felezjoo PI(the best pulse induction metal detector that I made until now)

[jladre]

"[QUOTE=Qiaozhi;224444]I have tested the Felezjoo with a 7" diameter mono coil, and the soda can was detected at around 16" max, but the threshold was very jittery. Basically, I'm still not impressed by this detector."


Seems like there something wrong in your setting or in your coil or even to your electronic parts.. even though it is not damp properly at 45" for a soda can it will give loud and clear tone..

[Geo]

I have tested the Felezjoo with a 7" diameter mono coil, and the soda can was detected at around 16" max, but the threshold was very jittery. Basically, I'm still not impressed by this detector.

Next I was going to do the same test with the same coil using an MPP. However, I'd forgotten that all my MPP boards have been modified during a series of experiments, so it wouldn't be a fair test to use one of those. Therefore I connected up a Geotech Baracuda REV-A, and the soda can was positively detected at 21". This was with the same 7" coil. I would expect the MPP to give a similar result if it was using the same TX pulse width and pulse rate.

Thank you Qiaozhi.

[Geo]

I have 65...70cm for soda can with a 40cm coil.
Sensitivity is at 16 because inside the lab there are many pests.
GB=5 Speed = 10
Coil.... not so good. A250 B645 C760 and auto delay =24us
Also i did not play with damping...

[ivconic]

I have 65...70cm for soda can with a 40cm coil.
Sensitivity is at 16 because inside the lab there are many pests.
GB=5 Speed = 10
Coil.... not so good. A250 B645 C760 and auto delay =24us
Also i did not play with damping...

Earlier i made 39cm coil and i got ~120cm on cola can.
But best distance in the air with that coil was 180cm on 30cm diameter thin silver coated copper plate.
In soil situation is somewhat different, same plate was buried at 80cm depth and covered with soil.
FelezJoo PI detected it with clear response and with coil 50-60cm above the soil surface.
Usually at such kind of testings i am not that neat to accurately measure the "depths" and "distances", that's why i am not exactly saying it is 100% accurate.
But it was obvious to me that FelezJoo PI is pretty deep PI detector, giving very satisfying results.
And GEB feature is just splendid! Very effective and very accurate. I like it. Something which i never had so far on my PI diys.
Coil for FelezJoo is delicate to make. We talked about that before.
I probably made dozen so far and just very few of those were good.

[ivconic]

There should be no wonder why FelezJoo PI is made that good and why it is performing that good too.
I don't recall if i mentioned this earlier but FelezJoo PI code is written mostly relying on whats elaborated in patent US5506506.
"METAL DETECTOR FOR DETECTING AND DISCRIMINATING BETWEEN FERROUS AND NON-FERROUS TARGETS IN GROUND" by Candy.
Read it and you'll understand the principles of how code works.

[ivconic]

In essence this is what's happening:

"...The detector samples at least three voltagemeasurements of the resultant voltage over three different
time periods to provide three sampled values which are then
processed by a microprocessor to provide three average
ground balanced signals from which at least two ratios are
calculated and compared against a pre-defined response..."

[Geo]

Earlier i made 39cm coil and i got ~120cm on cola can.
But best distance in the air with that coil was 180cm on 30cm diameter thin silver coated copper plate.

Ohhh fantastic depth.
My TDI has not this depth even at No GB mode....
It is time to start a new "good" coil.
Did you changed the damper??


Btw... what is happened??? did you visit Creta???

[ivconic]

Ohhh fantastic depth.
My TDI has not this depth even at No GB mode....
It is time to start a new "good" coil.
Did you changed the damper??


Btw... what is happened??? did you visit Creta???

No. I kept original values as on schematic.
This is been repeated several times on these pages; but all the "secret" is in to make most proper coil here.
Since code is written precisely to kind of LF357 authors had at the time; it will work better as coil and LF357 specs are closer to what author had while writing the code.
It is problem, i know. Pretty narrow and strict demands we are having here.
But those three debugging numbers are there to tell you how close you are with your coil.
Geo you have mentioned several times that you are having limited space at your workshop. I guess there also various kinds of noise sources too.
Maybe that's what making you additional problems?
Like i said; i made several attempts to make proper coil, all fails until i managed somehow.
Easiest for me was later when i made coil with pvc coated litz wire, 0.8-1mm (all litz together, of course).
And inductance was 0.35mH.
But i suggest you first to make several windings more, than attach coil to detector and check those debugging numbers and than subtract turn by turn each time again checking those numbers.
You'll notice how those numbers are changing. But be sure that your coil is away from any noise source in workshop.
...
Unfortunately i haven't visited Creta yet, since 1994.
I dream about that...

[Geo]

Also i did n't visit Adriatic at summer, maybe next summer so i moved our meeting for 8...9 months..
Now about coils... i made amother one 9" that gives very good numbers but not some special depth .
It has inductance = 320uH and 0.8R. A=350, B=745 C=840 . With these numbers it is better at disc but as i said not very good depth.
Anyway... winter is long.........

