Calibrating digital 5645 vs analog 645 servos at full range

Hi,

I’m not sure if this topic has been posted before but I’m a little puzzled when it comes to using digital servos at full range (180 deg). What I recall when using digital servos on Felix was that I had to recalculate how many pwm steps there was per deg compared to what I got from using analog servos.

As long as the digital servo (5645) are set to factor default (reset using the HFP-10 programmer) they has the same resolution (steps/deg) as the analog has. But after setting the EPA (end points) and neutral the steps/deg ratio changes.

Some result:

Standard analog 645:
pwm values vs. deg: 573, 1500, 2372 => -90, 0, +90 deg
Steps/deg = 0,10 (rounded)

Default 5645:
pwm values vs. deg: 750, 1500, 2220 => -77, 0, +77 deg
Steps/deg = 0,10 (rounded)

After setting the EPA to +/- 90 deg range on a 5645:
Center = +12
Left set = -180
Right set = +210
pwm values vs deg: 897, 1500, 2097 => -90, 0, +90 deg
Steps/deg = 0,15 (rounded)

This isn’t a really big problem, you just need to be aware of it when ex. using the Phoenix code and use another ratio for getting the correct servo position.

What I’m wondering about is if there is a way to program the 5645’s to full 180 deg range and still keep the the same Steps/Deg ratio as you have on a default servo or any analog servo?

good to know
i just notised the servo didnt go 500 - 2500us as i expected it to go … but all the same, as long as it goes 180 deg i am happy, after all its a quite simple and unexpencive ($45) servo whith surprisingly good speed and power :laughing:

Your testing perfectly matches what I have seen. I’m not aware of any way around it. :frowning:

Does this mean the “StepsPerDegree = 200” are “StepsPerDegree = 133” on the 5645 ? (in the phoenix-project)
… as i get 90 degree movment on HSERVO going from 0 to 12000

That sounds right, but I haven’t done much personally with the 5645’s so I’m no expert.

I can’t confirm this because I’ve not used full range 5645 on the ARC-32 yet. But your calculations seem correct.