Smart trainer ERG wattage question
#1
Smart trainer ERG wattage question
This winter, I've been playing with a Kickr Core and Zwift for weekday riding. As part of that, I'm pretty much in the habit of doing two tough ERG "workout mode" rides each week, and it seems to be suuuuuper productive in terms of building my threshold and my ability to sustain above-threshold efforts on climbs. My FTP is higher and trending upward. Digging it.
One weird issue though: the ERG resistance changes often don't quite nail the target wattage. Like it should be putting me at 255, but my actual wattage readout is fluctuating around 257 to 260. Or sometimes, the opposite: the workout increases the resistance to have me push at 340 for a minute, but I actually display wattages at like 334, 335 the whole time. Not a huge deal, and I could always make minor adjustments by adding or subtracting from the workout intensity by a percentage point or two, but just kind of weird that the trainer never seems to quite dial in the resistance correctly. And when I'm slogging through sweetspot 2x20's, four or five watts can be kind of significant. The issue seems to be independent of my cadence, and doing a spindown does not resolve it.
Is this just a totally normal margin of error? Do other people not have this problem? Anything I can do to address it?
One other burning question: do people say "E-R-G," or is it "ERG" as in "erg! arg! I'm a scary monster!" ???
Thanks!
One weird issue though: the ERG resistance changes often don't quite nail the target wattage. Like it should be putting me at 255, but my actual wattage readout is fluctuating around 257 to 260. Or sometimes, the opposite: the workout increases the resistance to have me push at 340 for a minute, but I actually display wattages at like 334, 335 the whole time. Not a huge deal, and I could always make minor adjustments by adding or subtracting from the workout intensity by a percentage point or two, but just kind of weird that the trainer never seems to quite dial in the resistance correctly. And when I'm slogging through sweetspot 2x20's, four or five watts can be kind of significant. The issue seems to be independent of my cadence, and doing a spindown does not resolve it.
Is this just a totally normal margin of error? Do other people not have this problem? Anything I can do to address it?
One other burning question: do people say "E-R-G," or is it "ERG" as in "erg! arg! I'm a scary monster!" ???
Thanks!
#2
Pointy Helmet Tribe
Join Date: Feb 2009
Location: Offthebackistan
Posts: 4,338
Bikes: R5, Allez Sprint, Shiv
Liked 627 Times
in
295 Posts
Are you getting the power reading from a powermeter or the smart trainer? When i use my Quarq to supply the power reading to Trainerroad, I get what you describe - probably because the app feedback to the Kickr is a little lagged and not able to match the small variations in power that i am actually putting out. When i use the power reading directly from the Kickr, it is stable, but i suspect some smoothening is happening there.
That said, I wouldnt fuss about a few watts here or there. While the FTP is obviously a single number, the physiological training stimulus is pretty much the same whether you are a few watts above or below.
That said, I wouldnt fuss about a few watts here or there. While the FTP is obviously a single number, the physiological training stimulus is pretty much the same whether you are a few watts above or below.
Likes For guadzilla:
#3
#4
Pointy Helmet Tribe
Join Date: Feb 2009
Location: Offthebackistan
Posts: 4,338
Bikes: R5, Allez Sprint, Shiv
Liked 627 Times
in
295 Posts
I get a little bit of fluctuation as well, but usually that happens when i change my cadence a little. It hasnt been enough that i have given it too much thought, to be honest.
#5
Lurker
I always assumed that this was just the Zwift display rounding the actual target wattage to the nearest 5 watts for intervals where the intensity is defined as a percentage of FTP.
#6
Duh, thanks, that makes perfect sense. I like how it never struck me as odd that the targets are always nice round numbers. Fog of war, I guess!
#7
serious cyclist
Join Date: Apr 2009
Location: Austin
Posts: 21,147
Bikes: S1, R2, P2
Liked 3,685 Times
in
2,028 Posts
It never quite matches for me; besides the rounding, there are issues like your cadence changing and the trainer moving the resistance to try to maintain wattage, but it lags. And then you usually don't know what the smoothing is.
