|
Post by skihnia on Feb 18, 2017 7:28:00 GMT
Hello all! Newbie here. My first motion simulator project. I have few questions and didn't know is it a hardware or software issue so started new thread for the project it self. I'm using linear actuators with AC servo motors and +-11V analog output from AMC1280USB. Just did my first tests and had one problem. I was unable to stop the motors. When I push kill switch I was left with 44mV signal from the controller translating about 5rpm on the motor. Also without position command from computer when I set the potentiometer to middle position (49-50) I get the same 44mV. Few other questions as I'm already here Since I'm using servo motors what by definition has position signal. Is there a way to use that signal as position signal for AMC1280USB? I need to rotate motor 60 times to get full movement from the actuator. Having hard time finding info about the jumpers. What do I need to change to get the end stops to work? I got the board with Jumper on S1-S2. Why analog endstops? Is that just indication how they work on the board or should I use something else than regular switches? Than you for reading!
|
|
|
Post by tronicgr on Feb 18, 2017 19:55:03 GMT
Hi,
I try to do my best to resolve your issues. Excuse my short message but I have been ill the last days.
It would be great help of you take some photos of the boards and the wiring, and photos of the servo drive wired and perhaps the motors themselves to have a better idea of what you try to accomplish. I mean these are linear servos?
Also a video of their motion would help.
About the signal conditioner, it has a small offset depending on the computer or power supply you use to power the amc180usb. When you use the kill switch it places the voltage to 2.5v (or 2.48v) that might not be exactly 0v for the servo ( 44mv as you said).
This will be fixed in firmware setting. My suggestion for now is not to use the KLM switch, but the Enable servo switch on the servo drive for now.
There is a way to set the amc1280usb to direct servo mode that outputs position on the analog out signal, instead of speed and direction. In this servo mode, you don't need potentiometer attached. You can activate this mode by placing a jumper on S1 if I remember correctly (near the analog inputs). I'll have to double check it... also the S3 jumper can be used as output to activate the servos when the controller is active and receives motion data. I'll get back to you on this as soon as I check a few things.
About the limit switches, they are not added in the firmware you have yet, will be added soon.
If not wrong most of the servo drives now support limit switches natively so you can hook up the linear actuator with no fear of runaway. Again, some photos and perhaps specs of the servo drives and actuators would allow better advice on this matter.
Thanks Thanos
|
|
|
Post by skihnia on Feb 20, 2017 7:20:24 GMT
Thank you for the quick reply! I'm using Exlar KM75 linear actuators and cheap Chinese servo motors. Link for the servo manuals I have: servo manuals
So actuator are hardware only (no position or electronics at all) and servos has no knowledge of their position, only it's own rotary position. This might be a problem using position signal instead of speed/direction, or is it? At the moment I'm planing to add extra potentiometer with 1:6 gearing at the end of the servo. I would imagine that to get my current setup to work without the potentiometers one should be able to use the servo position output and count the actuator position from that. I needed to clean up my work bench for now but I get you video soon as I do my next tests. You already seem to have answered many of my question so I'll wait what you find out.
|
|
|
Post by tronicgr on Feb 20, 2017 15:42:24 GMT
I think the best way to set up your actuators would be using external sensors for positioning. This way you just control the motors for speed and direction and let external absolute position sensor to handle the actuator position. You can use this string potentiometer solution I made: motionsim.freeforums.net/thread/230/3d-printed-string-potentiometerI have printed some extra that I can send you along the spring mechanisms to put together. You will just need to get a few 10-turn 10k potentiometers (cheap ones are fine too, no need for the magnetic I link). I'm going to do the changes needed for the Firmware too so you can use the 12-bit resolution DAC so you can set precise the signal conditioner to 0mv (instead of 44mv now). Thanks Thanos
|
|
|
Post by tronicgr on Feb 22, 2017 12:24:06 GMT
Hi Seppo Have a look at this video, it might help you setup some things. I didn't had the signal conditioner on hand to plug it, but I needed the 0-5v analog out anyways to show you a few things on the oscilloscope. youtu.be/ALFqsmKs384On the last part of the video I show you how to set the offset of the analog using the 6dof start position setting. Thanks Thanos
|
|
|
Post by tronicgr on Apr 9, 2017 19:59:09 GMT
|
|
|
Post by hexpod on Apr 10, 2017 19:03:13 GMT
Congratulations for the great development! Is the new 12bit analog resolution also available without extention board? (Teknic motors setup) Cheers
|
|
|
Post by tronicgr on Apr 11, 2017 11:17:04 GMT
Congratulations for the great development! Is the new 12bit analog resolution also available without extention board? (Teknic motors setup) Cheers There is 12-bit PWM outputs onboard the main AMC1280USB v2.1 that can be used with the clearpath motors. If I remember correctly the clear path motors only accept PWM, not analog signals.
