come abourd i need some bodies to build one

[simeon webber]

if your into designing metal detectors

i made one and you can view the schematic
by downloading the lite version of proteus vsm
www.labcenter.co.uk

then click this link http://homepage.ntlworld.com/vsm.libs/DESIGNS/METAL.DSN


http://homepage.ntlworld.com/vsm.libs/schematics.html

from the schematics page
of the vsm.libs site i run


i made a 4 coil radar design

gives 4 traces on a scope panel one for each tx or rx waveforms for each coil

coils can be disabled and the fireing order changed

i approched a few companies but there are blind to 1 coil or multitap designs

my coils are different shapes
interigation based on known metals tables
can reveal depth shape and metal type percentage showing chemical symbol

and likleyhood
on the display

a finds table can also be recorded to eeprom and all power levels and sensativity are adjustable and
sequencable giveing rise to many possibilities

i hope you like the diagram enough to help to deevelop this "metal engine" with me using c language for the micros
the radar scope section is complete and running with just minor adjustments to code to be made
the pwm and control i am slaving over night and day

pcb designs are avalible
psu section has been delayed by calculations of final currents
but i recon the design is on track for further test of this theory in a few weeks


if anyone after seeing it wishes to also work on it with me or help in any way
or want to see the code too date
let me know
at this email

[email protected]

below is some links also to a GIF file of the design



Click Here

[simeon webber]

Re: come aboard i need some bodies to build one

this is version 2 i added the pots chips

version 3 i plan to add extra features

notice that all gain control's double as waveform adjust also by different mode of display coils waveform
with the buttons for fine course stepping

three modes are planed for now

radar {detection mode}


coil adjust mode shows reflected pulse and foward tx pulse by time on a standard scope graticlue type
for each coil
this can be made into 4 panels displaying the different waveforms in each
this bit is done.......and was easy
using encoded pwm drive to the scope a/d

power level mode {the level of power can be adjusted for dry and wet conditions by using a simple small square of pcb using 1 cmos gate mounted on the back of your boot

when your in water the power can be varied as desired

many many aditions i know can and will be made

multiple coil does work
as the pwm signal dosnt arrive at each coil at the same time

but instead a step frame ahead of the last...



so this method is the best..... i beleve

using four different shapes
of coil first calibrating known metal types by scanning them at fluxuating freq to guage there metalic element
yes yes i know tin cans look like gold
also total rubbish yes to a round coil gold looks like iron in salt water

try using a triangle in salt water for pi it gives three distinctive nodes
there rx is good

so you can dip the point in the sand etc under rocks to gain valuble inches and feet


squares are used by lots of companies
but i never seen more than one

this is why i made this design

to plug the gap

and i cant think of a better start for a new design but
to wave all patent writes {if any} and pub it so greedy companies dont exploit it with the donkey and carrot senario of give em a little wick for there candle each time a little more


REMEMBER this method is highly experimental


so expect many delays and rethinks

but one thing i am sure of is it WILL work
on paper it is sound in theory too

so .... i leave it too you too decide

[Esteban Cabrera Grinok]

Re: come aboard i need some bodies to build one

Hi,

Very good metal detector project.

I'm curious about the capability respect the differentiation of various kind of metal.

Thanks

[simeon [email protected]]

Re: come aboard i need some bodies to build one

ok

to answer this question as briefly as i can without going in too deep

if you use a single coil pi type detector
we know that different coils respond to different size and shapes type of metals in different ways depending on how they get energised in the soil and re radiate for a few us the signal
becouse of the skin effect of the metal you get a kind of dopler shift in freq as the metal re radiates

these freq are easy too identify in detect conditions by first teaching the detector the patterns for the metal type

using a set of known object in known ground types

therefore by reverse of the theory it will then have a list of what metals {or any element} it knows

the best detection freq are basicaly 0 - 1 mhz but the 1Khz - 9 khz seems to be the band all detectors run on except richard brandson who uses 10 ghz and a baloon


this is a vlf pi basicaly becouse the minerals are so rich in some soils

given the vlf detectors are good for high mineral soil types it makes sense to start here
then increase cpu computation speed to get the higher end ...of the detection bands


the processor when it finds anything "looks up" the plot its getting using presise timing and power /gain level adjust {within tolerances limits }

to interigate the object too see what freq are happening and what arent so good what shifts etc

then simply in the cpu plots the result against a known chart

the scope can function as a display also so can easliy display graphs and charts {5" by 5" uses the same current as a two line beleve it or not {within a few ma}}
and as the code is fairly easy to understand and well documented it can easy have functions added

there is 256 K in a pic 877 i used a comadore 64 years ago had only 64 K

and 256K memory
in for instance the mark stuart design {i built it worked fine i can give help mail me}
it only uses 4 bytes of the 64K in the pic eeprom 4 bytes.....
a pic 877 has 256K eepro and easy to add more ram via 1 wire i2c you can then disconnect and read on a pc as a ram card you make and slot in for finds table and for analysis {useing a simulator to read back and corilate the results like the 3d cad for an atom spliting accelorator to map results and compair gleem for any new info etc }


so a pic is a powerful beast.... of a cpu if used properly


the whole point is it is low freq not so high "yet"

and highly experimental with freq typicaly in the range of 100 -1khz adjustable {but this is the goal for now}
i need help to complete this and want it too stay with the guys who love the art of detecting things.... for real

pcb;s i will break into three sections this board

fet stage
front end
cpu

everything else plugs into cpu board and front end with just interconections between each board
this keeps the clean area for the signal
the pot chips will be one on the fet board and one on the front end
i need feedback about what others think
remember roma wasn't built in a day
nor was the metal pi 4 coil detector
however a clear day will see the code finished for now
i need ideas on the circuit from all the forum guys i respect
it also needs a name
so any ideas on this
??? i dont know.....
the new pi on the block
maybe it will get one soon



another design i saw a while back
was a robot spider it had 8 legs and two claws a remote camera and a detector on each leg

for maping an area {looked very scary at the time}

but the 486 managed to gps map and survey the whole area

even underwater

i never heard of this guy again {perhaps the military got him}

[Al Arens]

Re: come aboard i need some bodies to build one

I did not see the gif file for the schematic. Can you send it to me.
Thanks, Al.

