Bicycle Power Meter DIY

Making myself a power meter for my road bike

Public Chat
Similar projects worth following

Making a power meter for my bicycle, because the once in store is too expensive for the components and technology behind it. 

Two different version, one for mounting on the inside of the crank and one to mount on the crank arm.

Send me a message if I should proceed with the crank arm mounted version or inside the crank.

Battery side needs to be solder by hand, while the other side comes pre-soldered.

Going to use a NRF52810 for sending over ANT to my Garmin Forerunner 245.

When I have finished the project, I will give out all my documentation. If you follow me on Github I will post some code and files along the way, but I won’t use a lot of time publishing before completed and success.

If you want a PCB card, send me a message and we can find a fair price.

  • Ordered components

    UltraReidar05/25/2021 at 18:44 0 comments

    Have just ordered the PCB boards. Order 10 pcs, so if anyone wants one, we can sort out a price that covers the cost + shipping.

    Also, order strain gauge and holders for cr2032 battery. Decided to not move anything on the PCB if I ever wanted the cr2032 on top of the board. 

    The plan now is to make a small 3D printed housing for the board which will be potted and clued onto the crank arm. The cr2032 will stay in a not potted box beside, for removal. Have not decided how I will get that 100 % waterproof. 

  • Checking CAD model

    UltraReidar05/24/2021 at 18:03 0 comments

    Making a test box for the electronics, to find out how it will look. First, it is very bulky, too bulky for the right crankarm. Therefore the coin-cell battery has to be removed from the top of the PCB, to the side. Since the battery holder does not come pre-soldered on the PCB, the current PCB file are ok (just solder some wire on) but I will probably change that since I haven't ordered it yet. If someone knows about a small battery that fits this project better, please contact me.

  • Price of PCB

    UltraReidar05/12/2021 at 12:51 0 comments

    Checking out the price of the coin cell version. 

    10 pack cost 115 dollar including shipping, but without custom.

    Going to check if I can fit it on both crankarms for more accurate measurements (measure both legs, instead of one and doubling up).

    Order going in in the near future.

  • Finished PCB card

    UltraReidar05/11/2021 at 20:44 0 comments

    Just finished the PCB card and ready for order.

    Two different version, one for inside the crank and one for the crank arm.

    Size are 50 x 10mm (lipo battery) and  40 x 20mm (cr2032 battery)

View all 4 project logs

Enjoy this project?



UltraReidar wrote 05/26/2021 at 19:43 point

First and foremost this is just for fun and learning new things. If this ends up working and with an accuracy of 10%, no problem. As long as it is showing more wattage than reality, compensating for some weak legs. 

Joking aside, I appreciate your comment. 

For the amplifier, it is just a blunder not adding one, but it could be easily sorted by adding it as a separate module for now. I'm working for my university this summer so I would probably find some cool amplifiers there. I will probably find some good adhesive there as well, but if it lasts a month, it lasts a month.

I haven't put much thought into the project yet, so the asymmetric cranks haven't struck me. Watced your video about it, and it does makes a lot of sense. Im using the 6800 crank myself, so could be worse I guess ;). Going to try to sort it as good as possible, but i could still put the strain gauge inside the crank like TeamZwatt, thoughts? The length of the wire would probably become an issue here. As you mention in your video, the software was for simulating the strain gauge was free, i would probably try this out to find the best position. I do have a pretty accurate CAD model of my crank so. 

I thought of having a third strain gauge between the two existing, but that would probably have little flex. But if you have two sets of two strain gauge one set of the upper section and one on the lower section of the crank. Then I belive it is possible to simplefie the measurements so we can linarize it. Or simulate a lot of spots on the crank arm, find a formula for the upper part and one for the lower part and, use that to compensate the measurements. But it is just a thought and it is possible complete garbage. 

Yet again, great input, and great youtube channel, a little goldmine for this project. 

  Are you sure? yes | no

kwakeham wrote 05/25/2021 at 21:11 point

Worked with strain gages and on several commercial powermeter products. Looks like you're missing either a gain stage or an integrated gain + ADC. You can't just run the strain gage inputs to a nordic and hope to get any result. You need gain of 60 - 200 because we're talking microvolts with traditional strain gages.  HX711 is trash garbage useless, avoid it. TI's ADS series is better with clear documentation and easy to get working. AD have better specs, but only on paper and in reality they come second or 3rd, and Maxim's parts except for the newest are poor. The newest are in high demand, hard to get, and have terrible documentation on how to use them.

As for strain gages, geometry of cranks affect readings, pedal offset does too if using the stages / 4iiiii's / pioneer / shimano on inner surface method a lot. It also leads to the pedal offset twist causing wonkiness in readings for offset, and axial sensitivity. How a company tunes their resistance affects thermal stability. But all that is nothing in comparison to avoid shimano's newest Dura-Ace to 105 cranks. If it' looks asymmetric you lost.

Cyanoacrylate is only good for short term installs. Long term there will be issues if you use it. Proper gage bonding adhesives require high temperature cure (Mbond 600 / 610) or are expensive one time use (AE-41) and proper surface prep and cleaning. Cyanoacrylate is okay for short term proof of concept but it weakens with moisture leading to creep problems. Proper Quality controlled stuff is best (Mbond 200). Off the shelve stuff can work but no gel. Gel superglues are not for strain gages. These will just ruin your gluelines and lead to zero's going off.

My advice I've given almost close to 100 people who've emailed me about powermeter design over the last 8 years. It is 100% cheaper to buy 2 or 3 of the most expensive powermeters in the world than building one yourself and getting it to work. So make sure you're doing this fun, not to save a buck.

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates