached to the motor via a chain drive. As soon as I give it pulses at the 40Hz that's asked for in the manual, the creeping carriage is easy to stop and I see the 32Volts drop to 24 indicating that my power supply can't keep up with the demand even though it should only be drawing perhaps 4 amps from a 20amp capable transformer.

I've set the current limit of the gecko to max and I've played with different gain and dampening settings with no effect. I don't have a scope to tune with.

One other symptom is that I can't get the carriage to go any other speed but very slow which equates to about 5 seconds per motor shaft rotation.

If anyone has any ideas of things to try, I'd really appreciate it!

[Alex_Cole]

I have these same motors running on Geckos. I have 3 G340 drives and 1 G320 drive. I have used them all on my Dyna mill. I have had some issues getting things up and running but now are all worked out.

I ran my G320 drive on my Z axis for a while and it did not exhibit this behavior when running.

I have not hooked up my multimeter to see what things are doing while running but I am getting reasonalble performance using a 24v 6amp switching type power supply.

At first I though I may be having issues with the drives running with these small motors but I eventually tracked down the real issues which were breakout board and noise issues.


I am new to all this stuff and hopefully somone more quallified will chime in and give you more info but I would guess based on what you have said that there is a tuning issue.

As far as the motor speed. These motors us a US digital encoder that has 500 counts per rev. These encoders have 4 quadrents per step so this really equates to 2000 steps per revolution of the motor. If you are using the parallel port to drive this then you will be limited by the pulses per sec the software/computer can run. This is why they offer the G340 drive with the built in pulse multiplyer.

I ended up getting the G340 drives because of the step multiplier and now my machine is running at the speeds I was expecting.

I am no expert but I will try to help you with what I can.

Link to my machine retrofit buildDyna 2400 Mill update NOT DIALUP FRIENDLY!
AC

[bitviper]

Appreciate the response. I don't think I need a step multiplier as I'm running things from my own microcontroller running at 8MHz so I ought to be able to send pulses at 12.5uS or so.

I have tested with 35uS on and 35us off and I'm still getting very slow rotation speeds (couple seconds per rotation).

very peculiar...

[Alex_Cole]

When I started this we were driving these from our own servo drives based off of a pic chip. We didn't encounter this issue but we did run into lost steps.

Are you sure your chip is sending a full 5v for the step and direction even when pulsing that fast?

Just a thought... when you get into the microcontrollers I only know enough to be dangerous...my friend who has been helping knows all of that stuff.

AC

[bitviper]

Interesting thought. I'm thinking about getting a scope to troubleshoot this stuff. Don't know the answer yet.

What's frustrating is that I had the system working perfectly with the old driver that came with the motor. It was pretty strong and fast. Seemed like it was compatible with my microcontroller voltage levels.

I'm thinking about removing the gecko and going back to the old driver with my improved power supply (the old supply could do 5 amps, the new one can do 20 amps).

[cjmerlincnc]

Originally Posted by bitviper Appreciate the response. I don't think I need a step multiplier as I'm running things from my own microcontroller running at 8MHz so I ought to be able to send pulses at 12.5uS or so. I have tested with 35uS on and 35us off and I'm still getting very slow rotation speeds (couple seconds per rotation). very peculiar...

I belive the Gecko's can step up to 250Khz which is what .. 4us? So the off time of your pulse should be less than 4us.

That should provide the correct pulse shape to the driver.

[Alex_Cole]

I think cjmerlincnc may have hit the nail on the head. Like I said I dont know too much about this when you start talking about the figures like that but to me it sounds like an issue with the pulse.

I don't have an rpm meter or else I would hook up my extra G320 drive and see what I get running Mach3 at full tilt with the 500 count encoder and let you know what I am getting rpm wise.

At the basic setting as I undersatand it should pulse 25,000 pps(pulses per sec) So if my math and understanding is correct then with this I would get

25,000 possible from computer.
2000 pulses required per rev for these motors with these encoders.
=12.5 rev per sec
*60 equals 750rpm

These motors if you have the gear reduction is able to do 2000rpm at the output shaft at 30.7 volt if I am not mistaken.

I could be way off base here but I am trying to learn and help if possible.


PS I think I have my pulse width set at 3us in Mach3
AC

[bitviper]

Well, I guess I should make a quick 555 timer circuit to do faster pulses to see if that makes the motor go faster.

What hasn't been figured out yet is why, when I'm going at the currently very slow speed (5 seconds per revolution at the motor shaft), there is almost no torque yet there is tons of holding torque when the motor is stopeed.

Perhaps the slowness and the low movement torque are related. Anyhow, I'll see if anything interesting is discovered with my oscillator circuit.

[Alex_Cole]

Any progress on your issue? Just wanted to know how things were going

AC

[bitviper]

Well, there is an update. When I last posted, I had tried using the new power supply with the gecko 320 and I was getting very slow rotation rate (5 sec/rotation at motor shaft).

I decided to go back to the old driver with the old power supply and I'm getting the same symptom!

That tells me that it isn't the gecko's fault nor the power supply's.

I've ordered an oscilloscope so I can see if the pulses out of my microcontroller are what I expect them to be. I have a feeling that the pulses are coming out much slower than what I'm trying to accomplish.

I'll post what I find out.

[cjmerlincnc]

Hi Good to hear you are progressing. Just a thought, Could you try changing the micros oscillator crystal to a higher frequency. Then you would know there is a problem with the chip or the code.



Kind Regards
John

[bitviper]

Wow. I now know enough to know I know nothing.

I just had a great conversation with Mariss from Gecko and he enlightened me as to the proper way to driver a servo.

One of the points he said was that it's good practice to keep the RPM fairly high. The equation he used was (1 - (continuous torque / peak torque)) * no-load RPM.

My motor has a no-load RPM of about 3000RPM and a cont. torque of 20oz-in with a peak of 170-oz. As such, I should aim for it to be 2600RPM or so to get the most out of it yet I was operating it at 300RPM or slower. I could go up to the 2600RPM and get a reducer that brings it back down.

Another interesting equation he told me was how to calculate watts of power from rotation speed at a certain torque: oz-in * RPM / 1351 = Watts of mechanical power. I need about 300 oz-in at 300 RPM so that means I need about 67Watts of mechanical power.

Anything less than 100 Watts is stepper territory and more than 200 Watts is servo land.

I'm going to read up a bit more on this. He has a nice document on the Gecko website that goes into detail with the steppers so I'll be checking that out tonight.

I'll post back how things go. I've got to say, it sure is nice talking with someone who A) Knows what they're talking about and B) Genuinely cares and takes the time to explain and answer all questions. Wish more companies were like that.