Sunday, May 13, 2007

Introduction

Note: If you only have time to view one clip, please see the *second* clip at THIS LINK (CLICK HERE). It runs about 5 minutes and tells much of the story.

This article contains video clips of the tests I performed of Homeland Safety International's (HSI) Sniffex explosive detector. These tests were conducted April 26-27, 2006 in Anaheim, California at the Seventh Annual California Safety and Security Conference. The conference was held for emergency response and law enforcement personel and HSI had an exhibitor's display booth at the meeting.

I asked the president of the company, Paul Johnson, and the Vice President for sales, Lee White, if they would allow me to do a double blind objective test of Sniffex using smokeless powder samples and table salt "controls". They agreed and, to their credit, cooperated fully with the tests which took about four hours. I saw nothing to suggest that they tried to "cheat" or in any way influence dishonestly the test results. On the other hand, several of Mr. Johnson's requests and remarks during the course of the testing suggest that he is not familiar with scientific method and double blind testing.

This blog is mainly about the video and audio clips I obtained during the test. I apologize for the less then ideal visual quality of the clips. I had to conduct the tests and document them at the same time which means that some of the time, I could not point the camera accurately or hold it perfectly steady. Still, a careful examination of the video and audio provides an accurate description of the tests.

For a much more comprehensive look at the history and an analysis of Sniffex, I strongly recommend that you read this excellent blog, done by another author:

"QUESTIONS ABOUT SNIFFEX" (this is a clickable link)

The author of that blog was also kind enough to post the full text of all emails and letters which I exchanged with Sniffex/Homeland Safety International about these experiments.

Following the complete failure of Sniffex to detect 90 grams of smokeless powder explosive and four loaded 9mm. cartridges, I expected some followup. I made several attempts to find out what additional investigation or tests Homeland Safety International had made to account for and correct the failure of their Sniffex detector. These attempts are documented on the blog linked above. The only thing resembling an explanation I received was a copy of brief email from "the inventor" saying the amount of explosive was small (it wasn't) and they should try more tests with the same gun powder. Although I offered to retest Sniffex with any amount of any explosive they liked, nobody contacted me to repeat the tests.

To this day (May 2007), Sniffex is advertised worldwide as an effective explosive detector suitable for such diverse tasks as screening aircraft luggage, large sports events, and even land mines. The evidence I obtained is that it would be extremely dangerous and totally foolhardy to use Sniffex at all for the detection of nitrocellulose/nitroglycerin based explosives. There is nothing I have found on the internet or elsewhere that suggests that Sniffex would be effective at finding any explosives of any type and there is much evidence to show that it is not.

The video clips run in length from just under a minute to just over five minutes each. The eight clips total about 27.5 minutes and cover the essentials of demonstrations and experiments which took about 4 hours in real time. I hope anyone considering risking their life and limbs using Sniffex to detect explosives will take the time to see and hear them all.

At the end of the blog, you will find some disclaimers about trademarks, some information about me, and an explanation of how the videos were obtained.

Note: The video clips are stored in Youtube. Youtube sometimes experiences short delays. If the clips do not appear, please be patient or reload the blog page.

I welcome official comment from Homeland Safety International (HSI), the manufacturer or the inventor of Sniffex and I will be happy to post any text or video of reasonable length that they provide me in response to the information below.

To contact me, simply leave a comment in the blog which includes your contact information and I will be in touch with you. To write a comment, please click on the link COMMENTS which appears after each posted segment. The link is in small type but you can find one at the end of each post directly after the time stamp on the post. Comments are moderated to avoid spam so expect some delay between writing and the appearance of your comment on the blog.

Sniffex "operating principles"

I approached Lee White, Vice President for sales of HSI, and asked him how Sniffex works. This is what he said:



Things to note: Sniffex uses no battery or other power source. It runs on "the human electric field" and uses the human body as an antenna. That makes no scientific sense. The claim in the clip is that Sniffex can locate most types of explosives within "a meter or so". The videos that follow prove conclusively that it can not.

Sniffex demonstration with known explosive locations.

After talking with Mr. White, I witnessed a live demonstration of Sniffex conducted by Mr. Ernest E. Griesser of Buena Park, California. His business card says he is a consultant. Mr. Griesser used Sniffex to locate "four bullets and a little bit of fertilizer" inside a closed steel box. Please see this clip:



I witnessed several demonstrations like this one and I was told that the demonstration was repeated many times over several hours without any failures or hesitation. The effect of the antenna swinging towards the target was so striking that the lady dressed in pink in the video exclaimed out loud "Oh my God!". The same four bullets used by Mr. Griesser for these demonstrations, with the addition of three ounces of gun powder, were used in my last experiment of the afternoon. In that experiment, Sniffex was completely unable to find these same explosives when the operator did not know their location.

