Tension arm with pot

Hi everyone.
I was recommended this site from the Arduino discord as I need a little advice.
Was told that I need something called a dancer tensioner for my 8mm scanner project. its to keep my takeup reel at the correct speed regardless of reel diameter. the hardware i can work out. but im unsure about the code and the electronics. I was hoping someone could explain electronicky how this dancer tensioner works. im guess i feed a voltage into the pot and if it varies either + or - adjust the speed of the takeup reel. Sorry for my bad spelling.

Best regards


@Creative_Samurai - welcome to the forum!

Concerning your question: there are quite a few options you might want to look into. Have a search for the keyword “tension” here in the forum - you will see that this is not an easy thing to achieve, and there are many ideas floating around. Using a potentiometer as tension measurement device is only one of the options. If you look at an old setup of mine,


you see that I am actually using tension arms with potentiometers.

There are two tension arms visible in the photo above which measure the tension left and right to the film gate. Here’s an old image from a test setup where you can see more details of the construction:

Let me tell you right from the start: this is quite a challenging approach. For one thing, the angle resolution of the potentiometer + tension arm setup, if they are connected directly to a DAC, is very low. I did not bother to design a conditioning circuit with opamps and I paid the price later with developing a lot of weird software routines. You need to make sure that the voltage applied to the potentiometer is absolutely constant. Otherwise, you will get a sensor signal which is very in low amplitude and very noisy.

The basic principle is that the tension in the film path changes the angle of the arm on which a roller is mounted. This is because a small spring wants to push the arm in one end position and the tension from the film pushes against that spring in the opposite direction. So a certain film tension corrresponds to a certain angle of the arm, which is measured by the potentiometer. That is because the axis of the potentiometer is actually the turning center of the arm.

So, a certain angle corresponds to a certain tension which corresponds to a certain voltage reading from your potentiometer. So you have realized a tension sensor.

You can pair this sensor with a PID-controller to control the windup motors. The PID controller gets a set point and will move the windup motors in such a way that the measured value stays always close to the set point.

That’s about all with respect to the functionality of a potentiometer-based tension system. But do not underestimate the challenges lurking in the background. If you want to use only the small 15m Super-8 rolls, the PID parameters might be easy to determine experimentally. If you however want to handle larger rolls of Super-8 film (as in the image above), it will be a challenge to pick the right motors and PID-constants to make this work.

Some people substitute the potentiometers with micro-switches directly connected to the windup motors. That is actually an approach which is simple and works fine, provided you stay with low scanning speeds. Many other people do not create their own film path, but simply reuse an old Super-8 projector for scanning. While the mechanical stress on your precious film stock is much higher than it potentially will be with your own design, you will get much faster nice results when reusing the mechanics of an old projector.

Somewhere on this forum, there are servo motors described which supply a constant tension all by themselves. Depends on your budget whether this is an option for you.

1 Like

@Creative_Samurai welcome to the forum.

@cpixip made excellent points from someone who used this method.

I have the electronics knowhow, but have not used a tension arm on my project. I will add to the above that potentiometers come in two forms, log and linear. If you decide to experiment, consider using a linear one, otherwise the variations at certain parts of the scale will be minimal.

Thank you all for your input. Iv ordered the correct pot from amazon and im currently testing with a very old coil wound 12k pot. I have a guy helping me with the math and has suggested using a pid library to help with error correction. If i have no luck with this. What other ways are there other than a friction clutch? Cos im running out of ideas lol

…a coil potentiometer only has very coarse discrete steps. The resolution is way too low, there is no way something like this would work reliably as a tension sensor. Also, in my post above, I gave you a few tips on alternative approaches that have been discussed/presented here on the forum. Please consider again using the forum’s search function - the forum members have amassed a wealth of information over the years which are here for you just to read and study …


@Creative_Samurai, I have implemented a dancing-arm test. You can see the context of the machine and the video of the arm test in this post.

@cpixip based on your posting on this thread, when I prototyped the dancing-arm I made it as short as I could, which it would translate into a larger angle, also a larger voltage change with tension. In regards to conditioning, I connected it directly to the ADC and a couple of small capacitors quiet the reading nicely.

1 Like