Direto XR problems with ERG mode

[JohandeWit]

One more possible clue, smartbiketrainers has HW Revision 003, as per screenshot of his upgrado app. Which HW revision you guys have? I have HW Revision 001 on my Direto XR bought 3 weeks ago.

I'm on HW 001 as well.

Regards,
Johan

[enesto]

Enesto, I am tempted to carry out some more steady state cadence drills at different cadences for the same gear.
I could then take the data sets and run an FFT on them to look at what fundamentals are in the data sets
Starting to feel that I am doing Elite's work for them!!!

You're right!!!
You are testing as a R&D department...

BTW I agree with your analysis, it makes perfectly sense, but I think Elite could reply with some tech info (n° of meas per stroke, best cadence/gear ratio, etc, power ceiling/floor), just to avoid all this "debunking" effort where we all here are falling (and let me add, it seems to me that we all are doing our best to get the same best from our devices!)

Now it is shiny clear that the results we saw on youtube/reviews (and that influenced us to buy this trainer and not another one) are far away from what we are getting here, so I don't think we have hundreds of hypotesis on the table... Or we are a group of crazy people with a hard dependence of scattering charts, standard deviations and smoothing average values, or something is not working on our devices...

I renew the invite to Elite to share more tech details and to explain the source of such differences in performances and to share if there is a way to perform some HW check/analysis on our trainers.

@JohandeWit to get excel you can export the workout from zwift ( in *.fit I think) import the workout in golden cheetah and then export the csv... or use the elite software where you can find the native csv export for the activity.

HW revision 1 for me too.

Hi guys, see you!
E.

PS this is the thread with more replies of the whole DIRETO section in the forum... Probably it's not a case!!

[jhankey]

Enesto, watch this from lama's test of a 'watt bike' the other day.
Watch from 41 mins in to say 43 mins.
https://www.youtube.com/watch?v=NakWDwAmqDU&t=2869s

This has the same sort of ripple we get as compared to some of the other trainers. He's getting quite a watt range here in erg mode.

[guiperpt]

Quick update. Just got update 73, and on a first quick test of 5 minutes, it looks definitely better, at least from the etraining app keeping the 2 as power smooth setting (6 introduces a huge delay in everything). Around 200W, most of the time, the fluctuations are between 190 and 210 or so.

I Will do a better testing tomorrow in Zwift, but at a first glance, ERG mode looks good enough for my level of riding.

[JohandeWit]

Hi all,

I took some time to read in the Drivo/Drivo II thread where they reported similar issues with EGR mode. I noticed that early this year Elite supplied some of the Drivo users a JSON file that they can upload to their trainer using the Upgrado app.

read here: http://forum.elite-it.com/viewtopic.php?f=35&t=78&start=130

When you look at the file you will see that is sets some parameters on some addresses in the trainer. Based on the names I assume it's setting the PID controller parameters for the PID controller that Elite implemented in the trainers to maintain the power required.
I notice that users pryswilliams and 4169 report back good results using the updated parameters. The picture 4169 shows the behavior one would expect and they also report back that with these settings the trainer is quick enough to respond to 10 second power jump intervals. That is what I want.

I can not imagine that the Drivo core is much different from our Direto XR core, so I expect that in our trainers it should also be possible to adjust the parameters of the PID controller.

So @eliteadmins, would it not be possible to supply us (or me) with 2 JSON files; a file with adjusted settings that give the trainer more responsiveness, and two a file that reflects the default settings as they are currently used (so that we can revert back to the current situation).

I am willing to test this.

regards,
Johan

[jhankey]

Enesto, looks like Johandewit is another engineer.
Now, this could be fun. As for PID control, that would be interesting to see. I wonder if they have used digital filters. Time to dust off my understanding of Z transforms. ( That's about 30 years of dust mind!!!!)

On one of my bikes I have been using a powerpod power meter. It is a calculated method for power rather than a strain gauge. One of the things they have implemented is 'dynamic' smoothing. So when you are at a consistent effort you get a smooth power number. But when you stomp on it, it very quickly changes to a dynamic mode to give you near instant readings. Its not like you are dependant on hardware, you could have a number of software algorithms.

[enesto]

Enesto, looks like Johandewit is another engineer.
Now, this could be fun. As for PID control, that would be interesting to see. I wonder if they have used digital filters. Time to dust off my understanding of Z transforms. ( That's about 30 years of dust mind!!!!)

This thread is like an "EngineerDemic"!
I hope Elite will help us on this, and moreover will take a position by telling us that everything is ok or not with our devices.

Yesterday I made the workout 1 of zwift academy, smoothing 6 and cadence following the indication, 85 or 100 on different intervals.
Not perfect, especially in transition, better than before, for now this is the best I can do, I will say barely acceptable (but not considering the cost of 850 euro).

In the FB group there are people every week asking for the same things, and I'm forwarding them to this post.

Keep waiting (and testing, next step will be connect a drill to my cranckset and set it to 85rpm )

Cheers!

[guiperpt]

Enesto, looks like Johandewit is another engineer.
Now, this could be fun. As for PID control, that would be interesting to see. I wonder if they have used digital filters. Time to dust off my understanding of Z transforms. ( That's about 30 years of dust mind!!!!)

This thread is like an "EngineerDemic"!
I hope Elite will help us on this, and moreover will take a position by telling us that everything is ok or not with our devices.

Yesterday I made the workout 1 of zwift academy, smoothing 6 and cadence following the indication, 85 or 100 on different intervals.
Not perfect, especially in transition, better than before, for now this is the best I can do, I will say barely acceptable (but not considering the cost of 850 euro).

In the FB group there are people every week asking for the same things, and I'm forwarding them to this post.

Keep waiting (and testing, next step will be connect a drill to my cranckset and set it to 85rpm )

Cheers!

Are you fine with the smoothing set to 6? For me, it introduces a considerable amount of lag, especially notable in SIM mode.
Definitely, it helps the ERG mode smoothing the power numbers, but for SIM mode it's unusable. With smoothing on 2, the default, the lag between putting the power down and see it on Zwift is probably around 2 seconds, which is far from perfect, with smoothing 6 it's at least the double. You always overshoot a sprint/catchup.

[jhankey]

Did the zwift ftp ramp test yesterday, 20w increments per minute.
Still greater variation than I would want , but for each value I was hitting the correct average power all the way up to about 320w.
Throughout, my cadence increased with loading form 90 to 100 rpm. ( I just spin faster as the load goes up but the change is gradual enough for the direto cope with) On the last level 320w, as soon as it went to this level, the direto dropped the power and zwift called for more power. I increased the rpm significantly but the power just never recovered, it should have dramatically increased. Gear wise I was in something like 34/ 17.
I will retry the zwift short ftp ramp test again in the near future. Maybe use a bigger gear. Elite still have some work to do here.
Additionally, and slightly off topic... has anyone got their xr to pair to a wahoo bolt and get the correct wheel speed? whilst connected to zwift. ( Note: zwift connection bluetooth, wahoo connection ant+)

[BelgianFlandrien]

Hmm, I had the same impression on my last Zwift session where I did a sprint workout (GCN Sprint! workout: https://whatsonzwift.com/workouts/gcn/sprints-sprint/).
During the first low wattage block, the trainer was constantly showing 20watts under the target power, even when changing my rpm. After lowering 1 gear this seemed to be fixed. After that, everything was quite normal for the lower wattages (still more variation than I want for a unit this expensive tho) but for the 20sec sprint-blocks the unit was acting weird. The resistance kicked in correctly, but as soon as I reached the target wattage, the trainer lowered the power, forcing me to increase my rpm significantly to be able to reach the power-goal.