[ivconic]

Authors own words:

"...A, B, C are equivalent to short, middle and long time constants in minelab patent.
A must be between 200 to 400.
B should be around 700
and C about 800..."

Geo you should go out and test it with Sense at 15,16,17,18...
You can't pull out maximum from it indoor due lot of noise and interferences.
Same situation i have here indoor.
But than again; completely the same problem with XP Deus too.
In room Deus is pretty erratic and crazy behavior. Total mess.
But if i lower the Sense; Deus also lose much on air "depths".
But outdoor both are performing very good!

[Geo]

OK..
I"ll try a better coil and when the weather will be better will go out for tests and will inform here.
I don't know what is happening at Serbia but here one day rains... the next stop and again from the start....

[Qiaozhi]

I decided to fiddle around with the settings a bit more to try and increase the depth.

The coil is 7" diameter with an inductance of 300uH. The detector seems to think this is ok, as there's no error message.
The damping resistor is 470R.

The ABC values are: A=175, B=493 and C=582.
According to the system settings, these values are a bit lower than expected. But the importance of these seems to have more to do with ferrous rejection than anything else. The opamp is an original, and comes from a reputable supplier.

Sensitivity is set to 20 (max). Delay offset is zero, and the menu indicates the actual sample delay to be 21us (which is fine for this test).
TX pulse width = 500us (max) and TX pulse rate = 150pps. The frequency (150pps) appears to be optimal for these settings. Input voltage = 18.5V.

The soda can (Coca Cola 330ml) was detected at 21" (53cm). This was the point where the audio tone was constant and the cola can was perfectly stationary. Beyond 21" the audio tone was intermittent and not continuous.

This result means that a cola can is detected at 3x the coil diameter (i.e. 21" with a 7" coil).
Ivconic achieved ~120cm with a 39cm coil, which is in fact also 3x the coil diameter.

[green]

I decided to fiddle around with the settings a bit more to try and increase the depth.

The coil is 7" diameter with an inductance of 300uH. The detector seems to think this is ok, as there's no error message.
The damping resistor is 470R.

The ABC values are: A=175, B=493 and C=582.
According to the system settings, these values are a bit lower than expected. But the importance of these seems to have more to do with ferrous rejection than anything else. The opamp is an original, and comes from a reputable supplier.

Sensitivity is set to 20 (max). Delay offset is zero, and the menu indicates the actual sample delay to be 21us (which is fine for this test).
TX pulse width = 500us (max) and TX pulse rate = 150pps. The frequency (150pps) appears to be optimal for these settings. Input voltage = 18.5V.

The soda can (Coca Cola 330ml) was detected at 21" (53cm). This was the point where the audio tone was constant and the cola can was perfectly stationary. Beyond 21" the audio tone was intermittent and not continuous.

This result means that a cola can is detected at 3x the coil diameter (i.e. 21" with a 7" coil).
Ivconic achieved ~120cm with a 39cm coil, which is in fact also 3x the coil diameter.

Because i highly suspect that your setup is optimal. Simply because i achieve much different result with my 8" coil. Cola can at ~45" with clear audio response.
Most probably you haven't made proper coil, that's the common reason for poor performances. (reply 1152)

45 inches with a 20cm coil, almost 6 times the diameter

[6666]

The ABC values are: A=175, B=493 and C=582.

For when I get my felezjoo-pi working what numbers are you trying to get to
or closer to ?

[ivconic]

Because i highly suspect that your setup is optimal. Simply because i achieve much different result with my 8" coil. Cola can at ~45" with clear audio response.
Most probably you haven't made proper coil, that's the common reason for poor performances. (reply 1152)

45 inches with a 20cm coil, almost 6 times the diameter

Smaller the coil = more immune against external noise and interferences.
7" coil size is very good choice for testing and experimenting indoor.
Unlike Geo's case and troubles with larger coil.
Smaller the coil = more dense and more "focused" is the em field too (if same current is applied).
It's the "rule of thumb" wins over the expected "linearity".
Funny but true.
"6x ratio" falls down to "3x ratio" and less as coil diameter rises.
Another proof is the case with 39cm coil and 30cm silver coated copper plate at 180cm in air.
Also this is all in air. Quite different situation is when target is buried in soil.
With same current at different coils you can't expect to maintain the same ratio.
And i think we can not establish any kind of "ratio" rule here, because of the too many "variables" involved.
Type of metal at target, coil size, current through coil, coil diameter, pps, pulse width, environmental noise and interferences etc...
So those "3x" or "6x" ratio is just plain coincidence.
One thing i am quite sure though; Qiaozhi can achieve better results if he make better coil, closer to desired debug numbers, avoid single wire and use litz and with inductance between 0.35 to 0.5mH.