Best way I’ve found is pair the power meter to the rollers, disable power match, and if you care about cadence, use the “cadence only” option for the pm paired to TR. That setup seems to be the best so far (and I believe it’s the InsideRide recommendation).
2 Likes
I pair my Quarq and InsideRide separately and use TR to powermatch. I did it this way initially because I didn’t read the instructions about pairing my Quarq to the InsiseRide when I first started. It worked so I never switched 
How weird. When I did it that way, I got single-digit target watts onscreen. You might be the Erg Whisperer. 
Since doing it like IR recommends and the way @crazyarm07 is doing it, all is mostly okay, but it really seems like the rollers are chasing the interval average, rather than instantaneous power, i.e., if your average power for the whole interval is low, it’ll clamp down a bit until the average is on target. Which is great for aerobic sessions, since you can zone out fot fifteen minutes and still hit the target watts dead on.
But for threshold? Good gawd, that’s no fun. (Do all smart trainers do this, btw? Shoot for interval averages, rather than momentary power?)
I was trying to play around the with App for this ride. The warmup[ and first intervals were done in resistance mode. I gave up and did the last 3 intervals and 20 minutes at zone 2 in erg. The 30 seconds into zone 5 are tricky but doable in erg mode. @Teuthis
1 Like
I should also state that I am connecting via ANT+ to my Mac for both the InsideRide and Quarq when using TR
1 Like
I use ANT too.
I also did some fiddling today. I had the IR app running on my phone whilst I ramp tested in erg, and it was interesting to see how the rollers adjusted continuously, bumping the resistance up and down a couple of percent. It went pretty well until the last few minutes, where I started chasing the watts; the power would dip a little, so I’d up my cadence to compensate, but then the power would dip a bit more. I ended up going 10-15 rpm faster than desired until I relented, dropped my cadence to where I wanted it, and let erg catch up. I dunno… Perhaps it’s a habit from the dumb trainer that I need to get over.
I think I’m the outlier here…I’ve been connecting my rollers via Bluetooth to TR. HR and quarq cadence to TR are also Bluetooth, and quarq power/cadence to Zwift is ant+.
Don’t use power match while at the same time pairing your power meter in the IR app, It will confuse the shit out of things. The E-Motion rollers are already sending power data to TR, so by using power match you are essentially sending the same data twice simultaneously…it’s like having two masters trying to tell you the same thing at roughly the same time. This defeats the whole purpose of the E-Motion unit auto calibrating with your own power meter, which is the genius of the system. If this weren’t the case, then of course you would want to use power match.
1 Like
This is interesting. You should try the way InsideRide suggests and report back to us on the comparison results. One thing you are not getting with the way you do it now is the E-Roller auto calibration functionality, which you are supposed to initiate before each ride by using their app, and riding at zero resistance for a few minutes while paired to your power meter. I personally think this is a pain in the ass and would welcome the opportunity to dispense with that routine all together! 
One benefit I can see by using power match in TR and not pairing your power meter in the E-Motion app, is that “supposedly” given TR doesn’t have a zero offset feature (why, I have no feakin idea!), I hear they do this in the background when your power meter pairs up with TR. I do not know if this is the case or not. I am just going off of what some techy guy posted in a forum, as he said he validated this by looking at the TR log file…he could see auto zero offsetting occurring…which disturbs me greatly, because zero off setting has to be done while un-cliped and not applying force on the pedals, and with some meter (e.g, Stages), the meter side crank has to be straight down, so how would TR know you are not pedaling and have your crank in the right position when it performs the zero offset in the background? It’s all so very confusing.
I’d love TR to chime in here and demystify this whole business about zero offsetting in their app. Zero offsetting is important, just like zeroing out a scale. Readings from your meter change slightly depending on temperature, humidity, etc.
“but it really seems like the rollers are chasing the interval average, rather than instantaneous power, i.e., if your average power for the whole interval is low, it’ll clamp down a bit until the average is on target.”
How do you know this? I mean, what are you seeing that leads you to believe that the rollers are changing resistance based on average and not a steady stream of data? By average do you mean that the rollers are sampling power data from your power meter? If they were sampling it at longer time intervals (e.g. 5s vs. 1s) then this could cause problems in fast burst sprints, etc.
1 Like
+1 To that.
I noticed in my fiddling (off the bike, examining the smart control on the rollers) that switching between erg and resistance on TR moved the magnet on the resistance unit, while the IR app showed 0% resistance in either case and the low limit light would illuminate.
In erg, the magnet was closer to the flywheel than in resistance, but still not close enough to incur any drag (I would guess).
I imagine IR will be tracking a lot of user data and doing more testing as time goes on…hopefully we’ll see some refinements and app updates to smooth many of these idiosyncrasies out.
