VLF MD with digital signal processing : Bee-Buzz 1

[dbanner]

"Generation-recombination noise" in a transistor, particularly due to the base current, refers to a type of electrical noise that occurs when electrons and holes randomly generate or recombine within the semiconductor material of the transistor's base region, causing fluctuations in the base current and subsequently impacting the collector current, leading to a noisy output signal; this noise is most prominent at low frequencies and often exhibits a "1/f" (pink noise) characteristic."

"Compared to transistors on integrated circuits, discrete transistors generally exhibit higher levels of flicker noise due to the less controlled manufacturing process, resulting in more imperfections and variations in the semiconductor material, which are the primary source of flicker noise (also known as 1/f noise) in transistors; whereas, modern integrated circuits with advanced fabrication techniques tend to have significantly lower flicker noise levels due to tighter control over the manufacturing process and smaller feature sizes."

-Google AI overview

[dbanner]

https://www.analog.com/en/lp/001/cus...ad-cortex.html

Unrelated but impressive.

[Detectorist#1]

Noise, noise, noise .... of the first stage in VLF and PI - now it is not problem in serious designs with now days OpAmps and new schematics. We have other problems in improving of our MDs.

[Aziz]

https://www.analog.com/en/lp/001/cus...ad-cortex.html
Unrelated but impressive.

Hi dbanner,

529-ball grid BGA package
Neural Cash Cow -oops- DSP.
Typical marketing B$.

Not enough pins to start with!
I wouldn't touch a cpu/dsp under LGA1155 pin count.

Don't dream of such machines. You can have much better machines without using soldering iron.
Cheers
Aziz

[dbanner]

Try Octave it's prob closer in capability.

Octave compares favorably with MatLab. It's also free and open source and has good support. Capability seems to be on par with most of the functionality.

[moodz]

Hi dbanner,

529-ball grid BGA package
Neural Cash Cow -oops- DSP.
Typical marketing B$.

Not enough pins to start with!
I wouldn't touch a cpu/dsp under LGA1155 pin count.

Don't dream of such machines. You can have much better machines without using soldering iron.
Cheers
Aziz

Aziz ... start a new thread ... the Bee Buzz should be focussed on the blue pill and the VLF development and what can be done with it.

[moodz]

Octave compares favorably with MatLab. It's also free and open source and has good support. Capability seems to be on par with most of the functionality.

Yes .. I used both years ago. Either should work for VLF sims.

[Atul Asthana]

Hi all,

in my special USB high latency case (Windows isn't an RTOS),
I have to decode even twice more to get the correct absolute phase lag between TX and RX. So I have to process the TX signal as well (besides the RX signal). Btw, any TX energy loss and frequency shift will be detected too. This gives more info for further processing.
For a true dual frequency VLF/LF detector with three narrow frequencies around the resonant frequencies, I need 12 decoders. This is no problem with CPU power on Tablet PC. But can be critical on embedded systems with micro controllers.

BTW, it is important, that the narrow band width should not be large, as we don't want to to operate the coils in the high Z region (Z is impedance of the LC-tank). We need some current flow through the TX coil of course. We are operating the TX coil in the low Z region around the resonant frequency fr.

Heavy mineralisation will increase the TX/RX inductance (and thus lower the resonant frequency) and targets will lower the inductance slightly (hence increases the resonant frequency). Magnetic field conduction TX -> RX occurs on ferro magnetic materials nearby the coil. Eddy current induction on metal targets too. All possible effects can be processed at the same time.

Aziz

Aziz,
YOUR EFFORTS will break the existing metal detector manufacturers

btw, for your information, Window's monolithic structure and its irregular internal integration is a nightmare for any real time data acquisition/processing.

Linux is much better : you can run it barebones to the minimum, and I know of radars, sonars and gun control systems happily running on Linux.

You should try with ChromeOS tablets, seems the 'ChromeOS' is more of a shell and VM/compatibility layer (for android) with underlying everything on thinned out linux.

