Help with a metal detector circuit

[showlyshah]

hello all

i want a help regarding a project. https://blog.podkalicki.com/attiny13-pi-metal-detector/

i build this one using attiny 85 . and it works good. but the problem is since it uses only 5 volts to power the coil .. its less sensitive to smaller objects. like a gold ring. for a gold ring the distance is about 4-5 cm at maximum sensitivity. but for gold bangle ..its about 10 cm.. for larger objects like a metal spanner.. it works about 15 cm.

i want the circuit to be more sensitive.. like 30 cm for a gold ring. ( actually my friend lost his pretty expensive gold wedding ring on some shrubbery area and i am trying to help him find ).

so my idea is to run the coil on 12-18 v ( 2*9v pp3 battery ) to get more power. but as you can see.. the sense wire is hooked directly to i/o port of attiny.  and when i measured the voltage through the sense wire .. ita almost as vcc voltage.. so if i run the coil on 12-18 volts.. it will feed back that higher voltage to the attiny and will fry the chip.

can anyone please help me to modify the same circuit to power the coil section at Higher voltage ? and still retain the original working of the circuit?

i know there are so many pi based circuits.. but this one is amazingly simple.. thats why i choose it.

please help me as quickly as possible.

thank you all

[StillTrying]

"so if i run the coil on 12-18 volts.. it will feed back that higher voltage to the attiny and will fry the chip."

All you've got to do to stop the 12V getting to the attiny sense input is to move the top of D2 to the attiny's VCC, in theory, and maybe increase the 10k value of R6.

[pwlps]

The "pulse induction" technique used here is based on the change of the decay time when the quality factor is affected by nearby metallic objects.  But the quality factor does not depend on the voltage so I don't understand well why you expect to get more sensitivity with higher voltage.  If this is merely to reduce the noise in the measured voltage I would first try averaging many measurements. Actually I don't see any averaging in the code ("measure_decay" function) so there is maybe quite a big margin left to improve things on the software side before changing anything in the hardware.

Edit
The sensitivity will be a rapidly decaying function of the ratio diameter/distance (asymptotically something like a 3rd power) therefore to increase the sensivity to larger distances you should rather use a larger coil. The original 5cm-diameter coil will have a vanishingly decreasing sensivity at 30cm or so, on the other hand a 30cm coil will have a smaller overall sensivity but it will decrease much less around 30cm.

Edit 2

Apparently the calibration code does not take into account the signal/noise ratio.  An optimal threshold calibration should choose the optimal pulse repetition rate in function of the measured S/N.