F16 TQS Electronic Friction Generator Experiment Successful!

 

I have not posted for awhile due to health problem, and recovery... Not much in this post, just to report the successful test of the TQS friction generator.

The idea is to read the angle of the throttle arm, and use it to dynamically generate friction using an industrial servo motor running in torque mode (not all servo controllers allow that). Now... there are two possibilities.

1. direct drive
2. convert the torque/force into friction using a disk brake or drum brake mechanism.

#1 has the advantage of simpler mechanism, but it is not exactly how we "feel" in the cockpit. That is.... for instance... let's assume the dumbest constant force response, i.e. at any position, you always get the same "preferred" friction force, just like what the original "friction wheel" does. But... remember, only if you move the arm should the motor generate that constant force opposite of the delta v. Now, if the pilot is not moving the TQS, then, it should maintain the position and counter any movement.

These above two requirements conflict with each other... theoretically they do not, but in practice, the do. The trouble is that very few industrial servo motor controller offers direct torque control. They usually offer position, or velocity mode, but use torque to control the motor. For instance, if you command position X, then, any deviation of position X will cause a counter torque to return it to the position X. The magnitude of the counter force will be proportional to the deviation of the position. But.... you may be able to tune how much the magnitude of the counter force be (rarely), you don't get direct control of the magnitude of the torque.

The controllers that do offer direct control of the torque... usually offer it as a torque mode, apart from the position mode. So, you can control both. This is because, they use the control of the current to control torque, and use torque to control the position. So, if they offer you the position control, they can't very well let you have control of the torque.

Therefore, I opted for using a bicycle disk brake and caliper to generate the friction, but use the servo motor in torque mode to operate the caliper. This, I have no position control.

The main problem with this "scheme" is... I need constant torque... and constant torque means operating the motor at stall at all time... stalling motors draw a lot of current... and stalling at all time is not very health for the motors.

The solution is to derate the motor.... when I choose the motor, I have to look at the constant torque values, not the peak torque values to chose the motors that need my need. I have considered several other solutions, including an approved robotic competition DC motor and controller... a huge 1Ω resistor (3" x 6" x 1.5"), etc. before selecting this current motor.

For instance, originally, the design had a 3D printed disk 1/4" thick, and no caliper but a 3D printed bar mounted on the motor axel and rest on the edge of the disk. The torque will rotate the bar, and the bar will generate a force against the edge of the disk, thus friction. There were two problems.

1. not enough force generated, even at 100% torque of the selected servo motor, it can barely generate enough friction to hold the TQS in place without moving under its own weight.
2. at 100% friction, generated using the vendor supplied software, it quickly exceeded the RMS Max value of the motor, and safety shutdown was automatically triggered.

After a phone chat with a vendor engineer, the conclusion was.

1. The RMS Max is built in into the motor controller firmware, it cannot be bypassed.
2. The RMS Max shutdown is there to protect the motor, so it cannot be bypasswed.
3. There is a built in cooling fan mounting thread on the back of the motor, mounting a cooling fan could potentially double the RMS Max.

My experiment showed that at 20% commanded torque, the RMS Max will rise up to 98%, but will never exceed the shutdown limit. So, if I can double it, and get to 40% or even 50%, that's be good.

Order and wait for fans...

Further experiments showed that #3 above about adding a cooling fan could potentially double the RMS Max is unfounded. It did nothing of that sort, although the motor does run a lot cooler. Cooling fan... keep.

So, ok... that might be ok... if I can get 20% of torque.... the vendor engineer did tell me that the internal PWM -> torque generation is either 10 bit or 11 bit resolution. So, if I get 20% of 10bit, I get about 1024 * 0.2 = 204 steps, slightly less than 8 bit resolution. That's not too bad for TQS. I mean, the ideal is to have the exact same resolution as the Hall Effect sensor, but 200 steps is good enough for practical uses of the TQS friction.

Now, the question is how much the motor hub radius should be as the lever length for the torque equation, and how much the caliper spring force to counter.... and design an adapter to mount the caliper... well... and shop for proper bicycle disk brakes and calipers for the design... I ordered 4 different disks, and 3 different calipers, 1 from Shimano with single activation, 2 from some lesser known brands but with two sided activation, one with hydraulic activation, the other with mechanical activation.

Long story short.... many moons passed, and many obstacles over came.... I had to disassemble the two side activation mechanically activated caliper apart and modify it so that the spring force is not that great... the more spring back force, the more of my precious torque is wasted in countering it. A few days ago, I finally achieved a satisfactory result...

That is... I was able to have the whole thing setup... hook the motor controller to an Amazon bought no-idea-which-chinese-made-it PWM generator for RC... powered up, and turn the knob on the PWM generator to 20%... "kazz"... the caliper clamped down... and the TQS stood upright without assistance. Then, I pushed it... and there is some friction against my pushing... turn the PWM knob to 30%... Oh... more friction... and it's about right.... Now, turn the PWM up to 50%... "clack"... and I can hardly move that TQS! Turn it to 100%, and I can move it for about a few mm.... Strong the force this one is!

Then, I left that setup energized there for the whole week.... at 30% PWM. The house did not burn down. And the RMS Max stayed at about 40%, green. The motor ran cool with the cooling fan running.

So, now, all I need to do is to write the firmware and finish the optical switches/HATs. Of course, the firmware will feature multiple curve profiles, and perhaps a REST API to allow communication between DCS and it.