# Bicycle pedal press power reading

Hi,
I'm a triathlete and a hobby constructor. I would like to build a power meter for my bike because the ones on the market are quiet expensive. The idea is to measure the power the rider puts into pedaling by measuring the tension between the pedal and the crank. They are connected with just one screw, so it is the single point of force transfer.
The question is - would a strain gauge mounted inside the screw do the job? I mean something like this:
http://www.tml.jp/e/product/strain_gauge/gauge_list/btm_list.html
It is meant to measure the tensile strain of the bolt, so how hard the bold is screwed in and how hard it is pulled, stretched - this is how I understand it.
I ask this question here, because I'm not sure how forces are distributed when a cyclist presses a pedal. One would think that if the pedal is screwed in tightly then this type of sensor should do the job. Do you think I'm right?

regards,

## Answers and Replies

Doug Huffman
Gold Member
When I have dreamed of building a power meter (what else does one think of at the end of a long day on the HPV), I would measure the tension in the top chain by deflection of the strain gauge sensitive element. That and angle rate ought to give power, certainly by unit analysis.

My primary design objective is avoiding a wireless connection to a moving sensor.

So where would you put the strain gauge? In/on the chain? I don't understand.
Anyway you haven't answered my question. Am I making some bad assumptions?

Doug Huffman
Gold Member
The pedal bolt tension does not vary through a revolution. It only resists the loosening of the tapered joint.

On information and belief, commercial power meters measure the deformation of the circumference of the female socket part of the joint, averaging eight sensitive elements.

Are you sure about the tension staying the same in the bolt? I drew this little picture. Doesn't it work this way that there is a pivot point which causes the tension in the bolt to increase?
http://prod.otn.pl/DSC_0213.JPG [Broken]
Thinking straight - if it was constant then any bolt would withstand after being screwed in. Pressing hard on the pedal wouldn't change a thing. So you could easily change the middle part of the bolt (the one between the "bolt" sign and the washer) to be made of anything. Let's say it's rubber. I can easily imagine it stretches. Wouldn't it? If so, then putting a strain sensor in this part should do the job, no?
Tell me where I'm making a mistake please and explain to me how it really is.

Last edited by a moderator:
CWatters
Homework Helper
Gold Member
I suspect the problem is one of calibration. If you take the pedal off and replace it you would have to pretension the bolt to exactly the same value and/or repeat the calibration process. Would the bolt also stretch over time (I think they call it "creep"?). Changes in temperature would also affect the tension but that could be compensated for.

CWatters
Homework Helper
Gold Member
Doug Huffman
Gold Member
There is no source of axial force opposing clamping tension in a crank bolt. Creep is a temperature dependent phenomena practically moot at STP and these practical dimensions. This is bicycle mechanics, not rocket science.

The phenomena of a loose crank is entirely due to poor practices, perhaps most frequently too much clamping force splitting the eye at the stress riser corners. Compare machine tool tapered joints, they don't necessarily have flats even.

I was in error above when I mentioned pedal bolt, it should have been crank bolt. Pedals depend n mechanical precession and should not be tightened beyond finger tight.

Last edited:
Thank you for the link, CWatters!
Do you agree with Doug on my idea?
If calibration is the only issue here then I can live with that - I'm doing it for myself anyway.

Ps: another idea just came to my mind - what If read the power even closer to the rider - I could put pressure sensors (mats) underneath shoe inserts. What do you think?

CWatters
Homework Helper
Gold Member
I think the bolt might work but I'd test it out before putting a lot of effort into it.

CWatters
Homework Helper
Gold Member
There are some very small and cheap Bluetooth audo modules available (intended for things like Bluetooth microphones). I wondered if possible to AC excite a strain gauge bridge at low audio frequencies and use Bluetooth module to transmit that to a smart phone. The hardware for that would be pretty small. Not sure about the software (I'm no programmer).

Unfortunately the bolt gauges are expensive and I don't have access to test them. Your BT idea should work, however I wanted something to work with my garmin.

