Drivo Erg Mode

[chrisdaniels]

Hi

I am having problems with my Drivo in Erg mode for workouts - it seems to overcompensate a lot and so swings way above and below my target power. I am riding at a steady cadence and without Erg mode i can ride to a much more steady power than with it which is obviously wrong. Sometimes it has also failed to raise the resistance for an interval for a few seconds. I use Trainerroad via Ant- FEC to control the Drivo from a windows laptop.

Is there any configuration i should look at or a firmware update?

Thanks
Chris

[sdrye]

I have the same problem, and it's mentioned repeatedly in the comments on DC Rainmaker's long-term review of the Drivo. It has made ERG mode unusable for me. It's very severe in the Elite Real software (150w->250w becomes 150w->400+w when riding a Power/Time training program). MyETraining in Bluetooth Smart mode has about a 15% overshoot.

It's less apparent in Zwift/ANT, but still there at about half of the overshoot seen in Real. The Zwiftalizer log analysis tool shows that I'm not having problems with ANT+ dropouts.

Any feedback on when this will be resolved would be greatly appreciated.

[Elite Admin S]

The oscillations around the target power are normal during the training. Those oscillations are due to the power variations between one pedal stroke and another and they happen to any cyclist (even professional cyclists have the same behavior, although with lower amplitude variations).
These oscillations are therefore real. Following, a graph that shows a comparison between the power measured and displayed by the Drivo and the power measured with an ergometer:



You could think that the oscillations are due to the trainer resistance variations, but they aren’t.
The continuous variations of the resistance done by the trainer motor are necessary to adjust the brake to the microvariations of the cyclist speed (indeed, on varying of speed, the brake must be adjusted to keep the power constant).
We have verified that the same kind of oscillations occur even if the brake is blocked, and this demonstrates that the oscillations are intrinsic to the pedaling and are not due to the trainer behavior.
However, we are working on some trainer parameters to improve this aspect.

The differences between the pedal strokes are due to the cyclist and don’t depend from the training mode. Below you can see two charts related to two different training modes: Erg mode and Simulation mode; the registered power is compared to the power measured by a SRM ergometer during the two trainings. The charts show clearly the power variations between the pedal strokes (that is the difference between the power value of one pedal stroke and the following one). These values are almost identical in the two training modes and they are very close to the SRM measurements.
We’d like to point out that some programs can modify the power data, smoothing it to give a more pleasant visualization. In these cases, only the displayed data is modified while the developed power remains the same.
However, if a user prefers to see a more smoothed value it is possible to use the Drivo smoothing option in the advanced setting of the app.



Regarding the failed resistance variations for some seconds, it may be due to some disturbance.
The various commands are sent by the software as data transmission packets, which could not arrive to the trainer due to interferences. If a packet is lost and the trainer doesn’t receive it, it will wait the following packet and this will cause a delay in resistance variation.
We’ve verified that TrainerRoad sends a packet approximately every 10 seconds, so a lost packet would explain the delay you had experienced. In case it would happen more frequently, it could be helpful to use an extension usb cable to place the ANT+ dongle closer to the trainer (decreasing in this way the number of lost packets).

Finally, regarding the firmware, by now we don’t have any available update.

[Elite Admin S]

Traduzione in italiano del post precedente

Le oscillazioni intorno al valore di potenza target durante l’allenamento sono normali. Tali oscillazioni sono dovute alle variazioni di potenza tra una pedalata e l’altra e si verificano per qualsiasi cilcista (anche nel caso di ciclisti professionisti si verifica un comportamento analogo, anche se con variazioni di minore ampiezza).
Queste oscillazioni sono pertanto reali. Di seguito, un grafico che mostra la comparazione tra la potenza visualizzata e quella misurata da un ergometro, dove si può vedere come le oscillazioni di potenza sono assolutamente sovrapponibili:



Si potrebbe pensare che le oscillazioni siano dovute alle variazioni di resistenza del trainer, ma non è così.
Il continuo aggiustamento della resistenza da parte del motore è necessario per adattare il freno alle microvariazioni di velocità del ciclista (infatti, al variare della velocità il freno si deve adattare per mantenere la potenza costante al valore di target).
Abbiamo verificato che lo stesso tipo di oscillazioni sono presenti anche con il freno bloccato, e questo dimostra che le variazioni sono intrinseche nella pedalata e non sono dovute al comportamento del trainer.

Stiamo comunque lavorando su alcuni parametri del trainer per migliorare ulteriormente questo aspetto.

