String Pot: Good, Bad or Mixed Bag?

[Joe Johnson]

Well it is that time of year again.... ...Time when the electrical/programming group finally have time to think about "how are we going to control this beast?"

It is best to think about these things before hand but, if you did not, now what do you do?

Another thread just reminded me about folks last year raving about string pots and how much they liked them on FIRST robots (Zondag, was it you?). I have no person experience whatever but they look pretty sexy. The folks I know that work in Delphi test labs swear by them.

Here is some data.

Picture:



Website with string pots on them.


More pictures. If you look at the pictures it seems like they come with enoder (digital) outputs as well as analog (potentiometers)

Does anyone have experience they wish to share?

Do Tell.

Joe J.

[Andy A.]

95 had wanted to use a string pot last year. The application was simple enough, keep track of how much of a telescoping arm had extended. The idea was that it made more sense to measure the distance the arm extended rather then measure the number of turns a gear made.

In the end, we couldn't find a string pot that met any of our criteria for size, output, price or availability. They just didn't seem to exist. Of course, the folks looking for them consisted of me and a mechanical engineer, neither of us qualified to know where to look or exactly what to look for. It seemed like they should be listed right near all the other kinds of pots with the same options for specifications. We considered making a string pot out of a pot string and such, but that seemed a little foolish.

They are just pots with a spring loaded string right? Why are they so expensive and difficult to find?

Maybe AndyMark should whip up some little adapter that can be stuck on a normal pot to turn it into a string pot.

-Andy A.

[kaszeta]

Originally Posted by Andy A.

95 had wanted to use a string pot last year.

I've used plenty of string pots at work with very good results. Only reason I haven't used them on an FRC bot is that every string pot I've used that I like was way too expensive to use in competition ($200, usually).

[Alan Anderson]

Originally Posted by Andy A.

...We considered making a string pot out of a pot string and such, but that seemed a little foolish...
...Maybe AndyMark should whip up some little adapter that can be stuck on a normal pot to turn it into a string pot.

If it's foolish for you to do it, why wouldn't it be just as foolish for AndyMark?

Go ahead and build one yourself if you want to use one. It'll be fun making it, you'll have something worthwhile to use, and you can publish a white paper on it when you're done.

[Chriszuma]

Our engineer built a couple of string pots last year consiting of a circular spring (what are those called?) that was unwound as the string. In the end it was too unreliable, so we switched over to digital encoders.

[Kris Verdeyen]

We used one last year on Chainzilla's elevator. They're very easy to integrate, it's just important to keep them from retracting too fast, or they'll fail. We also had at least one match where the pot caught on an opponent's robot, and was damaged. All in all, though, we were satisfied with it - we're using a shorter one on this year's robot.

I'll post the part numbers when I find them. IIRC, they go for a little less than $100.

[MattB703]

I had a dream about a low cost string pot last year. I didn't do anything with it, maybe somebody else will want to. I wanted to use a string pot for our lift mechanism, but I was running into all of the problems already mentioned, (cost, size, etc), then I looked at the little retractor thing that my work ID is hooked to. Wouldn't it be neat if someone would drill a 1/4" hole in the plastic case of one of these and glue the shaft of a pot or encoder to the inside part of this little retractor thing?

Hmmm

I'm sure its not as easy as that, but it is interesting.

Matt B.

[Kris Verdeyen]

McMaster part numbers:
10 inches 6440K1 $167.16
25 inches 6440K2 $167.16
50 inches 6440K3 $167.16

Last year, we used the 50" one, and this year, the 10".

[TubaMorg]

I am reviving this thread as we have a question about the string pots.
We are looking at these:
http://www.celesco.com/_datasheets/sp1.pdf

So the question we have is that these things are 10K ohms, but we should be using a 100k ohm potentiometer, correct? So if we just add in another 10k ohm resistor, won't that work?

[Alan Anderson]

The 100k requirement is only for variable resistors connected to a joystick port on the Operator Interface. On the robot, 10k is fine. As long as the resistance isn't so small that it overloads the RC's +5v supply, or is so large that EMI (motor noise) becomes a major factor, any potentiometer should work.

[TubaMorg]

Great, thanks for the clarification Alan!

What about this from the Guidelines manual?

G.6.4 Sensors
In general, the robot controller anticipates receiving inputs from sensors either in the form of contact closures (digital) such as those from a gear tooth sensor, or as 0-5V voltage levels (analog) at its input ports. In the case of analog signals, the robot controller supplies 5Vdc at 1 Amp for the analog inputs. As there are 7 input possibilities, each port should be limited to approximately 50ma. The input types, a combination of potentiometers, accelerometers, and gyro (yaw) rate sensors, should present high impedance to the analog input. All potentiometers must be 100KΩ.

Ok the last edit to this post. Just to confirm what Alan said (was there ever any doubt?!?), here is what the RC manual says about potentiometer inputs to the analog ports of the RC:

The analog inputs on the Robot Controller can be used to measure various conditions on the robot and
trigger automatic responses by the control program. There are 16 analog inputs available on the
ANALOG INPUTS connectors on the Full-Size Robot Controller.
Any sensor which outputs a variable 0-5V signal may be read with 10-bit resolution on these inputs.
Popular devices include potentiometers and gyro (yaw rate) sensors. Potentiometers should be wired per
the diagram below when connecting to the Analog Inputs. Use 250Ω – 100KΩ potentiometers. For
wiring yaw rate sensors, refer to the sensor manufacturer’s data sheet.

[Chipawa]

Many years ago, I used string pots for measuring linear motion of components on Navy ships. We called them "yo-yo pots." I can't speak for cost or availability, but the ones we used we're quite reliable.

One thing I would recommend is to check their linearity. After installation, we would monitor the pot's output at various points along the extension and calibrate our readings. Depending on how exact of a measurement you need, I recommend teams calibrate their yo-yos.

Chip Tucker
Team 1111

[TubaMorg]

Hey another Navy guy! I was in subs out of Pearl Harbor. Nuke mechanic!

According to the spec sheet for the yo-yos we are looking at they are very linear and the interior mechanism ensures the spin diameter remains constant. That being said, it definitely seems like a good idea to check them though. Has to be done anyhow to calibrate the conversion from output to distance.