Latency of the direto in ERG mode with Zwift

[raphnou]

Hello Elite,

Here is the graph of an easy 30x30 sceance in Zwift (155W / 210W) @ 100rpm

each point corresponds to one second

We can determine the Direto latency time of about 7/8 seconds at the ramp up, there is also a latency at power release.

With a precise power sensor, the latency is a decisive point to compare the trainers between them.

Why manufacturers do not communicate this information that can be decisive for structured training and predefined by the application used.

[raphnou]

Hi,

I repeat my graph above, to add a precision.

I took the selection # 1 which corresponds to the points in the Direto precision range which is +/- 2.5% or according to the requested power 210w. The range of points is between 204.75w - 215.25w

First observation, the average power of the selection # 1 is 210W on this range, for a requested power of 210W, it is really beautiful precision.

Unfortunately for this interval of 30 seconds, the points of the selection # 1 is only 25 seconds instead of 30 seconds.
No problem just add 5 seconds to the interval.

But if I decide to enlarge the selection to extend it to 30 seconds on the best points of the requested power in this range of the selection # 2
So the average is 206w, and more exactly 206.64w on 30 seconds for 210W asked, it's still in the precision range announced and it's excellent.

I looked at other intervals and it's still very good.

Finally on this test of 30x30 to 110w/210w, the latency is not problematic, it just results in erroneous graphics in garmin connect for example, which must be corrected on the timing.

It will be necessary to check on higher powers if the latency does not increase

[Elite Admin S]

Hi

Direto adjusts resistance using a stepping motor.
When requested power/slope changes, the motor moves a magnetic arm to match the requested value.
Of course this movement requires some seconds. Also, the software will add other seconds of delay (it has to verify the data, see the variation, send a command to the motor). At the end, the entire process can last some seconds (depending also on how big is the resistance variation - and consequently, the movement to do). A 5-10 seconds of delay is quite common.

[raphnou]

Hi

Direto adjusts resistance using a stepping motor.
When requested power/slope changes, the motor moves a magnetic arm to match the requested value.
Of course this movement requires some seconds. Also, the software will add other seconds of delay (it has to verify the data, see the variation, send a command to the motor). At the end, the entire process can last some seconds (depending also on how big is the resistance variation - and consequently, the movement to do). A 5-10 seconds of delay is quite common.

This is common to all Direto or Drivo is it more responsive?
I thought I read that the Drivo had 4 seconds of latency ...

[Elite Admin S]

It's not possible to tell how much is the latency, as it depends on how much is the resistance variations (the bigger the variation, the longer the the delay).
The motor in the Drivo is a little slower - BUT the power range is higher, so it is necessary less movement to make the same power variation.
In the end, they are quite similar.

[shop]

Please excuse me for hijacking this discussion, but i don't want to open an additional discussion, because my question aims in the same direction.

In the last weeks I did a trainingsplan on Zwift. When it came to short intervall trainings I always got problems to reach and hold the given power. If I am paddling at e.g. 180W and then a short intervall (10sec or 20sec) of e.g. 260W starts, I normally switch into a harder gear and increase my cadence, but my Direto does not react as expected, the pedalling feels pretty easy and the resistance does not increase significantly - in this time the shown power output in Zwift jumps up and down so you got almost no chance to hold it at the given level. The awful thing is, that the resistance increases very dramatically when the short intervall is over and the given power in Zwift drops to e.g. 140W and decrease my cadence while shifting in an easier gear.

It is frustrating that these intervall trainings does not work well with the Direto, because this kind of training is very efficient and I would like to integrate more of these intervall trainings in my trainings plan.

Is there a way to optimize the configuration of the Direto (and/or Zwift) to better fulfill these requirements ? Is there a chance that future firmware updates are going to solve/reduce these issues or is it a pure hardware problem. Does the Drivo II show a better performance under these conditions ?

[Elite Admin S]

@shop,

as we explained before, the Direto has an internal stepping motor that moves a magnetic arm to modify the trainer resistance.
This means that it takes some seconds to moving into the requested position. The bigger the power gap, the bigger the movement of the motor (and the delay). Also, this time is used to avoid overshoots on power reading.

Also, you have to consider that the software itself has its "time" (it must acknowledge that the first interval is finished, verify the power of the following interval, calculate the resistance command to send to the trainer and finally send the message). Also, the communication itself has its own delays (due to interference for example).

If you put together these two facts, you can easily understand that you can have several seconds delay - and a 10sec interval can be really too short.
Consider that 10 seconds are really a very short time also for you, to adapt your pedaling (and cadence) to the new power interval.
What I can suggest is to use bigger interval (let's say 30 seconds) and use that delay time to change gear and reach the proper cadence.


About Drivo II, indeed it is more fast than Direto. But it still needs some seconds to move for big excursions and it still needs some time to avoid overshoots on power. So, even with this, it may be that 10 seconds are still too little (depending on the power gap and speed of course).

I hope to have been clear enough

[shop]

@Elite Admin S

Thank you for the detailed explanation.

[sergey.maximenko]

If I didn't know it is Direto on the first graph I'd say it is my Drivo. The same 7 sec. delays to ramp up resistance. Slightly shorter delays when releasing resistance.
Maybe some Drivo hardware revisions have even slower stepping motor or control board.
Anyway it is so slow in ERG mode that it fails shorter Zwift intervals