### Author Topic: DIY cycling powermeter using Arduino and NRF24L01  (Read 5638 times)

0 Members and 1 Guest are viewing this topic.

#### stoppi

• Contributor
• Posts: 8
• Country:
##### DIY cycling powermeter using Arduino and NRF24L01
« on: November 20, 2016, 08:50:45 am »
Hello!

In one of my Projects I realised a simple and cheap cycling powermeter using an arduino, two NRF24L01 for communication and strain gauges from ebay for measuring the bending of the crankarm.

The four strain gauges arranged in a H-Bridge. The Signal is then amplified 330 times by a LF353. With this Settings I get about 1 Volt Output for Forces around 300 Newton.

To calculate the velocity of the pedal I use the gyroscope within the MPU6050. Within each circulation I add up the Forces and calculate the average. With this force-average and the speed of the pedal I know the Power P = F * v.

With the NRF24L01 the datas (power and rounds per Minute) are being sent to the Receiver fixed at the bar.

The results seem to be reasonable and this for less than 30 USD    I hope you like it

my YouTube channel: https://www.youtube.com/user/stopperl16/videos

The following users thanked this post: dSquare

#### casinada

• Frequent Contributor
• Posts: 600
• Country:
##### Re: DIY cycling powermeter using Arduino and NRF24L01
« Reply #1 on: November 20, 2016, 05:52:37 pm »
Very cool project. Use a piece of bicycle tube to cover the crank with the electronics.
How do you calibrate the power? it seems to be too high.
Can you add power to the other crank so you can tell the power difference between legs?
Keep us informed on your progress

#### Kleinstein

• Super Contributor
• Posts: 7653
• Country:
##### Re: DIY cycling powermeter using Arduino and NRF24L01
« Reply #2 on: November 20, 2016, 07:23:49 pm »
For the average force one should keep the sign. So for some time the force can be negative. From the shown curve this is not clear. An offset in the force signal could cause quite some error in the calculated power.

The amplifier LF353 has quite some drift - this could be a problem outside. There are better OPs available. One might even consider zero drift OPs, like AD8551 or MCP6V... .

Calibration is likely a force calibration, e.g. with a weight. Counting revolutions is pretty accurate.

#### Brutte

• Frequent Contributor
• Posts: 614
##### Re: DIY cycling powermeter using Arduino and NRF24L01
« Reply #3 on: November 24, 2016, 10:42:53 pm »
The four strain gauges arranged in a H-Bridge.

You have glued the gauges in the ~45% and ~55% of the beam and not where stress from bending is the highest, why is that so?

#### stoppi

• Contributor
• Posts: 8
• Country:
##### Re: DIY cycling powermeter using Arduino and NRF24L01
« Reply #4 on: November 26, 2016, 03:03:32 pm »
Quote
For the average force one should keep the sign. So for some time the force can be negative.

This is only when you're Spinning in the oposite direction and then you can't push with a lot of power because of the freewheel.

Quote
You have glued the gauges in the ~45% and ~55% of the beam and not where stress from bending is the highest, why is that so?

I thought that the bending would be quite uniformly and therefore I've glued them approximately in the middle of the crankarm

#### Brutte

• Frequent Contributor
• Posts: 614
##### Re: DIY cycling powermeter using Arduino and NRF24L01
« Reply #5 on: November 27, 2016, 04:01:19 pm »
Quote
You have glued the gauges in the ~45% and ~55% of the beam and not where stress from bending is the highest, why is that so?

I thought that the bending would be quite uniformly and therefore I've glued them approximately in the middle of the crankarm
I do not want to be picky but that is a compound stress state.
The beam is subjected to bending and at the same time to shearing (that is "beams 101"). And because the foot is not pushing in the plane of the rotation, there is also twisting.

Are you sure that in your arrangement:
Uout=0.98*bending+0.01*twisting+0.01*shearing
?

Did you do any calculations or was it "lets stick it in the middle and see the eevblog comments"?

And indeed in case of "pure bending 101" the stress in a beam of constant cross section the strain is uniform along the beam and the center is as good as any other place for sticking the gauge. But that is not in your case.

EDIT: Any thoughts on why this wrench does not have a "uniform" cross section:
wrench ?
« Last Edit: November 30, 2016, 09:29:09 am by Brutte »

Smf