That's part of the reason I just record from my PM on the trainer, so at least any errors are consistent for when I'm racing, and on the road.
And it's Erg like Urg.
That's part of the reason I just record from my PM on the trainer, so at least any errors are consistent for when I'm racing, and on the road.
And it's Erg like Urg.
#9
Kamen Rider
Join Date: Aug 2017
Location: KL, MY
Posts: 1,071
Bikes: Fuji Transonic Elite, Marechal Soul Ultimate, Dahon Dash Altena
Liked 277 Times
in
164 Posts
There would be minor fluctuations because you won't be fully 100% consistent with your cadence and pedaling cycle, so that's normal.
If the trainer is maxing out below the target wattage, it could be possible that it has reached its maximum resistance. This can happen if you are on a very light gear (e.g small ring in front, big ring on the cassette) and the app requires a very high wattage, or even low cadence on sufficiently high wattage; if that's the case, try shifting down to a heavier gear.
And yes, urg I'm dying, not erg I'm a scary monster.
If the trainer is maxing out below the target wattage, it could be possible that it has reached its maximum resistance. This can happen if you are on a very light gear (e.g small ring in front, big ring on the cassette) and the app requires a very high wattage, or even low cadence on sufficiently high wattage; if that's the case, try shifting down to a heavier gear.
And yes, urg I'm dying, not erg I'm a scary monster.
Last edited by atwl77; 01-24-20 at 10:20 PM.
#10
Totally normal and within the margins of error of a reasonably fast control loop bearing in mind how inconsistent you actually are with regards to torque application and cadence. Your zones are zones and not discrete targets so the small amount of under or over achieving shouldn't make much difference if any to your progression
#11
HarborBandS
I also use a Kickr Core with Zwift. The app is continuously adjusting the resistance based on power and cadence inputs from the trainer. If your cadence drops, the resistance increases. If your cadence increases, the resistance drops. There is a slight lag in this feedback loop, so that is why it is nearly impossible to keep the displayed wattage at the exact ERG mode setting. You'll notice it if you try to spike your cadence suddenly... The watts will briefly go up, and the trainer will quickly adjust to the new cadence to bring you back to the ERG setting. Then when you let up on your cadence, the displayed watts will temporarily drop below the ERG setting until the trainer adjusts.
#12
Senior Member
Join Date: Jul 2015
Location: reno, nv
Posts: 2,398
Bikes: yes, i have one
Liked 1,227 Times
in
720 Posts
I also use a Kickr Core with Zwift. The app is continuously adjusting the resistance based on power and cadence inputs from the trainer. If your cadence drops, the resistance increases. If your cadence increases, the resistance drops. There is a slight lag in this feedback loop, so that is why it is nearly impossible to keep the displayed wattage at the exact ERG mode setting. You'll notice it if you try to spike your cadence suddenly... The watts will briefly go up, and the trainer will quickly adjust to the new cadence to bring you back to the ERG setting. Then when you let up on your cadence, the displayed watts will temporarily drop below the ERG setting until the trainer adjusts.
#13
serious cyclist
Join Date: Apr 2009
Location: Austin
Posts: 21,147
Bikes: S1, R2, P2
Liked 3,685 Times
in
2,028 Posts
Not sure this is entirely correct, if i am reading the FE-C data sheet correctly from ANT+ the display tells the trainer what the target power should be and the trainer adjusts accordingly based on power provided by the rider, not the cadence of the rider. heck, my trainer does not even have a cadence sensor. because of the number of dropped ANT packets i see is large enough i think having the display involved in the feed back would be pretty bad.
