[wxqc] Looking for help with Station - Odd shift in analysis

Steve Meyers stevem at samcon.com
Wed Feb 17 16:52:14 CST 2016


Chris
What I was referring to was a routine. program, mode, operation that can be
applied to the PEET station to ensure the unit is running properly. If you
are having these kind of problems with your station why not rule out
issues/problems with your hardware first. Then, if there are no problems
with the hardware you can move to calibration issues.

Steve

On Fri, Feb 5, 2016 at 6:50 PM, Chris Miller <c_miller_1 at hotmail.com> wrote:

> Ok, just to be sure I have it right in my head, I will step my way through
> the questions.
>
> If you mean for a hardware analysis that I have checked the hardware for
> erratic operation or mis-calibration, here is what I have done that way:
>
> For the temperature sensor, I have compared it to an electronic
> thermometer to look for gross errors and it seems to be good.   A few years
> ago I did replace the sensor board due to a low temp fault that would show
> up below –20C.
>
> For the humidity sensor, I noticed that it tracked poorly with expected
> humidity when I replaced the sensor board above and recalibrated it to
> match another humidity sensor.  it now tracks a lot closer under most
> conditions although I find it does limit out under high humidity a little
> earlier than expected.  I also corrected an assembly error with the sensor
> case and the sensor array screen which had the one side of the sensor case
> pressed against the tray not allowing both faces of the sensor to be open.
> that improved the humidity response rate.  Assembly error on my part with
> the board change.
>
> For the barometer sensor, this has no default value on start up.  When
> powered up from a dead state you have to enter the initial pressure.
> During the summer conditions the unit had been giving smooth values, when I
> was not having to reset it after a power outage, and the unit seems to
> track the neighbouring stations I referenced, so it appears to be ok.
> However, I didn’t have a reliable sensor to compare it against locally
> until recently when I put an app on my work phone to access the onboard
> sensor.  So far, it seems to match, but it is a small random sample
> (Basically, when I think to cross check)
>
> For the wind speed and direction, I suffer from the usual siting issues
> from the forest and structures.  I hope to get a 30 ft tower setup in a
> more clear area but even at that the 50 to 100 foot forest that is within
> 50 to 100 feet of the station will shadow me either way.
>
> If I am completely off target for that question, oops!
>
> For the calibration, I think I have covered all the sensors above save the
> wind speed, but the shadowing is the biggest influence there so it is
> always going to be off somewhat.  With WUHU, I have done nothing with the
> data to shape or shift it.  Offsets are 0.0 and coefficients are 1.0 as my
> thinking was get a good long period of operation before messing around with
> that area and only if it absolutely needed it due to a non linear
> behaviour.  WD was for a trial to see if WUHU was doing some extra math in
> the background for the barometer that I was not seeing, but I think I have
> pretty well verified that it is not using temperature to correct the
> barometer data when it is sending out the data.
>
>
>
>
>
>
>
>
> *From:* Steve Meyers <stevem at samcon.com>
> *Sent:* Tuesday, 2 February, 2016 20:47
> *To:* Discussion of weather data quality issues
> <wxqc at lists.gladstonefamily.net>
> *Subject:* Re: [wxqc] Looking for help with Station - Odd shift in
> analysis
>
> Have you run a hardware analysis on your PEET station ?
> Have you set all calibration settings to default on PEET hardware,
> software and WD?
> From what I have read I would take these steps in order and know what the
> default values are.
> Also, I would backup all settings prior to taking any action.
> After set to default you can begin to introduce changes to calibration,
> incrementally.
> If you know certain settings are fine as is, write down tHe values of
> those settings and re-enter them after setting systems to default.
>
>
>
> On Tuesday, February 2, 2016, Chris Miller <c_miller_1 at hotmail.com> wrote:
>
>> Before I forget myself, thanks for all those who have helped be try and
>> bug this weirdness out so far.  It is appreciated.
>>
>> That is how I had the station setup initially, at least pretty close.  My
>> conversion factor was a bit of a guess at first, but I managed to get
>> closer to the figure with some of the aviation calculators.  About the time
>> that Ted Lum gave me a number of correction values, I ran across
>> http://www.srh.noaa.gov/epz/?n=wxcalc_altimetersetting and used that
>> correction to get my cellphone pressure reading, previously used to drag a
>> correction value for it from the airport, back to my station. I can make
>> the necessary correction to the MADIS value that Ted gave me, but it is
>> only off by about .2 mb right now for that.    Also, I am back to using
>> WUHU.  Hopefully, Win10 will let it run properly this time, although I
>> expect I may have to go back to Weather Display.  It is a shame that
>> Weather Display is not keeping my basic settings for the station
>> communication on restart, maybe that is a trial software thing, but that is
>> not for here.
>>
>> The sudden shift in expected pressure is what is driving me bonkers, I
>> guess, especially when all the stations are gathered together for
>> readings.  I think, when I was using the modelling to get my pressure to
>> spec, I ended up just entered the value to line.  This was because I
>> thought it was due to my station being too far from the reference stations
>> (Yes, I know, bad bad.)
>>
>> Usually, the two government stations and the two CWOP stations I
>> mentioned have good analysis, yet in winter mine shifts on average 6 mb
>> higher than my actual reading with no changes. Basically, MADIS has flagged
>> the station as Q rather than V lately.  So, I am presuming that something I
>> have set is giving MADIS bum data?
>>
>> I admit the jagged data, while trying to get the station back to trim,
>> made my standard deviation go wide a few weeks.  However, once I get it
>> settled, my station tracks the few stations I referenced in a nice tight
>> group, save when the storms are pushing through.   Which seems to be a lot
>> lately. [image: Smile]
>>
>> *From:* Monty Wilson
>> *Sent:* Tuesday, 2 February, 2016 00:42
>> *To:* 'Discussion of weather data quality issues'
>> *Subject:* Re: [wxqc] Looking for help with Station - Odd shift in
>> analysis
>>
>>
>> I’m not sure I can add much here other that ask if you over complicating
>> the process?
>>
>>
>>
>> You should be reporting (setting) your PEET pressure to QNH.  This is the
>> setting used by pilots when they set their altimeters.  Using this setting,
>> the altimeter will read airport elevation above sea level when the airplane
>> is setting on the ground.  It is reported by airports as the “altimeter
>> setting”.  The PEET Brothers weather station does not adjust this reading
>> for temperature, humidity, etc.  Since you have an airport within 35 or so
>> miles, a place to start is to set the PEET pressure to be the same as the
>> airport.  I would pick a day when the winds are calm and the pressure is
>> either steady or changing very slowly.
>>
>>
>>
>> Think of it this way…….on a steady pressure day, if a helicopter took off
>> from your closest airport with that airports altimeter setting and then
>> landed in your yard, its altimeter would read your (local) elevation.
>>
>>
>>
>> On your Weather Display uncheck the QFE box.  A QFE setting causes the
>> altimeter to read zero elevation when that aircraft is sitting on the
>> ground.  That is not what you want to report.
>>
>>
>>
>> I also am using the PEET Bros (U2100) and Weather Display software.
>>
>>
>>
>> Hope this helps
>>
>>
>>
>> Monty Wilson, AP677/NR0A
>>
>> javascript:_e(%7B%7D,'cvml','jwilson16 at cableone.net');
>>
>>
>>
>> *From:*
>> javascript:_e(%7B%7D,'cvml','wxqc-bounces at lists.gladstonefamily.net');
>> [mailto:
>> javascript:_e(%7B%7D,'cvml','wxqc-bounces at lists.gladstonefamily.net');] *On
>> Behalf Of *Chris Miller
>> *Sent:* Monday, February 01, 2016 6:56 AM
>> *To:* javascript:_e(%7B%7D,'cvml','gladstonefamily73.net at tedlum.com');;
>> Discussion of weather data quality issues
>>
>> *Subject:* Re: [wxqc] Looking for help with Station - Odd shift in
>> analysis
>>
>>
>>
>> Calibration shifts are usually after a data loss (power outage causing
>> battery death in the main panel) where I have to get a reading from CYYB
>> and then use few days to trim it to what seemed right.  (storms have been
>> running local power through the grinder here)  I wish the PEET had a better
>> battery life then I would drag it right to the airport.
>>
>>
>>
>> Then, to find a better way to get my data back running after a power
>> outage, I started using the barometer reading off my phone to get the
>> initial setting and tried comparing the phone to the local pressure at
>> CYYB, when near it, to cross check and it seemed good.  I then used the
>> stations either side of me to test the setting by taking the reads from
>> them and estimating the expected pressure and that seemed to hit similar
>> values when wx was stable.  Is this not a reasonable process?
>>
>>
>>
>> As far as I know the sensor in the PEET station does not send temp
>> corrected data on the fly and it does not have an altitude setting.
>> Neither did WUHU.  I only started using the Altitude setting in Weather
>> Display when I started not trusting my math for the altitude correction.
>>
>>
>>
>> Just to be sure, before I shift it back to the corrected value on the
>> PEET station,  with an altitude of 196m (643 feet i think is correct
>> conversion) and a sensor height of about 6 feet above that how much
>> pressure is the correction factor?  I keep getting 23.6 mb to be added, but
>> I am going to presume my sources are corrupt.
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> *From:* Ted Lum
>>
>> *Sent:* Sunday, 31 January, 2016 23:07
>>
>> *To:* Chris Miller ; Discussion of weather data quality issues
>>
>> *Subject:* Re: [wxqc] Looking for help with Station - Odd shift in
>> analysis
>>
>>
>>
>> Actually, half of it looks like temperature variation. The other half
>> looks like the calibration was being pushed all over the place.
>>
>> http://media.tedworld.com/AU770.Month.pdf
>> <http://media.tedworld.com/AU770.pdf>
>>
>> On 1/31/2016 10:42 PM, Ted Lum wrote:
>>
>> It's not agreeing with neighbors any more than a broken clock is right
>> twice a day. That chart is the last 180 days against the airport (CYYB) .
>> Your variance is a full 12 mb, which is among the worst I've seen, probably
>> since you're in Canada with a 109°F temp variation in the past 180 days....
>> and that's after I took out the garbage from your ill-fated calibration
>> attempts that were over 25mb.
>>
>> Suggest you set all elevations to zero, and calibrate to Altimeter
>> directly, at least until you can figure out where the correction is coming
>> from.
>>
>> On 1/31/2016 8:26 PM, Chris Miller wrote:
>>
>> setting in weather display is set for qfe checked and barometer on Peet
>> 2100 set to absolute pressure, however I have also tried setting on PEET
>> weather station set to corrected pressure on both weather display and WUHU
>> with same results both winters.  Right now I seem to slot in with the
>> neighbouring stations shown in the posting so I thought I was close.   I
>> don’t see any settings in PEET weather station that would be adding in the
>> temperature correction, or in the software as far as i can tell, but If
>> someone can point me toward the setting it would be appreciated.
>>
>>
>>
>> *From:* Ted Lum
>>
>> *Sent:* Sunday, 31 January, 2016 19:51
>>
>> *To:* Discussion of weather data quality issues
>>
>> *Subject:* Re: [wxqc] Looking for help with Station - Odd shift in
>> analysis
>>
>>
>>
>> You appear to be sending Sea Level Pressure (SLP) in stead of Altimeter.
>> SLP is temperature corrected, thus your error will vary with temperature
>> which varies with season. You need to fix the settings in the software, or
>> replace it.
>>
>> http://media.tedworld.com/AU770.pdf
>>
>> On 1/26/2016 6:19 PM, Chris Miller wrote:
>>
>> Good day to all in the group.
>>
>>
>>
>> I have a weather station, AU770, that I am trying to figure out what has
>> happened.  During the summer, I had been getting good analysis graphs,
>> thumbs up, and check marks for my barometer data and such after bugging out
>> station settings and battling power loss resets.  (May be visible in the
>> data as strong standard deviation spikes)
>>
>>
>>
>> However, once the fall weather presented itself, my analysis dropped away
>> and the chart started showing my barometer low by 6 mb on average (humidity
>> analysis dropped by 10 percent, but I am putting that one to issues with
>> local humidity being higher due to nearby river influence)
>>
>> I tried switching software, in case WUHU was not doing its work properly,
>> to Weather Display as a trial.  Things shifted to better for a while, but
>> then returned to the same behaviour.  I then tried setting my barometer up
>> to give the local pressure and let Weather Display do the math, in case my
>> estimate of the local pressure translated to sea level was wrong.  I still
>> ended up with the same number as I thought with no change in graphing or
>> thumbs and checks even with a trial period.