Sniffex test with "open" samples

In the video clip below, I explain to Paul Johnson and Lee White how I made the samples that we are going to use to test Sniffex. In the second portion of the clip, we move to a hotel hallway where we place the sample of salt and the sample of gun powder at a distance from each other as specified by Mr. Johnson.




Please note that again, when the experimenter knows where the explosive is, he gets a good deflection from Sniffex's antenna. In this test, with one ounce of smokeless powder, Mr. Johnson gets a strong positive result from the gun powder and no deflection of the antenna from the salt sample. Mr. White does not seem to reproduce the result. Mr. Johnson acknowledges that knowing where the sample is located can affect the results. I offer to perform the double blind test and Mr. Johnson agrees to it.

Sniffex test with one ounce (30 grams) of hidden smokeless powder

If you only have time to view one video segment, please start with the SECOND video clip below. You can return to the main article by using the BACK button on your browser or by clicking here.

In the first clip below, you can see that I placed 10 small manila envelopes along a hallway of the convention hotel, about 10 - 15 feet apart, in accord with directions by Mr. Johnson. Nine of these envelopes contained table salt. One contained one ounce of smokeless powder (about 30 grams). You should be able to see the envelopes on the floor, along the far wall in the video.

Although you will hear Mr. Johnson say that this is a "small amount", it's the same amount of explosive contained in about 15 bullets or in a small anti-personel landmine. In a later clip, you will hear Mr. Johnson and Mr. White claim that they found a single bullet in an entire prison ward. In several advertisements for Sniffex, it is advocated for land mine detection.



In the above video, you can see how Sniffex is used. You can also see that although the process of searching for explosives with this device is very slow, it gives no clear indication. Mr. Johnson repeatedly mentions that he is experiencing "user error" but does not explain what it is or how to eliminate it.

In the next clip below, Mr. Johnson is unable to explain why in an earlier demonstration of Sniffex, his associate, Mr. Griesser, was able to locate just four 9mm cartridges and now, Sniffex can not seem to find an amount of smokeless powder sufficient to fill more than 15 such cartridges. Finally, Mr. Johnson selects three envelopes as likely and one, number 9, as the most likely candidate. We take these three envelopes to another even longer hallway and spread them about 50 feet apart. We place all the "rejected" envelopes into a single plastic bag and that bag is also placed in the hallway. Mr. White then searches these four items for the explosive.

Finally, Mr. White and Mr. Johnson agree that one envelope, number nine, contains the explosive. The envelope is opened and contains salt. All envelopes in this set of ten are then opened to prove that one contained gun powder.

IF YOU ONLY HAVE TIME TO VIEW ONE CLIP, THE ONE BELOW IS RECOMMENDED.
It's a bit less than 5 minutes in length.



Sniffex clearly failed this test. You can return to the main article by using the BACK button on your browser or by clicking here.

Another test with with smokeless powder and four 9mm. cartridges and a story about Folsom Prison

Mr. Johnson, puzzled that he could not detect the gun powder sample, asked me to place four 9mm. cartridges (which he calls "bullets" in the video) under one of the envelopes for a quick retest. He only wanted to work with three envelopes. After explaining that this was too easy a test because of the small number of envelopes, I placed the four 9mm. cartridges, as requested, along with an ounce of the smokeless powder in one of three envelopes. These were the same "bullets" that Mr. Griesser had used to demonstrate Sniffex earlier in the exhibit room (see video #2 above). The other two envelopes contained table salt. The envelopes were spaced, as you can see, about 10 feet apart. I offered to move or relocate the samples any way the Sniffex users wanted to. Again, even with only three samples to choose from, and despite both users getting the same clear indication (middle envelope) they selected the wrong envelope.



At the same time as Mr. White was trying to locate the explosive with the Sniffex, you can hear Mr. Johnson telling me about a previous demonstration. He said that Mr. White had found a single ammunition cartridge (bullet) in an entire ward of Folsom Prison during a demonstration they gave to a warden's meeting. Mr. Johnson could not explain why that had worked but here, Sniffex seemed to fail each test.

