Mounting Angle, Background detection

Here's a previous theme with shorter data periods.

Experiment: To compare background readings with an RM-80 pancake detector using six different positions-
Vertical, (narrow profile towards sky/earth), mica window facing south, north, east & west
Horizontal, mica window facing up and down

Problem: Average readings of background tend to vary with the detector position.
There is the possibility that when the mica window is horizontal, facing up, short-lived radon daughter products are settling and increasing the count. Another possibility is that when mounted vertically, the edge-on profile of the detector is presenting a smaller surface area to cosmic rays, resulting in lower readings.

Equipment: Dedicated lap-top, RM-80 pancake detector, Aware AW-RADW software, Aware USB-MSP430 interface module, GK+4 Geiger Counter, LND7317 pancake probe, SD card, tin box, tape, Desk Top PC with Excel, Windows 8.1

Procedure: The RM-80 screw-on SS mesh shield was removed, washed, scrubbed and dried. The mica window was blown clean. Reassembled, the RM-80 detector was placed in a zip-lock freezer baggie, then taped in position atop a tin box, approximately 10 inches above the sill. The window was open about 2 inches.

Six one hour readings were downloaded from flash memory of the Aware USB-MSP430 interface module, from 15:00 to 21:00 as a CSV file. Data was imported to Excell 2000. Averages, standard deviation, minimum and maximum counts for each period were calculated.

The detector head was repositioned at the start of each 1 hour period.
Starting position, 15:00, vertical, facing south.
16:00 , vertical, facing north
17:00 , horizontal, facing up
18:00 , horizontal, facing down
19:00 , vertical, facing east
20:00 , vertical, facing west

In addition to the Aware RM-80 detector, a GK+4 was used as a control detector. A pancake probe was placed on the window sill, next to the tin box support. This probe was mounted vertical, facing north, and was connected to the GK+4 Geiger counter, which recorded data on an SD flash memory during the total experiment.
The GK+4 probe contains an LND7317 tube with characteristics identical to the LND7313-3 used in the RM-80.

Conditions: Weather was stormy, with heavy rain squalls all day and evening. No radioactive sources were situated in the room. A fairly steady draft came through the metal mesh window screen during the experiment. A subsequent check with the GK+4 pancake did not detect any elevated readings on the plastic bag exterior.

Data:
Reading 1: Detector vertical, facing south
"Average cpm 27.8667
"Standard Deviation 5.257655167
"Min cpm "16 cpm 1
"Max cpm "41 cpm 1
"Average microSv/hr 0.078608369

Reading 2: Detector vertical, facing north
"Average cpm 28.31667
"Standard Deviation 4.893758275
"Min cpm "17 cpm 1
"Max cpm "41 cpm 1
"Average microSv/hr 0.079877762

Reading 3: Detector horizontal, facing up
"Average cpm 34.2
"Standard Deviation 5.108318241
"Min cpm "21 cpm 1
"Max cpm "45 cpm 1
"Average microSv/hr 0.096473907

Reading 4: Detector horizontal, facing down
"Average cpm 32.93
"Standard Deviation 5.891293969
"Min cpm "19 cpm 2
"Max cpm "48 cpm 1
"Average microSv/hr 0.092900799

Reading 5: Vertical, facing east
"Average cpm 26.81666667
"Standard Deviation 5.610195521
"Min cpm "13 cpm 1
"Max cpm "39 cpm 2
"Average microSv/hr 0.07564645

Reading 6: Vertical, facing west
"Average cpm 25.4
"Standard Deviation 4.423358644
"Min cpm "17 cpm 2
"Max cpm "35 cpm 2
"Average microSv/hr 0.071650212

GK+4 6 hour Reading: vertical, facing north
"Average CPM 27.86157518
Standard Deviation 6.266790148
Min CPM "18 cpm 1
Max CPM "40 cpm 1

RM-80 Reading Averages-
Average CPM of 2 horizontal positions = 33.5667
Average CPM of 4 vertical positions = 27.1
Difference in average readings = 6.4667

Conclusions?
The variation in counts from the four vertical positions is small, while the difference between vertical and horizontal appears significant.
When the detector was face-down, a lower count may indicate the thick stainless steel shell and back provide some shielding against lower energy X-rays produced by cosmic ray showers.
Since the mica window was shielded from dust in the thick plastic baggie, air-born radon daughter products are not causing elevated readings when the RM-80 detector is face-up and horizontal. More likely, the skyward exposure of a larger capture area is producing higher counts.

excellent study!
I can understand the profile orientation could make a difference, but I'm not convinced that the plastic baggie would cut the radon progeny from affecting the results, nor that the stainless gm case could block cosmic rads.
It would be good to have another GC with similar sensitivity running alongside in a fixed position to compare against results. Windy, rainy days as well as seemingly stable days can bring unexpected variations in cpm from hour to hour...

Thanks for the feedback.
I did check the baggie after it had been around the RM-80 all night, but found no increased count. Baggies and balloons are great for attracting radon daughters if they get a static charge.
Day after the experiment, I set up the LND7313 and my new 7317 side by side. Weather was consistently blustry and raining. Since it was a large zip-lock baggie, I set it up over the LND7313 (RM-80) in a semi-inflated configuration. About four inches of air space above and all around the detector head. Both detectors were flat on the window ledge, horizontal, mica windows up for 24 hours.
The LND7317 was connected to the GK+4, and gave an average of 30 cpm, with no baggie. The LND7313 (RM-80) was still connected to the lap-top AW-RADW and gave an average of 31 cpm from inside the baggie.
I wasn't thinking the SS case could shield against muons, rather low energy secondary X-rays produced by prior collisions. These might not penetrate the 1/8" thick SS back, but they could easily penetrate the mica window.
Just a thought. I have a lead castle made from a 8 foot by 1 foot roll of lead around a 2" diameter annealed copper pipe. This was suggested to me by Steven Secilman for gamma spectroscopy . Muons and other cosmic ray products hit the lead and create secondary X-rays, but are largely blocked by the inner copper pipe, thus reducing noise.
I was thinking along this line as a possible explanation for the difference in readings back to front.
Brian
roll ofwrapped