What about Gradient Lag?

I wrote some in this thread last autumn, and like most of the writers I was very critical to Garmin about the gradient lag (and altitude lag, but the altitude, instant nor total, is not so important for me). I'm a little confused, this year I started to bicycle in May (I live in Sweden *brrr*) and then I thought this problem was solved, and in my opinion it was. Perhaps not perfect but way way better then how it was last autumn.

So, some are reporting that there is still a lag, up to 30s?
I just upgraded to 18.23 some minute ago and don't remember exactly what I had before, 17.something, I'm not in the BETA-program.
But with the 17.x version I think the gradient behavior was good, or perhaps acceptable. It was even detecting small quick changes in gradient, although it was shown some seconds after. But I can't say I reacted of some lag. I have an old 800 and now (with the 17.x version) I think 1040 was better than that, much better. 800 was perhaps as quick as 1040, or sometimes quicker, but 800 was flipping up and down. I think 1040 is a little afraid of showing incorrect gradient after all beating it got so it's thinking twice before showing the gradient, and that's fine with me.

I haven't read this thread since last autumn, I gave up about this. But was happy when I got out for some tours in May, and as I wrote I was thinking it was solved. So why does some people still have a big lag? Some BETA version? Not so good air as in Sweden?

Last autumn I was doing some testing and created a diagram showing altitude lags for 800, Fenix7 and 1040 on the same tour, I will do the same again with the new 1040 version. I'm right now struggling with creating an account on flickr to be able to post the diagram.

Here is the vid how Edge gives negative elevation readings when climbing. And a link to my workout, If activate Elev Corrections, 150m of total ascent difference
https://connect.garmin.com/modern/activity/11958403231
youtube.com/.../ANUqOyR4wGk

Several months back I would've agreed with you, but since beta firmware 17.?? the lag is down to 3-5 seconds for me. Regarding total elevation I've just enabled elevation correction for my last tour (total elevation 1500m) and got +400m.

BUT: Garmin themselves say that the elevation correction feature in Garmin Connect uses DEM data which isn't as accurate as barometric measuring. So, I've tried elevation correction with Strava, their's is based on a pool of recordings from users. Here, I've got a difference of just +30m what I consider within tolerance.

Also, total ascent and descent only differ about 1-2% on my rides. How do you calibrate your Edge? I use automatic barometric calibration at the start of each ride. 

Lastly I've looked at some of my older recordings with a Sigma ROX 11, elevation difference is negligible.

I've planned my tour with Komoot, here, too, elevation was nearly spot-on, meaning after the ride the Edge data corresponded with the plan.

To sum it up: A 3-5 seconds lag renders the device unusable for microclimbs but with anything else at least I don't see that much of a problem anymore. Although, I'd prefer no lag at all. Concerning elevation: With what I have at hand for comparion it looks okay.

And Garmin? They promised much more committed interaction and discussion starting with beta 17, but actually I don' recognise any change. Plus, Garmin Blake wrote some months ago they were working on something regarding the lag. No update yet. Very disappointing.

OliverS73 said:

since beta firmware 17.?? the lag is down to 3-5 seconds

There was one agile SW. But with 18.23 it is definitely not with in 3-5sec. For me it is now about 30-50sec.

Altimeter auto calibration 

Garmin watches work brilliantly even with microclimbs. Why Edge con not do the same is unknown

OliverS73 said:

They promised much more committed interaction and discussion starting with beta 17, but actually I don' recognise any change. Plus, Garmin Blake wrote some months ago they were working on something regarding the lag. No update yet. Very disappointing.

True

Lets see how the new 18.23 behaves.

I reposted this cause I figured out how to insert a picture...

I was trying to show the diagram directly, is that possible or it must be a link to flickr, someone knows?

Hm, the second half of your diagram is odd. During the first half the 1040 lags about some seconds like you said but ultimately recorded about the same values. But in the second half it's way off.

I had installed 18.23 for my last ride. Just from the feel I'd have said it's got even minimally snappier since 18.21b. - I don't know whether that influences  altitude measurement since it's supposed to be barometric, but I'm using the maximum GNSS setting (don't remember the name atm).

I know. I was testing the same tour some other times and that happen one of the time, but not as much as in the diagram, and one time it was accurate (but with the lag).
The "height" is a new bicycle bridge over a railway, perhaps the electric field is disturbing. But neither 800 nor Fenix7 behaves the same, none of the times I was testing. Perhaps 1040 is working more with DEM-data and the the bridge is new...
Or it's so simple that because of the big lag it decides to "cancel" the ascending and goes directly to the descending, strange.