Likes For Bah Humbug:
#14
HarborBandS
Not sure this is entirely correct, if i am reading the FE-C data sheet correctly from ANT+ the display tells the trainer what the target power should be and the trainer adjusts accordingly based on power provided by the rider, not the cadence of the rider. heck, my trainer does not even have a cadence sensor. because of the number of dropped ANT packets i see is large enough i think having the display involved in the feed back would be pretty bad.
I worded that poorly.
Last edited by HarborBandS; 01-31-20 at 03:24 PM.
#15
Senior Member
Join Date: Jul 2015
Location: reno, nv
Posts: 2,398
Bikes: yes, i have one
Liked 1,227 Times
in
720 Posts
How do you think it adjusts according to the power? It has a certain tension on the hub, and that multiplied by speed produces power. It doesn't need to know your cadence to know the speed the wheel (or simulated wheel) is spinning at, and because of the way gearing works, cadence=speed*gearing. If you set the trainer on erg mode at a certain power and vary your cadence suddenly you can see it react, overshoot, and then stabilize.
i am pretty sure cadence is not needed in order to make resistance changes due to changes in power. power should be derived from a strain gauge and how fast the cranks are moving is not related. i could be wrong on that but that is how we we employ strain gauges in the lab where i work (robotics). but again, this is my believe based on reading the ANT+ data sheets which the kicker uses for communication to my PC (no blue tooth). maybe when i get home tonight i will do an experiment to see how this pans out. i will remove the cadence sensor i do have and just run the trainer and swift only. then i will report back. i have a pretty good appetite for learning so i am prepared to eat my words.
-scott
#16
serious cyclist
Join Date: Apr 2009
Location: Austin
Posts: 21,147
Bikes: S1, R2, P2
Liked 3,685 Times
in
2,028 Posts
well, as i stated, i don't believe it requires cadence to derive power. nor do i believe it requires gearing info. i could be entirely wrong about the gearing but i do not recall ever inputting that data into zwift or into my trainer (wahoo kicker).
i am pretty sure cadence is not needed in order to make resistance changes due to changes in power. power should be derived from a strain gauge and how fast the cranks are moving is not related. i could be wrong on that but that is how we we employ strain gauges in the lab where i work (robotics). but again, this is my believe based on reading the ANT+ data sheets which the kicker uses for communication to my PC (no blue tooth). maybe when i get home tonight i will do an experiment to see how this pans out. i will remove the cadence sensor i do have and just run the trainer and swift only. then i will report back. i have a pretty good appetite for learning so i am prepared to eat my words.
-scott
i am pretty sure cadence is not needed in order to make resistance changes due to changes in power. power should be derived from a strain gauge and how fast the cranks are moving is not related. i could be wrong on that but that is how we we employ strain gauges in the lab where i work (robotics). but again, this is my believe based on reading the ANT+ data sheets which the kicker uses for communication to my PC (no blue tooth). maybe when i get home tonight i will do an experiment to see how this pans out. i will remove the cadence sensor i do have and just run the trainer and swift only. then i will report back. i have a pretty good appetite for learning so i am prepared to eat my words.
-scott
Apparently I'm not doing a great job of explaining it, but trust me, even though its not transmitted over FE-C, the trainer knows the speed it's spinning at.
#17
Senior Member
Join Date: Jul 2015
Location: reno, nv
Posts: 2,398
Bikes: yes, i have one
Liked 1,227 Times
in
720 Posts
It doesn't need to know the cadence, nor the gearing - it needs to know the speed the wheel or hub is spinning at. That's how it controls power. This is exactly like how it takes more power to ride faster under the same circumstances, and you have to either shift to a higher gear or increase your cadence to do so. If I only know that you increased your speed, I know that you increased your power, but I have no idea whether you did it by shifting or increasing your cadence.
Apparently I'm not doing a great job of explaining it, but trust me, even though its not transmitted over FE-C, the trainer knows the speed it's spinning at.
Apparently I'm not doing a great job of explaining it, but trust me, even though its not transmitted over FE-C, the trainer knows the speed it's spinning at.