Today I used the rollers on a virtual route in Rouvy. It is a virtual route of a real route that I have done locally countless times. I noticed something interesting on a 5 mile climb. On this particular route, at about 4%, if I am just doing a mid endurance zone ride, I’ll use my lowest gear (36t:28t), and at about 85-90 RPM my power will be roughly 190w. However, with the E-Motion rollers, my cadence was the same, as was my speed and average power; however, I was in 36t:15t !!! That’s interesting. I wonder why this would be the case. Do you think this is by design allow the rollers to handle super high grades without the drums going so slow with higher resistance that it’s difficult to stay balanced on them?
Aye! There’s the rub that makes calamity of… well, just calamity.
To answer your question, I don’t know for a fact. But I did an easy session yesterday on erg, just completely zoned out, and looking at the ride summary, the actual and target watts for each interval were bang on, or within a watt or two.
I’ve also had a similar thing happen on a more taxing interval, where I’m just barely hanging on, maybe running a few watts shy on average, and like I said, it’ll seem to clamp down if the little circle is below the white line.
Could this be all in my head? Well, considering all the unused space up there, it’s definitely a possibility. But tricks I played on my dumb trainer give me reason to doubt.
For example, if I was a little slow getting up to power on my dumb trainer, that deficit would drag the average down for the whole interval, and I’d have to overshoot at some point to correct. And either the Smart Control is doing that for me, or I’m doing it subconsciously out of habit now in erg.
I dunno. I’m still experimenting with the Erg ex Machina. But it might just be a loose nut behind the handlebars. 
I am going to test this, this week. I’m going to unpair my power meter in the IR app, and then pair it in TR and use power matching. It’s basically going to hand off all the decision making to TR. The rollers will only obey TR’s commands to decrease or increase resistance, but will not attempt to match power.
1 Like
Well, I have no inside information (pun sort of intended) on how IR designs things, but it sure seems like a good idea in either case.
And if I may digress a bit, I’d hafta say that the IR E-motions just hammer nails in the “rollers are only for leg speed drills and warm-ups” argument, this being a case heavily in point: 90 rpm in 34/28 at 250+ watts, and–shock and awe!–you don’t fall over. Ya just keep motoring. No biggie. (Side note: I’ve found it’s easier to go that slowly under power, rather than with no load.)
Anyway, to answer your question: I dunno, but I like it too.
Does anyone else have trouble with leg speed on the rollers vs. a fixed trainer? I did Huffaker this morning, where the recommended cadence during the work intervals is 100+…I can do that just fine on my KICKR (in erg mode) and hold 100+ throughout the intervals, but it seems like on my rollers (also in erg, so that shouldn’t matter), I don’t seem to be able to hold much more than 95-99ish (and it drops to 89-93 later in the workout as I fatigue). I wonder if it’s because I’m engaging more/different muscles to balance on the rollers than I do on the KICKR - which if anything gets me excited to use the rollers more often because it’ll train those muscles more realistically for outdoor rides.
Thoughts?
I actually prefer to do cadence work on the rollers. You get really good feedback if your form is sloppy. During the race season my coach has me do spin-up drills where I pedal 120-140rpm for 1 minute on/off x5. You can get really good really quickly with practice on the rollers. I had the opposite problem this morning. I was doing cadence pyramids (60, 80, 100, 110, 100, 80, 60rpm) and then the inverse for the back half. I found that if I drop my cadence below ~ 60-65 then the rollers lock up and you have to power out of it by putting out 400+ watts. You can see the power spike at around 18 and 38 minutes when I tried to hit 60 rpm. I put in a ticket to TR to see if it is an issue with the logic in the rollers or in the power matching with my power meter (Quarq) in TR (I use TR to power match). 
1 Like
TEST: TR PowerMatch Off and IR Paired with Pwr Meter VS. TR PowerMatch On and IR NOT Paired with Pwr Meter
I conducted a test this evening (workout images below). Basically, I created a short workout that started at recovery zone and increased by 5 watts until VO2max. I ran the workout first according to InsideRide’s recommendation, which is to pair my power meter in their app, let it auto calibrate in zero resistance for a few minutes, and then shut down the app, open TrainerRoad and disable Power Match. Note that I was paired to my power meter in TR, but was using the Cadence only mode.
I wanted to compare the power consistency in each test to see what system was more consistent and effective at using my power meter’s real time data to control resistance, TrainerRoad or the InsideRide unit.
One last note on the test before discussing the results. On my track bike, given the gearing, the low floor wattage of the rollers is roughly 160w, meaning the rollers will not start changing resistance in response to my power meter data until I get above 160w. Hence, knowing this, I wanted to see if the rollers truly ignored what I was doing until 160w, and so rather than riding at a consistent cadence until 160w, I decided to maintain my target power by increasing my cadence to see if the rollers reacted at all. Then after 160w, I started to maintain an average 90 RPM cadence for the rest of the incremental intervals, in order to test how the rollers, or TrainerRoad reacted to my power meter’s real time data.
