Let's made a PC-base metal detector with usb interface !!!

[gwzd]

Don't waste your time Sulaiman, he has no idea neither!
Hi Aziz! I'm back after working for real during the spring and summer.
What's up?! Still milking the german welfare?
RA

[Aziz]

Hello Sulaiman,

I am sorry, this project is not easy to follow and to build. It is not destined to offer this as a homemade project.

Aziz

[Sulaiman]

Dear Aziz,


I hope for you the best wishes in your life,
I'm interesting in this project from yesterday and I hope if you explain it and give me the summary about it.
Did you use laptop, special program and Electronics boards if I'm not wrong ?
What is the special program you use it, which platform you operate it on ?

BS,
Eng. Sulaiman

[Aziz]

Hi jamal,

HI AZIZ & my FRIND
I am ask you are complete made PC-base metal detector . and how distance to detect . If the sample 10 * 10 cm
thanks for you,,,,,,,,,,,,,,

no, it is not finished yet. It is destined for experiments only. Not for a real metal detector yet. The results of this experimenting platform will flow into a high performance metal detector later (very later).

I think, this project is becoming more or less a metal/ground anomaly detector development kit. It will help me to understand and learn the matter better.

As I am a very poor guy (no, I still do not want to work for krauts - never ever), the development suffers quiete due to expensive parts. Nevertheless, I will have this experimenting platform one day and the german government will pay this. Maybe I should apply for a government supported project, which could offer more benefits.

Aziz

[jamal]

PC-base metal detector

HI AZIZ & my FRIND
I am ask you are complete made PC-base metal detector . and how distance to detect . If the sample 10 * 10 cm
thanks for you,,,,,,,,,,,,,,

[Aziz]

Hi all,

I just have been told, that isn't a bug, but a feature!

The internal LTspice mosfets are not modelled in avalanche mode. Just use external mosfet models via ".include" command.

I am wondering, why they are then placed in the spice library?
Aziz

[simonbaker]

Hi PICON,



I had the version 4.04d (faulty).
Updated just a few minutes ago to version 4.04h and it seems to produce same faulty results. It seems, every n-mosfet in the internal spice library do not avalanche and delivers 1.6-2 kV flyback voltages (hey, what an incredible super target response!).

But my external spice parts seem to work properly.
And I thought, I hit the jackpot! What a disappointment.

Aziz

Thanks for warning us and let us know for replacements or how to fix.

Regards,

-SB

[PICON]

Hi Aziz,

Thanks for the reply.
I have designed a PI TX stage that can handle 3Kv flyback pulse but can't model it on LTspice. It would be nice to model it first but I think I will wait for the parts to come and just build it and then debug it.

[Aziz]

Hi PICON,

Hi Aziz,

I had the same problem with modeling PI TX stage in LTspice and flyback pulses being way larger than the mosfet breakdown voltage. I am new to LTspice and have attempted to edit mosfets model with no success. How have you over come this problem?

I had the version 4.04d (faulty).
Updated just a few minutes ago to version 4.04h and it seems to produce same faulty results. It seems, every n-mosfet in the internal spice library do not avalanche and delivers 1.6-2 kV flyback voltages (hey, what an incredible super target response!).

But my external spice parts seem to work properly.
And I thought, I hit the jackpot! What a disappointment.

Aziz

[PICON]

Hi Aziz,

I had the same problem with modeling PI TX stage in LTspice and flyback pulses being way larger than the mosfet breakdown voltage. I am new to LTspice and have attempted to edit mosfets model with no success. How have you over come this problem?

[Aziz]

Hi,

I want to mention, that the internal LTspice library contains some bad faulty spice models. If you look to the mosfets, some has different voltage breakdown voltages as specified in the data sheet and a lot of did not even have the breakdown voltage. So the pulse flyback voltage won't be clipped properly.

I have the newest update. What the **** has been changed?
Aziz

[Aziz]

Hi all,

regarding the novel method in the last post:
f1ck! f1ck!

Be very careful, where you get the parts spice model!!! I have found some bogus spice models in my library, which tend to show incredible results. Back to the roots and to the black board...
f1ck! f1ck!

Take care,
Aziz

[Aziz]

Hi all,

Project update:
No progress due to lack of means (money) and further delays for several month until I can afford the new parts.

I have been working on a novel method and it seems, I really hit the jackpot this time. The idea is very likely not invented yet (need to investigate that) and it promises much more performances, on which I will continue the work. I am sorry, that I can not publish any details of it and of course, it will delay this project significantly.

Take care,
Aziz

[Aziz]

Hi all,

I want to show you some FFT displays (visible up to 45 kHz using the better sound card). So you can make your own thoughts.

Aziz

[Aziz]

Hi all,

regarding the 24 bit HD sound cards and some comments on them:

There are good ones and bad ones. The two USB sound cards I already showed here (white and black one) differ quite.

They seldom achive 24 bit resolution due to heavy internal noises. You are quite lucky, if they even achieve 20 bit.
I have found, that my black sound card achieves 23 bits of resulution, if the input signal is attenuated as much as possible to reduce the noise of the sound cards internal amplifier. At 0 dB (no attenuation, 1Vrms input range), the resolution decreases down to 18 bits (noise floor at -110 dB).

The white one is much worser. The internal clock is not accurate. The black one is superior (very accurate clock and stability, lower noise). If the input signal is attenuated by 20 dB (x 0.1), the internal noise floor is at -130 dB (almost 22 bits). I even can set to a minimum input level (mixer ruler set to minimum possible), then the noise floor decreases down to -140 dB (23 bits). So the input signals should be rised in level from the external low noise circuit to use the full ADC resolution.

The big noise source is definitely the EMI (induced by the coil itself). So the conditioned four analog signals will be quite noisy. If they are somehow correlated, the software could compensate this (big advantage).

Aziz