Manticore Oscillogram

[Mark-VA]

Tried various damping resistors no stability.
Hold off to just a little more that period.
In order to get a half way decent shot I have to stop the scope.
A lot of, what looks like overshooting but there is a point of repeating.
Also did a FFT and looks like 3 distinct frequencies

[Carl-NC]

Visually, I see 2.6kHz, 7.8kHz, and 33.8kHz.

[Olly]

Depending on the operating mode, the frequencies according to the manual should be 5, 10, 15, 20 & 40 kHz, but in 'High Conductors' search mode the the lowest is below 4kHz, but is not specified.

[Mark-VA]

Visually, I see 2.6kHz, 7.8kHz, and 33.8kHz.

May I ask how you visually see the stated frequencies?
I must be interpreting the FFT completely wrong.

I would love to learn more about the correct and factual way of determining the true, and actual transmit frequencies that can be extracted from the oscillogram.

At your leisure please eliminate.

Mark

[Skippy]

"May I ask how you visually see the stated frequencies?"
A good question, it's a bit of an art.
Working out the repetition peroid of the whole waveform is the first step. That gives the first clues: the actual frequencies will be related to this in integer multiples., ie. f; 2f; 3f; 4f etc are possibles. The lowest, f, is unlikely, though can't be ruled out.
To find the lowest frequency, you have to 'low-pass filter' the signal in your head ... imagine all the highest frequency components are 'filtered out' , leaving only the slowest sinusoid that repeats a whole number of cycles during the total waveform period. For example on the Equinox, that is 7.8 kHz, and there are 3 cycles of that in the total waveform.
To find the maximum frequency component, apply a mental 'high-pass' filter, to just pick out the obvious highest frequency present.
The intermediate freqs ( if there are any ) are harder to spot, the FFT helps, as does the 'expected' values ... they are unlikely to transmit 15kHz AND 18kHz simultaneously, for example.

[Skippy]

So, in the Manti scope traces you posted, the repetition time for the waveform is 390 microsecs, ie. 2.56 kHz.

The highest freq can be seen to do 2.5 cycles in about 53 microsecs, ie about 21 usec period, or 47.6 kHz .... this has to be a multiple of 2.56k, so your choices are 18 x or 19 x the 'fundamental'. The correct answer is 48.6 kHz, just gut feeling that 18x is not likely, even though it could be seen as 3x3x2. Conveniently, your FFT confirms that 48.6kHz is in fact the answer. If you changed your scope timebase, you could measure the period of the high-freq burst with more precision, and come to a more accurate figure than I can.

For the lowest frequency: The entire waveform looks to have perfect anti-symmetry about the half-period, I just don't see any obvious component at 2.56kHz. The lowest I 'see' has 6 zero-crossing points, ie. 3 cycles per waveform. That is to say 3 x 2.56 kHz = 7.7kHz. This what I expected, as it's what the Equinox does, and it's 'confirmed' by your FFT showing '8kHz'.

For the difficult 'middle frequency': It's very hard to spot, as it is on the Equinox waveform. The Eqx has 18.2kHz, which is 7 x the waveform period , so one obvious guess is 18.2k, as they may have kept it the same as the Eqx. Other 'likely' multiples are 9x, 11x, 13x, 15x. My hunch would be that it's higher than the Eqx, because they have raised the upper frequency ( from 39kHz on the Eqx ) , so the mid-freq will go up too. It looks like 11x 2.56k is the answer, based on your FFT analysis.

So the three frequencies are in the ratio 3 : 11 : 19 . Compare with the Eqx waveform having 3 : 7 :15

I hope that made sense.

( for info, the FFT of the Manticore was posted elsewhere, some time ago ( maybe on DetectorProspector ) and they found a 49 -ish kHz highest freq component, too. )

I should also add:
Square-edged waveforms always create odd harmonics. So there will be some signals at 3 x 48.6, and 3 x 7.8, and 3 x 28.2. Plus numerous mixing products. It's safe to say: "All of these additional components are ignored"
They are weaker, not necessary, and not part of the design process. Plus, for example, 3 x 48kHz is way outside the limits of the coil itself, so there's probably not much good info comes back on the RX coil at that freq, anyway.

[Richard Ervin]

I have wanted to measure this for awhile, myself. What kind of test setup did you use? Did you open up the machine to get to the test points?
Thanks!

[Skippy]

Richard: you need to read Mark's previous threads, in particular this one:
XP Deus 2 Transmit frequencies - Geotech Forums (geotech1.com)

In summary: Use a pickup coil with a light load resistor on it (to kill off the ringing produced by L and distributed winding capacitance). This will indicate the coil drive voltage, as it forms a transformer with the TX coil winding.
And a shorted, or near-shorted pickup coil to give an indication of the TX coil current.

[Mark-VA]

I use a Rigol 1000Z series scope and a home made pickup coil.

[Richard Ervin]

I sort of figured I could couple to the coil with a homemade pickup coil; I had just not tried it yet! I saw this and thought you guys were inside the thing. That really interests me! I have been doing electronics for quite awhile and I still enjoy peeking under the hood…..Thanks for the info!

[ivconic]

I wonder if a digital oscilloscope of that class (low budget) is capable of "catching" even subtle transitions?
Carl sees only what the oscilloscope is able to "capture"... I'm afraid you don't see the whole "picture" of what's going on.