A few observations before I open this up for further discussion and testing. I like to use the sound wave analogy here: amplitude (the degree or extremes of variation in wave height) vs. wave length (the distance between each peak or valley across the wave form).
The first thing I noticed was that with TR PowerMatch Off, letting IR make all the decisions on what to do with the power meter data, was a longer wave length but similar variation in amplitude (power lows and highs). If you look at the power chart, it’s almost as if the rollers are only reacting to power meter changes in longer time increments, which means when the power changes it has some catching up to do and therefore results in more a more abrupt variation in the power amplitude. And you can even feel this when riding…resistance will feel fairly consistent for longer periods of time, but then it’s like the rollers will have a knee jerk reaction to correct with actual data is different in the next window that is samples it. Weird. Now with the test when Power Match was ON, and the power meter was not paired in the IR app, it is the opposite. Look at how tight the wave length is, but there is just as much variation. This is still in ERG mode, btw. So the amplitude or variation is similar, but the wave lengths are shorter, and also very consistent. The wave lengths with IR making all the decisions seems less consistent.
What does this all mean? I know what it means to me. Personally, when I do intervals, I want less abrupt changes and more consistency. If I want total consistency, I’d use my Wahoo Kickr in ERG mode, but that can’t be achieved with E-Motion rollers with power match off. The time delays or lag in changing resistance based on my power meter changes (longer wave length) irritates me. If it’s going to be inconsistent, then I want that inconsistency to occur in short time frames, because this is how it looks on the real road. So it’s like IR is trying to be as consistent as a Wahoo Kickr in ERG mode, but it can only do it for so long, and then it becomes erratic to get back on course. Whereas with PowerMatch on in TR, at least the inconsistency itself is “consistent” if that makes any sense. It’s like a consistent and short wave length of modulation of power over time. I can deal with that and actually prefer over the monotonous and unforgiving, and entirely unrealistic ERG mode in Wahoo Kickr.
The other thing to know is despite the low power floor of 160w, look at how different the power line is with my power meter paired to IR. It’s like it’s still doing something with resistance even though it can’t reduce the resistance to below 160w. What is it doing then?
TR Power Match Off / Power Meter Paired and Calibrated in IR App
TR Power Match On / Power Meter NOT Paired and Calibrated in IR App
3 Likes
Ah ha! I noticed the same! Thanks for submitting the ticket. I do a lot of cadence drills as well on my track bike and noticed that if I was spinning at 140RPM or so at high end endurance zone, then go abruptly into recover, the rollers are massively delayed in getting to that low power…as you say, you have to pedal your ass off again at high power to get the rollers to change resistance. That’s bizarre. So, I did not notice this in my test that I just posted, where I do not pair my power meter in the IR app, and then pair my meter in TR and use power match. At this point in time, I trust TR making all the decisions, not whatever is going on in the little magnet resistance black box. It seem to be going through some issues right now
I guess I’ll follow up too.
I run my IR like the instructions recommend (PM paired to IR), and like @krispenhartung, I roll easy for a few minutes before connecting to TR. What I’ve taken to doing, unmentioned elsewhere in this thread, it seems, is leaving the IR app running throughout the workout, enabling me to see what the IR is doing directly, and that’s been illuminating.
I did Kaiser last night, and I was able to see the Resistance % commanded as well as speed, cadence, and power in the IR app. In response to my previous post in this thread, yes, there was a loose nut behind the bars, and it was me. I was subconsciously tweaking my cadence to move the little circle like I used to do on the KKRM. Last night, eyeing my speed on the IR app (easier to maintain a consistent speed in tenths of a mph than use cadence), I endeavored to simply find a steady rhythm and let the IR handle the watts.
This resulted in a power trace similar to @krispenhartung’s top example above, and though not a dead flat line by any means, the interval averages came out within three or four watts, user error excepted (like a bobble recovery or other oops).
Being able to watch what the IR was doing on my phone kinda eased my doubts and let me roll, not constantly wondering if it was doing something weird. Indeed, it only got weird when sweat dripped on the screen mid interval, launching the resistance from 45% to 70%. That was not fun.
As to cadence weirdness like @TheBestMe mentioned, yeah, it seems there’s a learning curve. For example, if I raised my cadence substantially (from 95 to 110) to prepare for an interval, the resistance seemed slower to catch up than if I just held at 95. And if I drop to 60 abruptly, the resistance will go waaaay up to compensate. But if I’m going from 95 to 60, it’s because I’m getting out ouf the saddle to stretch, and I’ll upshift to the eleven first, drop my cadence on the way, and then rise. Since the change in cadence is a bit slower, I’m not getting ahead of the IR, and everything goes smoothly.
TL;DR: As I’m figuring out and adjusting to its quirks, I’m liking the IR more and having fewer confusions. 
2 Likes