Author Topic: AcuRite Learning Mode  (Read 1015 times)

0 Members and 1 Guest are viewing this topic.

Offline Neil_O

  • Member
  • *
  • Posts: 41
AcuRite Learning Mode
« on: December 27, 2018, 02:29:47 PM »
I have a question regarding the pressure adjustments made on an Atlas display after the Learning Mode is completed.

Here is my situation:  I have an Atlas outdoor unit, an Access connected to MyAcurite, and a municipal airport about 7-8 miles away that reports weather conditions.  I have entered the elevation of my location unit into MyAcurite.  It has been >45 days since the Learning Mode for the Atlas display was started, so it should be completed.  I have the display running on an uninterrupted power supply and I have a whole-house generator so there is no way that the display lost power which would have caused to Learning Mode to restart.  The Atlas display and the Access are in the same room, approximately 15 feet away from each other.

Historically, I get excellent agreement between the pressure readings from the Access unit and the airport.  About 40 days after the start of the Learning Mode, I needed only a small adjustment to get agreement between the Atlas display and the Access pressure reading.The good agreement between the display and the Access occurs most of the time.  However, whenever we are experiencing either high pressure or low pressure excursions from 29.92 in Hg, the agreement is lost.  The display reads high during low pressure conditions (< about 29.50 in Hg) and it reads low during high pressure conditions (> about 30.30 in. Hg).  It almost seems like the Learning Mode, in addition to correcting for elevation, is also attenuating the measurement before reporting it.  Does anyone know enough about the Learning Mode to know if this is possible?

Offline CW2274

  • Forecaster
  • *****
  • Posts: 4575
    • Conditions @ CW2274
Re: AcuRite Learning Mode
« Reply #1 on: December 27, 2018, 02:36:09 PM »
The airport is using the altimeter setting and I'm not sure, but I don't believe the Atlas is altimeter capable. My VP2 console, which also is not altimeter capable reacts exactly the same way.

Offline nincehelser

  • Forecaster
  • *****
  • Posts: 3105
Re: AcuRite Learning Mode
« Reply #2 on: December 27, 2018, 03:52:11 PM »
"Learning Mode", only used for the displays, is scaling against 29.92.  So as you do go to the extremes, there likely will be some "distortion", with the accuracy greatest near the center.

I'd imagine how much distortion is a function of how far you are above sea level, but that's just a guess looking at the equation.

Offline nincehelser

  • Forecaster
  • *****
  • Posts: 3105
Re: AcuRite Learning Mode
« Reply #3 on: December 27, 2018, 03:57:14 PM »
The airport is using the altimeter setting and I'm not sure, but I don't believe the Atlas is altimeter capable. My VP2 console, which also is not altimeter capable reacts exactly the same way.

The Access does altimeter.  The display does not have altimeter, but after the learning mode is over (well, anytime actually) you can manually adjust the reading +/- 10 inches of mercury as needed.

I'm not sure if manually adjusting the display turns the "learning mode" off or not.  That would be the best thing to do, but the Atlas documentation is pretty sparse on the subject.
« Last Edit: December 27, 2018, 04:00:32 PM by nincehelser »

Offline Neil_O

  • Member
  • *
  • Posts: 41
Re: AcuRite Learning Mode
« Reply #4 on: December 27, 2018, 06:53:51 PM »
"Learning Mode", only used for the displays, is scaling against 29.92.  So as you do go to the extremes, there likely will be some "distortion", with the accuracy greatest near the center.

I'd imagine how much distortion is a function of how far you are above sea level, but that's just a guess looking at the equation.

So does that mean that the Learning Mode does more than just estimate elevation above sea level?  I assumed that the absolute pressure measurement on the display was the same as the one in the Access unit and was then adjusted based on the elevation estimated from the Learning Mode.  Is that not correct?

Offline nincehelser

  • Forecaster
  • *****
  • Posts: 3105
Re: AcuRite Learning Mode
« Reply #5 on: December 27, 2018, 07:52:21 PM »
"Learning Mode", only used for the displays, is scaling against 29.92.  So as you do go to the extremes, there likely will be some "distortion", with the accuracy greatest near the center.

