GM Tube Info

Disclaimer: There are some links here to sources for tubes that I have not used myself. They are only for your information, and I'm not suggesting or recommending any particular source.

GM Tubes Evaluated:
This section should help you get a better idea about how some GM tubes perform, and may help in selecting a tube. It is by no means comprehensive. In fact, it's limited to only the tubes I've tested. When possible, I added the link to the specs and added my own impressions of each tube. Performance of each is on the comparison chart that follows.
  • SBM-20 - The quintessential Russian tube. Lower priced, and more sensitive to beta and gamma than most. This is still one of my favorite all around beta gamma tubes. The STS-5 is pretty much the same tube and may be cheaper.
  • M4011 - Glass tube made (new) in China. Claimed to be more sensitive than the SBM-20. Although I once got high readings, it now seems similar to the SMB-20. I am wary about this tube. It's quite possible this tube is light sensitive which would explain the initial high counts. There is a discussion about it's light sensitivity hereIt is reported that a good CPM to uSv/h conversion rate for this tube is 153.8.
  • SI-180G - Very sensitive to background (84 CPM) but less to other sources. This might be the best tube I have for background. (I like the way it looks too.) However it's similar to the SI-29G on the other samples. A user (Justin) said that a CPM to uSv/h rate of 321 gave a similar uSv/h as his GammaScout. 
    (I solder the wires to it like this.)
  • SI-29BG  - (specs: here) Good, smaller, cheaper, substitute for the SBM-20. This is a nice tube for compact projects. Fairly good sensitivity at a good price. In my tests this tube is about 70% as sensitive as the SBM-20. 122 might be a good ratio for uSv to CPM.
  • SBM-10 - (pulled it out of this ). It's tiny, but may be good for hunting Uranium glass. I wouldn't use it for much else.
  • LND 712 - The gold standard in end window alpha tubes. Good beta / gamma sensitivity and my best end window alpha detector. Nice size, but not cheap. You can purchase from LND directly. They normally have a minimum order value $100, but you could try writing them to see if they will make an exception. The only other source I know of is this one. Most people use a conversion ratio of 108 CPM to uSv.
  • SBT-9 -  Nice alpha tube for the price if you can get one. It was used in Russian space vehicles. A good poor mans substitute for the LND 712 if you're looking for an alpha tube. The  SBT-9 seems to be about 50% as sensitive as the LDN-712 for beta and gamma, and about 35% as sensitive for alpha.  (85 might be a good ratio for uSv to CPM if compared to the SBM-20 ratios) This tube and the STS-5 was used in Russian spacecraft.
  • LND 7317 - alpha pancake tube - Yippee! finally have one. At 60 cps/mRh for 60Co the CPM to uSv/h ratio for the kit comes out as 360. However, the "Inspector", which uses this same tube, seems to use a ratio of 330. (or 3300 for mR/hr)
  • RFT MKD VA-Z-115.1 - This is an East German glass beta/gamma tube. Reasonably priced, and small (~2"/~5cm). Overall, it seems to be the most sensitive of the beta/gamma tubes.  Unlike most glass tubes, this one is painted black - probably to reduce the effects of ambient light. atomic.dave runs this tube with the kit at 450V with at 2M anode resistor and a CPM to uSv/h conversion ratio of 227.
    Note: I also bought its little brother -
     VA-Z-114NR. What a difference one digit makes!  This tube is not sensitive at all - background is 4-5 CPM - thorite sample ~440 CPM vs. 16.3k CPM for the VA-Z-115.1.  Lantern mantle ~72 CPM. Not recommended.
  • SBT-11A - My favorite alpha tube. It's a sensitive alpha tube, comparable to the LND 7317 in some ways. It also seems sturdier and is certainly cheaper. The data sheet shows the Recommended Operating Voltage as 400V and it gives: 44-49 cps/uR/sec. Therefore the CPM to uSv/h conversion ratio becomes 318 CPM/uSv/hr. (thanks Brian) Pin-out diagram is here.
    [12/8/17] I learned that there is also an SBT-11 (no 'A') version of this tube that is reportedly less sensitive than the SBT-11A. My understanding is that tubes with thinner mica (as measured after production ?) receive the 'A' designation. To what degree the thicker mica affects alpha sensitivity is unknown. Since the non-A version looks identical to the A version, some are sold as SBT11A. If you want the 'A' version it's best to ask the vendor. Recently I heard that there may be counterfeit "A" versions. Great! I am not clear at this point on how to tell.
  • JAN (LND?) 5979 - I got this NOS tube from LeedsRadio on Etsy. (He has lots of other cool stuff.) I also see that Electronic Goldmine now carries this tube. It's an end window alpha tube. I don't think LND makes it anymore - the one I bought is from the 70's. It's a solid thing! It was added to the comparison chart for alpha tubes below. Note that it's pulse amplitude seems to be pretty low, so you may get fewer clicks than counts with it. (This is due to a difference in sensitivity between the two circuits.) I run this tube at 700 - 750V. I'd estimate the CPM to uSv ratio to be about 44.
  • SBM-19 (STS-6) - Well here's a fun tube! It's huge, cheap and very sensitive. I'm getting background in the range of 160 CPM. If you can spare the space this is a great tube for background. I found a few datasheets here and here. It's now added to the comparison charts. Compared to the SBM-20, it gives 5x the background (because of it's size), but only ~1.5x the readings for the other test samples which are more point sources. If I were using this tube for background, I'd use 5x the SB-20's rate of 175 or 875 CPM / uSv/h. For point sources I'd use 1.5x the rate or 263 CPM / uSv/h. (just my best guess )
  • SI-8B / CI-8b - Sort of the Russian version of the LND 7313. It seems much sturdier due to a thicker mica window. It's also about half the price. I just got this tube but it's now on the comparison charts. Here is a nice translation of the spec sheet and here is a pic of how to wire its socket. I have heard that people are using a conversion rate of 430 - 450 for this tube. I did find this tube to be light sensitive and you should also run it within the specified (narrow), voltage range - 360 - 440V. Higher voltages give suspiciously higher counts. I also found these specifications (from the manufacturer?) that have some new info.
    Regarding the load resistor: The Russian spec sheet lists the max current through the tube as "18.2 mkA". This has been often translated as 18.2 mA. That would make the load resistor only 22kΩ at 400V. However I found the Russian use of "mk" is actually "micro". This makes the load resistor 22MΩ at 400V which is more reasonable. Note that the diagram in the manufacturer specifications (linked above) show the load resistor at 5.1MΩ. However, I went with the original spec and I increased the load resistor from 5MΩ to 20MΩ. I got ~5% more counts at 20MΩ so I left it at that. I think the less current through the tube the better. So the short answer is that I recommend 20MΩ.

