Calabrating the servos

[rust710]

I have a servo that will not go to it's full range. ie 0 should be all the way to the left but it will only go to a forth of the way there from center but it goes all the way on the other side. Is there a way to calibrate a servo?

[Greg Ross]

I don't think there is a way to "callibrate" a servo. The best you can do is to unscrew the little "propeller", and center the blade you're interested in. (Or get it positioned wherever you want it for a particular output value.)

Keep in mind that the kit servos do not rotate a full 360°. I don't know, off hand, how far they do rotate -- maybe 210° - 300°?

[D.J. Fluck]

Ive done a little research, and according to the FIRST 2002 handbook, one of these 2 servos are used in the kit of parts. The parts list didnt give me enough info to figure out which one it is, because the specs on the two are exactly the same. It looks like the deluxe can handle a little more stress, and weighs 2 oz lighter.



http://www.hitecrcd.com/Servos/hs311.htm

http://www.hitecrcd.com/Servos/hs322hd.htm


I couldn't get the pdf's to work but maybe its just me

[rust710]

Ok, but my servo won't go the full possible range it has. ie I can turn it by hand about 180 degrees but it won't go that far when I set it to it's max through software.

[D.J. Fluck]

Then you might have a bad servo...

[rust710]

Ok, I will check it against a good servo. Thanks for you help.

[Lloyd Burns]

Your troubles may be due to the servo not wanting go more than 180 degrees, which might compromise a model aircraft's turn.

If you take the top off (carefully), you can see the stop on one of the gears; note how far it can go.

You say that the motion is not symmetric. This can be calibrated out with the Victors; they are quite complicated and thus often don't need recalibrating.

The servos are simple devices that must be trimmed at the control's (joystick trim tabs, eg) input; this alters the signal sent to the servo, to fit the servo's character.

If you can't trim it out with the trimmers, you can program it out. (There are at least two ways to do this.)

[kmcclary]

Originally posted by rust710
I have a servo that will not go to it's full range. ie 0 should be all the way to the left but it will only go to a forth of the way there from center but it goes all the way on the other side. Is there a way to calibrate a servo? You can't calibrate the old fashioned PWM style RC servo. The very expensive "DIGITAL" servos you can calibrate, but they're still fairly new devices.

To work, standard PWM RC servos require a 5V positive going pulse, with a pulse width of roughly 1.0ms - 2.0ms, repeated at about 50 Hz. The pulsewidth encodes the "target" position. 1.0ms = full one way, 1.5ms = the "center" position, and 2.0ms = full the other way. The circuit board inside the servo compares the incoming pulse against one it generates itself based on its feedback pot. If they're different, the difference is stretched out and drives the motor the right way, creating a little "jerk" toward the target position. That's why you need to keep refreshing the servo with pulses, to "walk it" to where you wish it to be, and KEEP it in position against any backdrive forces.

There's nothing to calibrate in this system. Electrical parts COULD go bad, but that's rare. Most of the time either your mechanics fail, the driver transistors burn out, or the pot gets munged up, but I've yet to see the timing circuit screw up from a bad capacitor. The timer is a single IC.

The pot inside the servo CAN go bad, which COULD cause your symptoms, but it is MORE likely a problem with your transmitter CONTROL'S calibration, and not the servo itself.

Try this: Disconnect anything attached to your servo (to eliminate any "mechanical stops" from the thing you're driving). Now if your servo doesn't run the full travel, check your transmitter. They often have trim pots either beside the sticks, or under a cover somewhere (often on the back, or under a label plate) for setting a servo's "travel range" and "offset".

If you wish to test beyond that, here's Ahmet Onat's page on how servos work, including a cute little 555 based tester box you can build to check out servos:
- http://www.geocities.com/BourbonStreet/3220/servobasics.html


BTW... BE CAREFUL with Onat's wiring of the servo! The RED is the +5v, the BLACK is the Ground, and the OTHER COLOR (typically white or orange) is the PWM Signal lead that connects to the transistor, as is shown in the schematic.

However, in OUR servos, the white is on one EDGE, and is NOT the middle conductor! The wire ORDER for OUR servos is normally Orange/White(Signal), THEN Red(+5v), THEN Black(Ground), which is DIFFERENT than that schematic!

For more information, see this thread, called "Pwm":
http://www.chiefdelphi.com/forums/showthread.php?threadid=14565&highlight=servo


Good luck, and Happy Holidays!

- Keith

[TF8]

mechanicly servos only rotate ~180 degrees before they hit the stops internally. electrically the feedback loop on the servo will only allow the servo to travel ~120 degrees. Altered servo travel and continuous rotation are possible but require hacking the pot and internals which may not be first legal.

Tyler

[Adam Y.]

Technically you could glue the gut the stops and glue the pot in the center position.:) That's all you really need.
Step 1
Remove case of servo to expose the gear train, motor, and potentiometer. This is accomplished by removing the screws on the back of the servo case and separating the top and bottom.
Step 2
File or cut off the nub on the underside of the outpout gear that prevents full rotation. This may mean removing some of the gears so remeber how you put them in.
Step 3
Remove the pot (or find the center position on the potand glue it there) and replace it with two 2.7K ohm 1 percent tolerance resistors. This fools the servo to make it think it's at center.
Step 4
Ressemble.
There generic hacking directions. Of course who knows if it's legal or not to hack a servo.