I've got an SRM, some wahoo KICKRs, a few powertaps and a Quarq and while there are differences in how often they internally update their data and externally broadcast that data, in general it does seem like they're all lagged by 1-2 seconds or so. I've not had garmin vectors in my test area yet, but it sounds similar. 3 seconds would be on the extreme end of things but I think you may be overestimating that time.

I'll actually measure the differences sometime soon, but I play with these devices every day and initially nothing has stood out as dramatically better than any other in this regard.

PS, Some power meters update their outward facing data only once per crank revolution, so if you're cranking away at 50rpm the latency might be much worse than at 100rpm. Others power meters broadcast at a fixed frequency. I don't know which way the garmin vectors work.

Thanks for the info. I just took some video timings, and the latency is even higher than I expected: around 4s from not peddling at all to seeing watts show up on the garmin. It seems a little faster (3s) if I go from peddling to peddling harder. In the one test I did of peddling to no peddling, and then waiting to see 0 watts, it took just over 5s. This is with instantaneous power showing on the garmin, and peddling at 90-100 cadence.

You'll see slightly more delay going to 0w, as most power meter head units will wait a second or two to validate a power drop before showing it (plus it often takes a second for the PM itself to validate it).

---

-
My tiny little slice of the internets: dcrainmaker.com

Yes, exactly. There are 2 separate scenarios going on here. The ANT+ standard is to display 0's after 3 seconds of stale or no data, so the "return to 0" test isn't a very good one.

Meanwhile, when you're pedaling the head unit will typically see data sent at 1hz or once a second. There's still some inherant lag there, but it's not going to be 5 seconds. If you're really seeing 3-4 seconds from 0 to "something" that does sound a little worse than everything I've seen.

PS, dcrainmaker, I think we're trying to meet up with you in Paris in a couple of weeks if you're available.

Vector is one of a bunch of power meters that does cycle-based updates (this group includes SRM and Quarq).

• If you change from being stationary to pedaling then it may take up to two full crank rotations (from your perspective) for it to register the first full rotation.
• Until the first full crank rotation completes it will not change its messaging output because it doesn't yet have a complete cycle.
• Once the first full cycle happens, it may take a couple of ANT messages before the first message with a valid torque and cadence period comes out. This can add 1/2 to 3/4 second delay.
• Once the Edge gets that, the Edge knows you're now pedaling, but it uses that first message to re-set its internal baseline for the ANT datastream.
• Once you get the second full crank rotation (and the pedals have had a chance to transmit that, and the Edge has received and processed it) the Edge can calculate power based on the difference between the two messages.

Now the Edge can update your display and you can see what you're doing. But it may not update until its next internal one-second tick and that could be another second.

---

Less is more.

Not SRM but I'm sure I read on the Power2Max thread that there is a 2 second delay in the figure that is shown on the head unit.

I do a decent amount of TT'ing and I always see this on my P2M with the first 2 seconds after the push off always zeros - this is on an Edge 500.