PI Power Output

[nick_f]

Hi Tinkerer,

I can hardly see the circuit, it is too small.
Did you try adding an electrolytical cap in parallel on the V+ to AGND?
Can't you use an external power supply (linear and regulated with a LM317) to provide 15V? If it is only for testing, that would be ok,

Regards,
Nicolae

[Tinkerer]

Hi Tinkerer,

I can hardly see the circuit, it is too small.
Did you try adding an electrolytical cap in parallel on the V+ to AGND?
Can't you use an external power supply (linear and regulated with a LM317) to provide 15V? If it is only for testing, that would be ok,

Regards,
Nicolae

Here is the Zipped EAGLE file.

Tinkerer

[B^C]

Thanks for the explanation Aziz, It is hard to model all the parameters & mostly due to them being unknown at this stage.

I have a lot of data that doesn't seem to be able to be reproduced when not in the field for some reason?

Like mentioned it would be good to have everybody test with the same things.

Target response signal reading, what is the best way to do this, i have tried numerous ways & also including just having an O-scope probe across the test piece--target but the results vary.
I have gold nuggets to get some readings from but as yet i do not like the way i have been testing these or the results.
I want to test this individually away from the MD circuit, basically i want to use just an adjustable one shot circuit with inductor charge time & voltage variable, trigger this to pulse the target & read the target response with no other circuit connections, whats the best way to get good results.

TX on time, if it is the transient spike that is going to give us the largest target inducing pulse then the TX on time is really only the inductor charge time--correct?
If i am correct then all the test results will vary from individual to individual because all the inductors will vary so the intensity of the inducing pulse will also vary & so the results.
We will need to find the inductor saturation time as it would seem going beyond this point may be a waste of time & power?

As a test std why don't we just use say some 18AWG winding wire as it's readily available & we can just wind a suitable diameter inductor for testing, at least this way we can share comparitive results.

I was reading an article where a university tested a lot of detectors including a ML prototype, with air testing of the targets all other detectors where better by far than the ML. The second test was done with the target in the ground & the ML won hands down & all the others where not comparable--food for thought!

Tinkerer, it's hard to see the schematic, can you post it again?

[B^C]

Tinkerer,

I unzipped the file you posted but it seems it's in a different format than what i can open, is there a program i can open it with?

[Tinkerer]

Tinkerer,

I unzipped the file you posted but it seems it's in a different format than what i can open, is there a program i can open it with?

Hi B^C,
the EAGLE is V5.4. It seems that the different versions are not compatible with each other.
I have changed the schematic into a PDF so that should work.

I like to use just a front end for the testing. Then I know what I see. Once the signal has been processed, one does not know anymore how it looked like at the beginning.
I feel the most important part is the front end. If the input is bad, even the best processing will not make it good.

Tinkerer

[Tinkerer]

Forgot to attach the PDF

It just does not want to upload. will try once more.

Tinkerer

[Tinkerer]

Once more

Tinkerer

[nick_f]

Hi Tinkerer,
I suspect you refer to the +12V output of the LM317. When the output is noisy, can you check the signal at its input, if it also gets noisier?
I don't know if it helps, but can you try a large capacitor between V+ and AGND? (2200uF maybe).
What's the voltage at the input of LM317? Do you have at least 3V more?
I also think you can put a small cap in parallel with R4 (isn't one in the recommended circuit?).

Regards,
Nicolae

[Unregistered]

Tinkerer, i tried the pdf but all i get is a message saying insuffient data for image, not sure?

I should have explained more about the target response signal, i need to read this direct from the target itself, read the eddy current strength directly from the target.
I want to eliminate everything & then once i get a signal from that i can then see how everything else affects this pure signal if you get what i mean.

I have a lot of data that is from very bad ground conditions, if i can get a pure signal from a target then i can compare these signals & put numbers to the ground noise interference etc.
I have no issue sharing the info but i have to get the pure target comparitive figure as a constant before i can proceed with anything.

I have tried many methods but still there are questions regarding the tests so obviously i am not getting it correct.

[B^C]

Sorry forgot to log in, my message above---unregistered about the target signal.

[nick_f]

Hi B^C,
The pdf works, I opened it. Maybe you need a newer version of acrobat reader?
How can you read the eddy currents directly from the target? Unless your target is a very tiny coil, followed by a preamplifier?

Regards,
Nicolae

[Mustafa]

Tinkerer's schematic with more resolution

[Tinkerer]

Tinkerer's schematic with more resolution

Thanks for all the help.

I have traced the problem.
Q1 and Q2 do not provide enough current.
The switching frequency is limited to 10k because of the ICL7662 (not ICL7660)

I am changing the capacitor C6 for 6800uF. (Amazing all the old stuff forgotten in the parts box)

Tinkerer

[Tinkerer]

HIGH POWER AND LOW GAIN

Changed C10 to 1500u, C6 to 6800u, C7 to 3300u.
Changed PPS to 900
Changed TX pulse to 72uS at 12V, that comes to 2A coil current at switch OFF.
Sample taken at 2.5uS after Switch OFF

Gain of preamp 69

S&H

Low pass R/C filter (R4k7/C940n) at output of S&H

Dynamic testing, swinging the target suspended from a string, across the coil.
Results:
1" square alu foil detected at 25 cm
0.55 gram silver ring at 30 cm
0.5 gram gold ring at 25 cm

I feel the results are good.
Does anybody have some numbers to compare with?

Tinkerer

[hobbes_lives]

Tinkerer,

Your findings in post #59 above are quite interesting. When I get back to it, I will find some of the simulations that I ran and perform some similar tests. I recall some differences in target response by varying the pulse width, but I'll have to dig through my notes, or recreate the tests, to figure out the details.