Observations Electronic trainer vs Power meter vs speed/cad sensor

(L Read) #1

After several hundred miles with Zwift riding a Computrainer vs Stages PM (with a mag trainer) in the game (sometimes back to back), I have observed that the ride is significantly more difficult with the Computrainer. With good calibration, it is easier to maintain a given power and draft another rider using Stages and a dumb trainer. It was also easier to go downhill much faster with Stages plus the dumb trainer. Now I see how riders can go much faster than me down the hill while using less wpkg when I am on the CT.

Drafting is still unrealistic with computrainer. I have to maintain nearly as much wpkg or sometimes more to stay with the lead rider.

Perhaps one issue is that a dumb trainer can not simulate realistic momentum compared to an electronic trainer. Probably no way to correct this.

I could see that if someone is using a dumb trainer with just cadence/speed sensor, the ride would be highly unrealistic.

We should have a way to see next to a rider’s name what class of equipment is being used- ET (electronic trainer), PM (power meter), Sn (sensors), etc. That way , I know why I cant even suck their wheels going downhill.

(W ish I was Outside) #2

I agree there’s something weird going on with the CT and descents. I’d complained about AI riders coasting down the hill not pedaling at a rate faster than if I was putting in hundreds of watts on the CT, maybe these issues are all related? https://support.zwift.com/hc/communities/public/questions/201895895-AIs-and-descents

Overall the watts recorded by the Zwift/CT are very very close to my garmin+Powertap on efforts in the sub-500W range. I’ve been running them simultaneously on all my recent rides and comparing. Once efforts are over 700W I’ve found the Zwift/CT is reporting 50-100W less than the Powertap, and it doesn’t seem to be just tire slip although that may be factoring into it, I tried increasing the roller pressure on the tire and it didn’t seem to help (within reasonable limits e.g. still below 3.5 calibration). There also seems to be some latency within Zwift/CT compared to the Powertap data that affects the gameplay, and this may play in to why drafting or catching on to a powerful train is hard since you may hit 800W almost instantly according to the Powertap but Zwift will stay at 250W or whatever you were at for several seconds and by that time a big gap has opened.

(Peter Lin) #3

There seems to be significantly more lag with USB ANT+ than with bluetooth LE or Garmin head unit. I record the rides using cyclemeter using BTLE and I do notice a different in the number of data points. The data I get from BTLE with iPhone + cyclemeter is definitely better than with the USB ANT+. I’m close enough to my workstation that I don’t see the signal dropping, so I’m guessing it could be just the sampling rate of USB ANT+.

(Shawn DeBoer) #4

I assume you cannot control resistance with Stages/dumb trainer combination. I have the Cycleops Powerbeam Pro (smart) and Zwift does not yet support it for resistance control. Big disadvantage for power.

(James Prosperi SASCC) #5

I decided to put my Vectors on to do a comparison and for the same ride intervals, comparison gave a typical overall 10% greater VP as opposed to the vectors.

I use Vectors and a Cycleops Fluid 2. Tyre pressures are all good and roller contact nice and tight.

Values were reported after initial fluid warm up, maybe 10 minutes or so.

The instant values seemed to correlate more closely after 30 plus minutes into ride and I did take into account any smoothing 3s, 10s and lag etc.

The Vectors read power at your pedals and take crank length x force to calculate torque and power when using the cadence or rpm. Would I be correct is saying that an overall 10% lower Virtual Power reading would be mostly due to drive train losses due to the VP readings coming from the trainer at the back wheel.??

Software is awesome and I remember thinking to myself years and years ago, I wonder when someone will do this!! Now you have done it. Well done.

I hope the transition from Beta to final release goes well.

I am also REALLY KEEN to see the training packages and how they are managed and work.

How far off are these? Will they be offered for Beta testing also?