HobbyKing.com New Products Flash Sale
Reply
Thread Tools
Old Aug 27, 2014, 09:52 AM
Registered User
jhaywood's Avatar
United States, VA, Suffolk
Joined Nov 2007
1,072 Posts
Discussion
Servo position

I am testing some servos for accuracy. I already have a controller that I can use to send the servo to an exact spot. The problem is that the servo does not come back to center everytime. Right now, I have a 4" pointer attached to the arm and just mark the spot at center when powered up.

I would like to electronically measure the stopping point. Could I hook at pot to the servo output and then read the voltage that the pot outputs, or something like this??
jhaywood is online now Find More Posts by jhaywood
Reply With Quote
Sign up now
to remove ads between posts
Old Aug 28, 2014, 02:49 AM
Registered User
Scottish Borders
Joined Mar 2006
578 Posts
The resistance of the potentiometer could be measured directly with an ohmmeter.
With a stabilised voltage source applied to the potentiometer the centre voltage could be measured.
Could you tell us more of your test set-up and your results so far and the servos you are testing?

Dave
orraman is offline Find More Posts by orraman
Reply With Quote
Old Aug 29, 2014, 10:40 AM
Registered User
jhaywood's Avatar
United States, VA, Suffolk
Joined Nov 2007
1,072 Posts
I am using the Pololu Maestro Servo controller. I program it to go to center on start up, move one way, back to center, then the other way, back to center, then a full sweep, and return to center. I let it run for 2-3 minutes to make sure temps have stabilized inside the servo. I am using a 5v power supply in an effort to keep that the same. I need to find an adjustable power supply so that servos can be tested at their designed voltage...standard vs HV, etc.

With one brand of servos that I have tested(2 high end, one standard) the high end servos returned to exactly the same spot everytime as indicated by a 4" needle like pointer. The standard servo would vary up 1/16 of an inch at center. The high end servos are coreless, digital and metal gear while the standard is a cored, 3 pole, plastic geared servo. One of the high end servos is OLD, and been beat around, while the other is brand new.

Another brand's high end mini was only marginally better than the standard servo mentioned above. It is a brand new coreless, digital, metal gear servo. This was quite a shock actually..


The first brand is double the price, but so far it seems that you do get what you pay for.

I am not mentioning brands until I have true qualitative data to hand out, if then. I am not here to cause arguments and Ford vs Chevy debates, just would like to find out the true precision of these so called super precision high dollar servos.

I am working on using a 5" degree wheel bolted to the output horn now with a stationary pointer. I am trying to keep it light so that inertia has little to no effect on stopping points. At least that will give a number and you can see in any videos that I make.
I would prefer a better way to measure the position of the output arm in relation to the commanded position, but I am not an electronics guy by any stretch....
I am open to suggestions!
jhaywood is online now Find More Posts by jhaywood
Reply With Quote
Old Aug 30, 2014, 02:09 PM
Registered User
Scottish Borders
Joined Mar 2006
578 Posts
Thank you for your detailed reply.

The specification of your servo controller and the centring accuracy of the better servos indicates that your demand position is reliable.
First thought would be to arrange the outer end of your pointer to be a vertical blade over a mirror scale to improve your reading accuracy as in a quality analogue meter.
A potentiometer on the servo output would allow an automatic recording and averaging over a number of cycles but only within the accuracy of the potentiometer and the coupling.

However your requirement to reduce weight so that inertia has little to no effect on stopping points surprises me.
I believe that while poor centring of unloaded servos gives a superficial indication of quality I consider that the real test should be for centring under increasing loads within the manufacturers specifications.

You mention the possibility of producing videos of your testing and I would regard this as more convincing than numerical statistics.
A related subject is the failure of sail servos to return to centre and meet the manufacturers specifications under load in practical tests.
The author of the link below has done considerable practical testing and the video at the foot of the page shows manufacturers unreliable expectations.

