STUART, DELTAPULSE and other

[Philippe]

Hello every body,
A wink from FRANCE.
MANY thanks for all the informations that i have found on your forum, very very interresting.
I want to build a PI metal detector: as you can know there is a lot of schema on the web.I want to have some information about someone who has build a PIC stuart or a delta pulse or a white PI 1000.
The stuart has a very simple analogic schema as the other are more complex: filter and auto tune.
I want to experiment a design with a 16F876 :
one analogic detector
one numerical program with two mode :
one mode to drive all the different pulse for the analogic detector and an other mode to make the analyze of the pulse decay for trying to discriminate.
Can you help me by sending information about your realization : depth of detection, simplicity to make it, to tune it...
bye bye
regards
philippe

[simeon webber]

Re: STUART, DELTAPULSE and other

i am working on a similar system
if you look at 4metal engine design
using proteus vsm {there is french version avalible also}

then i got this far so far

i fully agree it is so easy

i draw some up have a look

i am interested to get some people also to help and design as a team

i use as well as standard pi technique using a single coil

multiple coils doing time spaced samples to detect centering of the object and depth by a dopler shift calculations at each sample and resample point
you can be analyse {burst sampling multifreq pwm} using the individual coils etc
guage also the exact find metalergy {well lets see ....... if this part is achievable but i know it is by the laws of physics of emf}

and enough cpu power or cleaver tasking and portloaded pwm on int

of course i use two pic 877 to get the speeds and one as a radar type scope {normaly osciloscope} but i use in x and y mode to display a radar trace of the object below


take a look at 4 metal even for some ideas for you

thankyou
email me if your interested too work together...

[simeon webber]

Re: STUART, DELTAPULSE and other

hello again all

a little experience of mine
and
further to my research into pi detectors avalible as kits and comercial ones

in monitoring the current they consume
at various points of the circuit
i realised something

pi detectors that pulse the wave with a shark tooth shape tend to detect lower metal resistivity types not so well but detect high > 2R types much more easily than others will
with gold in around this point of the dip

i further noted the current each stage consumes as a peek on several types of detector both pi and ib
as the voltage is dumped pinged to produce the desired flux fields from the coil


the fact is
all the real designs from manufacturers where exactly the same
when the voltage spikes the current node is in the coil not the battery where it should be
with the pump or oscillators driving very hard to reposition it like a watch pivot also a little voltage dip was noticed at the fet drain


this is wrong and will lead to the shark tooth being destorted to a sine wave anyway
a monitor loop made from a few turns on a scope probe near the coil gives the real radiated waveform

and it isnt pretty
it more of a interupted sine than a shark or a square wave as you would expect

but again
when you add correction for the current node to be in the battery {by tuning the whole detectors resistaces carefully {i use a meter at the battery {till the dip is pinging in a half weekened cell ...biggest you can tune to } by altering the size of dia and the turns {or the reactance of the coilwith its respect to its impedance}
}

and the damper resistors in parallel with the coil these are there to increase the bandwidth and give a flat responce

this fixes most detectors i tried
giving much more drive

if you do this with three coils and a pinger {i have this running now the dip is quite large for the pinger but the trade off is good
reporting a 2 p copper coin at 1.5 meters depth in sand {as far as i could be bothered to digdown}

this using a very high saturation of anti current rich emf

so by moving the coils current node to the battery in the middle of the bandwidth of the dectctor you get a big BIG increase in depth

to finish i can safly assume that
1} whites pi 1000 2000 3000 and fisher 1260 and later pi
all never take into account the current dip and node but only the voltage dip and nodes

two entiry different beasts
afterall it is the charge of current on the skin of a metal object that gives the biggest re radiated emf or flux field variation

so..... the better the current antinode and higher the voltage nodes at the object the better it re radiates


so let the 4 coils and a current rich pinger work for you too

[Philippe]

Re: STUART, DELTAPULSE and other

Hello
thanks for your reply but i don't understand some informations:
Can you give us some diagram for more explanation.
You speak about the curent node to be in the battery ??? what's mean?

