Design > Engineering board

Electric Motor Control - severe Geek speak

<< < (4/4)

AMV:

--- Quote from: Howard Rush on February 15, 2022, 01:16:05 PM ---Technical merit of this discussion aside, sensors have sure come a long way since 1980.

--- End quote ---

...and to me, 1980 forever remains 20 years ago.  LL~ LL~
The other day I was watching a video interview of some average Joe recounting his memories from his youth in the 1950s. The interview was done sometime around 1980.  It dawned on me -- at the time of his interview, WWII was as fresh to him as 1980 is to us now.

Time flies like an arrow; fruit flies like a banana.

Mark wood:

--- Quote from: Howard Rush on February 15, 2022, 01:16:05 PM ---Technical merit of this discussion aside, sensors have sure come a long way since 1980.

--- End quote ---

That's very true. I did consider using one of the Kalman integrated sensors but the reality is for the job it is way over kill. Besides, without a data collection flight to understand the impact of tethering on the data I don't have enough understanding to use one. The bottom line is that all you need to know is which way is the gravity pointing in relation to the direction the nose is pointing. That's a simple solution which can be determined easily enough using the Gx and Gz vectors. It works quite well even in the wingover maneuvers. As the airplane crosses over the pole Gx changes from positive to negative which drags the gain sign with it cause the output to go from adding thrust to reducing thrust. The only problem is preventing a divide by zero crash as the wingover involves a Gz = 0 and at the pole Gx and Gz are zero. So I added 0.01 to the dividend to keep the 1/0 daemon away.

Mark wood:
Here is a first glimpse at the control functioning. After the first couple of successful flights, I had to mess with it. Needless to say that wasn't exactly a good idea, although it is ultimately necessary to bring the control to it's potential. The next feature I am going to incorporate is the RF terminator which will expand to inflight gain trimming. That will begin as soon as the PCBs arrive.



 

Mark wood:
Here's a slightly better video with some discussion of what is happening with the control. My heater is running in the background since it is RF cold outside right now here. It's about 7 F and my shop isn't super well insulated.

Mark wood:
We didn't get any video of the control in the GeeBee flying yesterday but the performance is becoming very good. Several of us flew the airplane and the consensus is very positive. The turning from high to low is gaining a bit too much energy, as could be anticipated, otherwise the power and velocity are quite trimmable. I have a bias solution to cut back some of the maneuvering gain during the high turning low maneuvers. The trouble is that the attitude gain lags noticeably and doesn't catch up until just past the exit of the corner which is too late to prevent the acceleration.

I have some PCB boards coming to build a few of these in a more "professional" package.

Next time we have the gym for flying, I'm hoping to have the bias gain working. I'll be certain to get some video.

Navigation

[0] Message Index

[*] Previous page