Last edited:
Would be great I someone else could say something. At the moment the risk of buying sensors just to check if it works seems to high. The ones I proposed at the beginning of this thread cost about $250 for a batch of 10 (can't buy a single piece) which is OK if you know it will work. Or maybe, Doug, you could try to convince me you're right? For now you haven't provided any explanation - only statements. Or maybe you were thinking about crank bolts all the time and now you can tell me something else about my idea? Thanks. Hi, I'm a triathlete and a hobby constructor. I would like to build a power meter for my bike because the ones on the market are quiet expensive. The idea is to measure the power the rider puts into pedaling by measuring the tension between the pedal and the crank. They are connected with just one screw, so it is the single point of force transfer. The question is - would a strain gauge mounted inside the screw do the job? I mean something like this: http://www.tml.jp/e/product/strain_gauge/gauge_list/btm_list.html It is meant to measure the tensile strain of the bolt, so how hard the bold is screwed in and how hard it is pulled, stretched - this is how I understand it. I ask this question here, because I'm not sure how forces are distributed when a cyclist presses a pedal. One would think that if the pedal is screwed in tightly then this type of sensor should do the job. Do you think I'm right? regards, If you're just looking for how much power you're producing when you're pedaling, may I offer an alternative solution? If you have an odometer that tracks calories while you're biking, go out for a bike ride for a set amount of time, say 10 minutes, somewhere you can maintain a constant speed (ie won't have to stop for traffic). Then convert the number of (dietary, since for nutrition purposes we're actually using kilocalories) calories you use to joules by multiplying by 4184. Then divide the number you get from that by the duration of your ride in seconds, and you'll get your average power output in watts (hence why I said somewhere you can maintain a constant speed). I think there are some bike odometers that even have this function built-in. Hehe jack, this would be tad inaccurate :) A much better solution to calculate power without actually measuring it would be to measure wind/elevation change/acceleration and then use that data to calculate power. But even that is still very inaccurate, does not take many things into account, like road resistance, tire pressure, etc. I really need to measure power directly. But for that I need to confirm my theory. Hehe jack, this would be tad inaccurate :) A much better solution to calculate power without actually measuring it would be to measure wind/elevation change/acceleration and then use that data to calculate power. But even that is still very inaccurate, does not take many things into account, like road resistance, tire pressure, etc. I really need to measure power directly. But for that I need to confirm my theory. Oh, I figured this was for training purposes where a rougher measurement would be acceptable. In that case, I wish I could be more help, sorry. CWatters Science Advisor Homework Helper Gold Member Yikes. Ordinary strain gauges are more like$10.

When you press hard on the pedals the frame of a bike tends to twist. I wonder if you could put a regular strain gauge on the frame somewhere and use that to indirectly measure to pressure on the pedals? Would also avoid the need for a wireless link. Does the frame twist significantly for other reasons that would cause a lot of noise?

That's too far from the power source I'm afraid (too much noise) plus it would require me to damage my frame! I don't even want to grind my crank! It's all expensive stuff...
I've read somewhere that people make their own strain sensing bolts. But I don't know how to do it ... yet :)

Doug Huffman
Gold Member
[ ... ]Or maybe, Doug, you could try to convince me you're right? For now you haven't provided any explanation - only statements.[ ... ]
Yep, positive statements, that can be falsified or verified at your pleasure, and no ad-hockery from me.

CWatters
Homework Helper
Gold Member
I've read somewhere that people make their own strain sensing bolts. But I don't know how to do it ... yet :)

See...
http://www.morevision.co.uk/mvweb/test.htm

After disappointing performance with commercially available bolt sensors MoreVision devised and produced special test bolts by introducing flats on the shank of the bolts and attaching standard linear gauges (photo on the left).

Looks promising - thank you! :)

The downside is... if the pull (strain) on the bolt is not equal across it and it probably isn't, then putting sensors on sides will produce "varying" results because the force to the pedal is applied at varying angles between the pedal and the crank (the bolt) as the pedal rotates around the bolt AFAIK. The sensor in the middle was the only solution I'm afraid.
PS: now I'm starting to wonder if this is going to work at all. If there is no thread on the bolt in the part between "bolt sign" and the washer (picture) (i.e. if the bolt is not screwed in to the crank) then it might work, but I'm not sure anymore... Will have to check later at home.

Last edited:
CWatters
Homework Helper
Gold Member
Had another another idea. See diagram.

Suppose you drilled a hole right through the length of the bolt and put a wire pin through it. Then weld the tip of the wire to the bolt at the thread end only. When the bolt is tensioned/released the pin should move out/in of the head end very slightly. If a thin (spring?) steel strip was fixed across the head as shown the pin should bend it slightly and perhaps that could be measured with a regular strain gauge.

The bolt would need to be put under maximum tension when the pin is cut/ground down to length and when the steel strip is fixed to the head. When the tension is released the pin would deflect the steel strip. Increasing strain on the bolt would reduce strain in the steel strip.

I think this approach might even compensate for temperature effects.

I suppose you might even get away without the steel strip, just fix the strain gauge to the head with silicon sealer or something similarly flexible?

I like this idea - clever thinking :)
Unfortunately the pedal is constructed differently. I was wrong about it from the beginning :( The pedal itself is threaded and is screwed in to the crank.
Still thinking about something different, but now the idea from the tutorial you pasted seems really the best solution.