Dave

http://www.sailservo.co.uk/techinfo.html
orraman is offline Find More Posts by orraman
Reply With Quote
Old Aug 31, 2014, 01:24 AM
Registered User
EloyM's Avatar
SantaAna CA
Joined Aug 2003
316 Posts
I don't remember where I got it, but I have a 12 inch wide protractor like drawing that is made
for servo testing. It includes a servo location and uses a 5" arm.
Mine is mounted on a piece of 1/4 : plastic with short feet to provide clearance for the servo which protrudes from the bottom. Well marked in degrees.
I have an extra I can have copied & shipped in a tube. More if you are in Canada or Argentina..... Direct please: eloymz@att.net..... (in CA)
EloyM is offline Find More Posts by EloyM
Reply With Quote
Old Sep 02, 2014, 11:46 AM
Registered User
jhaywood's Avatar
United States, VA, Suffolk
Joined Nov 2007
1,072 Posts
Dave, I never thought about the measuring pot having built in imprecision. That could lead to tolerances ganging up and making things look worse than they are. It would be the same imprecision for all servos tested, but not sure I want to have to explain that portion.

The reason that I am trying to lower intertia is simply because I am trying to test the servo in its raw form. From that, maybe building a simulated control surface could happen. That could also lead to better than expected results on the bench. The weight of the control surface could cause the servo to return to the same spot, which will be against the the next tooth in the gear set. Gravity will cause the slop to be "eliminated" in a static environment. I guess I could build a vertical surface, ie, a rudder type assembly and that will eliminate gravity's role in this. Then you have the question about hinge friction, ball link friction, etc.
That is why my first theory is to make a light degree wheel or pointer that will not cause the servo to overrun its commanded position.
It could be that the imprecision may not be felt in the air due to friction, gravity, turbulence, the fact that we are almost always moving a control surface while flying, etc...
That all being said, maybe I will add weight to the degree wheel, or even maybe a friction pad to make the servo "work" a bit more and check the results.....opinions???


I never bought into the high dollar coreless stuff until 20 years ago, I swapped in a set of good servos for a set of Futaba 148s. As JR used to say, "Feel the Difference". The plane flew MUCH better.

Eloy, do you have a picture of your setup?
jhaywood is online now Find More Posts by jhaywood
Reply With Quote
Old Sep 03, 2014, 03:26 AM
We want... Information!
Bruce Abbott's Avatar
Hastings, New Zealand
Joined Jan 2001
5,185 Posts
For frictionless angle measurement with negligible inertia, attach a small magnet to the center of the servo horn and use a rotary Hall effect sensor to measure its orientation.

The MLX90316 has a resolution of 0.08 degrees. Here's an experiment showing how to read it with an Arduino:-

http://interface.khm.de/index.php/la...nsor-mlx90316/
Bruce Abbott is offline Find More Posts by Bruce Abbott
Reply With Quote
Old Sep 03, 2014, 02:26 PM
Registered User
Joined Sep 2003
334 Posts
Mount the servo on its side so the servo arm travels up/down. Hang a load weight on the arm if you want.

To the servo arm attach a laser pen.

Attach a ruler or tape measure to a distant wall.

The deflection error angle =(new spot height"-old spot height")/ (servo shaft distance to the ruler").
Then push the calculator 2nd function key then Atn or Tan-1 key.

For small deflections this method should be very accurate. For large deflections the distance from servo shaft to ruler changes . Use a^2+b^2=c^ to find the new servo shaft distance to ruler, then proceed.
Plantoflap is offline Find More Posts by Plantoflap
Last edited by Plantoflap; Sep 03, 2014 at 02:41 PM. Reason: Addition
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion Hobby King servo HKS-9650 - center position moves 45 deg w/temperature increase cphdrider Fuel Plane Talk 5 Mar 10, 2014 02:41 PM
Discussion Servo arm out of position=4in1 or servo failure?? misfire Coaxial Helicopters 5 Sep 27, 2007 07:16 AM