Sunray CE

[WhatATest]

What do you mean?
Isnt voltage drop basically linearly correlated with load? Then you can correct voltage readings based on that.

To me that seems to work, I wish I had a screenshot of how it looked before.

 

[EinEinfach]

Isn't getBatterVoltage a real measured value?

If you add always an offset based on motorCurrent, then the result isn't real measured value. The oscilated voltage is real measured value, what you goal is to filter voltage for chart. You don't want to have oscilation in your charts even is that there. Therefore my suggerstion to add an offset if it neccassary and not always.

 

[WhatATest]

Yeah, getBatteryVoltage() is the real voltage.

Unless Im missundertanding something that doesnt represent the actual SoC no?, you get a voltage drop from the batterys internal resistance. Once motors stop running the voltage goes back up to the floating battery voltage.
I dont know if this just happens to me, but I had to set the charging voltage cutoff to 28.6 for it to actually be at 28 when the current gets disconnected, for example. Same for the docking voltage.

 

[Mr. Tree]

theres a misalignment is suppose... if relais is closed.. the read voltage of charger connected or not would be the battery voltage? or am i missing a diaode? I would say the charging voltage is somehow the battery voltage if connected to charger an charging. if the charger connection is interrupted, only the disable relais would show you the real voltage of the mower connectors, else it just reads battery voltage. Just my finding, or i have some disfunctioning hardware. To get the soc however is quite easy for liIo over voltage, ... for lifepo it will not work.

 

[WhatATest]

I dont think your hadware is the problem, from what Ive researched seems like its not possible to read the batterys "real" voltage when its either charging or loaded.
The method I posted gives a good approximation, though you would still need to take into consideration things like temperature and the PCBs current draw.

 

[Mr. Tree]

yes, this makes a lot of logic unlogic, if you would assume charger voltage is actually the charger voltage. I didnt know that, until i discovered that it´s absolutely neccessary to open the relais and then take a "look" at the connectors voltage.

 

[Whatson]

Lost my account.
Anyway, this years improvements:

• Lots of state estimation improvements
  • Better position and heading fusion in a mini kalman like filter
  • Continous integration of imu and odometry
  • Im getting basically perfect docking now with no gps.
• Better tracking
  • Proper acceleration and deceleration for linear and angular movement
  • Continous loop
• Autotuning
  • Wheel diameter
  • Wheel base
  • Feed Fordward coefficients
  • Variances
• Dynamic rear traction
• Optimizations
  • 2.5 ms loop when mowing. Dont know what regular sunray takes.
  • Non blocking temperature reads.
  • Reduced ublox payload
• Probably other stuff

I dont know why I keep improving the firmware, its been basically flawless for me for a while now...