>>
>>
>>
>> So, now I am wondering if the error is related to a few of the local
>> stations having weird barometer readings?
>>
>> My nearest stations, CYYB, CYWA, and AV065, AR943 seem to track
>> appropriately usually bracketing my graph for the most part to within a
>> mb.  However, E2423 and E2429 usually report about 13 mb lower than the
>> expected pressure.   So, I wonder if that is skewing my estimate graph and
>> the thumbs and checks, not knowing which stations they are using for my
>> local group for L3 filtering?
>>
>> I wish more of the government weather stations reported pressure, and
>> reported it with the same consistency in the METAR data, as this would
>> probably fix some of my skewing that I see.  I expect my local group is not
>> very local for the L3 check and MADIS checks.
>>
>>
>>
>> I realise that it is not a big deal that the analysis graphs are off, as
>> that is based on the lumped data with no L3 filtering of bad stations, as I
>> understand it.  But the MADIS check is also showing bad, so I may be being
>> removed from the data group when the data is actually good.   And since the
>> stations are very sparse up here, I am trying to keep my station on spec as
>> much as I can.
>>
>> Now, if I could get in contact with the owner of E2423 and get those
>> stations all on spec, we would have a significant gain in good pressure
>> readings.  He has 15 stations at least, spread all over Ontario, all with
>> pressure and location issues.  Although, I think some of that may be
>> related to barometer offset altitude.  Regrettably, there is no way to
>> indicate the height of the sensor relative to your surface altitude in the
>> system, that I have noticed.
>>
>>
>>
>> If it is not that, then what is it that I have not set correctly on the
>> station that causes this skewing of the data seasonally, or over a period
>> since the chart cleared for a while after the software change?
>>
>>
>>
>>
>>
>>
>>
>> Thanks in advance for any assistance,
>>
>>
>>
>> Chris
>>
>> AU770
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> --
>> This message has been scanned for viruses and
>> dangerous content by *MailScanner* <http://www.mailscanner.info/>, and
>> is
>> believed to be clean.
>> ------------------------------
>> _______________________________________________
>> wxqc mailing list
>> Post messages to
>> javascript:_e(%7B%7D,'cvml','wxqc at lists.gladstonefamily.net');
>> To unsubcribe or change delivery options, please go to:
>> http://server.gladstonefamily.net/mailman/listinfo/wxqc
>> To search the archives:
>> http://www.google.com/coop/cse?cx=008314629403309390388%3Aknlfnptih9u
>>
>> The contents of this message are the responsibility of the author.
>>
> ------------------------------
> _______________________________________________
> wxqc mailing list
> Post messages to wxqc at lists.gladstonefamily.net
> To unsubcribe or change delivery options, please go to:
> http://server.gladstonefamily.net/mailman/listinfo/wxqc
> To search the archives:
> http://www.google.com/coop/cse?cx=008314629403309390388%3Aknlfnptih9u
>
> The contents of this message are the responsibility of the author.
>
>
> _______________________________________________
> wxqc mailing list
> Post messages to wxqc at lists.gladstonefamily.net
> To unsubcribe or change delivery options, please go to:
> http://server.gladstonefamily.net/mailman/listinfo/wxqc
> To search the archives:
> http://www.google.com/coop/cse?cx=008314629403309390388%3Aknlfnptih9u
>
> The contents of this message are the responsibility of the author.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://server.gladstonefamily.net/pipermail/wxqc/attachments/20160217/12a1c1b3/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: wlEmoticon-smile[1].png
Type: image/png
Size: 1046 bytes
Desc: not available
URL: <http://server.gladstonefamily.net/pipermail/wxqc/attachments/20160217/12a1c1b3/attachment.png>


More information about the wxqc mailing list