I'd imagine how much distortion is a function of how far you are above sea level, but that's just a guess looking at the equation.

So does that mean that the Learning Mode does more than just estimate elevation above sea level?  I assumed that the absolute pressure measurement on the display was the same as the one in the Access unit and was then adjusted based on the elevation estimated from the Learning Mode.  Is that not correct?

The "learning method" doesn't actually estimate an elevation.  It learns your average pressure through sampling for a month, then uses that average scaled against 29.92 to adjust the current absolute pressure reading. 

It doesn't have any interaction with the Access.  The displays use their own internal baro sensors for the calculation.

The actual equation is printed in an early 5n1 spec sheet, but I don't have it handy on this computer I'm using.  I'm sure I've posted it somewhere in this forum, though.


Offline Neil_O

  • Member
  • *
  • Posts: 41
Re: AcuRite Learning Mode
« Reply #6 on: December 27, 2018, 08:58:12 PM »
"Learning Mode", only used for the displays, is scaling against 29.92.  So as you do go to the extremes, there likely will be some "distortion", with the accuracy greatest near the center.

I'd imagine how much distortion is a function of how far you are above sea level, but that's just a guess looking at the equation.

So does that mean that the Learning Mode does more than just estimate elevation above sea level?  I assumed that the absolute pressure measurement on the display was the same as the one in the Access unit and was then adjusted based on the elevation estimated from the Learning Mode.  Is that not correct?

The "learning method" doesn't actually estimate an elevation.  It learns your average pressure through sampling for a month, then uses that average scaled against 29.92 to adjust the current absolute pressure reading. 

It doesn't have any interaction with the Access.  The displays use their own internal baro sensors for the calculation.

The actual equation is printed in an early 5n1 spec sheet, but I don't have it handy on this computer I'm using.  I'm sure I've posted it somewhere in this forum, though.

Based on this definition, I made the assumption that the Learning Mode is calculating altitude:

https://www.acurite.com/learn/glossary/learning-mode

"Learning Mode is a time span during which AcuRite products utilize a unique algorithm to analyze changes in pressure and temperature over time to determine your altitude." seems pretty unambiguous.  Does Learning Mode actually work this way?

Thanks.

Offline nincehelser

  • Forecaster
  • *****
  • Posts: 3105
Re: AcuRite Learning Mode
« Reply #7 on: December 27, 2018, 09:00:31 PM »
"Learning Mode", only used for the displays, is scaling against 29.92.  So as you do go to the extremes, there likely will be some "distortion", with the accuracy greatest near the center.

I'd imagine how much distortion is a function of how far you are above sea level, but that's just a guess looking at the equation.

So does that mean that the Learning Mode does more than just estimate elevation above sea level?  I assumed that the absolute pressure measurement on the display was the same as the one in the Access unit and was then adjusted based on the elevation estimated from the Learning Mode.  Is that not correct?

The "learning method" doesn't actually estimate an elevation.  It learns your average pressure through sampling for a month, then uses that average scaled against 29.92 to adjust the current absolute pressure reading. 

It doesn't have any interaction with the Access.  The displays use their own internal baro sensors for the calculation.

The actual equation is printed in an early 5n1 spec sheet, but I don't have it handy on this computer I'm using.  I'm sure I've posted it somewhere in this forum, though.

Based on this definition, I made the assumption that the Learning Mode is calculating altitude:

https://www.acurite.com/learn/glossary/learning-mode

"Learning Mode is a time span during which AcuRite products utilize a unique algorithm to analyze changes in pressure and temperature over time to determine your altitude." seems pretty unambiguous.  Does Learning Mode actually work this way?

Thanks.

Not according to Acurite's published equation in the spec sheet.  There's nothing about determining altitude or any temperature variable.

Offline Neil_O

  • Member
  • *
  • Posts: 41
Re: AcuRite Learning Mode
« Reply #8 on: December 27, 2018, 09:03:44 PM »
"Learning Mode", only used for the displays, is scaling against 29.92.  So as you do go to the extremes, there likely will be some "distortion", with the accuracy greatest near the center.

I'd imagine how much distortion is a function of how far you are above sea level, but that's just a guess looking at the equation.

So does that mean that the Learning Mode does more than just estimate elevation above sea level?  I assumed that the absolute pressure measurement on the display was the same as the one in the Access unit and was then adjusted based on the elevation estimated from the Learning Mode.  Is that not correct?

The "learning method" doesn't actually estimate an elevation.  It learns your average pressure through sampling for a month, then uses that average scaled against 29.92 to adjust the current absolute pressure reading. 

It doesn't have any interaction with the Access.  The displays use their own internal baro sensors for the calculation.

The actual equation is printed in an early 5n1 spec sheet, but I don't have it handy on this computer I'm using.  I'm sure I've posted it somewhere in this forum, though.

Based on this definition, I made the assumption that the Learning Mode is calculating altitude:

https://www.acurite.com/learn/glossary/learning-mode

"Learning Mode is a time span during which AcuRite products utilize a unique algorithm to analyze changes in pressure and temperature over time to determine your altitude." seems pretty unambiguous.  Does Learning Mode actually work this way?

Thanks.

Not according to Acurite's published equation in the spec sheet.  There's nothing about determining altitude or any temperature variable.

Hmmm.  Interesting inconsistency since the definition I posted is also on the AcuRite web site.


Offline nincehelser

  • Forecaster
  • *****
  • Posts: 3105
Re: AcuRite Learning Mode
« Reply #9 on: December 27, 2018, 09:05:52 PM »
"Learning Mode", only used for the displays, is scaling against 29.92.  So as you do go to the extremes, there likely will be some "distortion", with the accuracy greatest near the center.

I'd imagine how much distortion is a function of how far you are above sea level, but that's just a guess looking at the equation.

So does that mean that the Learning Mode does more than just estimate elevation above sea level?  I assumed that the absolute pressure measurement on the display was the same as the one in the Access unit and was then adjusted based on the elevation estimated from the Learning Mode.  Is that not correct?

The "learning method" doesn't actually estimate an elevation.  It learns your average pressure through sampling for a month, then uses that average scaled against 29.92 to adjust the current absolute pressure reading. 

It doesn't have any interaction with the Access.  The displays use their own internal baro sensors for the calculation.

The actual equation is printed in an early 5n1 spec sheet, but I don't have it handy on this computer I'm using.  I'm sure I've posted it somewhere in this forum, though.

Based on this definition, I made the assumption that the Learning Mode is calculating altitude:

https://www.acurite.com/learn/glossary/learning-mode

"Learning Mode is a time span during which AcuRite products utilize a unique algorithm to analyze changes in pressure and temperature over time to determine your altitude." seems pretty unambiguous.  Does Learning Mode actually work this way?

Thanks.

Not according to Acurite's published equation in the spec sheet.  There's nothing about determining altitude or any temperature variable.

Hmmm.  Interesting inconsistency since the definition I posted is also on the AcuRite web site.

You have to take the marketing materials with a grain of salt.  I'll see if I can dig up the formula posted here.


Offline nincehelser

  • Forecaster
  • *****
  • Posts: 3105
Re: AcuRite Learning Mode
« Reply #10 on: December 27, 2018, 09:34:35 PM »
Hmmm.  Interesting inconsistency since the definition I posted is also on the AcuRite web site.

Here we go:

 [ You are not allowed to view attachments ]

Offline Neil_O

  • Member
  • *
  • Posts: 41
Re: AcuRite Learning Mode
« Reply #11 on: December 28, 2018, 09:18:27 AM »
Hmmm.  Interesting inconsistency since the definition I posted is also on the AcuRite web site.

Here we go:

 [ You are not allowed to view attachments ]

Thanks for finding this and sending it along.  That's quite a basic method for pressure adjustment - nothing particularly complicated.  Interestingly, the form of the equation does seem to allow for what I am seeing.  Because the equation adjustment is basically done as a fraction or percentage (29.92/average pressure during Learning Mode) and the calibration adjustment on the display is done as an absolution number (in. Hg), the correction will tend to slightly under-report pressure at high pressures and over-predict pressures at low pressures.  Playing with the numbers, I wouldn't expect the magnitude of attenuation that I'm seeing, but the trends now make sense.  It also supports your original statement that that any distortion caused by this correction would be most noticeable at the extremes.  Interestingly, my location is only about 250 feet above sea level.