Le variazioni della potenza tra una pedalata e l’altra non dipendono dal tipo di allenamento ma del ciclista. Di seguito riportiamo i grafici di potenza del Drivo, confrontati con i dati misurati con un ergometro SRM, durante due diversi tipi di allenamento: Erg mode e Simulation mode.
Come si può vedere, nel grafico sono riportate le variazioni di potenza della pedalata (cioè la differenza del valore di potenza tra una pedalata e la successiva). Questi valori sono praticamente uguali nelle due modalità e simili ai dai dell’SRM.
Ricordiamo che a volte alcuni programmi possono alterare il dato, “smussando”, per una visualizzazione più piacevole. In questi casi, non cambia la potenza effettivamente sviluppata ma solo il dato visualizzato.
Cmq se un utilizzatore volesse è possibile abilitare un po’ di “smussamento” sul Drivo abilitando l’opzione relativa nelle impostazioni avanzate.



Per quanto riguarda il fatto che a volte la resistenza non varia per alcuni secondi, potrebbe essere dovuto a qualche disturbo ambientale.
I vari comandi sono infatti spediti dal software tramite pacchetti radio, che possono non arrivare al trainer a causa di interferenze. Nel caso in cui un pacchetto venisse perso e il trainer non lo ricevesse, esso attenderà il successivo pacchetto e questo causerà un ritardo nella variazione della resistenza.
Abbiamo verificato che TrainerRoad spedisce un pacchetto ogni 10 secondi circa e questo giustificherebbe il ritardo da lei descritto nel caso di perdita di un pacchetto. Qualora il problema si verificasse più frequentemente, potrebbe essere utile avvicinare la chiavetta ANT+ al trainer utilizzando una prolunga USB (e riducendo in tal modo la possibilità di pacchetti persi).

Infine, per quanto riguarda il firmware, non esistono per ora aggiornamenti disponibili.

[Elite Admin S]

@chrisdaniels,
@sdrye,

our customer care will contact you anyway to analyze further the situations.

[chrisdaniels]

Hi

Thanks for the reply, my issue is not the recording of the power showing variation - i realise that there will be many small variations in power - it is the response from the turbo to it. I am experiencing really quite big shifts when at near constant cadence and well within the power range i can comfortably sustain. These variations are causing the unit to be very hard to ride - 80W when at 300W is a very very big range to be swinging about

2086729-400x300.jpg (16.57 KiB) Viewed 110111 times

, i can ride much much closer than that with my power meter or a watt bike and have been riding with power for 4 or more now so am familiar with my readings.

[Steve Rode]

Yes, I see the same thing with the power swings and that is with a very steady cadence. I did not see this oscilation on my last smart trainer (I returned it and purchased the Drivo as a replacement), it was quite stable. I can also feel the difference in the resistance, its can get 10 watts from target plus or minus. Comparing my Power2Max data vs the Drivo power meter it shows the same oscillation. It seems to be less with the pedaling smoothness set on 5 instead of 1 or 2 (it came set at 2). This is using Zwift to drive my intervals, I have not tried other applications for intervals. This is the only complaint I have about the drivo for power management, everything else is very very nice.

[Elite Admin S]

@chrisdaniels,

80W are indeed too much. It is necessary to investigate further the situation.
We've already sent you an email, please check your mailbox.


@Steve,

10W means a very steady power, comparable to pro cyclist pedaling
The power smoothing option affects only the shown data and doesn't affect how the trainer brakes.
The higher the power smoothing value is, the more the power will be smoothed.

Regards

[Steve Rode]

Well I always appreciate compliments! However perhaps I wasn't clear - the issue is the trainer resistance varies in a wave like pattern slowly going up and down around the target instead of adjusting to the target output - the constant over correction makes the power output look like a sine wave. The issue is amplified at higher wattage targets/outputs. As pointed out, my previous smart trainer didn't do this, it corrected to the target output. I like that the Drivo is very quick to respond to grade changes using Zwift in Erg mode, however when doing intervals this sine wave response to a target output is not desireable fixing this would make this a perfect trainer in my opinion. My experiences are echoed by a number of people on bike forums and in reviews/blogs, so it's not unique to me.

The reason I adjusted the output power smoothing was I believed the power output might contribute to the over correction and having a more stable number from an average might reduce the response from the trainer ( a typical controls system damping technique) and it did have a noticeable difference comparing a setting of 1 to a setting of 5, but the wave action amplitude continued in either case as measured by my crank based power meter viewed on a Garmin computer.

[Elite Admin S]

Hi Steve

It could be interesting to check the results of your test, to see this sinusoidal behavior.
We will contact you via email to ask you further information and analyze your specific case.
Take a look at your mail box