IF YOU HAVE MAKE , TRY THIS .

[sisco]

Hi all

If somebody want detect gold with minipuls must programmed ATMEGA8 with

this program (but LCD don't work and only can use HP or SP)

$regfile = "m8def.dat"
$crystal = 8000000
Config Timer1 = Timer , Prescale = 1
Stop Timer1
Config Portb = Output
Config Portd = Output
Portb.1 = 0
Portd.6 = 0
Portd.7 = 0
Dim A As Word

Startline1:
Start Timer1
A = 0
Portb.1 = 1
Do
A = A + 1
Loop Until A = 27
Portb.1 = 0
A = 0
Do
A = A + 1
Loop Until A = 4
A = 0
Portd.6 = 1
Do
A = A + 1
Loop Until A = 12
A = 0
Portd.6 = 0
Do
A = A + 1
Loop Until A = 53
A = 0
Portd.7 = 1
Do
A = A + 1
Loop Until A = 12
A = 0
Portd.7 = 0
Do
A = A + 1
Loop Until A = 158
Stop Timer1
Goto Startline1

[sisco]

after test many timing , I find these good timing for detect :

1-IRON (not gold...)

2-ALL METAL (GOLD+SILVER+IRON+...)

3-ONLY GOLD

if use IRON.HEX you can detect all metal without gold.

if use ALLMETAL.HEX can detect all metal.

if use GOLD.HEX you can detect ONLY GOLD.

I try write a program can use selector key for choice IRON,ALL,GOLD.

and other program automatic change delays and DISC your object.

DON'T HURRY , STEP BY STEP

[sisco]

I forget tell , I test this timing by 13cm dia COIL.

[sisco]

Only replace R52 with key and programming your minipuls-3 ATMEGA8 with

this HEX file and select between 16 difference choice for best detection .

but must know with this software only can use small coil and with large coil

don't answer.

be successful

[Geo]

Large coils needs large tx pulse, 150.... 600us. If you use hight frequency maybe this value to be out of the pulse cycle.

[sisco]

Large coils needs large tx pulse, 150.... 600us. If you use hight frequency maybe this value to be out of the pulse cycle.

today I understand why don't work with large coil , don't work because

frequency is HIGH. I test with 100Hz and 50Hz and see work.

[sisco]

I use 50Hz with 350us PW and work very good with good sinsiviti.

[Geo]

I use 50Hz with 350us PW and work very good with good sinsiviti.

With 50Hz you can put the Tx pulse width at 1000us.

[sisco]

Geo , I test your suggest.

use 50Hz with 1000us PW and :

with large coil : work better about distance for big object near 25%

with small coil : work better about distance for small object near 10%

THIS PROJECT FOR TEST TIMING IS BEST PROJECT

BECAUSE YOU CAN TEST YOUR IDEA ONLY IN 2 MINUTE WITHOUT CHANGE

ANYTHING IN CIRCUIT.ONLY MUST CHANGE SIX NUMBER AND REPROGRAMM

YOUR MICROCONTROLLER.

(1-ON PW,2-DELAY,3-FIRST PW DELAY,4-INTERDELAY,5-SECOND PW DELAY,6-OFF PW)

[Great_Alex]

Hi Sisco

you choose a nice project , just keep on try ...

could anyone disassemble the Hex file of ATMEGA 8 in orginal assembly language ?

[aft_72005]

Hi Great_Alex
Please attached mini pulse disasm file.
Study this file for how it works minipulse is difficult.
Regards.

[aft_72005]

Hi again Great_Alex
Sorry I cannot upload rar file with 10 kb capacity .

[sisco]

Max , you think this idea is true or not ?

I think maybe can use test point more than 2 location for example 8 location

and test amplifier output voltage and with compare between this difference

voltage understand this object is what metal .

I try do this but can't , because internal micro A/D burn 65us-250us for

every convertion and this time is very high so must use external A/D

converter with high speed.

can do other way for understand about disc?

THIS CHART ONLY FOR EXAMPLE.

[Max]

Max , you think this idea is true or not ?

I think maybe can use test point more than 2 location for example 8 location

and test amplifier output voltage and with compare between this difference

voltage understand this object is what metal .

I try do this but can't , because internal micro A/D burn 65us-250us for

every convertion and this time is very high so must use external A/D

converter with high speed.

can do other way for understand about disc?

Hi,
the idea is good... even if then in the field there are other factors that can alter your elaboration cause of wrong samples due e.g. to magnetic viscosity...

Problem is that you need a very fast hardware to perform well.

A solution could be a fast hi-res ADC (16-bit min) then an FPGA and then the "slow" MCU. But is really complicated.

Maybe a fast pic or a newer atmel with integrated reatively fast ADC could be used... I think you can gain e.g. 20uS conversion time that way.

Kind regards,
Max

[gef12]

Hi 1843 ,

you make this or not ?

can help me for make better this ? (like AD844)

and can test AD8055 in this detector because I try two month for find Ad8055 but can't find this.

Would an AD797 be better choice .. ultra low noise better than ad8055
cheers