LuxShots Films
lvl.3
United States
Offline
|
DJI Mindy Posted at 3-26 03:31
Hi there, thanks for your feedback, our engineers will keep optimizing this function to make you have a better experience, please note the future update, thank you.
Hello Mindy.
The biggest issue with your engineer's response is simple. I am an Electrical Engineer so please take my input as something that should be given to your engineers for consideration. A gimbal is a machine that is designed to operate in a loop. The whole purpose of a loop (be that mechanical, electrical or in software) is that some external input will attempt to move it from it's set point. If the loop is designed properly, it has to return to it's set point if external forces aren't stronger than the compensation mechanism (i.e. the motors); and in a product such as a camera gimbal, should do so more smoothly than rapidly.
In a typical PID control loop(Proportional, Integral, Derivative) , some simple math will dictate how rapidly that occurs. But the problem may be that a PID loop was not implemented by your engineering team, and a PD loop was sued. The classical flaw of a PD loop is when the integral component is not considered, you get a build up of errors over time. PD loops are fine for a lot of products, but a camera gimbal is not one of them. The good news is that moving to a PID loop is just some simple math (at least simple for engineers). If you guys are using a PID loop, then your math is wrong. Get out some graph paper, a control theory book, and go fix it.
As far as your drifting and joystick calibration issue, this may be another matter. It is clear that the control you are using doesn't not have the appropriate tolerance to be used in the way it is in this system. Instead of having the user calibrate the joystick when (not if) it drifts off, the control should automatically zero the joystick periodically. Most controls utilize a standard wheatstone bridge circuitm and yours probably does the same. The issue is that you are probably using too high of a resistance in the bridge (probably to reduce power use and improve battery life), and as such, a small imbalance in the resistor tolerance can result is a large enough change to the measured bridge voltage.
The second issue with this is the bridge is probably measuring voltage instead of current. This is a trademark of inexperienced engineers because on the surface, they should work the same way. But in practice, voltage is much more susceptible to noise from nearby circuits contaminating the measurement, where current measurement has much more immunity to such issues. Of course I haven't cracked my new Ronin-S open to tell if this is truly the case. My other Ronin-S was garbage, but this new one seems to be acting right. In any case, you may have a harder time addressing the joystick drifting issue than the PID loop.
I hope this helps.
|
|