[simeon webber]

Re: STUART, DELTAPULSE and other

even though a pi detector is just a type that uses pulse induction

it still radiates rf from the coil in the form of emf rather than a modulated or plain carrier

so realy it is pulse rf causeing induction

like all rf device the coil is a tuned circuit
it just happens that the coil you use is tuned to the band you want to cover with it
also giving a flat response and good bandwidth


so like any coaxia rf carrier even though it is emf will still carry a set of voltage and current nodes {with respect of pulse width and coil coax or wires etc} along the voltage rails and coil connections as far as the battery}
that you need to carefully tune inorder to avoid placing a node at the fet junction and in the coil for any freq you are on {you need to move the gnd of the fet and coil with respect to actual gnd of the battery and detector pcb while maintaing any power rails}

it is just that all detectors i tryed have this node in the coil or the fet and not the battery where this node belongs
@ 10khz say .....
each node will be a sin factor of the voltage lead lag timing as the fet switches to polerise the coil
along the lengh of the down cable and also by factors of induction calculation

to give a tuned circuit at 10K

so use the rf equassions like
1/2pi *2rt LC
or modified equasions by many others few docs on geotech for that
the fet the down wire and the coupling factor of the turns on the coil + any diodes or etc

all worked in makes the resonant freq of this tuned circuit

so you get maximum power transfer {of voltage.... a find takes voltage and convert to current this is handy becouse some metals convert to current esp oar's

so by positioning carefully the node set you can more accuratly derive the metal type and positions
compairing against known sets at per frequency


so these rf positional current and... {current will lag voltage by 90 deg relative}
to the pulse direction " nodes" and "antinodes"




so for square waves makes a triangle and tangental factors on the cables to the coil {very important becouse they behave differently allowing for tuning {its not sometimes a matter of a bigger coil but a smaller wire resonant lengh is the last thing you need this can be take 3 " from the cable 20% bigger node in the battery 40% more depth to most detector homebrew or even most comercial } and the coil its self also square waves at 10K + the rectified current from any metal object as a cos of the lead lag of the current and voltage as it reflexes from the find}

much like an ora the detector carries a field of resonace { or it should }
see kerellian photography for a ht pwm scanning device used to diagnose health
used by professional doctors in ukrane and russia for years now....

have a specific use in metal type
also
the amount of current drawn dictates the power returned in a different form or as the same form at a different phase angle

[simeon webber]

Re: STUART, DELTAPULSE and other

i must point out
not to go and cut your cables up by 3 "
but to think using a multimeter and a half dead cell

set the multimeter to peek hold to get peek current
and
with a spare cable you make a little longer choose your desired center band the coil is tuned for and tune the cable for max current dip for that coil

so

please dont cut your original if you see the last message and think what a good idea but one to note......

[Philippe]

Re: hello every body :STUART, DELTAPULSE and other

hello
can i have some other remarqs from other personne about my first question?

thanks
philippe from FRANCE

[Philippe]

Re: HELLOOOOOO,STUART, DELTAPULSE and other

Hello every body!!
Does i think that there is only Simeon and Carl who work on PI?
I need your help to understand realy how work a PI: In the large idea i understand but now What's happen exactly?
Here is what i understand : IF you are not agree answer me...
We drive a coil with current which generate a magnetic field around: first question how much is field around the coil not only on center or mediane( spacial distribution) ?.
After we cut the current in the coil : like Mr FOUCAULT said , there is some current generate inside the piece of metal which is locate on the decay field, and this eddy current generate field around the piece of metal field which decay as the magnetic parameter and resitive parameter of the piece of metal. As the coil receive this decay field also it generate a voltage: the problem is to isolate the voltage which is generat by the cut of current and the voltage generat by decay field of the piece of metal we want to detect.
The two voltage are not of the same level:
cut of current :100 to 500 Volts
eddy current : some µvolts !!!!! ( i think because i don't make any measurement..)

The two voltages are very dependant:
the first( cut of current) UL = - L Di/Dt
and the second ( eddy current from the piece) UP = -D(phi) / Dt as phi= mu0*I/(2*pi*R) in the center of the coil.
In the first case Di is not good because it increase the voltage of the current cut UL and in the second case Di increase the voltage generate by the target because much higher is Di much higher is D( phi) much higher is UP.
Actualy we use the effect that the eddy current don't stop immediately and we try to decay more rapidely the voltage due to the current stop.