#18
Senior Member
Join Date: Jul 2015
Location: reno, nv
Posts: 2,398
Bikes: yes, i have one
Liked 1,227 Times
in
720 Posts
i took the time to do a very quick experiment, one with, and one without a cadence sensor. with a cadence sensor because i neglected to take it off the crank but it makes for a good before/after test.
the first image demonstrates with cadence. i purposely slowed down to detect increased resistance with stable power out. the second shows a steady cadence of 43 which is what it was when i removed it and moved it into the house, although the cadence never deviates from 43 power is certainly controlled. i also decreased my cadence but since there is no sensor there is no evidence either, however, the power remains controlled. the increase in power at the end is the next increase during the warmup period.
the first image demonstrates with cadence. i purposely slowed down to detect increased resistance with stable power out. the second shows a steady cadence of 43 which is what it was when i removed it and moved it into the house, although the cadence never deviates from 43 power is certainly controlled. i also decreased my cadence but since there is no sensor there is no evidence either, however, the power remains controlled. the increase in power at the end is the next increase during the warmup period.
#19
HarborBandS
I’m not even sure what you are arguing any more. I admitted my poorly worded post caused confusion. There is no adjustment of resistance based on input from a cadence sensor.
However, there is adjustment of resistance based on trainer flywheel RPM, which is a direct relationship to cadence if you stay in the same gear. Note it’s resistance, not power. They are not the same thing. A higher RPM with lower resistance can be the same power as a lower RPM with a higher resistance.
You can easily feel this if you slow your pedaling cadence towards a stop with ERG mode on. At some point the resistance goes so high that you can no longer spin the cranks. Zwift calls this an “ERG mode death spiral”. To recover, you have to switch ERG mode off so you can get the cranks spinning again, then turn it back on once your cadence is up again.
However, there is adjustment of resistance based on trainer flywheel RPM, which is a direct relationship to cadence if you stay in the same gear. Note it’s resistance, not power. They are not the same thing. A higher RPM with lower resistance can be the same power as a lower RPM with a higher resistance.
You can easily feel this if you slow your pedaling cadence towards a stop with ERG mode on. At some point the resistance goes so high that you can no longer spin the cranks. Zwift calls this an “ERG mode death spiral”. To recover, you have to switch ERG mode off so you can get the cranks spinning again, then turn it back on once your cadence is up again.
#20
Senior Member
Join Date: Jul 2015
Location: reno, nv
Posts: 2,398
Bikes: yes, i have one
Liked 1,227 Times
in
720 Posts
Was not my intent to argue with you. probably more so with BahHumbug but i think he was really arguing my point exactly. anyway, i said i would do the experiment and i wanted to do it mostly because i did not want to come off as someone who says "i'll do it" and then doesn't. besides, i am learning how to program the trainer and this is good info to collect.
ride on!
ride on!
Likes For spelger:
#21
serious cyclist
Join Date: Apr 2009
Location: Austin
Posts: 21,147
Bikes: S1, R2, P2
Liked 3,685 Times
in
2,028 Posts
Was not my intent to argue with you. probably more so with BahHumbug but i think he was really arguing my point exactly. anyway, i said i would do the experiment and i wanted to do it mostly because i did not want to come off as someone who says "i'll do it" and then doesn't. besides, i am learning how to program the trainer and this is good info to collect.
ride on!
ride on!
#22
Senior Member
Join Date: Jul 2015
Location: reno, nv
Posts: 2,398
Bikes: yes, i have one
Liked 1,227 Times
in
720 Posts
You can argue with me, but I'm saying the same thing as HarborBandS, not you. You're confusing the RPM of the crank with the RPM of the resistance unit, but I'm not going to keep trying to explain it. As you use the trainer more you'll probably realize what we're trying to say.
that aside...tomorrow will be sunny and 66F. i will not be on the trainer tomorrow.