I wonder how many CPD settings are adjusted to compensate for other settings.
3 Likes
We also use an old Autofarm A5 system. It gets calibrated without machine in the field. After calibration it almost always runs very smooth with all machines. Which means completely straight lines (by looking) and crosstrackerror on the screen of maximum +/-2.
They only have 2 or 3 values to play with in a range of maybe 1 or 2 “points”. So very narrow adjustment necessary.
I still think the steering angle should always be displayed correct and it would be better to create a new parameter like steering angle gain. It’s easier to understand for someone who is new.
1 Like
Maybe I’m wrong 
, but, with dual antenna or good heading sensor we know :
- heading
- time
(so we should know the rad/s) - speed
So with rad/s with the speed tractor (and maybe wheels distance) we should know the steering angle without wheels sensor ?
This virtual steering angle is maybe better because ignore slippery grounds ripping.
- use only this without real angle sensor may have a problem if it’s too slipply (if the tractor go straight even steering is full turning)
- use fusion of real angle and virtual angle can be good (can correct the ripping)
Maybe I’m just completely stupid 
Hello,
Yes I know, voltage on AIN pin is limited to VDD+0,3V.
And yes our sensor is generally 0,5V to 4,5V so we only use positive count range from 0 to +32767 (15 bits).
I suppose the FSR range of ADS is set to 6,444V, so LSB size is 187.5µV.
So 4.5V is 24000 counts (5V is 26667 counts).
I look on the Arduino code and what I haven’t seen before is that the 2 LSB are thrown, so the code only used 13bits, so max count is 8191.
So with FSR at 6.444V, LSB in 13bits is 750µV (6.444V/2^13). So 4.5V is 6000counts (5V is 6667 counts).
More clear for me now.
Yes I undestand what is CPD. As we can have tractor with different wheel angle and ss our sensor mounting never fully operate the full sensor range (ideally, wheel full left the sensor output is 0,5V and full right is 4,5V, but this is never reach), we have to adjust CPD for each mounting.
But like @darrenjlobb, on what I was surprised and what I found strange is the fact that CPD is a parameter that we have to consider to adjust for optimising the steering.
For my it was a “fix” parameter: CPD is set one time to obtain wheel angle in AOG = real wheel angle and then it’s fixed. Then the other parameter have to be adjusted to optimise the steering.
Yes I understant that we can have a wheel pointing at a certain angle and a a different value displayed in AOG and we don’t know.
But I was thiniking that to have a good autosteering, AOG shall know the real wheel angle. I found that strange that AOG steering algorithms can sometime works better with a wheel angle value in AOG different from the real wheel angle ?
Math
I think it s because the angle is not always the real direction that your rover gonna take in real . drifting , mud , weight in front or rear of tractor … so many reason …
1 Like
Isn’t it the difference of steer angle and heading? 
the speed of the heading change would rather be the answer to the steer angle?
Yes I think heading change in radian/second. And speed tractor. At 5km/h or 10km/h the heading change will be x2 for the same steer angle.
I think
i think maybe a factor is the amount of toe in of the tractor and how agressively it is set - or if there is toe out or neutral. it takes a tiny bit more to correct on some tractors, and maybe that is the difference. Just a guess though.
Hello @BrianTee_Admin
Question, in the .INO it’s writed that: “1288 counts per 1 volt”, line 6 on USB 4.3.10 INO.
How did you obtain this ?
ADS1115 is 16bits, with Two’s complement, so only 15 bits for data, and in the INO, the 2 LSB bits are thrown, so only 13bits for data.
So for positive values, max count is (2^13)-1=8191 counts.
So as the ADS range is 6.144V, counts per volts is 8191/6.144=1333 counts per volts.
Or have I miss something ?
Math
Maybe an old comment that isn’t right?
Don’t know 
If I’m correct with 1333 count per volt, still in .INO 4.3.10, line 433 and 439 this comment aren’t correct too:
“0 to 6640 is 0 to 5v” → 5V is 6665.
Not a big change but if my calculations are correct, perhaps update the commentary in the futures .INO (if the commentary are reported in v5), to avoid any confusions.
Math
Bit shift by 2 gives 0 to 13610 for 0 to 5v in most recent v5 ino.
And, oh look, the comment is even correct 
6805 is subtracted to give a zero center, ±4000 are used for WAS offset.
1 Like
Hello,
Sorry Brian, I don’t understand !
I see line 406 of 5.1 INO:
steeringPosition = (steeringPosition >> 1); //bit shift by 2 0 to 13610 is 0 to 5v
Now bit shifting is 1, so still one bit for sign (always 0 in our case) and 14 bits for value. So max count is (2^14)-1=16383 count.
So 16383/6.144 = 2667 counts/volt.
So 5V = 13335 counts. So why 13610 in INO ? There is somthing that I miss ?
Math
2 Likes
There is a gain amplifier set to 6.14 full scale so that takes it down when only using 5v
Hi team,
My landrover style sensor gives a voltage range of .2v to 4.8v over a travel range of 84 degrees.
Looking at the 1288 counts per 1 v, I believe this gives me a count range of 5925, and a CPD of 70/71.
Am I understanding this correctly ?
The autosteer wizard gave me a CPD of 161 in comparison, which feels like a huge difference in calculation.
161 is a number between 0-255, CPD adjustment range, nothing to do with actual CPD. In this range 180 CPD would correspond, then there is the adjustment on that, which we know works better in a lower range than the one found by the wizard. And this is how the 161 are calculated, which are approximate and you must test in the field.
5925 / 84 = 70.53
70.53-------->100
x -------->255
x = (70.53 * 255) / 100
x = 179.85
2 Likes
dont worry about the numbers if it works
on my fergy with linear sensor the numbers are 166 right and 92 left and it works fine
1 Like