I don't speak about the capacitance of the coil, the rf noise read by the coil , the noise of the amplifier ... it's another history ! aswell as the noise is praticly egal to the voltage generat by the eddy current.
I just try to understand what we want to mesure????????


QUESTIONS:
Does somebody had make some relevement( with values) of the decay curve from different shape and metal?
If yes would it be possible for us ( and me ...) to put it on the forum?
Does somebody had make the calculation of the spacial distribution of the field around a coil?
Does somebody had make a pi TYPE STAURT or DELTA PULSE OR WHITE 1000 and if yes what is your depht of detection that you obtain?

It is a very long email but I want to undertsnad exactely what we measure whit a PI detector.
SORRY for my very POOR english ...
bets regards from FRANCE
Philippe

[Robert Hoolko]

Re: HELLOOOOOO,STUART, DELTAPULSE and other




Philippe

You seem to have a good understanding of how a PI works already.

In general the target signal decay can be described as an infinite sum of exponential decays. Fortunately many targets we are interested in can be described by less than an infinite number of terms. Small low conductivity targets can be described by a single exponential decay, and many other target can be described well enough by the sum of two exponential decays. These do not give the exact results, but metal detectors do not usually make very precise measurements anyway. Many rings and coins have decay time constants in the range of 5 micro seconds to 100 microseconds.

I do not have a collection of decay curves because until last month I did not have any PI device to play with. I plan to collect some decay data, but I will be doing that over the next year. I do not expect to have any data soon.

I use the calculation above to find the strength of the magnetic field around the coil. This only gives the vertical component of the field. You would have to do a similar calculation to get the horizontal component. I am usually interested in the field far from the coil and near the axis of the coil so the vertical part is all I need.

R is the radius of the coil
d is the distance from the plane of the coil
x is the distance from the axis of the coil

The integration is done around the circumference of the coil, but because of symmetry I only have to integrate half way.

But this only gives the strength of the coil's magnetic field. If you want to know about the received signal you would have to calculate how much current this induces in the target. Then you would have to calculate the field around the target due to the target current. Then you would have to calculate how much of that field cuts through the receive coil. These calculations are not easy to do. So I think you will only find results for points along the coil axis. And even then a lot of approximations will be used.

I have never built any of the detectors you are interested in so I cannot tell you anything about them.

Robert

[Philippe]

Re: HELLOOOOOO,STUART, DELTAPULSE and other

Robert
Many thanks for your answer.
What i understood too ( as you confirm by your mail) is that it will be very difficult to assume a calcul of the signal receive by the coil from the target.
You speak about 5 µs for the time decay of little object: that's mean that if we want to detect it we need to have the time to slow down the voltage due to the current cut for about less 5µs.
What do you think about the idea of Simeon to put some electronic switch to modifie the decay circuit?
For me i think it's a good idea.
Have you some other equation to calculate the parasit capacitor of a search coil ?

Philippe

[Robert Hoolko]

Re: HELLOOOOOO,STUART, DELTAPULSE and other

Philippe

One useful formula is v = (k*r/(d^2 + r^2))^3. This gives the signal from a small target on the axis of the coil, r is the radius of the coil, d is the distance from the coil to the target, and k is hard to calculate. But if you can measure the signal with the target at one distance then you can calculate what the signal will be at other distances. It is easiest if you measure the signal when d is 0 (the target is at the center of the coil). Then you can calculate what the signal will be when the target is 20 or 30 cm away. Make a graph of this formula to see how fast the signal decreases with distance.

Simeon has made many suggestions and I do not know which one you are talking about. Very short decay targets are probably best detected by a low inductance coil with low current and a high pulse rate. The low current and inductance will let the magnetic field collapse quickly, and the high pulse rate will let you add the signals from many pulses together. The high pulse rate partially makes up for the weak magnetic field. For slow decay targets a lower pulse rate and higher current and inductance would be better.

I have not tried to calculate the coil capacitance. I have not been looking at that part of the design yet.

Robert