Pressure Wackiness (solved)

@dsj I’m really struggling to understand what’s going on here. As you can see from the chart, at 7:30 my pressure was 30.346, virtually identical to my Davis VP2 and surrounding, trusted stations (including NWS/FAA sites nearby). But then, all of a sudden, it was off to the races as the pressure increased to 30.54 for absolutely no reason. Why would the CL take over in a situation where every trusted station is reporting essentially the same 30.34-30.35 readings? What would be the purpose of increasing pressure to 30.54 which no station in my area is even close to…even ones I know are whacky.

At some point (my station has been up for 2 weeks now) doesn’t the CL process say, “Yup, that reading is essentially correct as I can see from trusted sources (my Davis is supposed to be one of them) and I’ll leave it alone”? There’s got to be a gap in my understanding of what’s going on here. BTW, my new Air behaves essentially the same way except for more false lightning strikes than my first Air. However, in general, my original Air seems to be doing somewhat better overall.

Dave, can you ‘splain’ this weird barometric pressure behavior and when, if ever, does this settle down? :wink:

1 Like

You keep mentioning the CL but I’m betting this has little to nothing to do with the continuing learning system.

Gary, you may be right, but I’m at a loss to explain it. My other Air exhibits this behavior too, but it’s newer and the curve is even steeper. The fact that it’s newer and has a steeper curve, led me to believe it was due to the CL system, since it was earlier in the learning process and the changes/adjustments might be more radical.

IOW, I’m searching for an explanation.

Here’s the graph from my newer Air, for the same time period, exhibiting an even more radical change. Again, at 7:20 it was identical to my first Air and both were bang on accurate. Then the fun began.


That may have to do with it needing calibrated. While calibration and the “continuing learning” system are related they are not the same.

Gary, that might make sense. I just looked at other WF units on Long Island that have been up for awhile, and they show very gentle pressure changes over this same time period as would be expected with our current weather. The interesting exception is a unit showing even more drastic changes than mine, but that’s a friend of mine who received his unit a few days ago.

I guess I would have thought that after 2 weeks my calibration would have settled down more than it has.

Still trying to understand the system. :slight_smile:

Calibration is applied over a LONG period of time.

The “continuing learning” system is code that runs on the WeatherFlow servers. This code cycles through EVERY station and examines the data. Then it decides if a calibration on a device is needed. Your calibrations are done in small amounts and it may be two weeks between each one.

OK, again, what you say makes sense. With that said, I’m seeing these relatively steep changes every day as oppose to every 2 weeks. I understand mine has been up for only 2 weeks, so I don’t know if these daily swings are to be expected at this stage of the ‘maturation process’.

What confuses me more is why the calibration process needs to overshoot the correct mark by such a degree only to eventually come back to correct ballpark readings.

First off, you may not have received a calibration.

But Gary, if so, then what’s causing that daily fluctuation? If it’s not calibration and not CL, then what is it? I have 2 units that do that as well as my friend’s Air that exhibits the same behavior. Something is causing these units to move from a correct reading to significantly overshoot the mark and then, hours later, return to a ballpark reading.

It’s the sensor reading a change. The question is, “Why is it reading a change?” Now, maybe you are not seeing this on other devices because they are slower to report sudden changes and balance the readings. I can give a dozen reasons why one shows rapid change and another won’t.

No, I can tell you with 100% certainty that it’s not a question of other stations not keeping up with the changes. I very carefully look at NWS sites with very expensive professional devices, properly sited, as well as FAA locations and trusted CWOP sites. All will show a fluctuation of perhaps .01-.02" during that same period and all reading within .02" of each other. They are in total agreement and can remain so for hours while my 2 Airs show this relatively large increase. In fact, the graph clearly indicates that both units rose to readings that were .2" or more above actual readings. In fact, subsequent actual readings at all trusted locations were falling a bit, certainly not exhibiting a significant rise in barometric pressure.

No, I’m convinced this must be something else. If you’d like to see what my friend’s new Air looks like, take a look at this insane increase in pressure today. He’s been getting this since he placed it in service a few days ago. It’s the same type of response, but to a much greater extent. Again, actual pressure at this point is decreasing. His rise in pressure is crazy. My increases when the unit was first placed in service was somewhat similar to his. There is something else going on. IMG_1307

One other thing. The hardware is different. So a consumer grade sensor is never going to perform as a commercial grade sensor. I doubt the Smart Weather Station will ever perform as a Vantage Pro…

Just like a Volkswagen Beetle will not preform like a F1 car.

Ken, give us a few days to take a deep look at pressure data from your two AIR units soon. This does not look like a CL issue but something’s definitely going on.

1 Like

Gary that may be, but being off my a bit is very different than this odd behavior.

Dave, thanks. If you can look at my friend’s data, that would be great. Same issue, but a larger magnitude.

It could simply be a hardware issue. My point, Don’t be so quick to place blame on the “continuing learning” system. Give WeatherFlow a chance to investigate.

Gary, I’m certainly not ‘placing blame’ on any facet of the system, I’m simply trying to understand what’s going on and what’s causing these fluctuations.

Can you post the wind graph for that same time period? I’d like to see what your wind was showing.

Here ya go: