Parameter to scale the Power module Voltage Divider & amps per volt based on Throttle percentage

[Arshish Maneckji]

Upon testing different settings with the voltage divider and amps per volt parameters I noticed that the current draw does not match a connected wattmeter current draw in a linear manner. 

I started off by setting these 2 parameter @ 50% throttle in Mission planner. 

Then I connected a wattmeter to the quad with the props inverted and swapped one position right. The quad was connected wirelessly to MIssion planner. The further up I went from 50% throttle bigger the difference between the mission planner reported V & I and the Wattmeter V & I. Then I set the voltage divider and amps per volt parameters based on 80% throttle position. Again the test revealed the same results where the Mission planner values and the wattmeter values matched at 80% - 90% throttle but were way off upto 60% throttle. The issue with this is that when your land warning goes off based on a particular mAh consumption you either come down with too much left in the battery or too little left in the battery. I have cross checked this by checking the amps that the battery takes back in. If I had set the parameters based on a 50% throttle, it takes in more than what I would expect it to take based on the MAH COMSUMED as reported by Mission planner & Vice Versa if I set the parameters @ 80+% throttle

Now from trail and error on my system I know that with a fully charged LIPO, if I fly for 10.5 minutes I come down with 6% battery left. I have verified this multiple times based on the MAH that the battery takes back.

So, rather than tune the amps per volt and voltage divider for a particular throttle setting to compare with a power analyser, I tuned the amps per volt parameter such that the Mission planner shows the exact capacity used as is being put back into the battery on a subsequent charge. I had to do this multiple times to get the value to match exactly with an accuracy of close to 20mAh on a 5000mAh battery. 

Now coming to the subject of this post - Would it be useful to have a parameter that scales the voltage divider and amps per volt parameters based on the throttle position?

[Tom Pittenger]

What you're describing is a non-linear slope of the current/voltage sensor. Everything is more accurate if you allow for non-linear, or multi-linear points. This is indirectly addressed in https://github.com/diydrones/ardupilot/pull/3573 but it is directed at plane throttle (copter is easy to adapt though I'm sure). I know there are other tactics already implemented for Copter but I'll let Randy chime in on that.

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Luis Vale Gonçalves]

This might help, although indirectly:

http://copter.ardupilot.com/wiki/current-limiting-and-voltage-scaling/

There are other Power modules referenced on the wiki, and with one of them that uses Hall effect the linearization is very good.

To get more accurate parameters with the "normal" 3DR Power module I would recommend measuring the current at an average value that represents real flight. Example: on one my copters it uses aprox 45 Amps flying, so I calibrated at 45 Amps :)

[Luis Vale Gonçalves]

But the method you describe is quite valid.

[Randy Mackay]

 

     I wonder if perhaps a hardware solution would be simpler?  It seems they address the non-linear issue directly and the price is really not bad.

          http://copter.ardupilot.com/wiki/common-mauch-power-modules/

 

-Randy

[Arshish Maneckji]

Very Interesting.. I am tempted to pick one up. At 22$ it's even cheaper than the standard power module. Any reason why this is not the standard recommendation? looks like a win win to me.

http://www.mauch-electronic.com/apps/webstore/products/show/6626750

[Arshish Maneckji]

Reading Randy's & your response - The Hall effect sensor is the way to go..

Maybe this needs to become the recommendation?

Thanks Everyone!

[Andy Little]

I have used the ACS758  Hall effect current sensors for some time. They have many advantages and no disadvantages. They can use an isolated power supply ie ground not connected to motor ground. They have a very low resistance so dont get hot even at high currents. They can be used at 3.3 V too though accuracy not guaranteed then.

I have previously just soldered the raw ACS758 i.c in the +ve ESC battey lead close to the ESC ( I solder a 0.1 uF smd cap between the 5V and gnd sensor terminals for decoupling) as shown in the picture ) The snor would be a good cndidate to include direct onto the ESC PCB