Offline havtrail

  • Contributor
  • ***
  • Posts: 129
    • Haverford Weather Station
Re: AcuRite Learning Mode
« Reply #12 on: December 28, 2018, 10:56:37 AM »
Hmmm.  Interesting inconsistency since the definition I posted is also on the AcuRite web site.

Here we go:

 [ You are not allowed to view attachments ]

Thanks for this - very interesting. I have had the suspicion my barometer (not an Acurite PWS) has been reading systematically low, between 1 and 2 tenths of an inch. Armed with this average of 29.92, I just looked at the last 2 quarters of data (every 5 minutes), with over 22,000 readings each. One quarter averaged 29.75 and the other 29.83.

Rich K.
« Last Edit: December 28, 2018, 10:59:22 AM by havtrail »
Onset HOBO RX3003 Cellular
https://weather.havtrail.com
WU KPAHAVER17

Offline BeaverMeadow

  • Senior Contributor
  • ****
  • Posts: 223
Re: AcuRite Learning Mode
« Reply #13 on: December 28, 2018, 11:10:28 AM »
 Too bad Acurite didn't eliminate this feature in their new touch screen Atlas display. Entering accurate altitude info is easy to determine thanks to gps, google or else via a USGS topographic map. I guess redesigning features on the display is something that Acurite did not want to get into, or else they are happy with the current, somewhat inaccurate method of determining pressure-- since it's the trend line that matters. 

Offline nincehelser

  • Forecaster
  • *****
  • Posts: 3105
Re: AcuRite Learning Mode
« Reply #14 on: December 28, 2018, 11:50:13 AM »
Too bad Acurite didn't eliminate this feature in their new touch screen Atlas display. Entering accurate altitude info is easy to determine thanks to gps, google or else via a USGS topographic map. I guess redesigning features on the display is something that Acurite did not want to get into, or else they are happy with the current, somewhat inaccurate method of determining pressure-- since it's the trend line that matters.

Just ignore it then, and make a manual adjustment.

Remember, this is just for the LOCAL display value.  Nothing else.

Anything reported doesn't use this at all.  It's been that way for years now.

Offline CW2274

  • Forecaster
  • *****
  • Posts: 4575
    • Conditions @ CW2274
Re: AcuRite Learning Mode
« Reply #15 on: December 28, 2018, 02:44:10 PM »
Hmmm.  Interesting inconsistency since the definition I posted is also on the AcuRite web site.

Here we go:

 [ You are not allowed to view attachments ]

 Playing with the numbers, I wouldn't expect the magnitude of attenuation that I'm seeing, but the trends now make sense.  It also supports your original statement that that any distortion caused by this correction would be most noticeable at the extremes.  Interestingly, my location is only about 250 feet above sea level.
Since it has been clarified that your Atlas console does not support the altimeter, you can play with the numbers all you want, but the best you'll do is set it with the airport around 29.92" and deal with the extremes at the top and bottom. If you want a more fair comparison, you'll need a pressure measurement that's not the altimeter setting (i.e. MSLP)...or buy a console that does measure it.

Offline nincehelser

  • Forecaster
  • *****
  • Posts: 3105
Re: AcuRite Learning Mode
« Reply #16 on: December 28, 2018, 04:04:04 PM »
Hmmm.  Interesting inconsistency since the definition I posted is also on the AcuRite web site.

Here we go:

 [ You are not allowed to view attachments ]

 Playing with the numbers, I wouldn't expect the magnitude of attenuation that I'm seeing, but the trends now make sense.  It also supports your original statement that that any distortion caused by this correction would be most noticeable at the extremes.  Interestingly, my location is only about 250 feet above sea level.
Since it has been clarified that your Atlas console does not support the altimeter, you can play with the numbers all you want, but the best you'll do is set it with the airport around 29.92" and deal with the extremes at the top and bottom. If you want a more fair comparison, you'll need a pressure measurement that's not the altimeter setting (i.e. MSLP)...or buy a console that does measure it.

What extremes, and why wait for 29.92?

Offline CW2274

  • Forecaster
  • *****
  • Posts: 4575
    • Conditions @ CW2274
Re: AcuRite Learning Mode
« Reply #17 on: December 28, 2018, 04:19:24 PM »
Hmmm.  Interesting inconsistency since the definition I posted is also on the AcuRite web site.

Here we go:

 [ You are not allowed to view attachments ]

 Playing with the numbers, I wouldn't expect the magnitude of attenuation that I'm seeing, but the trends now make sense.  It also supports your original statement that that any distortion caused by this correction would be most noticeable at the extremes.  Interestingly, my location is only about 250 feet above sea level.
Since it has been clarified that your Atlas console does not support the altimeter, you can play with the numbers all you want, but the best you'll do is set it with the airport around 29.92" and deal with the extremes at the top and bottom. If you want a more fair comparison, you'll need a pressure measurement that's not the altimeter setting (i.e. MSLP)...or buy a console that does measure it.

What extremes, and why wait for 29.92?
29.92" of mercury is standard barometric pressure and that which is higher or lower is considered just that. As the altimeter setting gets further from 29.92 (if that's where you set your console in the first place) the error of the altimeter setting and that of a non-altimeter capable console becomes greater with the spread. As I stated above, my VP2 console reacted exactly as the OP stated originally, as it's not altimeter capable either :roll:. I personally prefer dealing with the altimeter, so as a fix I bought a Vue console which does extrapolate the altimeter. Problem solved...for me.

Offline nincehelser

  • Forecaster
  • *****
  • Posts: 3105
Re: AcuRite Learning Mode
« Reply #18 on: December 28, 2018, 08:51:20 PM »

What extremes, and why wait for 29.92?
29.92" of mercury is standard barometric pressure and that which is higher or lower is considered just that. As the altimeter setting gets further from 29.92 (if that's where you set your console in the first place) the error of the altimeter setting and that of a non-altimeter capable console becomes greater with the spread. As I stated above, my VP2 console reacted exactly as the OP stated originally, as it's not altimeter capable either :roll:. I personally prefer dealing with the altimeter, so as a fix I bought a Vue console which does extrapolate the altimeter. Problem solved...for me.

OK.  I wasn't sure what mode you were talking about.

In the "Learning Mode", there's a superstition going around that you should wait to 29.92 before starting the learning period.  (Acurite support actually perpetuates it).  Anyway, it doesn't matter mathematically.  The averaging levels everything out eventually.

Most Acurite users usually have a SmartHUB or Access which does have altimeter.  For those who want everything to "match", I suggest adjusting the display barometer to match the SmartHUB/Access altimeter reading.  That seems to keep them reading pretty close over time, or at least enough that they are happy.

That works well for the Access, but some SmartHUBs tend to have a fairly noticeable fixed bias in the baro reading.   The only way to fix that under myAcurite is to fiddle with the the elevation reading to get an accurate altimeter reading.  However, since SmartHUBs are going away, that issue is dying off.
« Last Edit: December 28, 2018, 08:56:11 PM by nincehelser »

Offline CW2274

  • Forecaster
  • *****
  • Posts: 4575
    • Conditions @ CW2274
Re: AcuRite Learning Mode
« Reply #19 on: December 29, 2018, 12:46:18 AM »

What extremes, and why wait for 29.92?
29.92" of mercury is standard barometric pressure and that which is higher or lower is considered just that. As the altimeter setting gets further from 29.92 (if that's where you set your console in the first place) the error of the altimeter setting and that of a non-altimeter capable console becomes greater with the spread. As I stated above, my VP2 console reacted exactly as the OP stated originally, as it's not altimeter capable either :roll:. I personally prefer dealing with the altimeter, so as a fix I bought a Vue console which does extrapolate the altimeter. Problem solved...for me.
OK.  I wasn't sure what mode you were talking about.
Just for clarification, I have no "mode". I select what pressure measurement I wish to display, enter my altitude MSL, and hit "enter". Done.
The bottom line is comparing an altimeter to a console that's not capable will be inaccurate compared to something that's a comparable measurement. It's simply nothing more than apples to apples.

