The Bugs Continue...

After all the testing I have done I find that I am still not 100% happy with how the calibration works on the DESC.  It works great for awhile then for no apparent reason it looses the EEPROM value on startup.  Maybe I am just being to hard on the board but I want it to be 100% reliable.  The hardware is sound but the software still leaves much to be desired.  

 

A New Method

This problem has lead me into an entirely new direction for the board.  I am now experimenting with calibrating every time on startup.  Initial tests look promising.  The first 1/4 second after the board receives a good signal it will calibrate based on stick position (this means sticks need to be at neutral).  After calibration it may still be slightly off but it can easily be adjusted using the trims on the radio.  I have found that it is usually only one or two clicks out of center, this is within the deadband so the motors will not spin at idle after calibration.  Should power be lost on the robot during combat it will re-calibrate to a bad position; however this should be a rare event.  It can also be fixed without touching the robot because I have it setup so that switching the transmitter on and off will also re-calibrate.  

 

New Feature

Since I am no longer using the calibration jumper I am writing a sub routine for on board mixing.  I have the mixing working on my test bot I just need to write code to toggle it on or off.  If the old calibration pin is not jumpered then there will be no mixing.  If the pin is jumpered then mixing will be enabled.  Since I still have another 40 or so boards left the silkscreen will not be updated just yet but that is a minor issue.

 

Light at the end of the tunnel

Nearly 3 years after starting this project I am almost at the point where I am 100% happy with this.  It has been a rough road especially as of late because I thought I was there before these calibration bugs came up.  It is my intent to do even more through testing during this holiday break and start selling the boards again in early December.  I have a few great things planned once I have this board all worked out so keep an eye out.  

The Bugs Continue...

After all the testing I have done I find that I am still not 100% happy with how the calibration works on the DESC.  It works great for awhile then for no apparent reason it looses the EEPROM value on startup.  Maybe I am just being to hard on the board but I want it to be 100% reliable.  The hardware is sound but the software still leaves much to be desired.  

 

A New Method

This problem has lead me into an entirely new direction for the board.  I am now experimenting with calibrating every time on startup.  Initial tests look promising.  The first 1/4 second after the board receives a good signal it will calibrate based on stick position (this means sticks need to be at neutral).  After calibration it may still be slightly off but it can easily be adjusted using the trims on the radio.  I have found that it is usually only one or two clicks out of center, this is within the deadband so the motors will not spin at idle after calibration.  Should power be lost on the robot during combat it will re-calibrate to a bad position; however this should be a rare event.  It can also be fixed without touching the robot because I have it setup so that switching the transmitter on and off will also re-calibrate.  

 

New Feature

Since I am no longer using the calibration jumper I am writing a sub routine for on board mixing.  I have the mixing working on my test bot I just need to write code to toggle it on or off.  If the old calibration pin is not jumpered then there will be no mixing.  If the pin is jumpered then mixing will be enabled.  Since I still have another 40 or so boards left the silkscreen will not be updated just yet but that is a minor issue.

 

Light at the end of the tunnel

Nearly 3 years after starting this project I am almost at the point where I am 100% happy with this.  It has been a rough road especially as of late because I thought I was there before these calibration bugs came up.  It is my intent to do even more through testing during this holiday break and start selling the boards again in early December.  I have a few great things planned once I have this board all worked out so keep an eye out.  

Older Post Newer Post

0 comments

Leave a comment