uRADMonitor â€“ Online remote radiation monitoring station / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

uRADMonitor â€“ Online remote radiation monitoring station

first previous 4 5 6 7 next last

Move Thread

LAN_403

Proud Mary

Thu Oct 18 2012, 10:35PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

I thought you would like to see how very much you have improved the accuracy and usefulness of your survey meter by using PIC:

1350599189 543 FT111231 Table Of Measurements

This resolution down to 0.1 mSv will give your annual rates a very useful degree of accuracy, when you look at the dose rate ranges
you are likely to find in practice (and without a nuclear disaster, of course!)

1350599212 543 FT111231 Table Of Measurements 2

radhoo

Fri Oct 19 2012, 08:47AM

Registered Member #1938 Joined: Sun Jan 25 2009, 12:44PM
Location: Romania
Posts: 699

How about the integration time constant? Is there any criteria for choosing an optimum value? I'll be building graphs, with selectable time interval, but they should be a collection of cpms collected over the given interval.

Conundrum

Fri Oct 19 2012, 10:41AM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4059

Interesting thread.
Does anyone here have any experience with the "micro" Russian tubes, got two here that are still waiting for a suitable driver and host.
I got one working about a year ago with a makeshift supply based on a personal attack alarm which was enough to prove they worked but were highly inefficient.

It did get me thinking though, winding a very fine copper coil around the tube and passing about an amp through it should increase the sensitivity, and could also double as a detector if connected to something like a MK484 as the slightest ionisation event should then be detectable.
Set it to an unused frequency like 40 kHz so it doesen't detect Droitwicz or Rugby.

Another interesting idea, make a primitive induction driver so that the tube is running with 40 kHz and then detect the change in current corresponding to an ionisation event.
Essentially the tube then becomes part of a tuned circuit so the slightest change in impedance then results in a pulse.

-A

Proud Mary

Fri Oct 19 2012, 11:24AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

radhoo wrote ...

How about the integration time constant? Is there any criteria for choosing an optimum value? I'll be building graphs, with selectable time interval, but they should be a collection of cpms collected over the given interval.

I have seen 4 hours given as the optimum integration period for background radiation counting, but cannot immediately find the references for it.

However, 4 hours may be too long to identify and quantify some common causes of background variation. For example, Rn-222 is released from cracks and fissures in rocks in response to changes in temperature and atmospheric pressure. This released radon is not evenly distributed through the lower atmosphere. "Under early morning still air conditions, Rn and its daughters can be trapped close to the ground in temperature inversion layers."*

I guess the most useful approach for a PIC based monitor is to have all your count/time data in a bin, so that you can ask any counts-versus-time questions you like about it. Data can then be tested against variables such as sunrise and sunset times for every day of the year, which impact on diurnal background variation. Maybe even the phases of the moon may have some small effect on the rate of diffusion of Rn into the atmosphere from soil and rocks, as it does with the sea and its tides. Who can say? Times during which rain is falling (a rain detector sensor?) may also produce interesting variations in count rate.

* Minty BRS, Luyendyk APJ, Brodie RC, Calibration and data processing for airborne gamma-ray spectrometry AGSO Journal of Australian Geology and Geophysics, 17 (2), p. 53

radhoo

Fri Oct 19 2012, 10:11PM

Registered Member #1938 Joined: Sun Jan 25 2009, 12:44PM
Location: Romania
Posts: 699

Indeed 4Hours is too much, but I will export data on a shorter interval, and offer the option to calculate an average for the chosen time interval . So we'll have both a higher resolution, and a better integrated result, depending on the user's interest. I'll detail the algorithm later.

Regarding the rain sensor, I'm also expecting an increase in dose in the first minutes of falling rain, but we'll see. The rain itself should be detectable using the humidity sensor, combined with readings from the pressure sensor (I didn't receive it yet, I'll have to add it later).

The good thing about microcontrollers is that they offer a lot of freedom for processing all the data, exactly like a small (but limited) computer.

@Conundrum: what tubes are those?

Update:

- box ready, completely made of PVC pipe plastic, after transforming a piece of pipe into a sheet and cutting it when heated.

I also created sockets, from the same plastic source, for the geiger tube:

The orange plastic is great, the finished box actually gives the impression of a very good quality plastic / case, similar to older materials! Funny considering what I've used.

Proud Mary

Sat Oct 20 2012, 10:00AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

First Prize for the plastic box! Wonderfully inventive!

radhoo

Sat Oct 20 2012, 11:43AM

Registered Member #1938 Joined: Sun Jan 25 2009, 12:44PM
Location: Romania
Posts: 699

I love that plastic. I think I'll be using again. Need to find some better tools for it, as it has potential.

I'm preparing to mount the enclosure outside. Last modifications on the software - but will be able to change it again if required (via a longer cable).

In regards to my last post, my idea for the data export is as follows:

Given a number of pulses, and a timer counting seconds, one can choose to compute the CPM using various intervals for time integration:

What my counter proposes to calculate, is the following function:
f(t) = p
Where t is the time and p are the pulses.

for the given time values (with 0 as reference):
t1 = 0 .. 1 minute (length 1 min)
t2 = 1 .. 2 minute (length 1 min)
...
t10 = 9..10 minute (length 1min)

We'll have the corresponding values f(t1)=p1 ... f(t10) = p10

Computing the CPM's as cpm1 = f(t1)/t1, ... , cpm10 = f(t10)/t10 will give use highly variable results, as Proud Mary kindly indicated, a value of 4 hours is indicated for CPM integration.

Nevertheless, we have the option of computing cpm1, ...cpm10 (each on 1 minute interval) and export them .

This data can be further process, and the time integration can be extended to any higher value as follows:

cpm_10minutes = (cpm1+...+cpm10) / 10

cpm_4hours = (cpm1+...cpm240) / 240 = (f(t1)/t1 + ... f(t240)/t240)/240 =
= ( because t1 = t2 = ... = t240 = 1minute) = (f(t1) + f(t2) + ... f(t240))/240 =
= (p1+p2+...p240) / 240
Resulting that making the average of smaller time value cpms for a given total time T, gives the same result like counting all pulses during T, and dividing that to T.

So I'll leave the option to choose the time integration constant for the post processing part or make it selectable by the user.

For now I will export CPMs calculated on a 1minute basis, or 10minutes , depending on the total data size estimate (in bytes) , so my server storage can handle it for a few years.

radhoo

Sat Oct 20 2012, 03:43PM

Registered Member #1938 Joined: Sun Jan 25 2009, 12:44PM
Location: Romania
Posts: 699

update: I drilled a hole in one of my home's walls, and mounted the enclosure outside. Installed power cable and crimped ethernet cable. Data is getting in to my server already.

Looks like there's a nice 22* temperature outside with normal background radiation, so time for a break for a little walk :)

Proud Mary

Sat Oct 20 2012, 04:08PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Have you thought about a Java real time plotting/charting graph so users could look at dose rate trends - past 24 hours, past week, past month, past year?

This is exactly what people will want to see next time there is a nuclear accident. Is it going up, and by how much?

radhoo

Sat Oct 20 2012, 05:29PM

Registered Member #1938 Joined: Sun Jan 25 2009, 12:44PM
Location: Romania
Posts: 699

Yes, this is exactly what I'm planning to have! I'm currently working on this part, and will probably have a first version by tomorrow.

first previous 4 5 6 7 next last

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob