PeteHski

The only time I’ve ever seen the semi-transparent trainer is on the start line before a race. Is that what we are talking about? I just check for the lightning bolt ⚡️to make sure I’m not smashing myself racing a Z-Power rider 😂

himespau 

I actually wasn't aware that transparent trainer was a notation for zPower until RChung said it. I've seen it drug around the roads a few times, but very rarely.

Since I'm a C, I'm not as worried about ZPower racers at my teammates who are As who will see them go off the front at crazy sustained power (I did an A race once because my team needed finishing points and the guy who "won" it held 6 W/kg pretty steadily for the whole 45 minutes and was gone off the front after the start - he didn't seem to understand it was a points race, or couldn't push harder than that with his trainer, so I actually took some segment points from him). At the C level, if you go over power (whether that's real or pretend), you will cat up before too long and I don't have to race you anymore. Unless there are series points on the line for my team, I am just racing pixels. Cheaters (or bots in the case of indieVelo) or real people putting out real efforts, either way, it's a jersey in front of me to chase down. As long as I'm in a race with Cat Enforcement, there aren't quite as many sandbaggers blowing up the front of the race (and making me blow myself up chasing them).

PeteHski

I’ve never seen anyone actually dragging one of those trainers along the road in the 4 years I’ve been using Zwift. It’s the kind of thing I would notice!

I’m a middling Cat B rider, which is a bit frustrating because I have zero chance of winning a Cat enforced race. I’m not a sprinter or a lightweight climber either, so I get well beaten over pretty much any course profile. I agree about racing pixels too. I don’t really care if the guy I’m racing neck and neck is lying about their weight. But I try to ignore those who are really taking the piss!

RChung

That's interesting. So they allow Snaps but not on-bike PMs even when the latter can be validated. I guess it's obvious that I'm not au courant, and my interests in power data aren't the same as theirs.

RChung

A friend was recently looking for a used Snap and at least on local sites and the fleabay he was seeing around $200. I just bought a spare Powertap for $150 from someone who was switching to pedal-based power. I've seen dumb trainers for $25.

But it's absolutely true that I value on-bike power data more because I live in a climate that allows me to ride outdoors the majority of the year, I don't do e-racing, I am interested in measuring actual CdA and Crr, and I do very little hyper structured training. So while I may not see the value proposition for a Snap, you've reminded me that others may very well be able to.

unterhausen

I'm confused that they don't allow racers to use power meters. Is that because a crank based meter is going to read higher due to losses?

PeteHski

I don’t know if that’s the reason, but they do usually read higher. Mine did anyway by around 4-5%, which ties in with drivetrain losses. Enough to make the difference between being competitive and an also-ran in Zwift races.

msu2001la

Your crank power meter reads 4-5% higher than your trainer power meter?
My stages crank arm PM seems to be very close to my Snap (when calibrated w/ spin down). I've always assumed the crank arm power meter would be more accurate since there's fewer variables involved, and I just use that for Zwift and don't even hook up to the trainer PM.

PeteHski

Yes the crank power meter always read a little higher, as expected since my trainer (Elite Direto X) has an optical power meter measuring torque at the rear wheel. Drivetrain losses are typically in the order of 5% so it made sense to me.

In your case I expect your crank arm PM is more accurate and it's just coincidence that your Snap is reading a little high for power at the rear wheel.

spelger

not being an mechanical engineer...wouldn't drive train losses be a constant? stating a % implies increasing losses as more power is provided by the rider.

RChung

No. However, it's often convenient to treat them as *close to* proportional to power. For example, we think that drive train losses are close to proportional to chain tension, rather than proportional to the product of chain tension and chain speed (i.e., power). The "normal" range in chain tension is greater than the "normal" range in chain speed, so the relationship between power and chain tension tends to be higher than the relationship between power and chain speed, so using power as a proxy for chain tension is sorta kinda ballpark okay.

spelger

Ok, that makes sense. a more taut chain would exhibit more forces in the link bearings and require more power to overcome the increasing friction. thanks. (although i'm sure there is more to it than that.)

RChung

As an aside, measuring chain tension and chain speed was the basis for one of the early on-bike power meters. The inventor originally used a pick-up from an electric guitar to measure the tension. Clever as hell. That power meter had problems but the idea was neat.

unterhausen

I'm trying to imagine how a guitar pickup measures strain and failing. On a similar tangent, I was going to make a wheel dishing gauge that used a guitar string, pickup, and tuner. But I decided it was too silly.

RChung

In this case, when the chain passes by a magnetic sensor it generates a field that can reveal its vibration frequency. That frequency varies with tension.
https://patents.google.com/patent/US6199021B1/e

PeteHski

As RChung explained, it is not a constant. Drivetrain losses are usually quoted as a percentage of power output in tests. My own simple testing of crank power vs hub power showed a pretty consistent percentage difference at different power levels. I measured the difference in power at 50W intervals from 100-500W.

unterhausen

Thanks for the patent. On my bikes, the chain vibrates at frequencies mostly related to the chain interacting with the chain rings and rear cassette cog. Seems to me that discriminating all that from its natural vibrations would be a really difficult task. Probably made harder on bikes maintained by slobs like me.

On edit: they say all that can be "easily filtered" from the signal. They even have source code. So we could all build one if we wanted

RChung

It does depend on the free unsupported length of the chain, the chain weight, and how dirty the chain is but in practice these were handled pretty well -- the practical problem turned out to be not the way they measured chain tension but the way they chose to measure chain speed. There are several ways to do that, and they chose one that had higher "resolution" but was (unexpectedly for them) more sensitive to noise. If they'd survived they might've gone to a different more robust way to measure chain speed. As I said, the idea was clever; the execution of the idea wasn't ideal.