A few weeks ago Wahoo released updated Speedplay pedals. I don’t ride Speedplays but I have tried some “clones” and those clones were not my cup-o-tea. I didn’t like the clip in action which was tough and not adjustable. I’m a fervent believer based on some unscientific testing that those claiming tension on Look and SPD-SL style pedals is required for efficiency on pull up are in the Velominati BS category or have worn their cleats past a useable life. These are the same people claiming zero float cleats are the goal regardless of professional fitter advice. *sigh*. However, this is the first instance I wish I could have adjusted it. It’s way too hard. So the fake Speedplay experience left me unimpressed. I had a real Speedplay users try them and they indicated similar but stiffer/harder clip in but that could just be because they were new — they were unsure.
But there are diehard Speedplay fans and I can get behind some of the principles like two sided clip in. I like my old Aluminum RD5800’s because they hang with the right angle to make clip in easy. My Looks, Vector 3, and Xpedo don’t with their composite bodies. Composite = injection molded nylon with glass fibers (short). Marketing…
So Wahoo teased an image which could easily be brightened to see what it was. Two days latter the FCC filing went public, and a few days after that it went confidential. I’m not 100% sure why but this seems to happen to Wahoo’s filings while others don’t get such bad luck. Well their bad luck. I get to look at the insides which I enjoy. So watch the video for more details.
What Wahoo got right
- USB-C Charging (well sort of, dumb USB-C but this is so low powered it’s meaningless to talk on this)
- Able to use Allen wrench
- 120mah of Rechargeable battery = very good battery life potentially, and potentially replaceable!
- Maybe Gyro? So oval ring compatible (if only we knew if oval rings did anything)
- Serviceability!!!!! Holy cow!
- Strain gage physical (only) arrangement
What I’m critical off though is bit more force sensing
- Differential Bending beam as ONE sensor
- Replicating Stages (exactly?) ADC
I’m not 100% sure on some of this but what it boils down to is compromise (well just made worse) in signal quality due to the differential bending beam as one sensor. In a sensitive Wheatstone bridge, say half bridge, one gage goes up when the other gage goes down. In a differential one gage goes up but the other goes up too, just not as much and dependent on the spacing distance. Instead of improving signal quality it takes it away thus desensitizing it. What’s more ideal is to keep these separate but that doubles the number of inputs and either needs a precision reference of a full Wheatstone bridge which would be more expensive to wire. Also it makes calibration at least twice as hard and if a company doesn’t know how to properly do experiments (*cough* One Factor at a Time, OFAT, is not a proper experiment).
You can normally get away with differential bending beam (desensitizing the bridge by 80 – 95%) or Successive Approximation Registers (SAR) ADC. See SAR ADC’s are not really ADC’s so much as a comparator and a DAC. So it using a Sample and hold at a point (Lots of noise potential because you can’t know if you’re high or low on the noise without oversampling!!!) then runs that into a comparator (lossy) and compares it against a voltage from a DAC that counts from 0 to the reference voltage. When the comparator trips you get you data conversion. These are only good for 10 – 14bits max. and with strain gages and high gain you’re lucky to get 10bits on a good day. If you do the 4 bridge design this is fine because the strain on a pedal is very large and you get amazingly clean signal!
These are fine separately but using them together is like shooting yourself in one foot then deciding you need the holes to match. Sure you can walk if it’s not too bad, but you can’t run.
That’s what Wahoo did. They shot themselves in both feet unnecessarily. The cost of a much better Sigma delta ADC that would have a built in gain stage, better response, programmability, and built in Multiplexer would have likely been the same price or maybe a little more than the re-use of their 2010ish Stages design. Yes, Wahoo designed and built the first stages PCB’s and firmware on contract to Stages. And yes, it looks identical except for a mux switch.
It’s no secret I’m involved with Bodyrocket. They aren’t blowing in a tube and making guesses drag, they are measuring it — directly. That means load cells everywhere! But that’s hard. Every requirement is just dialed up to eleven. I look at high speed data, compression, and stuff like Bodyrocket is doing and I think about what Wahoo could have built for 3 dollars more in electronics and twice as much calibration effort — and calibration effort is the bigger cost.
Doesn’t mean they can’t built a Gen 2 or higher spec if any future stuff catches on. That serviceability is a roadmap for evolution too. Swap in a new electronics setup. But it’s really a 2013 – 2018 product and not a 2022 product. Like Garmin Rally is just Vector 3 (same FCC filing) so we’re just getting a 2018 product again there too.