I made the comparisons below by attaching each tube to the Geiger Kit and testing with the same set of samples. These are non-scientific tests. 
For some consistency, the test parameters were:
  • only CPM is compared, so uSv conversion ratios are not a factor 
  • the sample was placed directly on the tube when possible with the goal of getting maximum counts
  • I did not try blocking alpha when testing alpha tubes
  • counts were recorded when they stabilized with a small variance
  • tube voltage was set within the approximate operating range
  • the on-board 4.7M anode resistor was use unless noted
  • For the LND-712, since its spec'd for an anode resistor of 10MΩ, an additional 4.7MΩ resistor was added at the anode of the tube. Mine died at some point, so values for the thorite, radium and Cs137 samples in the rest of the column were made with a tube from the DR-M3 probe. It's very similar to LND712 in size and design but slightly less sensitive. It was tested in a housing with a metal grid. Samples were placed on the center ~8mm above the mica. The LND 7317 was also used with a Bicron Analyst with very similar test results for all samples. (Yay Kit!)
  • Since the SI-8b / CI-8b seems to be light sensitive, readings were taken in a darkened room. It was run at 400V with at 5MΩ load resistor (Later changed to 20MΩ, see above)
Click on each chart below for a larger image. The Excel spreadsheet I used to make the charts below can be downloaded here.

The chart below shows the CPM readings I got with "non-alpha" GM tubes. High energy samples samples were not included on this chart so that low energy sources would show up better on the chart.

The chart below shows CPM readings with "alpha" GM tubes. High energy samples samples are included on this chart.

The chart below shows CPM readings for all GM tubes tested. High energy samples samples are included on this chart.

