Order Tray | Contact Us | Home | SIG Lists

[wxsig] Hobby Boards Barometer

William Beals will at beals5.com
Wed Apr 30 03:04:06 UTC 2008


I did find the coding error.  Long story, but boils down to not handling the
remainder right on some fixed point math for calculating metric BP.  It only
affects the fractional part, and the way the error manifests itself, only
decimal values of .0, .1, and .2 are possible!  I'd say that pretty much
describes the problem!

I'll work on the fix over the next week.  The tricky part for me will be
testing it given my wx station is down right now, but there are ways.  I've
had some other features coded up but never released and so there will be
more changes than just the BP fix.


-----Original Message-----
From: William Beals [mailto:will at beals5.com] 
Sent: Monday, April 28, 2008 9:57 PM
To: 'TAPR Weather Station SIG Mailing List'
Subject: RE: [wxsig] Hobby Boards Barometer

Dave and Ken:

Eeuw, this problem now sounds very familiar!

I went back in my own archives and this was reported by Ken Brown back in
March 2006!
I thought I had fixed this problem, but I still have it on an old "to do"
list and I don't see the fix listed anywhere in my release notes.

I'll go look at my code and see what I can figure out.  It has been a while
since I've looked at that fixed-point math!


-----Original Message-----
From: wxsig-bounces at lists.tapr.org [mailto:wxsig-bounces at lists.tapr.org] On
Behalf Of Dave Barton
Sent: Monday, April 28, 2008 7:50 PM
To: TAPR Weather Station SIG Mailing List
Subject: Re: [wxsig] Hobby Boards Barometer

   Mr. Brown stated the problem very clearly and succinctly. Thank you,
   To use empirical data,  as noted in my previous email, you can view
   the data of multiple T238 stations that are reporting barometric
   pressure.  Here are a few:
   The vast majority of the pressure readings end with a zero tenths
   value.  However, a small number end in xxxx.1 or xxxx.2.  These values
   seems to occur when the value is trending up or down, as in

   1023.0 1023.0 1022.2 1022.1 1022.2 1022.2 1022.0 1022.0 1022.0

   The APRS standard allows the tenths digit, the english inches of
   mercury values show a decent resolution, and there is even one T238
   station that  reports over TCP/IP with good resolution.  I am just
   stumped to figure out how we can occasionally have xxxx.1 or xxxx.2
   values, but not xxxx.3 through xxxx.9.  If we had the .3 through .9
   values, presumably we'd have a smoother graph.  Since we have
   acceptable resolution for the inches of mercury, it seems like we
   should be able to convert to millibars.
   Inches       Millibars

   29.91 1012.9
   29.92 1013.2
   29.93 1013.5
   29.94 1013.9
   29.95 1014.2

   My last assembly programming was 25 years ago and although I have
   looked at your code, I doubt that I can be of any help, but I am duly
   impressed with the effort that has gone into the code for the T238+!
   Regards, Dave
   [6]Ken.A.Brown at noaa.gov wrote:


The reason for the scaling needed on the barometer was explained
well on the Jennings web site.  It was indeed a pain to get the
Bray barometer calibrated for my elevation and took several weeks
to do so with fine adjustments to the span and offset values.  I
did a comparison for several weeks with four local aviation reporting
stations and it now falls in well when I use the LCD BP value.

What the sequence of previous message concerned is the conversion
of the LCD BP value in inches of mercury to the APRS output value
in millibars.  The LCD BP value seems to increase and decrease
smoothly where as the millibar value seems to "stair step" in
mostly whole numbers instead of millibars and tenths.  Using
conversion tables from the Smithsonian Meteorological Tables
shows that the conversions do not seem consistent.  I was never
able to get the LCD to display millibars correctly so it is
always indicating altimeter setting inches of mercury values.

