Unexplained Readings

[Richard Wood]

I had two unexplainable high readings on my bGeige over two days, one was not high enough to cause alarm but I had the same type of occurrence the next day that did have a high reading.  The meter seems fine, I did a few rather hard shake and shock tests and reruns to try and repeat the readings with no resolution.

Log 36112, record 30230526.log

I have not got the QGis app working but could find and edit the readings out of the file. ?
I do not like deleting data from my records, I understand why it is done for the presentation portion but I save this data.  I go to other sources to see if they saw something similar so do not want to lose data myself.
Advice?  I need the record approved for the rest of the data on there.  It completes a tour of northwest Oregon for the team!!!

[Richard Wood]

Of Note, The drive app was also logging so this does not seem that start up error? Those files are also loaded and not yet approved... which was what sparked a question about duplicate records.

[Jan Helebrant]

Hi,

my opinion is that it is not valid measurement point because there is only one point with such a high value:

according to the speed you were walking so in case of some real anomaly you would probably have more higher points around. In my measurement I would probably remove that point. We had a measurement from a guy who probably met someone after radiotherapy going home from hospital but in that measurement he had several higher points about few microSv/h and decreasing when he moved away.

Jan

[ferdo]

People who receive radiotherapy with iodine-131 are released from hospital when they emit less than between 10 and 43.5 microSv/h on a 1 meter distance.

(http://www.nuclmed.gr/magazine/eng/may13/25.pdf, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4937693/)

I think that's significant.

[Mat Schaffer]

The one point with 1000+ counts in 5s has me wondering if a wiring problem could cause incorrect counts.

-Mat

matschaffer.com

[Jan Helebrant]

I also suspect something like that. We had a false anomaly here:

https://api.safecast.org/en-US/bgeigie_imports/30378

and it was faulty bGeigie which occasionally logged such higher values. Our technician checked the device, did some "tech magic" :-) (maybe he just replugged the cables or soldered something better) and the device is OK now.

Jan 

Dne úterý 5. června 2018 8:20:43 UTC+2 Mat Schaffer napsal(a):

[Wood]

The meter had a smaller spike the day before, I am watching stationary readings and did some shock test to boot.  If it happens again I'll take the unit apart and check my solder work over, so good reminder to pack a soldering iron.

If I have bad readings can I edit those lines out of the files and submit, should I note it?  Will those missing lines cause other problems for the record?
Sorry, I asked this question elsewhere.  I suppose it was answered as QGis submits edited records?

[Jan Helebrant]

Hi,

If I have bad readings can I edit those lines out of the files and submit, should I note it?

I would be careful what to consider "bad reading" but especially in your LOG mentioned above (single elevated point close to points with normal values) I would remove that point.

Will those missing lines cause other problems for the record?

No, I am sure :-)

I suppose it was answered as QGis submits edited records?

Yes, we often use QGIS to edit the LOG data - to remove parts of the track or to split the LOG in several smaller ones in case of combined measurements. For example these three uploads (1x car drive, 2x walk) came from one source LOG file:

https://api.safecast.org/en-US/bgeigie_imports/36253

https://api.safecast.org/en-US/bgeigie_imports/36254

https://api.safecast.org/en-US/bgeigie_imports/36255 

The QGIS plugin was designed to keep the LOG formatting and not to let the user alter the data values in any way. The API accepts the LOG as any other.

Jan

Dne úterý 5. června 2018 9:10:05 UTC+2 Richard Wood napsal(a):

[Lugul banda]

to make shure, try something, dont use the drive app, it may be only me, but i think its the culprint for some strange measurings, if the bluetooth connction ges interupted by any case,, 

second thing, i noticed the bgige, has in its warmup phase some elevated measurings, so try to start it somewere around that area and then move in , and the other way someoneels has to revesite the place to make shure the measuring is valid ... and be shure the device can reach to the sattelites up in the air and nothing is blocking the signal.... 

[Wood]

I use both data sets, a drive record is a subset of the meter log file.

I have not found a difference in the data, and that is how the logging works

[Lugul banda]

Am Sonntag, 3. Juni 2018 19:08:13 UTC+2 schrieb Richard Wood:

  I need the record approved for the rest of the data on there.  It completes a tour of northwest Oregon for the team!!!

have no fear my oldest aprovals are from around 22 may 2018 ;-)

 

[Richard Wood]

Jan,
I just reread your post and hope the comment was the only issue and think I corrected that. It was just another record and a few points in thousands.  It was not repeatable with location or instrument stress tests so it was a UFO?  I am loath to delete data when correlation is such a part of the data.   The primary display of Counts per Five Seconds as counts per minute are at the edge of these detectors precision curve, shoulda been 600volts?   If you look at my recent records, I think I'm getting better at this.



On Tuesday, June 5, 2018 at 1:07:28 AM UTC-7, Jan Helebrant wrote:

Hi,h a

[Jan Helebrant]

Hi,

I checked your LOGs and found strange thing - how did you got the two files 3023-0526.log and 30230526.log? Because it looks like they both share some data:

and the part with high value

there are several points with same timestamp and coordinate e.g. wrong data. If you look on the 5 second raw data then there are values within the range 1-5 pulses (per 5 second interval), then there are those points with same date/time and coordinate with value 1436 pulses and after that you have again values 1-6. So there are many reasons to drop this part. Even with much more expensive instruments you sometimes get some outliers - no instrument is perfect.

Meanwhile I removed the wrong points (outside America) from:

https://api.safecast.org/en-US/bgeigie_imports/36114

so you can resubmit the file if you want and delete the old one :-)

Jan

Dne pondělí 18. června 2018 22:23:45 UTC+2 Richard Wood napsal(a):