DISCRIMINATING SIGNAL PROCESSING

[moodz]

version 23 hex for 30f4011 .... 4 sample pulses.

works fine on serial port

I dont use the LCD or rotary. ( tho it is in the code ).

moodz

[Tinkerer]

The one currently loaded is at the link below, remember to do the 10mhz crystal mod to make it run. The crystal goes across pin 9 and 10 and two 22pf capacitors from each pin to ground. Also you have to remove the satlock circuit and pullup resistor from pin 10.

http://www.geotech1.com/forums/showp...7&postcount=67


Here are the output pictures the first one is 100us per div and the bottom is 50us per div to give you a better look at the output of the THAT1510, the rails as mentioned before are 9 volts for the TX and +/- 5 volts to run the amp etc. It also has a separate +5 volt reg for the logic circuits.




Uploaded with ImageShack.us



Uploaded with ImageShack.us

Hi mickstv,

Thanks for the pictures.
How much gain do you have on the preamp?
Is this an induction balanced coil?

Can you show just 20us before TX switch OFF until 50us after switch OFF?

I am going to ask you to send more pictures with more details. At this link,http://www.irfanview.com/main_download_engl.htm you can download a totally free software that works real well for reducing the pictures to a practical size that is easy to upload directly to the forum.
Use "save for web", chose a size of 50 to 100k and 800 to 1000pixel image width and the pictures load up fast and are sharp.

all the best

Tinkerer

[mickstv]

Hi Tinkerer,

the amp gain is set to about 500, and the coil I'm using is a spidercoil using approx 26 turns of twisted pair mono coil and the coil has an internal diameter of 18cm, the coil then connects to a THAT1510 amplifier the input circuit uses two 1k input resistors from either side of the coil to the amp and 1N4148 diodes for protection.

I can detect on the output of the amp whilst monitoring with the cro a 0.4 gram nugget at about 10cm which I consider pretty good given that the amp is only running a gain of about 500.

With regards to pictures I can upload some better shots, but on the downside I cant get any better view than the second one using 50us per div, but I will try again and see what happens, if I can it will be in about 6-7 hours as its past 1am here time for some sleep.

Thanks for the replys.


Cheers
Mick

[Tinkerer]

Hi Tinkerer,

the amp gain is set to about 500, and the coil I'm using is a spidercoil using approx 26 turns of twisted pair mono coil and the coil has an internal diameter of 18cm, the coil then connects to a THAT1510 amplifier the input circuit uses two 1k input resistors from either side of the coil to the amp and 1N4148 diodes for protection.

I can detect on the output of the amp whilst monitoring with the cro a 0.4 gram nugget at about 10cm which I consider pretty good given that the amp is only running a gain of about 500.

With regards to pictures I can upload some better shots, but on the downside I cant get any better view than the second one using 50us per div, but I will try again and see what happens, if I can it will be in about 6-7 hours as its past 1am here time for some sleep.

Thanks for the replys.


Cheers
Mick

Mick,

a 0.4gram nugget at 10cm with only the preamp is good. So maybe you could just show me the cro picture where you can distinguish that.
Are you using a stick to present the nugget to the coil? Sometimes the coil picks up the empty hand very well.
With the twisted differential coil, the discrimination manifests itself at a different time than with the IB coil, so the sample timing given above, does not apply.

A good way to look for the FE discrimination, is to look at the RX signal wave form while presenting a crown cork or bottle top to the coil.
Presenting the crown cork in the vertical position strongly shows the FE reactive response.
The response of the crown cork in the flat position is dominated by the resistive response that is of opposite polarity to the reactive response.

Tinkerer

[mickstv]

Here's some pictures of no target and target, sorry for the poor quality again. Both shots are at .1volt/div and 2us/div. at about 8cm tried to take a photo at 10cm distance from target but the picture quality is too poor to pick up the difference but I can see it.




Uploaded with ImageShack.us



Uploaded with ImageShack.us

[mickstv]

I forgot to mention yes the nugget is taped to a 4 foot length of dowel and on the opposite end I have a 10gram lead sinker for testing. I checked the coil with hand movement but it showed no response on the cro at all. This coil hasn't even got any shielding on it and is poorly wound picture below..

[Tinkerer]

Excellent.

Now, there is some oscillation on the wave form.
Is this an artifact of the coil winding, or does this show that the coil is slightly under damped?

If you take a wide sample that covers several oscillations on this wave form, it will average the oscillations.

If you are looking on a specific spot, yo need to sample only that spot, that is a narrow sample.

Have you tried the crown cork?

Tinkerer

[mickstv]

Sorry I should have mentioned the oscillation is being caused by the powersupply I'm using, normally I run the circuit from battery but it was flat when those shots were taken. I haven't tried a crown cork yet but some more target testing will be happening shortly.

The main thing I was going to try and tie down first was ground response and mineralisation reduction. The area I want to test at has a high level of mineralisation, it would be good to see If I can make it's operation as quite as possible.



Mick

[Tinkerer]

Sorry I should have mentioned the oscillation is being caused by the powersupply I'm using, normally I run the circuit from battery but it was flat when those shots were taken. I haven't tried a crown cork yet but some more target testing will be happening shortly.

The main thing I was going to try and tie down first was ground response and mineralisation reduction. The area I want to test at has a high level of mineralisation, it would be good to see If I can make it's operation as quite as possible.



Mick

I believe that the ground response is mostly due to iron. Therefore, if you can discriminate FE (iron) then it should help with the ground response.

Tinkerer

[Midas]

Here's some pictures of no target and target, sorry for the poor quality again. Both shots are at .1volt/div and 2us/div. at about 8cm tried to take a photo at 10cm distance from target but the picture quality is too poor to pick up the difference but I can see it.

Don't be so sure
A little bit of scaling, some distortion correction to get all the grid perfect. Followed by some colour filtering for signal distintion and you get the following. White for no target, green for gold nugget. Actually knowing where to look you can just make it out the un-altered shots but this technique makes it really clear.

[mickstv]

I believe that the ground response is mostly due to iron. Therefore, if you can discriminate FE (iron) then it should help with the ground response.

Tinkerer

Thanks Tinkerer, I'll get back to target testing and post up the results.



Mick

[mickstv]

Don't be so sure
A little bit of scaling, some distortion correction to get all the grid perfect. Followed by some colour filtering for signal distintion and you get the following. White for no target, green for gold nugget. Actually knowing where to look you can just make it out the un-altered shots but this technique makes it really clear.

Hi Midas, did you use photoshop for that ?



Mick

[Midas]

Hi Midas, did you use photoshop for that ?

Mick

I sure did. But if you want an infinitely cheaper option then Gimp would probably be able to do similar things.