Offline nincehelser

  • Forecaster
  • *****
  • Posts: 3105
Re: AcuRite Learning Mode
« Reply #20 on: December 29, 2018, 01:16:58 AM »

What extremes, and why wait for 29.92?
29.92" of mercury is standard barometric pressure and that which is higher or lower is considered just that. As the altimeter setting gets further from 29.92 (if that's where you set your console in the first place) the error of the altimeter setting and that of a non-altimeter capable console becomes greater with the spread. As I stated above, my VP2 console reacted exactly as the OP stated originally, as it's not altimeter capable either :roll:. I personally prefer dealing with the altimeter, so as a fix I bought a Vue console which does extrapolate the altimeter. Problem solved...for me.
OK.  I wasn't sure what mode you were talking about.
Just for clarification, I have no "mode". I select what pressure measurement I wish to display, enter my altitude MSL, and hit "enter". Done.
The bottom line is comparing an altimeter to a console that's not capable will be inaccurate compared to something that's a comparable measurement. It's simply nothing more than apples to apples.

Meh.  It's plenty accurate for a device that isn't reporting data anywhere except your eyeball.





Offline CW2274

  • Forecaster
  • *****
  • Posts: 4575
    • Conditions @ CW2274
Re: AcuRite Learning Mode
« Reply #21 on: December 29, 2018, 02:23:23 AM »

What extremes, and why wait for 29.92?
29.92" of mercury is standard barometric pressure and that which is higher or lower is considered just that. As the altimeter setting gets further from 29.92 (if that's where you set your console in the first place) the error of the altimeter setting and that of a non-altimeter capable console becomes greater with the spread. As I stated above, my VP2 console reacted exactly as the OP stated originally, as it's not altimeter capable either :roll:. I personally prefer dealing with the altimeter, so as a fix I bought a Vue console which does extrapolate the altimeter. Problem solved...for me.
OK.  I wasn't sure what mode you were talking about.
Just for clarification, I have no "mode". I select what pressure measurement I wish to display, enter my altitude MSL, and hit "enter". Done.
The bottom line is comparing an altimeter to a console that's not capable will be inaccurate compared to something that's a comparable measurement. It's simply nothing more than apples to apples.

Meh.  It's plenty accurate for a device that isn't reporting data anywhere except your eyeball.
Seriously? That's lame and has nothing to do with the OP's question of what's occurring, which I've supplied an informed answer.

BTW, it maybe just your "eyeball" that see your data, not mine.

Offline nincehelser

  • Forecaster
  • *****
  • Posts: 3105
Re: AcuRite Learning Mode
« Reply #22 on: December 29, 2018, 02:42:51 AM »
Seriously? That's lame and has nothing to do with the OP's question of what's occurring, which I've supplied an informed answer.

BTW, it maybe just your "eyeball" that see your data, not mine.

The OP's question was about Acurite's learning mode.

Your sophistry is tiresome.
« Last Edit: December 29, 2018, 02:44:25 AM by nincehelser »

Offline CW2274

  • Forecaster
  • *****
  • Posts: 4575
    • Conditions @ CW2274
Re: AcuRite Learning Mode
« Reply #23 on: December 29, 2018, 04:18:41 AM »
Seriously? That's lame and has nothing to do with the OP's question of what's occurring, which I've supplied an informed answer.

BTW, it maybe just your "eyeball" that see your data, not mine.
Your sophistry is tiresome.
Where?

Offline BeaverMeadow

  • Senior Contributor
  • ****
  • Posts: 223
Re: AcuRite Learning Mode
« Reply #24 on: December 29, 2018, 10:53:06 AM »
I tried changing my display's barometer compensation adjustment to the correct relative pressure but even at max comp it is still not in agreement (and never was).

Was it too difficult (or expensive) for Acurite to rectify this weakness in their Atlas display or do they think it is a brilliant feature that is too good to lose?

I understand that for some people entering altitude manually to match their location may be beyond their capabilities but I doubt it is much of a problem for anyone purchasing a 5in1 or an Atlas.

 

anything