Ken Brown, n5kuk

----- Original Message -----
From: William Beals [7]<will at beals5.com>
Date: Monday, April 28, 2008 0:57 am
Subject: RE: [wxsig] Hobby Boards Barometer

Dave (and all):

It took longer than expected, but here is what I was able to dig up
from my
archives on barometric pressure.

The root issue being bantered around at the time revolved around
resolutionof the A/D converter.  The problem was that we didn't
have or wanted to
afford an A/D that could cover the entire range of BP values from
sea level
to 10,000+ feet and do so with enough resolution to be able to
accuratelydiscern trends.

The problem is if you can only count on about 9 bits of data (512
steps), if
you decided to digitize all possible values at any altitude, any given
person at a fixed altitude would have only a very small number of
thosesteps available for local measurements.

This wasn't a revelation of mine (though I was having similar
concerns), it
was from Jim Jennings.  From his design and the folk that copied
him, the
answer was to use resistors and op-amps to take the output of the
sensorthat could cover all pressures and normalize the output of
the sensor so
that it will put out a specific voltage if the local pressure
compensatedfor sea level was around 27 inches.  In this way, the
analog electronics
compensate for the differences in altitude and then the A/D
converters range
only deals with local variations from that baseline.

In my code, the equation is Baval= 27.46+(Bavad*0.429).  I am
pretty sure
(but didn't document well), Bavad has a range of 0.0 to 5.0.

This is actually a very way of doing things!  Changes due to local
pressurechanges are MUCH smaller than changes due to altitude, so
"nulling it out"
in the analog domain means we get to use a much cheaper (and 1-wire
powerable) A/D converter and as it is based on local conditions and
neverchanges, kind of a waste to use such an A/D for something that
can be
calibrated out.  Also, what it also meant was that for programmers
like me,
we got to write one piece of code and it will work everywhere
without the
need for compensating for different locations.


-----Original Message-----
From: Dave Barton [[8]mailto:dbarton at nettally.com]
Sent: Thursday, April 10, 2008 5:47 PM
To: [9]will at beals5.com; TAPR Weather Station SIG Mailing List
Subject: Re: [wxsig] Hobby Boards Barometer

I tried searching the TAPR archives and found one reference to the
problem  from Mr. Brown, N5KUK, but nothing further.  No rush, Mr.
not a high-priority, I am just curious.  The APRS specs allow for
digit, I assume, since it is displayed on the log files.


[10]will at beals5.com wrote:


There is a lot of history on this!  Like everything else it was a

conpromise. I'm on a trip now, but can see if I can dig them up this

Sent with thumbs from a PDA

-----Original Message-----
From: John Yost [11]<k3yjp at yahoo.com>

Date: Thu, 10 Apr 2008 10:10:12
To:TAPR Weather Station SIG Mailing List [12]<wxsig at lists.tapr.org>
Subject: Re: [wxsig] Hobby Boards Barometer

I am using a T238+ with the TAPR X1W-1 barometer kit.

I see the same behavior, the local display in inches
of mercury seems to be ok but loses the resolution
when sent to aprs.

my .02


--- Ken Brown [13]<Ken.A.Brown at noaa.gov> wrote:

The problem with the pressure reading only reporting
whole millibar
readings is not in the barometer but in the T238
code that converts
the reading received from the barometer to the
millibar reading
transmitted.  At my station, n5kuk, I have two T238s
each with a
Bray barometer. One is on APRS and one is a test
unit on the bench.
Both exhibit the same problem. The problem is seen
on the APRS output
but not on the displayed value in inches of mercury.
 I use the APRS
T238 as a standalone with it's own tnc and radio.
On another APRS
station with WinAPRS the nice weather graph clearly
shows the stair
step of the pressure reading. My units are T238 and
not T238+.  On
the bench unit the reading were noted on the rad
RS-232 output that
would go to a tnc in APRS service.  Both T238 units
running the latest

Ken, n5kuk

Dave Barton wrote:

More information on this issue.  This web site,


has logs of

T238+ weather stations.  I've only checked a

handful of T238+ stations,

but I have found a couple of stations which seem

to to have the same

issue as mine, N5KUK and K3YJP-4.