I considered using my ChromeOS tablet before focusing on Bee-Buzz 1, but didnt have enoigh time to read thru the ChromeOS and analyse it for the real time requirements.

[Aziz]

Hi all,

just want to let you know, that my new USB sound card has been delivered and it works nice. Slighty more noise floor above 60 kHz. Between 70 - 80 kHz there is a low pass filter corner frequency. But I can still detect and process signals up to 95 kHz with some attenuation. Nice.
Aziz

[dbanner]

Real-Time Linux preemptive.

[dbanner]

Hi all,

just want to let you know, that my new USB sound card has been delivered and it works nice. Slighty more noise floor above 60 kHz. Between 70 - 80 kHz there is a low pass filter corner frequency. But I can still detect and process signals up to 95 kHz with some attenuation. Nice.
Aziz

I own one of these Alienware alphas. https://www.cnet.com/reviews/alienware-alpha-review/
Can I adapt it to do what you are doing?

[Aziz]

I own one of these Alienware alphas. https://www.cnet.com/reviews/alienware-alpha-review/
Can I adapt it to do what you are doing?

Nice Box!
---

Oh oh oh!
I'm not happy with my new sound card (USB Creative Sound BlasterX G6). Too much noise in the upper frequency range (48 - 96 kHz). SNR down to -100 dB at Line out -> Line in loop back connection. This gets even worse sometimes.
Below 48 kHz signal range it is acceptable (-110 dB). But I have better other sound cards.

Found another issue: USB-Power supply does not seem to be good filtered in the sound card and is producing noise in the upper frequency range too.
I have to buy or make an USB-Power filter. I hope, it gets better then.

I have to redesign the old detector software (it has 1001 features ). A light weight, less power consumption version would be really nice for field testing and detecting.

Cheers,
Aziz

[moodz]

If you want a compact development platform just get hold of an android phone with decent specs eg LGV50 has a really excellent codec onboard ..... sound cards are from 10 years ago.



Android is way easier to code than windows anyway.

Also you still have not done a signal flow / system diagram ( with maths ) and you have missed one big GOTCHA ( dont worry ... the commercial manufacturers missed it also - but that is a story for a later time ).

[Atul Asthana]

If you want a ............... and you have missed one big GOTCHA ( dont worry ... the commercial manufacturers missed it also - but that is a story for a later time ).

Great observation, but unless the observation is revealed, its as good as not being there.

And, If commercial manufacturers, with an Army of developers, researchers and umpteen resources missed something important, I think I can be pardoned, being a lone researcher, with very little time for all this, and completely out of electronics / signal processing / metal detector field for many decades.

though, I am not sure that the commercial manufacturers would have missed out something important, it msy be that they havent revealed or found it too insignificant or have found better alternatives.

[moodz]

Great observation, but unless the observation is revealed, its as good as not being there.

And, If commercial manufacturers, with an Army of developers, researchers and umpteen resources missed something important, I think I can be pardoned, being a lone researcher, with very little time for all this, and completely out of electronics / signal processing / metal detector field for many decades.

though, I am not sure that the commercial manufacturers would have missed out something important, it msy be that they havent revealed or found it too insignificant or have found better alternatives.

It has been said that a true scientist is a cynical opportunist ... believe in something absolutely until something new comes along . Another quote that comes to mind from Lord of the RIngs is "much that once was is lost, for none now live who remember it"

Nowadays engineers apply lego block principles .. this ADC that opamp this CPU .... it leads to mistakes due to assumptions that are made at the block level when what is needed is a continuum signal analysis. When it is done this way there actually is no difference between a PI and VLF or even a GPR detector ... its all just numbers. The question is what numbers are important and what is the relationship over time those numbers have.

In short if you sample any waveform ( PI VLF even GPR ) with a high enough bandwidth ( sample rate ) and high enough resolution ( efffective number of bits ) then all the detector modes ( VLF PI GPR etc ) can be performed using the right math transforms on the sampled data. This is not what the missing ingredient is though ... there is always something that can be made better hence we have patents ( though some are rubbish ).