From the results, I made these observations . . .
  • Results depend heavily on the physical size of the sample compared to the detecting area of the tube. A sample that only covers a part of a larg tube and all of a small tube will weigh in favor of the small tube.
  • The SI-180G kicks butt when it comes to measuring background. However, surprisingly, it's more like the SI-29BG when it comes to the other samples. It may have to do with the surface area of the tube. The large surface area picks up more background - which is all around the tube. But with a sample that only touches part of the tube, the surface area no longer plays a roll.
  • The ratio between counts from the LND-712 (with alpha blocked) and the SBM-20 can be calculated. For selected tests these are:
    background = .78 (statistically a small sample), uranium ore = .85, smoke detector pellet = .82
    Note that the commonly accepted conversion factors for CPM -> uSv/hr are 123.14 for the LND-712 and 150.51 for the SBM-20. The ratio between these is about .82. This compares nicely to the ratio of counts between these tubes for the selected tests. 
  • The LND 7317 is the hottest tube I tested. Since it is an alpha pancake style this is to be expected. 
  • The East German VA-Z-115.1 looks like the most sensitive beta/gamma tube. In tests where there was lower counts than other beta/gamma tubes (lantern mantle) it was likely due to the smaller surface area.  
  • The SBT-11A gives about 70% of the readings of the LND 7317 when the samples are smaller than the active area of the tube.

Notes on Other GM Tubes:
Some kit builders have reported success with other tubes. This is what I have heard from them. 
Since I have not worked with them myself I can not guarantee how they will work.
  • SI-1G - Glass tube more sensitive than the SI-3BG. Working Voltage ~450V
  • SI-3BG - This is not a very sensitive tube. I'd avoid it if you can.  It is reported that a good CPM to uSv/h conversion rate for this tube is 22.5.
  • LND-7231alpha tube
  • LND-7616.- alpha tube I also heard from two people who had success with this high voltage tube (750-950V). It's really over the max for earlier versions of the kit (650V max) .However, it seems to working at the max voltage. (Newer versions of the kit with the v4.0.2 board are ~940V max.) Note however, that the anode resistor is spec'd at 1M for this tube and it is necessary to have this value. Best solution is to short R7 (4.7M) and put a 1MΩ at the anode connection. One was purchased here.
  • Victoreen/ Anton 6107. - alpha tube Severial builders have reported the kit working with this 700V tube. I've heard it's a similar tube to the LND-7616 mentioned above, but I've also heard contrary - so I don't know for sure. I've heard that the value (and placement ?) of the anode resistor (10MΩ) may be critical to get this working properly. Here is one users experience before and after adding the anode resistance:
    "I would turn up the HV until counts started...and I'd then try to find the plateau (the activity when the counts first started was way too low)....but the Plateau was basically not'd raise the voltage, and the tube would start false counting-and of course, much above that, avalancing would start. With the extra resistor added, the plateau WAS there, and adjusting the HV worked well. At that point, the tube seemed to work fine. So I'd say that, at least on my kit, and my tube, it works well." Another builder confirmed this tube works with the 10M anode resistor at the tube and 640V applied.
  • Anton 6993 / OCD-D-103 - (Whichever is in my CDV-700 Model 6.) This tube has a very low pulse height < 600mV and requires ~900V to run. Because of the low pulse height it will not count on the stock Geiger Kits. However, I found that if you remove R9 from the board will count. I have not done extensive testing with this setup so I can't recommend using the kit with this tube. 
  • Raytheon CK1026 - As reported: "From the Raytheon data sheet the anode resistor is 1 Mohm and by experiment my tube operates at 960 volts. The background count varies from about 12 to 40 CPM running averaged over 30 seconds and with a 1 uCi Cs137 source the count (gamma + beta) exceeds 2700 CPM."
  • Phillips 18504 - alpha tube This tube can detect alpha, beta, gamma and neutron emissions.  (See the Gallery for a counter made by Tomy that uses this tube.)
  • SBT10-A - alpha tube A good tube for measuring food? This report from Pedro (thanks for keeping us in the loop):
    "Finally the replacement for the faulty SBT10-A tube arrived and after connecting it to the arduino the counter shows a 160/190 CPM with all the segments connected to a 10Mohm resistor. The anode voltage seems to be critical for this tube as it creates avalanches if anode voltage is over 370v. Care must be taken in the voltage calibration as most voltmeters will show erroneous reading due to intrinsic impedance, so in my case the adjustment was made to 330v (digital readout) but the real voltage without voltmeter load was 370v. In my case this tube detects 2400 CPM for a lantern mantle at 10cm from the mica window without alpha blocked and 690 CPM with alpha blocked."
    [7/22/16] A customer reports that 10MΩ is too low for this tube. He had expected results with as high as 30MΩ. He used that resistance on every anode of the tube. He also said that the sensitivity to the HV setting described above is explained by too low of anode resistance.
    [9/8/18] Another customer found that a 10M anode resistor (for all segments tied together) was about right. Going to 20M reduced counts. background for him was 120-160 CPM.
    For the conversion ratio, I calculated using Method 2 below, and got a ratio of 1641 CPM / uSv. Seems high, but is a sensitive tube.
  • FHZ76V - A subassembly produced by Frieseke & Hoepfner containing Philips/VALVO 18550 GM tube. Used with the kit in this review. The tube was used in F&H FH40 GM counter - German manual here. (thanks to Reinhard for this info)
  • MC6 - another giant tube! Even bigger than SBM-19 - 260mm (10.25 inches)! This is a glass tube and is reportedly sensitive to UV light. One user was getting 2000 CPM background with the tube exposed to light. After covering the tube, the background went down to a more reasonable 125 CPM. There is a review on it here.
  • CDV-700 RP - is reported to work fine. He ran it at 925V.
  • Ludlum 44-9 - (LND 7311) - this probe is reported to work. Nice video on the probe here.

Converting CPM to a Dose Unit (uSv/h, mR/h): 
This is a little messy but I don't think it's my fault! Over the years I have read a lot about conversion rates on the web and the information on this subject varies widely. It's interesting to read the list compiled by 'paguyu' of the various conversion rates used by different sites. (PDF here) So as nice as it would be to state one method of making the conversion, I have yet to find what I feel is  a single authoritative method. So instead I will simply report the various methods I have run across. I am not married to any of them frankly. 

We know that a GM tube provides pulses that correspond to "events" in the tube that occur from it's interaction with ionizing radiation - gamma rays, beta, and alpha particles (for some tubes). The events are counted by the Geiger Counter over a time period and eventually result in counts per minute - CPM. This is the raw data, but in some ways it is the most accurate due to it's simplicity. If you know what the CPM are for 'normal background' with your tube, the difference measured will give you a good sense of the intensity. Personally, I always tend to think in terms of CPM.

However, different models of GM tubes vary greatly in their sensitivity. This makes it difficult to compare readings between different counters or to published rates. Enter the "dose unit". Now at this point the subject can get very complicated, and I will avoid getting into detail. Suffice to say, that a dose unit defines how much radiation is absorbed. ("absorbed dose" and "equivalent dose" will be considered the same here.)  Both uSv/hr and mR/hr will be considered to be dose units here.

The main point is that the spec sheets for most GM tubes define the CPM (or cps) that are equivalent to some dose unit (usually mR/hr). Since GM tubes also have varying sensitivity to different isotopes, there may be several values listed. Lets look at a part of the spec sheet for the SBM-20: 

Gamma Sensitivity Ra226 (cps/mR/hr)29
Gamma Sensitivity Co60 (cps/mR/hr)22

The object here is to convert this information into a ratio of CPM to dose unit - so how many CPM = 1 uSv, for example. These ratios are then set via the menu in the Geiger kit. The kit actually supports storing of 2 of these ratios, with the one in use selected by a switch. As shipped, a default ratio for each is provided - 175.43, the most common for the SBM-20 tube, and 100, common for the LND-712.

The following describes the methods of calculating this ratio that I have run across. They use the SBM-20 example above, but they should apply to other spec sheets as well.
Method 1:
This is the most straightforward method and the one I originally used. It only uses the Co60 value from the spec.: 22 cps = 1 mR/hr
That's counts per second so multiply by 60 to get CPM: 22 x 60 = 1320 CPM / mR/hr
If your dose unit is uSv, convert from mR. The common conversion is 1 mR = 10 uSv. So divide the CPM by 10: 1320 / 10 = 132
(You could have just multiplied the 22 cps by 6 to get the same effect.)

So 132 is the ratio that can be entered into the menu of the Geiger kit if you are using an SBM-20 GM tube.

Method 2: (thanks to Mike)
This starts by simply taking an average of the two cps values for the two isotopes. So: 29 + 22  / 2 = 25.5 cps
Multiply counts per second by 60 to get CPM: 25.5 x 60 = 1530 CPM

Now we introduce the absorption rate. There is a lot that can be said about this, but in a nutshell, it is the rate at which ionization events are absorbed into a material. (Jorge M. has written me with some very good details on the subject. You can read that here. and there are references to others using it it here and here along with Japanese Wikipedia article hereIt appears that also built into this rate is a conversion from mR to uSv (1mR = 10uSv). Frankly, I am not too clear on how it is integrated!

