Hello World Programs (+ problem with Motor C Channel)

[Blake Hament]

Hello,

So I am doing initial calibration and testing all components with the compiled haptifabrikenapi and chai3d-master libraries.

Which programs do you recommend for getting started? I found many sample programs in /chai3d-master/bin/lin-x86_64/, including "13-primitives", the teapot program featured in many WoodenHaptics demo pics and videos. So far "01-mydevice" has been most helpful. I can see a cursor that changes position and orientation as I move the interface and a white vector showing my speed. 

Do you use these chai3d example programs to test motor actuation? I have been able to read encoder values as I move the motors manually, but I have not yet been able to activate any motor actuation. I've tried activating potential field and dampening, but no luck yet.

On a side note, all motors and driver chips work fine on channels A and B, but I am not seeing any encoder signal from channel C when running chai3d examples. Soldering looks good and the motor driver chip gets a blinking green light as in the other channels, so I am wondering if there is some software parameter I should look at?

Thank you and best regards,

Blake

[Blake Hament]

My issue with channel C was due to a bad connection at the quad differential line receiver. Now I can read encoder counts, send analog signal to enable motors, and send analog setpoints to the motors.  

I'm finding that Chai3D was helpful to identify big problems (like chan c encoder not counting), but then it was more helpful to go back and debug just at the lower s826 SDK or HaptikFabrikenAPI levels. 

I still don't have "good" force rendering yet, but forward kinematics seem ok. Continuing to try my own experiments and tweak the parameters in haptikfabrikenapi.h

[Jonas Forsslund]

Hi,

If you have the 310009 motor and reading position seems ok, you should only have to change the line configuring woodenhaptics_v2015() 

 (around line 300 in haptikfabrikenapi.h):

0.0259, 0.0259, 0.0259, 3.0, 2000, 2000, 2000,0,0,0,

to:

0.0538,0.0538,0.0538, 3.0, 2000, 2000, 2000,0,0,0,

since you have 53.8 mNm/A torque constant (which is much better than the 310007). Let me know how that feels!

Jonas

--
You received this message because you are subscribed to the Google Groups "WoodenHaptics users and developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to woodenhaptic...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/woodenhaptics/cefb7f82-8e22-4ccb-9921-f10bdbbbd26d%40googlegroups.com.

[Blake Hament]

Hi Jonas,

I changed the torque constant as you suggest. I've also tried adjusting the max current (1.74 A for 310009). I played with the "...aligned..." params so that my position and angles change as I expect, forward kinematics look good. 

But I still don't really feel the forces I am trying to render. Using the existing spring force lines in haptikfabrikenapi main.cpp, I expect to feel an increasing repelling force as I move away from the origin. I really don't feel anything except a quick jerk now and then, usually during calibration when the motors are first enabled. If I take my hands off the WH device, sometimes I see the motors push it in the correct direction, but it is a very weak torque.

I've tried just directly sending current to the motors with setCurrent and then observing the device motion with my hands off. I've gone up to 3A (max current for 310009 for short operation), but that's not even enough to overcome the gravity on Lb and Lc with motors B and C. Motor A can rotate body A with a little push to overcome static friction at 3A, but this is well over the suggested max current for continuous operation, 1.74 A. I've also played with the max_linear_force parameter, but I believe it is not used, I see "5" hardcoded in setForce() as the force limit in haptikfabrikenapi.cpp.

My thought now is that the 310009 motors do not have enough torque for the WH device. However, looking at the speed/torque/current graphs for 310007 and 310009, it looks like they should be supplying very similar torques at their respective max nominal currents (85.6mNm vs 89.7mNm). A quick, rough calculation of the torque necessary to overcome gravity on fully extended arm B with motor B (for now considering arm C not attached) is torque = mass_b*(9.81 m/s^2)*Lb = 320 mNm. Of course it will be less if arm B is closer to vertical, but still, this is easily 2 or 3 times bigger than the ~90mNM torque the motors are rated for.

Do you still get strong force transmission with the 310007? Sometimes I get huge frictions from the cables overlapping, but even when I'm very careful that the joints are all rotating freely, I don't experience much force transmission. Appreciate any ideas you have for what to try next, especially if you think these motors should be strong enough.

Best Regards,

Blake

To unsubscribe from this group and stop receiving emails from it, send an email to wooden...@googlegroups.com.

[Jonas Forsslund]

Hi Blake,

I am using the same motors in some devices and they are very powerful. What power supply do you use? A powerful laptop adaptor would do. 

Perhaps you can share pictures/videos of your device? Try to avoid overlapping cables, that increases friction. 

You can try setCurrent() and you should defenitetely feel something. You can also connect a usb cable to one escon and verify live in escon studio, when running the device, what current set value it produces.

Jonas

To unsubscribe from this group and stop receiving emails from it, send an email to woodenhaptic...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/woodenhaptics/a092063a-5fb2-4534-9751-9362ade169e5%40googlegroups.com.

[Blake Hament]

Hi Jonas,

Ok, sounds like the torque should be plenty if done right. I'm using a generic laptop power supply, 19.5V 3A.

I monitored the motors and drivers using ESCON Studio like you suggest, and I found the issue is related to "Motor I2t Limitation". At low voltage and amps, the motor driver sends the analog out amp setpoint like I expect. But if the analog in setpoint commanded by software goes over about .2A, the I2t limitation kicks in and the analog out amp setpoint drops to almost 0A. 

I'm reading more on I2t and may get in touch with Maxon to see how they recommend to compensate. Perhaps the 310009s are less tolerant to the heat build up during stall? I'll let you know what I find out.

Best Regards,

Blake

To view this discussion on the web, visit https://groups.google.com/d/msgid/woodenhaptics/a092063a-5fb2-4534-9751-9362ade169e5%40googlegroups.com.

[Jonas Forsslund]

Hi,

Please go through the setup guide again and check that you have the right values compared to my screenshots. 

As I said, I have used the ..09 and they are great. 

Jonas

To unsubscribe from this group and stop receiving emails from it, send an email to woodenhaptic...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/woodenhaptics/dd55d528-4876-4575-a397-be63708224c2%40googlegroups.com.

[Blake Hament]

Yep! I had entered the wrong row from the spec sheet in the setup. The driver was limiting current to the motor because my nominal current was set too low. I'm feeling much stronger torque from the motors now, thank you! Still seeing a few weird things, but definitely improved!

To view this discussion on the web, visit https://groups.google.com/d/msgid/woodenhaptics/dd55d528-4876-4575-a397-be63708224c2%40googlegroups.com.