There is another station, KA1GEU*, *that has a

normal range of first

decimal places.  However, his station is reporting

via the Internet, not


I'll try to find more examples after work today.

Now I wonder if it's

the modem board?


Gerry Creager wrote:

I'll have to get some sleep, then check but I

think the APRS spec, and

therefore FindU, resolve to 1 mB.

More's the pity.


Dave Barton wrote:

I am puzzled by something.  I recently bought

and hooked up a Hobby

Boards barometer (B1-R1-K) and it seems to be

working OK.  One thing

I am seeing is that the barometric pressures

reported by FINDU almost

always have a .0 for a decimal millibar value.

In other words,

1016.0, 1017.0, 1018.0.  I have occasionally

seen xxxx.1 or xxxx.2,

but never xxxx.3 through xxxx.9.  My FINDU plots

(AI4GF-4) have a

very jumpy barometer chart.  The barometer is

supposed to have a

resolution of 0.007 in.Hg.

I looked at the raw packets and the anomaly is

reflected in the raw

packet.   I have the latest firmware, 1.15.8R.

I saw a suggested mod of a resistor in parallel

with R1 but have not

done that because I don't quite understand the

mod.  Could this

problem somehow be related to not having done

the mod?

73, Dave
Tallahassee, FL

wxsig mailing list
[15]wxsig at lists.tapr.org


wxsig mailing list
[17]wxsig at lists.tapr.org


wxsig mailing list
[19]wxsig at lists.tapr.org


Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around

wxsig mailing list
[22]wxsig at lists.tapr.org



wxsig mailing list
[24]wxsig at lists.tapr.org

wxsig mailing list
[26]wxsig at lists.tapr.org

wxsig mailing list
[28]wxsig at lists.tapr.org


   1. http://www.findu.com/cgi-bin/wx.cgi?call=k3yjp-4
   2. http://www.findu.com/cgi-bin/wx.cgi?call=w0dvm
   3. http://www.findu.com/cgi-bin/wx.cgi?call=w0maf-1
   4. http://www.findu.com/cgi-bin/wx.cgi?call=wa2gug
   5. http://www.findu.com/cgi-bin/wx.cgi?call=ai4gf-4
   6. mailto:Ken.A.Brown at noaa.gov
   7. mailto:will at beals5.com
   8. mailto:dbarton at nettally.com
   9. mailto:will at beals5.com
  10. mailto:will at beals5.com
  11. mailto:k3yjp at yahoo.com
  12. mailto:wxsig at lists.tapr.org
  13. mailto:Ken.A.Brown at noaa.gov
  14. http://www.wxqa.com/callsminmax/index_callsminmax.html
  15. mailto:wxsig at lists.tapr.org
  16. https://lists.tapr.org/cgi-bin/mailman/listinfo/wxsig
  17. mailto:wxsig at lists.tapr.org
  18. https://lists.tapr.org/cgi-bin/mailman/listinfo/wxsig
  19. mailto:wxsig at lists.tapr.org
  20. https://lists.tapr.org/cgi-bin/mailman/listinfo/wxsig
  21. http://mail.yahoo.com/
  22. mailto:wxsig at lists.tapr.org
  23. https://lists.tapr.org/cgi-bin/mailman/listinfo/wxsig
  24. mailto:wxsig at lists.tapr.org
  25. https://lists.tapr.org/cgi-bin/mailman/listinfo/wxsig
  26. mailto:wxsig at lists.tapr.org
  27. https://lists.tapr.org/cgi-bin/mailman/listinfo/wxsig
  28. mailto:wxsig at lists.tapr.org
  29. https://lists.tapr.org/cgi-bin/mailman/listinfo/wxsig
wxsig mailing list
wxsig at lists.tapr.org

More information about the wxsig mailing list