TEFLON MULTI STRAND WIRE

[kt315]

..

I forgot something: Do you know the very informative article about PI coils of bbsailor?

http://www.geotech1.com/pages/metdet...s/FastCoil.pdf

[daverave]

yes ive read the artical several times...one of my favorites....i didnt realize it was written by bbsailor.

[daverave]

Dave,

The function of a Lock-in Amplifier works by having the sample time tied to the TX pulse as it always samples the pulse in the same place and feeds the integrator circuit along with the longer delay to eliminate the Earth's magnetic field. See Carl's explaination in the initial Hammerhead article.

Lock-in amplifiers separate a signal burried in noise because noise is random but by sampling the response of a target in the same window many times, a better signal to noise ratio can be had. However, coils move over the target and have limited time over a target so the TX frequency needs to be high in the 3000PPS range to be effective. A microprocessor controlled PI may already have an integrator circuit which functions like a lock-in amplifier. I mentioned the Lock-in amplifier to show how these signal recovery methods have been incorporated into PI designs to allow better detection of weak signals.

Joseph Rogowski

I hope this helps

when you say that the sample is tied to the tx pulse do you mean that the tx pulse is used as a reference and compared with the signal frequency so noise is eliminated....i know about the earth's magnetic field cancellation....which i believe is common mode rejection if im right....i love your artical on fast coils !!!

[Mechanic]

Hi Dave,

Tied to the tx pulse basically means that every time the tx pulse finishes there is a set amount of time before the sample is taken, which is the same every time.
Example:
start
tx pulse on 100us
first sample delay 10us
sample 1 on 10us
second sample delay 20us
sample 2 on 10us
end; back to start

It just goes on and on and on, always sampling at the same time after the tx pulse. The target response will always be the same,t he noise will always be different, hence the noise gets averaged out, hopefully leaving the target response behind.

The earth field cancellation is basically subtracting out an offset voltage. The earthfield response is a low frequency response that happens at sweep speed 0.1hz to 10hz. by having the sample pulse widths even, there is practically no difference from the EF between the samples, therefore it is subtracted out. The samples are looking for fast decaying signals in the range of 100khz to 1khz depending on the sample widths and delays.

Cheers Mick

[WM6]

Hi Mick
what is here TX pulse duration (time-point of TX Pulse start and TX pulse end)?
Mean this that TX pulse start right after sample 2 end or with some delay?
Maybe in drawing to imagine better.


start
tx pulse on 100us
first sample delay 10us
sample 1 on 10us
second sample delay 20us
sample 2 on 10us
end; back to start

[Mechanic]

Hi Wm6,

Well that is just a theoretical pulse train, in this instance the tx starts after the sample 2 has been on for 10us and is turned off. You might want to have a delay after sample 2, you might want to add another sample after sample 2, it was more just to show the sequence of things.

The tx in this example turns on at start, the duration is 100us, then turns off, the a delay of 10us etc.

Cheers Mick

[daverave]

hi guys....are you saying that the first sample pulse occurs 10 us after the tx pulse finishes...and is 10 us wide...then the second sample pulse starts 20 us after the finish of the first sample pulse and is 10 us wide....also how does the width of the sample affect detection ????

[Mechanic]

Hi Dave,

Yes that is how it goes in my example, and then after it finishes, it all starts again and does everything at exactly the same time each cycle. That is how the noise is reduced.
Yes sample width does effect the detection. I found that 8 to 12us was good for really tiny targets, with a sample delay as small as possible. To figure out the minimum sample delay I looked at the output of the front end and looked for when the coil had settled after the tx and flyback and added a little bit more time(1 or 2us) then took the first sample. I then fiddled with the sample width to get the best response on my desired target size.

Cheers Mick

[Jose]

Mescanico.
What you say is like Fig. The distance between the first and second sample is only 20 us?.
If so, this affects the sensitivity of the detector?
Thank you. Jose

[kt315]

also how does the width of the sample affect detection ????

in Surf PI width of sampleS is fixed and that is right. Surf PI is submersible detector, for diving, foremost. if you add a tune pot you waste a money, that is a rule of any company - maximum earnings minimum commitment. ANY waterresistance thing costs in 3-5 times ABOVE any landing using thing. so whites did specific schematic for specific application. you begin to copy surf pi FOR WHAT? surf pi has TWO POTS, it is enough. Dave, leave this project, begin something of Carl Moreland. there is tuning those things like width, delay etc.... all you can probe and expirience.


PEOPLE DO NOT BEGIN MAKE SURF PI IF YOU DID NOT LEARN THE BASE OF PI THEORY. READ THE CARL'S ARTICLE BEFORE.

[kt315]

surf pi, like it looking.

[daverave]

thanks kt315 for that and the photos....i have already made the carl morland hammerhead detector but not tested the pcb as been so busy on the surf pi project...tell me on the hammerhead can i switch the audio vco/non vco using a toggle switch ????

[Mechanic]

Mescanico.
What you say is like Fig. The distance between the first and second sample is only 20 us?.
If so, this affects the sensitivity of the detector?
Thank you. Jose
[ATTACH]27012[/ATTACH]

HI Jose,

Yes if the second sample is close to the first sample, the sensitivity/ depth on larger targets will be reduced. However on small targets a short delay would not matter.
Lets try a little example, lets assume that the ground is giving no response. With no target, sample 1 is 0v sample 2 is 0v. S1-S2=0v
Now we put a small target under the coil, this causes sample 1 to go up to say .5v and sample 2 is still 0v because the small target eddy currents have decayed away by this time.
So s1-s2= .5-0=.5v. So now we have a response of .5v.
If we were to delay s2 further there would be no difference on this small target, however, lets look at a larger target that takes a lot longer for its eddy currents to decay away.
S1=1v
S2=.8v
S1=S2=.2v
So the larger target now has given a lesser response than the small target.
So for larger targets it would be best to have a longer delay.

Cheers Mick

[Jose]

Mechanical
Thank you for your extensive and clear answer.
regards
Jose

[daverave]

on the hammerhead revision D how do i select the active mosfet drive as there are three holes on the pcb....which is the active option ???? i also want to be able to switch the audio modes so i can choose vco/non vco....there are six holes on the pcb for this....can i use a toggle switch to select which audio i want ????