|
|
|
Post by skihnia on May 21, 2017 7:41:27 GMT
Tried to change to 12-bit with od behavior. Cut the 8-bit connections and soldered 12-bit connections. Changed setting to 12-bit output. Now all movements were as slow motion. Made some measurements on output test mode (dip: 1 down, 4 up). When measuring from +-11V signal conditioner output voltage isn't even near what it should. Measuring signal conditioner 5V pins it gives 2.4-2.6V. Same 5V pin signal conditioner removed gives ~0-5V as I think it should. Video from signal conditioner 5v pinsVideo signal conditioner removed
|
|
|
Post by tronicgr on May 21, 2017 13:44:58 GMT
Tried to change to 12-bit with od behavior. Cut the 8-bit connections and soldered 12-bit connections. Changed setting to 12-bit output. Now all movements were as slow motion. Made some measurements on output test mode (dip: 1 down, 4 up). When measuring from +-11V signal conditioner output voltage isn't even near what it should. Measuring signal conditioner 5V pins it gives 2.4-2.6V. Same 5V pin signal conditioner removed gives ~0-5V as I think it should. Video from signal conditioner 5v pinsVideo signal conditioner removedIs the pads that connect the LM324 amplifiers cut? Take a close photo of the 6dof board without the signal conditioner on it. Can you measure the voltages on the big black part on the signal conditioner? You should measure +14v and -14v on two of the legs (if its easy to reach). Also you will have to choose the 12bit analog output in the menu as well to activate it. And in normal operation you will have to increase the max motor speed from 25% to higher value to get the full voltage output. Thanks Thanos
|
|
|
Post by skihnia on May 21, 2017 19:50:27 GMT
I'm running the simulator with the 8-bit output for now and it's quite laborious to get the board off. So new measurements when I'm confident enough to try the 12-bit again. On the +-11 outputs it gives right voltages with the 8-bit but low (below +-1V) with 12-bit. Could this be some software setting that might need a reset? I recently upgraded to 2.3 firmware. Just remembered I might have read that someone needed to reset some settings.
Everything worked as expected with 8-bit before and after the 12-bit test. I'm sure I changed the output setting to 12-bit and quite sure I had max motor speed at 50%. What does the Steps % setting do in the motor speed settings?
Another topic... How does the USO strings work? I have the home position set low (10%) to get in easily. Now when I start x-sim it jumps quite harshly to middle position. Could I use some startup string to ease the jump? Edit: Or could I have some script on the computer to easily modify the home position from the drivers seat? Fabians configurator didn't seem to work well with the 2.3 firmware.
|
|
|
Post by tronicgr on May 22, 2017 3:07:35 GMT
I'm running the simulator with the 8-bit output for now and it's quite laborious to get the board off. So new measurements when I'm confident enough to try the 12-bit again. On the +-11 outputs it gives right voltages with the 8-bit but low (below +-1V) with 12-bit. Could this be some software setting that might need a reset? I recently upgraded to 2.3 firmware. Just remembered I might have read that someone needed to reset some settings. Everything worked as expected with 8-bit before and after the 12-bit test. I'm sure I changed the output setting to 12-bit and quite sure I had max motor speed at 50%. What does the Steps % setting do in the motor speed settings? Another topic... How does the USO strings work? I have the home position set low (10%) to get in easily. Now when I start x-sim it jumps quite harshly to middle position. Could I use some startup string to ease the jump? Edit: Or could I have some script on the computer to easily modify the home position from the drivers seat? Fabians configurator didn't seem to work well with the 2.3 firmware. The steps setting introduces virtual steps to force smaller steps between two points when the platform moves from online mode to offline mode or home position. Usually is best to set to 99, but it might come handy to play with it a little, it might remedy the Jump you get during power up. Fabians configurator is missing many of the new added functions and options of version 2.3 firmware. Still compatible with functions of version 2.0 firmware if you deactivate the "motion only data" setting lock, for PID calibration perhaps. What is your PID settings btw? I'll look into the 12bit issue tomorrow morning and take a video with actual voltage measurements at points you need to do to verify that all was set correctly. Thanks Thanos
|
|
|
Post by skihnia on May 22, 2017 13:59:58 GMT
