The Mighty "Fisher CZ5 Quicksilver" and it's secrets

[Altra]

"What would the 5th harmonic look like if we used a simple 3kHz square wave for TX ?"
The third harmonic has amplitude = 0.333 relative to the fundamental; the fifth harmonic has amplitude = 0.200 relative to fundamental.

For a theoretical 'CZ' square wave that was amplitude +10 Volts / -10 Volts , the amplitude of the sine wave constituents is:
Fundamental = 12.73 V (9.0 V RMS)
3rd harm = 4.24 V (3.0 V RMS)
5th harm = 2.55 V (1.8 V RMS)

Thank you!

[Skippy]

I will try and play with some f / 3f waveforms tomorrow.

My thoughts regarding the XP 3-level waveform, versus these PWM-style ones:
The 3-level waveform does vaguely resemble the ( y = sin x + sin 3x ) on which it is clearly based. I think the 3f demodulator will produce a cleaner signal, with less cycle-to-cycle ripple , when compared to what is produced by demodulating the PWM-style waveform. This could help reduce the filtering needed ( after the demod, or 3f channel filtering before the demod ). This would help achieve a faster responding detector at the low 4.7k/14.2k freq mix. FMF = fast multi-freq. No point making a fast-responding machine at higher freqs if it becomes sluggish at lower freqs. I suspect this 3-level waveform plays a part in keeping the Deus2 fast, even in the lower freq modes.

You will observe Minelab's Equinox doesn't encounter this problem. And the reason could be: their lowest transmitted freq is 7.8 kHz, fast enough to avoid the issue. The Eqx coil is certain to work at 5 kHz, so ML could've run a lower-freq mix ( eg. 5.1k / 11.9k / 25.5k ) , they may have tried it, and discovered this problem.

Inspired by this thought, I wondered if the 3-level waveform could equally be applied to a f / 5f mode. And yes, I think it would be possible. I played around with that 'Desmos' graph drawer, and could easily produce a 3-level waveform that would do the job.
If the Equinox coil was used for this, the upper freq limit is probably over 40 kHz, but it's unknown. So using f / 5f with a 40kHz limit naturally leads me to 2 of the actual Eqx frequencies of 7.8kHz / 39kHz as the higher limit, and perhaps 4.8k / 24k as a lower limit.

[Skippy]

Here's a three-level waveform that's intended to mimic: y = sin(wt) + sin(5wt)
that has similar fundamental and 5th harmonic amplitude, with minimal 3rd, 7th etc harmonics.
In this example I chose 5kHz / 25kHz to illustrate it, but it's not important, it applies equally to 3kHz/15kHz or 8kHz/40kHz as you want.
Attached are the waveform, with 160 steps per cycle. This is a multiple of 20, meaning that precise I & Q demod signals for both fundamental and 5th harmonic can be created along with the TX.
And the spectrum analysis, showing 5th harmonic levels 6% greater than fundamental level. The actual levels are shown half what they should be (shrug), they are actually 3.2 V and 3.4 V for the +5V/-5V waveform.
So the TX mimics: V = 3.2 sin (wt) + 3.4 sin (5wt)

[dbanner]

PDM is what I think maybe used to generate the dfx waveform type. It is a pulse train in which it's pulse duration is modulated by a sine wave using PDM techniques. Derived from mathematical formula, to write in assembly might be pretty straightforward for those who are skilled in the art.

[dbanner]

The carrier wave might simply be a pulse train of fixed duty cycle generated via PWM, then modulated using PDM. Perhaps the modulating wave can be triangle or sinusoidal.
It think it has to be a simple method well known in Telecom and satellite industry.