[simeon [email protected]]

adendum

adendum

if you download proteus vsm lite shareware version
you can navigate the design by using F6 F7 F8 keys

on the keyboard when vieing the schematic form of this design

the links are in previous post on the topic

[simeon [email protected]]

Re: come abourd i need some bodies to build one

Here is a multisheet version of the design
i dont want to anoy anyone by posting metafiles
i have just started so it will take a few hrs to get it all arranged


use prolite.exe
direct link
www.labcenter.co.uk/download/prolite.exe 10 mb
you need keys for full functions but it will run happily as shareware witha 5 sec nag screen on boot

use f5 f6 f7 f8 to navigate and click design on the menu bar to change the sheet


here is the link to the mrk 1 multisheet design version
for those wishing to keep close track

as i save it auto uploads to my site

http://homepage.ntlworld.com/vsm.libs/DESIGNS/multicoilmultisheet.dsn

[simeon webber]

Re: come abourd i need some bodies to build one

if you are still unsure after reading other posts

the principle can be explained

you pulse .01 10ms second intervals of pwm "pulse width" on each coils {or very much less } and a recieve window within fully adjustable anyway....
even the slop of all the front ends timing cut off is easy to program {sweep half the speed of a standard detector

so..... many coils is easy as long as the discharge and rx times as a total for each dont overlap {unless you use the pinger .....!!!!!!
the center 100T coil is a simple ping!!!!!!

the ping can also be auto triggered by any find


the time difference between the rx signal on the first ring and the rx to the outside square coil when above a find is
used to center then the machine will analyse it by burst firing each coil at different freq looking for a qualifier amount curve {known not to be the same as any but the metal it is "thinking off" for at this part of the analysis process}
this process takes seconds three or four at the most
so it homes in
scans and identifies by cross interpliration of data sets and known lookup.... by ground type

as soon as you use it on different ground type it switches mode

so if it detects too much mineral hiss level on a spot freq 1khz then it will change the gain and soil type auto

if the spot freq is clear it switches to another detection mode depending on the second spot 6Khz

so is a kind of auto descrimination but using its intelegance {known soil types list}

a 20 mhz cpu is quite capable of churning away on the lookup while the s&h is in detect {runningmode} the cpu dumps data accross at each stage using shift out and in commands on an interupt
so you need only teach it the tables for your area and conditions

underwater i am sure this is a valuble tool

and a coil can be placed on top of historic finds metal or ciment plastic or anything even soil can be analysed using pwm systems

i will attempt to add a slew circuit {not hard i have one in mind} to the pwm signal to make the fets slew a bit the waveform
this as someone has said is a good thing also ... so this is only one more pot ic single pot so not much to add
one more line of code



with the rectifiers and s&h etc not getting buzzed up by other freq
mearly switched off and the rectifier will dump any induced voltage
the coils i made different shapes to change there respose curves and dips over the band
and as i pointed out some shapes are better for side nodes from the windings bend points {for square and triangles}

the coil dimentions are ones i personaly have used on the same detector single and all at once via a resistor network
so i know them well

[simeon webber]

Re: come abourd i need some bodies to build one

the link to the newset version
i named the detector "THE 4METAL ENGINE"

HERE IS THE NEW LINK TO MONITOR
pcb's etc i will optimise with opinions any part you feel would need it





Click Here

[Nicko]

Re: come abourd i need some bodies to build one

Simeon,

I tried to look at your circuits, but to be honest its a major pain - Proteus is just horrible and difficult to make prints and to read - its completely non-windows-standard in its output (wouldn't a "print page to fit" be nice?).

If you could post a SIMPLE (JPG/GIF etc.) schematic, I'm sure you'd get much more feedback.

Nicko

[SIMEON WEBBER]

Re: come abourd i need some bodies to build one

nicko

it is 8 sheets+ design most are a0 size

so... learn to use proteus
its only as friendly as you set up too use
choose a comfortable res like 1024 /768 or if you have a 17" monitor use 1280/1024
use f5 f6 f7 f8 keys f6 and f7 to zoom in and out to there the mouse is pointing
it cant be simpler.....

to print use resize factor change the paper orientation to the correct size
but realy a0 is a massive size... so... you should learn too use proteus
after the frustrating bit
it is a real dream.....
i do sympothise... but even 1 metafile is nearly 1mb or more...

so... gif loose quality jpeg i couls make
but i prefer to wait till the design completes then make a pdf ebook

showing all theorys ........


proteus is a nice cad
and you get what you put in in the way of effort to learn out of it...
as you can see by looking at the designs i do it is a quite nice cad with many facets i dont use the pro version so.. i dont need to add simulation charts etc.... or run very very complex sims
simply 1k or later full no code limits{v6 is coming so ill spend then}

runs most things
cad being more important too me for now
as i am still learning more each day about it myself....

so... give it another go and dont give in to the ide....