Universal PI Micro

[1843]

Thanks 1843.
That has helped me understand the discrimination. I can see where all I need to sample.
However,
What do the numbers indicate? are they relative time delay values? [5.85, 8.3, 34.1]
And what is the significance of the T[0....5]?

You need 2 sample, depends on your coil. If your coil is fast enough like the diagram you can use T5 and T4 (both in microsecond). The diagram timing is the best positions for proper DISC and keep in mind that the samples width should be 1 to 5us.

If your DISC doesn't work OK with this timing, then it means that your opAmps gain is not set properly. You should try to get this timing and then set the samples gain.

Regards,
1843

[bugwhiskers]

Hi Atul Asthana & 1843,

Thanks for that pic, it makes it a lot easier for people to understand the underlying principle.

You can clearly see that the better the conductor the flatter the curve.

The "T's" relate to the timings of the samples but it will only apply to their coil/preamp configuration, just experiment to find the sweet spot.



regards
bugwhiskers

[Atul Asthana]

Discrimination : Decay diagrams

Thanks 1843.
That has helped me understand the discrimination. I can see where all I need to sample.
However,
What do the numbers indicate? are they relative time delay values? [5.85, 8.3, 34.1]
And what is the significance of the T[0....5]?

[1843]

UPIM Principle

Hi Atul Asthana,
Its principle is here:



Regards!

[bugwhiskers]

Hi Atul Asthana,

It's not possible to give you absolute locations in terms of microseconds out from TX off to take your samples due to the spread of coils, pre-amp types and feedback caps on the pre-amp.

It's all a matter of adjusting the locations and widths of the two samples to get the maximum signal difference between targets of differing conductivity.
The best place to put the first sample is where the decay is more horizontal than vertical, just past the knee of the curve.

regards
bugwhikers

[Atul Asthana]

Discrimination

Bugwhiskers,
Thanks for the reply.
Can I request you to post decay diagrams and timings for various targets, which I could utilise in understanding the discrimination process and setting sampling delays.
Thanks once again.

[bugwhiskers]

HI Alexis,

The method has been shown to work both in air tests and with a target covered by undisturbed ground (lateral hole in bank of a creek).

The method could be applied to any PI that can take 2 samples in the knee of the curve, not just the evaluation board.

If you take the 2 samples and adjust their widths so the earlier sample is equal in amplitude to the later sample then a flatter decay curve (eg from Silver) will cause the later sample amplitude to be larger than the earlier sample. Optimisation of the above principle will be achieved through experimentation with different sampling positions to find the "sweet spot".

regards
bugwhiskers

Hello Bugwhiskers ,
For discriminate, did you test in soil , did you have some difference with air test because of the complex electromagnetic pattern generate by the coil in soil ,
can you tell us your experience in Australian field where mineralization is hard
we will apreciate very much your conclusion and this will stimulate the purchase of the evaluation board for us in Mexico,Thank you BW
Saludos from Mexico
Alexis

[bugwhiskers]

Hi Atul Asthana,

Ignore all the posts prior to the introduction of the Evaluation Board.

The discrimination method has been well documented in a number of posts but I will reiterate.

The decay curve shape is influenced by the conductivity of the target with better conductors producing a flatter curve and worse conductors a steeper curve. If 2 samples are taken to look at the shape of the curve (it's flatness) then the user has an indication of the conductivity and therefore the metal type with Silver producing the flattest and Steel much steeper.
Orientation of the target will have some effect but not to the extent of Steel looking the same as Aluminium.

regards
bugwhiskers

[Atul Asthana]

Complete documentation of UPIM required

Everyone involved in UPIM design and improvements has done tremendous amount of hard (and intelligent) efforts. And we need to thank all those involved.

However, the efforts' knowledge is distributed in many mails and drawings. This makes it difficult for any one to understand the entire UPIM.

I'll request if the knowledge could be put together in form of a document (please do not use docx[MS OOXML], which most of the people will not be able to open), it'll help a large number of people.

And, I'll be further grateful if the theory of operations (specially, discrimination) could be included.

[1843]

Hi bugwhiskers,

I think that for proper discrimination by this PI we need to a fast coil and low pulse width about 50-75us to allowing a fast decay.
Sample1 should be taken about 8us or less for disc between metals and sample2 should be taken at about 30us for the response range.
I don't have any idea for ground balance. What idea do you have for it?

Regards!

[Oscar]

Bugwhiskers,

Thanks for update!

Lets hope we get the support of all people interested in this project. I will try to do some research on your project as it is very educational and interesting. Now that I have obtained all available books about AVR programming both in assembly or C languages and also many references in hardware digital design. But I need to refresh my assembly language skills first before I can be any real assistance.

I also hope people interested in this project, get somehow involved and contribute depending on their capabilities. I do understand not all of us have electronics engineering backgrounds but being an open-source project, it is good opportunity for everyone to participate thanks to BW. I have been personally working full-time and studying part-time for last 5 years but I will try to find time for this project and do some experimentations.

I also like Hammerhead project as it is very modular design and educational.

Thanks everybody,

Oscar

[bugwhiskers]

Hi 1843,

The UPIM is a platform containing all the essential components for a PI detector. It can take 8 samples of any part of the decay waveform. It is left up to the user to make simple circuitry on an extender board to do other things like discriminate, ground balance, further amplification etc. There are many posts and some circuits on how the discrimination works based on the conductivity of the target.

I have been using the UPIM for the last few months doing some experiments with my own extender board.

Here are the results:
Coil: Round 10 inch 300uH
Total amplifier gain: 100,000
Target: Aluminium block (6060-T5) 100mm*50mm*10mm
Discrimination: Full Ferrous/Non-ferrous
Signal at 27 inches = 0.1v (100mV)

The method used above is not a "normal" PI which shows just how versatile the UPIM is in allowing the user to do their own thing.


regards
bugwhiskers

[1843]

Discrim sch

Hi 1843,

Please read post #202.


regards
bugwhiskers

Thanks!

Now the next step is to program an ATmega8 for this PI! I will be doing it when I get free time.

Best regards,
1843

[bugwhiskers]

Putting it all together.

Hi Oscar,

The most up to date info is as follows.

post #225 Eagle format schematic and pcb
post #234 .PDF of schematic and pcb
post #256 Firmware
post #260 Rotary Encoder info
post #262 BOM in text format

regards
bugwhiskers