[wxsig] T-238 Humidity Calibration

[Ken Brown]

Will,
I have been looking at the graphs on the QC site for many weeks.
http://weather.gladstonefamily.net/cgi-bin/wxqchart.pl?site=ar301
My sensor is a TAPR X1W2 or whatever and it is in a Davis
aspirated radiation shield.  I have taken everything apart and
cleaned the shield and sensor several times.  From doing QC on
the local official stations at the NWS office I believe the data
in the qc report to be reasonably accurate so it seems that it
must just be my sensor reading high.  It would take a while of
"curve fitting" during the right weather conditions to make it
look good but I think it might work.  I can also use a sling
psychrometer for quick calibration but would need to watch it
over a period of time to see how it looks.

I think I running the most recent software prior to the release
you made last week.  I'll try to track the raw data on the LCD
for a while and see what it looks like.

Ken Brown


Will Beals wrote:
> Ken:
> 
> One thing I can do is add some diagnostic data to occasional APRS packets.
> I'm doing this now with the gust data to help isolate if I have a sensor
> issue or a firmware issue.  Once that trial runs it's course, I can add a BP
> diagnostic string next.  I like it as the data gets appended to an APRS
> string that you can then go look at a few days worth of data on findu and
> see what happened.
> 
> I have the raw humidity data displayed right now on the LCD diagnostic
> screen starting with 1.15.? (forgot).  It shows the raw sensor data.  The
> LCD is handier for issues that are either easily repeatable or happen for a
> while (like saturated humidity sensors).
> 
> How confident are you the offset is a constant over both humidity and
> temperature?  My concern with simply adding an offset is that it makes the
> 100% case look good, but in fact may make less obvious readings (ones harder
> to verify) even more innacurate.
> 
> will
>  
> 
> -----Original Message-----
> From: Ken.A.Brown at noaa.gov [mailto:Ken.A.Brown at noaa.gov] 
> Sent: Friday, September 02, 2005 6:14 AM
> To: will at beals5.com; TAPR Weather Station SIG Mailing List
> Subject: Re: RE: [wxsig] T-238 Humidity Calibration
> 
> Will,
> 
> I understand the problem with the BP sensor. You need to use a very accurate
> and high resolution volt meter to read the voltages directly on the board
> and even then it takes a lot of fooling around to get the span and offset
> correct.  I had to calibrate mine several times. Each time I would determine
> how many millibars in error it was, convert that error to meters of
> elevation using the Smithsonian met tables, and compute new voltage values
> for the settings.  I'm still a bit off but it is real close and tracks well.
> I'm still troubled with out lying spikes in the indicated pressure that are
> many millibars high, somewhat like the wind gust problem you described.
> Most days I get none but other days there may be one or a few, hours apart.
> I have never been sitting in front of the
> T238 when it happens and only see it in the APRS output packets.  I'm sure
> it isn't corrupted packets because I see it on my local APRS station and
> also in the QC data on the web.
> 
> For humidity my sensors indicate saturation when the RH is only around 95%
> and they seem to track about 5% high all the time.  I thought maybe scaling
> the voltage from the sensor to the one-wire device might work but haven' 
> tried it yet.
> 
> Ken Brown