Building from Scratch Test Results

[jlsilicon]

I am posting from Moodz original Design and Schematic.

https://www.geotech1.com/forums/atta...0&d=1293274805

Setup for first Test :

1. Twisted Coil (about 1.5ft diam, 15 Twisted / 30 SingleStrand).
2. IRF740 MosFet & 10000uF Cap.
3. Diode Limiters & 1K Resistors.
4. TC4471 Driver.
5. THAT1510 (without the 20uF cap) - not yet used - just mounted the Datascope across the Diff Inputs for this first test.
6. 78L05 & 79L05 Pos & Neg VRegs.

Notes:
* I think I blew the IRF740 a few times,
- first I misunderstood that the TC4471 is Inverted (Low=EnableMosfet / High=Disable) ,
- then I understood the Timing needed (I used 10ns High on 100ns Period - Not good).
See: moodz: - "I have tested from 100 us to 60us pulses at 4 - 6 Khz and works OK."
So, I changed Timings to: 50ns HIGH on 5000ns Period , 10ns HIGH on 500ns Period , 2ns HIGH on 100ns Period ;
- I setup ann Arduino to Outp\ut these Pulses to the TC4471.
Best results with 50ns/5000ns so far.

[jlsilicon]

Posting the First Tested WAVE Photos !

Some Results are listed below.
- So far I can not recognize a difference for coins -or even a handful of (6) Quarters !
Could be just that the Datascope is limited to 1us and 10mV - I can see the curve but it is Limited to the lower 10% of the screen.

Wave photos are for : NoTarget , Hammer , Keys&Keychain , TinCan - each placed in Center of the Coil & Hovering
(via extended on cardboard - because the hand static made the wave jump 10% of screen).
- sorry about the lousy quality - you can still see the differences in waves though !

[jlsilicon]

And, the 2nd Test WAVE Photos with placing object in center of the coil :
- Notice the Bottom Shape and Size of the Waves , and that they tend to match the corelating Waves above ...

[jlsilicon]

maybe this can help.
It is a grouping of all of the images (shrunken).

[jlsilicon]

Both Schematics from Moodz Front end Metal Detector :

https://www.geotech1.com/forums/atta...0&d=1293274805


https://www.geotech1.com/forums/atta...7&d=1293394319


- from :
https://www.geotech1.com/forums/show...t-model-T-quot

[jlsilicon]

- sorry, that was 220pf cap removed from across THAT1510 p3..p4 - not 20uf :

>

5. THAT1510 (without the 20uF cap) - not yet used - just mounted the Datascope across the Diff Inputs for this first test.

=> 5. THAT1510 (without the 220pF cap) - not yet used - just mounted the Datascope across the Diff Inputs for this first test.

[jlsilicon]

New update photos from today !
Taken at the Diff-OpAmp Inputs.

Seem to have aproblem with the 78L05 & 79L05 Regs giving Output for the DiffOpAmp Vs+- though.

[jlsilicon]

One set of yesterday tests - measured at 1us intervals / 10us total, at 1mV interval / 10mV total.

[jlsilicon]

Second set of yesterday tests - measured at 10us intervals / 100us total, at 1mV interval / 10mV total.

[jlsilicon]

And, comparison of yersterday tests :

Comparing _My_MD_Inv_All-NoTarget-Keys-AluminumTin-TinIronCan-Hammer-Keys_10us_20220603-17000-18000-d.jpg :
- Iron has a Broader Curve , but quick Decay - shorter compared to Curve size.
......... __
...... /
... /
. |
- Aluminum has a Smaller Curve compared to Delay size , but Long Delay - longer compared to Curve size.
... ___-----^^^^
.. /
. |
(excuse my graphics ex )



-

Comparing : _My_MD_Inv_All-Aluminum_1us_20220603-17000-18000-d.jpg :
- notice Targets (Aluminum for ex) just shifts Up and Down for Distance and size.
- worth noting that Larger Targets shift noticeably Up and Down More - and so be calculated as Larger else smaller Targets.

[jlsilicon]

I have tried to hookup the THAT1510 and the INA217, but get zero results.
I think original problem is that the 78L05 & 79L05 dont seem to be functioning.

I skipped around this and just wired the V- to Gnd, with some results.
The best I could get was a vague shape like the Coil output to OpAmp Input leads scoped out,
but no Gain - same voltage out on Diff-OpAmp as VIn+ and VIn- on the scope.

[jlsilicon]

Ok, today's results:
I added a ne5534 in place of the Diff-OpAmp (which has not been successful).

Below are the resulting photos :
-- I think the BIAS is off - giving the Inverted Result ; Wave deflates with a Target instead of Peaking.

NoTarget : - The Peak decaying slowly to the Right :



Hammer : - The Sharp drop that Curves up and to the Right :


AluminumTin 3inch : - The small Double Peak :


Keys : - The Little Peak :

[moodz]

While since I looked at the fully differential circuit ... so Bravo.

First I will suggest ( IMHO ) you move away from the 5534 as it has this input which does not behave well in a PI ...



and move to a more robust input amplifier .... eg LM6171



the input on the 5532/34 start conducting current during flyback surges ...


moodz.

[jlsilicon]

Hey Moodz !

Lm6171 - 100MHz cool ! - should be more useful.

My goal here is to work on the MicroController side of the project.
- I am a professional programmer, and I found a few ways of speeding up the Microcontroller input.

But first, I need to get the Analog side working ... I am a Digital electrical guy ...
- analog is always slow for me, the tricks needed never seems to make sense to me.

- I ordered a batch of the Lm6172 (dual of the Lm6171) - should be here in a week.


Meanwhile, trying to get the project to propery work for myself :

-- so far, all I can recognize are large pieces of Metal, Coins are noticeable yet.

* trying ICL7660 with the INA217 / THAT1510 - instead of using the 78L05 / 79L05 which does not seem to function for me.

* trying alternate Coil Sizes - maybe i can improve Sensitivity with more Windings / Inductance ...
- - Currently using a 15inch Coil with 15 Twisted pair Windings (about 75feet Twisted / 150feet single wire).

- just using the Diode Limiter right now - pushing aside the Mosfet Limiters for later.

jlsilicon

[jlsilicon]

On the Microcontroller Side :

I am currently working on Arduino & Esp32 projects :

* Currently, arduino is just providing the 10us pulse at 5KHz.

* Using TLC5510 integrated to both Arduino & Esp32.

* Working on Arduino - to input between 3Msps - 3Msps.

* Working on Esp32 - to input between 10Msps - 20Msps.

jlsilicon