Minelab Equinox Challenge

[bklein]

Sorry to be off topic a bit here but say you do use a 1 ohm resistor in place of the TX coil. Would a CTX or Equinox still "run"? I'm wondering if then you could get a more accurate TX waveform. Also, another wacky idea, if the detector still runs (no "no coil detected" error or similar), could you experiment with your own TX drive circuit to actually drive the TX coil? (Does the rx circuitry/code need to synchronize with the TX signal?

[Qiaozhi]

What kind of grass do you have on your lawn? It's making you creative.

I've just got your meaning after a short temporal delay!
It wasn't "wacky baccy", if that's what you meant.

[Qiaozhi]

Quote:"While mowing the grass"
What kind of grass do you have on your lawn? It's making you creative.

How are you determining 'Power' , when there's no resistive object in your model to get hot?

The coil has a series resistance of 1ohm. It's not shown on the schematic.
V1 is a zero-volt voltage source (a shortcircuit) that's used to monitor the coil current. This is used as input to the H element (current-controlled voltage source). Then an E element (voltage-controlled voltage source) has an attached polynomial that multiplies the voltage at the X node with that at the Y node to calculate power. The graph facility will then let you run an FFT on the Z node.

[Skippy]

Quote:"While mowing the grass"
What kind of grass do you have on your lawn? It's making you creative.

How are you determining 'Power' , when there's no resistive object in your model to get hot?

[Skippy]

Quote:"When I open the text file I get a bunch of A's"

The file contains 386 single-precision floating-point numbers, one after the other, no seperators like comma's etc, no End-of-file marker. The numbers are 4 bytes each, so the file size is 4 x 386 bytes in size.
I named it .txt , but could've chosen .dat, I suppose. It will just be blanks and random characters when viewed with WordPad etc.
It should easily import into programs like MatLab, or any proper programming language, and is among the varied formats supported by the ScopeDSP software.

[Qiaozhi]

The thing that's been bugging me is why the voltage and current waveforms show peaks in the FFT spectrum that are at odds with the numbers published by Minelab, and the troughs seem to be aligned with the expected values. While mowing the grass I came with an idea, so here's the result ->





If you generate an FFT of the instantaneous power (and not just the voltage or current) the peaks are located as shown above.
There are actually several peaks at 5.2kHz, 10.4kHz, 15.6kHz, 20.8kHz, 26kHz, 31.1kHz, 36kHz, 41.6kHz, 46.6kHz, etc.

The ones of interest are (of course) 5.2kHz, 10.4kHz, 15.6kHz, 20.8kHz and 41.6kHz.

Since LTspice won't allow you to create an FFT result from a waveform generated by the graphing functions, I had to calculate the power in the coil by using a nonlinear controlled source. Hence why the power is actually displayed in volts at node Z. You need to ignore the values shown anyway, as these are just arbitrary.

[green]

Curiosity got to me.
With QuickBasic now on my current machine, I gave it a spin. I read off the transition times from Carls current chart, but not the 'current'. Using a Basic program, I entered 386 points of data as either Hi or Lo, giving them the arbitrary numerical value of +8 and -8, so I have a reconstruction of the coil Voltage waveform. This was saved as a Binary data file consisting of a string of 386 single-precision floating point numbers (attached).
This file was then loaded into a trial version of 'ScopeDSP' by Iowegian. This produced a spectrum analysis , showing the 3 : 7 : 15 multiples of the 'basic' 2.6kHz waveform repetition rate, the '15' multiple, of 39kHz being the strongest, the 7.8k and 18.2k less strong.


Attached .txt data file, waveform reconstruction, spectrum analysis.

When I open the text file I get a bunch of A's. I know this is a detector form and not a class on computers but since I'm computer stupid maybe you could give me a little help. Thanks

[Altra]

On the Advanced page, add your attachments as normal, then click on the paperclip icon in the top row of the menu. There's a drop-down box from which you can select images that you want to be displayed.

Thanks, I'll give it a try next time.

[Skippy]

I notice that Sled came up with 8.3kHz, matching the figure of 8.26k shown on Carls scope screenshot.
I assumed that the 8.26k figure would be a jumpy value, that depended on what part of the waveform was showing on the screen at any given time. Scope's aren't good at repeatably triggering on complex waveforms.

[Andi68]

8,26kHz

[Skippy]

I hope it's in the visible spectrum.

[Andi68]

I only see one frequency

[Skippy]

Curiosity got to me.
With QuickBasic now on my current machine, I gave it a spin. I read off the transition times from Carls current chart, but not the 'current'. Using a Basic program, I entered 386 points of data as either Hi or Lo, giving them the arbitrary numerical value of +8 and -8, so I have a reconstruction of the coil Voltage waveform. This was saved as a Binary data file consisting of a string of 386 single-precision floating point numbers (attached).
This file was then loaded into a trial version of 'ScopeDSP' by Iowegian. This produced a spectrum analysis , showing the 3 : 7 : 15 multiples of the 'basic' 2.6kHz waveform repetition rate, the '15' multiple, of 39kHz being the strongest, the 7.8k and 18.2k less strong.


Attached .txt data file, waveform reconstruction, spectrum analysis.

[Qiaozhi]

For comparison purposes here is a scope shot of the Whites DFX. Ignore the frequency readings on the scope it's confused.

Very interesting.
It looks like the Minelab Equinox is an extension of the same concept.

[Qiaozhi]

How do you guys get the images to insert on the page? I have to go to advanced and upload

On the Advanced page, add your attachments as normal, then click on the paperclip icon in the top row of the menu. There's a drop-down box from which you can select images that you want to be displayed.