The value for the absorption rate to be use is 8.77 which I understand is for air, and 9.56 is for soft tissue. 
So finally the 1530 CPM is divided by the absorption rate of 8.77 to get ratio between CPM and uSv: 1530 / 8.77 = 174.46
(Or multiply 25.5 cps by 6.84)

This is the ratio that can be entered into the menu of the Geiger kit if you are using an SBM-20 GM tube. It is also very close to 175.43 (or it's inverse .0057) that is commonly found on the web for that tube. To be consistent with this common ratio the Geiger kits use 175.43 as the default setting. 

Method 3:
This method is similar to Method 2 except the two isotopes are not averaged and only Co60 is used.
Therefore: 22 x 60 / 8.77 = 150.51
(Or multiply 22 cps by 6.84)

Method 4: (thanks to Rick)

It is more common to calibrate to Cs137 than Co60 In the process of truly calibrating an SBM-20 with a lab sample of Cs137, my friend found the conversion ratio turned out to be 65% too high based on the expected dose-rate. It turned out that this is related to the fact that ratio is derived from Ra226 and Co60, which are the only only two isotopes specified for most Russian tubes.(If your spec sheet has a value for Cs137 than just use Method 3.) From several sources, he found that "The energy response for a GM tube will always be different for Cs137 compared to Co60, so an instrument will always show a higher reading at Co60 energies compared with Cs137 energies for a given dose rate."  and he also found that the value for this difference is 65.5%. (I don't have his sources at this writing.)

So to begin, we start with the ratio for Co60  already calculated using Method 3: 150.51 CPM/uSv/hr

To calibrate for Cs137 multiply that ratio by 65.5%: 150.51 x .655 = 98.58 CPM/uSv/hr. That would be the ratio to use for Cs-137 samples.
(Using this ratio reportedly gave him readings of within <1% of measurements expected in his lab test.)

So there you have it! Four methods, four ratios! Take your pick.

Using one of these methods did you actually calibrate your Geiger counter? I would say no. First there are variations within the same model of GM tube, based on factors like the age of the tube, and the bias voltage used. Moreover, to truly calibrate your counter you must use a standardized check source of the isotope you want to calibrate to (usually Cs137 or Co60) and use procedures which include such things as distance from the source, tube geometry, etc. to fine tune your ratio of CPM to the dose unit you want to use. However, what you have done is to get a ballpark estimate that hopefully will be comparable to others using the same conversion methods. 

The Anode Resistor, Stray Capacitance, Cable:
To be honest, I used to be pretty cavalier about these things - connected the tube - it clicks - yippee! But now I'm a more informed (and frightened ;) person. 

The suggested value for the anode resistor is usually given in the datasheet for the tube. The kit uses 4.7M for this value. This is about right for the SBM-20 (5.1M) but the LND-712 needs 10M. So I put another 4.7M in series right at the tube. It's  best to have the anode resistor right at the tube anyway, as it isolates the stray capacitance from the tube. 

So what happens if you don't have the right value? Some have reported problems where the tube is in constant avalanche. In one case, the problem was solved by lowering the anode resistance. (This seemed counter intuitive to me, but Ed straightened me out (again) and said "I know it seems the opposite should happen, pulse should get bigger, and it does to a point, but the more capacitance, the more the Plateau tilts up vertically, and then you get to the point where it’s getting smaller because  the operating point voltage swing is getting shrunk, which means lower magnitude pulses , and lower dead time.") So if your experiencing this type of problem with an uncommon tube, you might look at the spec sheet for the recommended anode resistor. 

Stray capacitance can also cause problems. I've read it can increase the dead time of the tube which results in lower counts. Stray capacitance also makes it "harder" for the tube to discharge. This, in turn, can shorten the life of the tube. 

Having the anode resistor at the tube, as previously mentioned, is perhaps the best defense. If you want to put all the anode resistance at the tube, you can easily do this by  jumpering R7 on the board (RL_JMP) and installing the appropriate resistor right at the anode clip of the tube. Careful! The HV at the screw terminals may now bite when touched. Keeping the wires to the tube short and separate also helps. 

If you're going to use a cable,you should also  keep in mind that the kit uses the "cathode sensing" technique. This means that events are sensed from the cathode (-) side of the tube. This is the preferred method of sensing by several accounts. However, it means that the cathode is not at ground potential. Therefore it's  a mistake to run the cathode side of the tube on the shield of the cable. Better to run with a 2 conductor cable. I did a quick experiment with 10 feet of an old multi-conductor shielded cable (26 AWG) and found no appreciable loss of counts with a 450V tube. The anode resistor was still on the kit and I did not ground the shield.