In the last videos below, you will see how Sniffex fails again, using two different instruments operated respectively by the president and vice president of the company and using more than 90 grams of smokeless powder-- a substantial and dangerous amount.

Sniffex tests with 3 ounces (90 grams) of smokeless powder and 9mm. cartridges

The clip below starts with Mr. Johnson explaining why it's "easier" to find explosives with Sniffex when you already know where they are (!). He goes on to describe, apparently unknowingly, the "ideomotor effect" (click this link for a description of it.) Meanwhile, Mr. White is using Sniffex in an effort to locate an envelope containing 90 grams of smokeless powder and four 9mm. loaded cartridges. Again, these are same four cartridges they used in their earlier demonstrations (see clip #2 at the start of this article). In this test, one envelope contains the explosives and the other nine contain 3 ounces of table salt each. This test is blind to the Sniffex people however I had to prepare the samples spontaneously during a break so I have a rough idea which one has the explosive. This does not seem to influence the test which turns out to be another complete failure of Sniffex.



At the end of the above clip, both Mr. White and Mr. Johnson agree on the sample envelope they think contains explosive. It contains salt.

In the final video clip, Mr. Johnson and Mr. White again search for the bullets and gun powder envelope, this time among only nine remaining envelopes, eight of which contain salt. Using two different Sniffex instruments, they can not get any indication of where the explosive is and after about a half hour more, they give up and admit that Sniffex did not work in these tests. Mr. Johnson promises to find out why. He briefly speculates that they have a "bad batch of devices" but then admits that the very same Sniffex devices worked fine for them before.



In the end, Mr. Johnson signs a receipt for a one ounce sample of the gun powder which we used for the test and which I provided for him. He promises to find out why Sniffex failed the test and to get back to me. He also says he is going to get additional independent testing, possibly with Sandia National Laboratories. As far as I know, no further testing of Sniffex was done. I wrote a final letter, sent by certified mail, spelling out in detail the risks to people and structures if they continue to sell Sniffex without being able to prove that it works reliably to detect explosives. I received no reply. I will be happy to publish one here if I ever receive it.

How test samples were prepared.

I purchased a pound of smokeless powder (Hodgdon H110) at a gun shop.
This product is used to reload pistol (magnum) and shotgun shells. Smokeless powder contains nitrocellulose or nitroglycerine or both and is exactly the type of explosive that Sniffex is advertised to detect in small quantity.

I weighed out three one ounce portions with a postal scale and placed them into small plastic bags which in turn, I placed into bubblewrap-padded opaque manila envelopes.



I then did the same for 18 envelopes which contained plain table salt. To avoid contamination, an all fresh setup and disposable plastic spoon were used to make the salt samples. I made two sets containing 9 salt envelopes and 1 gun powder envelope each. All envelopes were labelled only with a letter or number and sealed. After I was finished I did not recall which envelope held the gunpowder. For the experiments, I placed the envelopes where requested by the testers. I handled and placed them upside down so that nobody, including me, could see the labels.


I made two additional "known" samples -- one salt and one gunpowder, each of which was clearly labelled with the contents (below is the gun powder sample). These were used to show that the Sniffex devices were working in their customary manner so that nobody could later claim they were defective.

Saturday, May 12, 2007

DISCLAIMERS & GENERAL INFORMATION

All videos in this blog were recorded by me using a full size Sony DSC-W5 digital camera in plain sight, either at eye level, or hanging from a neck strap. The videos were made in public areas of the convention hotel during the time the convention and equipment demonstrations were in progress.

As you will hear on the video, I stated many times to Mr. White and Mr. Johnson that I would be writing a report. In my last letter to Mr. Johnson sent by certified mail with an email copy to Mr. White, I also told them that, unless I received a satisfactory response from them, the information would be placed on the internet. I received no response at all to that letter.

Video clips are edited for brevity and clarity. If someone needs the full unedited recordings for a legitimate research or documentation activity, please contact me by leaving a comment on the blog, including your contact information.

Trademarks used in this blog are not mine and obviously belong to the trademark holders.

I have no stock position in Sniffex or Homeland Safety International (HSI) and have never held any.

I received no payment or compensation of any kind from any source for travelling to the meeting nor for performing and reporting these tests.

My background is of little importance because the videos are self-evident. I have experience in the detection of explosives using mass spectrometry and gas chromatography as a result of past military service. I have conducted many experiments in which double blind method was necessary to avoid bias. I have a passing acquaintance with statistical methods sufficient for simple tests of the type described above.