Phew. It’s taken a bit, but in theory – and in individual component measurement, I am down to just about 5ma of current. On a CR2450 battery this should be almost 70 - 100 hours depending. It’s slightly out of spec for current consumption but it’s close to acceptable.
How does it break down?
The nRF51 is maxing out at 2ma. I’m hoping to get this a bit lower, but I’m not 100% sure it’s easily possible. After realizing I was drawing 18ma to the protoboard with the ADS1247, I realized that I wasn’t just powering the ADS1247, I was actually powering a nRF24AP2 Module. After sorting it all out, I measured the ADS1248 which is what I’ll be using and it fell in around 0.62ma, rounded up to 0.65ma.
Putting a resistor in series such that the voltage drop is around 0.5v across the Wheatstone bridge which gives a current of 1.21ma measured. This was about 20 counts of noise on the 16bit, until I placed a capacitor across the bridge (the Voltage input, not the output) and it stabilized it back to the 0/1 counts I was experiencing already. I’m hoping I can push this even lower, possibly 0.3 - 0.5ma per bridge. That would get me down to 3.5ma. This is seemingly safe coin cell operation area for the CR2450 and I think reasonable.
So, there hasn’t been much mention of the gyro. It’s likely out for now. Sorry. I’m making provisions for future testing. Essentially I’m putting in the ability to add a recharge circuit, rechargeable lithium battery input, and the gyro in the circuit board. However, they won’t be included in initial prototypes.
There are gains to accuracy to be had by using the gyro, but in terms of cost and development time, I am making the executive decision to aim for coin cell operation and thus a magnetic switch. 1.0% to 2.0% accuracy is achievable over the pedal stroke, it’s just when you get into high rates (every 45 or 90 degrees for instance) that there would be loose in accuracy.
I’m after low cost implementation with some advanced feature sets. In order to get to the minimum viable product I’ve had to make this decision. Otherwise it’d be September / October before I could have prototypes ready with a gyro.
In other news, I’ve been informed that the programmers that I wish to include with the beta’s are currently having circuit boards built. This is good news as it’ll only be a few weeks before I can test them out.