BTW a very good source with technical information on these two subjects can be downloaded here.

The Effect of "Dead Time" on Counts :
Dead time is the time after an event in which the tube will not register a count. It's like the tube is resetting. Most specs on tubes list the dead time (in uS).

Someone pointed out the formula for calculating the counts lost to deadtime based on the observed count and the published dead time for the tube. (Thanks Al!) I thought it would be worth writing it up here. You can also read more about this subject here

To use an example from the chart above, I got 5253 CPM from the mantle on the SBM-20. The dead time for this tube is listed as 190 uS. The formula is:
Time is expressed in seconds, so counts are counts / second, and deadtime is in seconds. So the fist step is to make these conversions . . .
5253 CPM / 60 = 87.55 CPS  and 190 uS = .000190 seconds
Plugging this in, we have . . .
ACTUAL COUNTS = 87.55 / 1 - (87.55 x .000190)
which is . . .
ACTUAL COUNTS = 87.55 / 1 - .01663
or . . .
ACTUAL COUNTS = 87.55 / .98337
or . . .
ACTUAL COUNTS = 89.03 CPS or 5342 CPM

So ~89 CPM was lost due to deadtime (5342 - 5253) - a 1.7% loss due to deadtime.
Not much at this lower count rate but it becomes significant at higher rates.

Some Videos:
The thorite sample came from eBay. To protect from small pieces breaking off I found that placing the sample in a parts zip-lock and blasting it with a heat gun makes a nice thick shrink wrap. Below is a video of a small piece of thorite tested with six different GM tubes. You will see that there is a difference between tubes, but that it's not as great as the difference you get when measuring background.

The video below shows both the LND-712 and SBM-20 connected to the kit at the same time. It's easy to see the LND-712s response to alpha.

Connecting Multiple Tubes in Parallel:
Someone asked if this was possible. So I strung 7 SI-180G tubes together (rather sloppily) and got some nice high readings. I got about 7x the normal background for that tube.
So 450-600 CPM for background, 6744 CPM for a mantle, and over 24,000 CPM for the thorite.

Experiments on Range and Directionality of GM Tubes:

And speaking of connecting multiple tubes together!

The picture at left is one of the experiments Utsunomia-san is doing.
He also has two videos on testing the range and directionality of various tubes and configurations.

You can view these videos here and here.

Inside the SBM-20:
Ever wonder what's inside those Geiger tubes? Marek sent me these pics of the insides of an SBM-20 that arrived smashed in the post. 

I thought it would be handy to have a cutaway of an SBM-20 tube that is that is a bit more in tact. So off to the belt sander!

And while I'm talking about the SBM-20, here is a chart of the types. (really all the same size) ...

Experiments You Can Try:
Experiment #1:
If you have a tube that responds to alpha . . .
Stretch a piece of cloth over the end of your vacuum cleaner (shop vac). Let it run for about 10 minutes.
If you have a basement, do it there.
Measure the cloth. I got about 240 CPM! (using an LND 712)
I let it sit for about 12 hours and didn't get a noticeable reading anymore.
Thanks Daniel More info on this here.

Experiment #2:
This is much simpler. After a rain, run a paper towel over your car. (assuming it was outside!) 
I wiped about 1 sq. meter of the car and squeezed out the excess water. 
Measure the towel. I got 480 CPM average with an SBM-20 tube.
(About what I get with my Uranium ore sample!)
I measured it again 2 hours later and only got about 100 CPM.
At that point I also checked for alpha with the LND 712 but didn't get a significant reading.
I measured it again 20 hours later and it was down to 41 CPM.
I was more surprised about how fast the count went down than about the high count in the first place. The effect is likely due to washout of radon daughter products;  218Po, 214Bi and 214 Pb which have half lives of 3, 20, and 27 min respectively. This is as opposed to the radon itself, who's half life is 3.8 days. (thanks for the link Sheldon)

Experiment #3:
"A magnetic field affects radioactive rays differently depending on the type of ray. Alpha rays (heavy, positively charged particles) are deflected slightly in one direction. Beta rays (light, negatively charged electrons) are deflected strongly in the opposite direction." Gamma rays are not deflected.
 Here is a video that someone did of this experiment. 
More interesting is this high altitude balloon experiment regarding the shielding effects of magnetism at  high altitude.

Links on GM Tubes and Radiation:
Some have been mentioned above, but these are useful links that I've collected..