I'm running the simulator with the 8-bit output for now and it's quite laborious to get the board off. So new measurements when I'm confident enough to try the 12-bit again. On the +-11 outputs it gives right voltages with the 8-bit but low (below +-1V) with 12-bit. Could this be some software setting that might need a reset? I recently upgraded to 2.3 firmware. Just remembered I might have read that someone needed to reset some settings. Everything worked as expected with 8-bit before and after the 12-bit test. I'm sure I changed the output setting to 12-bit and quite sure I had max motor speed at 50%. What does the Steps % setting do in the motor speed settings? Another topic... How does the USO strings work? I have the home position set low (10%) to get in easily. Now when I start x-sim it jumps quite harshly to middle position. Could I use some startup string to ease the jump? Edit: Or could I have some script on the computer to easily modify the home position from the drivers seat? Fabians configurator didn't seem to work well with the 2.3 firmware. The steps setting introduces virtual steps to force smaller steps between two points when the platform moves from online mode to offline mode or home position. Usually is best to set to 99, but it might come handy to play with it a little, it might remedy the Jump you get during power up. Fabians configurator is missing many of the new added functions and options of version 2.3 firmware. Still compatible with functions of version 2.0 firmware if you deactivate the "motion only data" setting lock, for PID calibration perhaps. What is your PID settings btw? I'll look into the 12bit issue tomorrow morning and take a video with actual voltage measurements at points you need to do to verify that all was set correctly. Thanks Thanos I have had the step setting at 50% and that makes sense why the move from online to offline(home) is very smooth and exactly what is hoped. The problem is that when going from offline to online and the x-sim takes over jump from low home position to middle is so hard that something needs to be done. Solution would be that I could raise the home position to middle before going online. Desired function would be that I could have two home position with different buttons. Another problem with home position. I'm forced to use 1-2-3-4-5-6 setting instead of 4DOF at the moment because of inversions of the second actuator. Now (in 4DOF mode) if I set home position to anything other than 50% second actuator goes to opposite direction. If I had regular potentiometers I could just reverse the polarity but I'm using hall potentiometers and can't reverse it. Suggestions? Have to test the configurator again. When I tested it gave me some errors and it was impossible to set the home position. In 8-bit mode I have increased (P)ID value slowly from 5 to 20. I and D 0 so far. In 12-bit mode when movement vent slow I tried P value up to 100 but that didn't help. Thank you! Yet another topic. I haven't noticed any impact on the motor max speed settings. I think I have tried anything between 10-100%.
|
|
|
Post by tronicgr on May 23, 2017 11:37:53 GMT
Hi, I need to edit the firmware again, the max motor speed is indeed broken for unidirectional setting, everything else seems to work fine in 8bit analog output: youtu.be/7VuzPR5A2vcMade a video too showing the modification for redirecting the 12bit analog output signal on the 6dof ext board: youtu.be/iIhfP9lUzeMYou are right, there is an impendance issue on 12bit output to the signal conditioner. I'll modify the output filter with resistors in parallel until it gets correct voltages and frequency response. youtu.be/yOJOb0sdduYI'll let you know soon. Thanks Thanos
|
|
|
Post by skihnia on May 23, 2017 14:14:09 GMT
Hi, I need to edit the firmware again, the max motor speed is indeed broken for unidirectional setting, everything else seems to work fine in 8bit analog output: youtu.be/7VuzPR5A2vcMade a video too showing the modification for redirecting the 12bit analog output signal on the 6dof ext board: youtu.be/iIhfP9lUzeMYou are right, there is an impendance issue on 12bit output to the signal conditioner. I'll modify the output filter with resistors in parallel until it gets correct voltages and frequency response. youtu.be/yOJOb0sdduYI'll let you know soon. Thanks Thanos This is good news! It means that I didn't screw anything. How about the home position of the second actuator? Could there be an option in the firmware? For the life of me I can't figure out how to setup the yaw. The problem is that I can't figure out how to get yaw acceleration information only exact heading where the car is pointing in the game. I have position information on the yaw axle but at the moment firmware doesn't know the shortest way to correct position. Example (this is only theory. haven't dared to test it in real world yet): I do donuts in the game and platform follows the movement at slight delay. For example yaw@359deg platform@356deg(and closing)... yaw goes over to 1deg... platform still @359deg... AMC now gets command to turn full speed to opposite direction (because it doesn't know there is a shorter way over the 0 to correct position).
|
|