Oh dear. This absolutely explains some of the ‘feel’ issues I have with my Drivo. It’s ‘accurate’, but that doesn’t mean smooth. I’ll get through this trainer season, then maybe revisit a new trainer.
We will never get updates from Elite
I’ve had drivo since winter 2017 (the one with the hole around the cassette, unlike the one GPLama is using). And while I sometimes also experience the over-underadjusting, this was definitely worse in Zwift compared to what I experience in TR. If an interval is commencing and I happen to apply watts too early it will oscillate a bit, but then recovers itself rather quickly. Accordingly, I’ve learned to ‘feel’ the trainer a bit and can now usually ease in nicely into an interval.
Here’s and example from Bluebell (note that I have TR power smoothing = 0 and elite power smoothing set to 6 (a remainder of my zwifting days)):
and a bit of Taylor:
but then this happened on my latest ramp test (not too happy about that):
But what if those variations are actually a feature and not a flaw? I’ve said this before, but I think the Drivo is a bit of a slave of its own accuracy. I can’t imagine a human being able to put out a constant wattage on every pedal stroke. And this variation in our power output will be impacted by whatever our state of mind is (fatigued, amped, later in the workout…) Yet, rather than smoothing things out a bit, which is what I expect other trainers (and now the drivo II according to gplama video) are doing, the drivo is programmed to adjust to every single slight variation we produce.
In accordance to this, we all know the flat Kickr power lines are artificial, but nobody complains about them. We just feel good and think we produce smooth power. In contrast, Elite chose a very low number for their default power output smoothing (I think it is originally set at 2), and we all complain because it doesn’t look smooth.
In that sense, I could even see why Elite are a bit on the fence about updating. They surely would be reluctant to smooth their ‘beast of accuracy’, to them it would be like putting a speed limiter on a Ferrari just because all other cars can’t go that fast.
That said, I definitely also wasn’t happy with the power roller-coaster I sometimes encountered with the Drivo in ERG mode in Zwift, but I have to say that this experience has been much better since I came back to TR. So it might even be a client software thing…
Nevertheless, given the stir this is causing Elite should at the very least OPENLY communicate about this and provide us with their view on things.
I have never adjusted the smoothing in the Elite app. I don’t care how smooth the drawn line appears in a chart — does the app ‘smoothing’ number noticeably affect the pedal feel ? Maybe I’ll give that a whirl.
I’ve had my Drivo September 2017.
If it goes up and down by a few watts and averages what it should average over and interval - isn’t that all that is needed?
Agreed the Drivo does look a bit twitchy (more on some than others) - but who can pedal with exactly the same power on every crank rotation?
And what is better - real time, slightly variable power, where your variations are there for all to see, or an artificially smoothed line that doesn’t make you more conscious of your own inconsistencies?
Ive generally found that my Drivo does ok when I use its power meter but in power match its a lot worse, always over shoots interval and then over corrects anything less than a 1 min interval is way under the prescribed watts.
theoretically it should not (according to the interwebs - but I guess only Elite knows whether this does something more that just smooth the output), but then there is something in between the number we see on the screen and our legs: our brain. We’re all very much slaves to that. Brain sees too high wattage on screen, micro-adjusts pedal stroke, but brain doesn’t take into account number is slightly delayed etc. and off you go into this seesaw pattern.
anecdotally I find I can achieve less wiggly power lines when not constantly watching the power number on the TR interface.
Hence the reason that I suggest people ignore power and focus purely on their cadence when using ERG mode. With a good app (TR = check) and a good trainer (Drivo = ?) you should be able to pedal at a steady cadence and let the system hit the power targets.
I see too many people trying to ‘help’ the system and hit power targets by changing their cadence and power input. But that simply leads to more issues than it solves.
Hold as steady and predictable of a cadence as you can and take what you get.
Exactly. Here are some data from a ride last night with a Direto. 0 power smoothing set in the myEtraining app.
I personally haven’t had any issues with my Elite trainer. Using Bluetooth with the macOS app.
“Control Systems - and how not to design them” would be my takeaway.
@GPLama is the man
Ive been in contact with Elite support as my Drivo has just started making a ‘knocking’ sound at low watts. Have sent the youtube link to them and asked if a firmware update will be issued. Will keep you posted…
fwiw, I put the power smoothing in the Elite app to 0 yesterday (was 6 before). It did not make any difference in ride feel. My powerline was obviously more squiggly than before. All targets met, did not feel any oscillations in the trainer.
However, when looking at the graph now in my past rides, it actually looks like TR is effectively applying some smoothing to the powerline.
Conclusion: being data freaks we 1) have a tendency to overanalyze things, 2) seem to live under the assumption that different software all treat our data identically, and 3) seem to have a strong believe that what any software presents to us is the ‘raw’ data. 2) and 3) are definitely false beliefs. Therefore, although it’s a difficult process (being a datascientist by training) I’m trying to convince myself that the truth is somewhere in the average across all datasites I’m looking at. What’s more, I’m actually moving away a bit from looking too much into these metrics and instead use them as ballpark guidance numbers while trying to rely much more on feel. My drivo feels right when I ride it, I feel that doing TR workouts are making me stronger, I feel that things are going easier when riding outside… It’s a difficult process, but I’ll get there at some point (after I’m done browsing for a power meter for my outside bike ).
For me its the time it takes to get to a wattage targets (avg. 8 secs) - if you gun it there are the overshoots and recovery issues which is the oscillations I see, I can ride around that a bit by keeping to a constant 85-95rpm cadence throughout, i don’t really care too much about how smooth a graph looks, if the interval targets are met, which on shorter (30-40secs or less) is not the often the case, so i switch off ERG for those… I used the ‘intro workout smart’ and played around with both the elite smoothing and TR smoothing settings, not much of a variance to be honest, 5 in TR and 2 in Elite. I got my Drivo at a really good price (£600 halfords) in Nov and really like the it after my BKOOL… so i’m not too upset but do wish elite would sort out their lack of F/W update capability, no idea which version i have or how would change it… how come Drivo II is better (according to GPLama) is that F/W or H/W… we may never know !
As people have said it’s too easy to get fixated on the numbers, I don’t race, doing TR on my Drivo is making me faster. However, I think the ramp test in ERG may actually be an issue.
I started a thread on whether to use 8min test or ramp test around Christmas as I had another disappointing ramp test. Ramp test I got FTP of 248 which I felt was low so next day average 317 and 310 for the 2 x 8 min tests. When looking at the data from the ramp test on 3 separate occasions I did 380 watts for 10 seconds. The first of which was when it should have been at 180 watts just 10 mins in. Is this why my legs exploded?
That’s definitely not desired for the type of testing. It’s essentially mini sprints and going to take a toll if that happens at the start of each step. Definitely not something I would want in my test.
never had that spike issue with ramp tests on my Drivo, the increments are small really so maybe there was a different issue, drop out which it tried to recover from ? … (PS: I also feel the operation is less solid via Ant+ versus Bluetooth… i had no issues with my previous turbo on Ant+ (same position etc.) yet this one does comes up with message “pedal to resume” for a mere second a few times in a 1hr workout… i switched to iPad and its been fine.
That does sound like dropouts. Is cadence supplied by the Drivo? The “pedal to resume” appears to result from a dropout in cadence input from the trainer to the app.
Thinking of picking up a Direto 1 second hand thats on a good deal.
I know there are some issues with this trainer but would I be silly to get this rather than an old kickr? Kickr1 second hand is roughly $100-$140 more