Coil and ground plane

see how

erial

al_

com

s the

ing

h,

se

So far it's not possible to make a stupid comment/suggestion for our technology, our approach is too embryonic and has yielded unexpected results. Even naysaying teaches here. Every comment, suggestion, takes you to a new perspective as you try to comprehend the implications of the suggestion, then even further as you try to adapt/expand the suggestion, etc etc.

Only one detector at a time can look into the material. It gets very crowded at the surface. And down below, you subtend a very small region.

The diameter of that detector was a default 1/4 inch size. Default, because my fingers are too big to make much smaller.

Even though the detector is relatively large diameter, the system's 'horizontal' resolution is comparably about 5 times better than what i've seen out there. Down below the 2-5 mil range! It's just that I can see the rounded edges and want more! And, with the low noise floor and linearity of my system, I have room to manipulate mathematically! Still have better than 200:1 dynamic range left. When you consider the field going down attenuates at the inverse cube function and what gets reflected back up is also attenuated at the inverse cube function you can see that a 10 times improvement in signal to noise ratio only gains you 50% more depth - a lot of battle for small gains.

The image I sent you the probe was held as vertically as possible against the material, yet there was still a fixed approx 8 degree tilt off perpendicular, which shows up in the data by making ghost like edges to the images looking down alongside the 'tubular' rivets as you view down through the material and the material appears to turn transparent. It's very visible, as I shift the 'depth of view' display. Slide up and down into the material so to speak.

Effectively what I said.

e:

d see how

aterial

real_

gh.

n.com

=A0As the

ng

lking

pth,

hase

n)

en

p://

now

How many turns? Could you do something on fiber glass or flex? As you make the coils smaller do you lose signal?

Is that from the wire thickness?

It sounds and looks nice. A new product? They put a new gas line in outside where I work last year. Maybe two foot diameter. Before they dropped it in the hole there was a guy running some bit of electronics around the pipe. Similar idea?

George H.

which shows up in the data by making ghost like

see how

material

_real_

consultingwww.wescottdesign.com

the

talking

depth,

phase

I don't know if you can see full headers, but my proper email address is there. I would like to have the data just to find out if i can find a = way to "see" it.

?-)

ee how

rial

l_

om

the

ng

way

The original data comes in at around 30Mb/s and even on a 500MHz WinXP, the C++ code I wrote can keep up real-time! Plus, we're talking some major DSP activity on all that data, too. I had to abandon any attempt at using .dll library. The code I wrote went 6 to 30 times faster, so the program gets the calculations done in time.

Thank you for the offer. With an extra brain noodling on this problem, might make significant progress, but sorry, cannot share the data, too proprietary. Even losslessly compressed 30:1 the data files tend to be around 28 to 30 MB I've got some that are 128MB.

Yes, can do something with ANY material, just a matter of frequency.

Yes, wire thickness. Can't seem to violate those basic laws of physics. somehow they're all couched in reality.

Not familiar with testing of gas lines. He could have been doing an inspection of any welds, or a leak check with a 'gas sniffer'

way

It seems YOUR email address is munged by my google access

all I get is the sbcglobal.net after the AT sign and the joseph_barr... at the start. Bet EVERYBODY else can read it though

Click on the ... then prove you're human and google will reveal that secret.

could see how

material

_real_

though.

consultingwww.wescottdesign.com

=A0As the

going

talking

depth,

phase

depth

gain)

between

back...system

was

and

send

time

is

a way

The file size doesn't bother me. But the proprietary issue seems insuperable. Maybe Jason's method will help if direct contact becomes worthwhile.

?-)