Digital BFO-like (but not really) - discussion on feasibility

[Teleno]

Sensor characteristics:

- Mono coil, part of a tank oscillator.
- Sinus wave, 300mA, 40V, 90KHz. Oscillation amplitude is held constant by servo loop wit mV resolution.

Measured parameters:

- Frequency changes (X component), sub-Hertz resolution.
- Power changes (R component), 300uW resolution.

Would these characteristics be sufficient for such a detector to perform similarly to an IB detector in terms of depth and discrimination?

[homefire]

Why re invent the wheel ?

E - NANDBFOTXT.txt

[Teleno]

Because the existing wheels have low sensitivity and worse discrimination.

If PI's and VLF's can be improved, BFO's can also be improved. Your design can't separate the X and R components.

Today we have tools like MCU's that didn't exist in the days of your design. Control loops are also absent in the old designs. The oscillators are also very weak.

[homefire]

LOL Good One. BFO's will always have a place in my heart. I started off with one around 1972? Yes I think they have room to grow. I only posted the one I did because how Simple the thing is. I think your going to find the same Road Blocks with a Micro Controlled BFO as any other . Varying Ground Signal being most issue. Sub Hertz resolution is no help if everything around a target is changing in Hertz. How do you think you could improve on a Crystal controlled L/O and L/C search Oscillator ? Interested here.

[homefire]

Weak Tx power is of no real issues with detectors. Distinguishing Target from Background noise is the Main issue. More Power Produces More Noise. Sorta the same as the Diminished Returns of adding power to a PI detector. It takes a offal lot of power to realize a tiny bit of gain in target signal.

[Carl-NC]

Would these characteristics be sufficient for such a detector to perform similarly to an IB detector in terms of depth and discrimination?

Bottom line: no. An IB design is fundamentally more sensitive than proximity effects (frequency shift and power loading). Also, proximity effects are dependent on distance (depth) meaning you can't reliably discriminate.

[Teleno]

Bottom line: no. An IB design is fundamentally more sensitive than proximity effects (frequency shift and power loading). Also, proximity effects are dependent on distance (depth) meaning you can't reliably discriminate.

Existing BFO designs that measure changes in amplitude use crude rectifiers. In the Falcon MD-20, for example, a 0.1mV target response would get buried in the 20V common mode. Using a deadband amplifier instead can reduce the common mode to single figures around the peak. Depth can be vastly improved.

Regarding changes in frequency, a 16 bit counter driven by a 16MHz clock (eg. Arduino Uno) can capture a 1Hz shift of a 80KHz oscillator in only 5ms. Sub-Hertz resolution also means improving the detection depth.

Both effects depend linearly on distance (depth). The ratio of the increments dF/dW would be constant for most of the detector's range and depend on the particular metal.

These technology upgrades could potentially boost a BFO's depth and discrimination capabilities to the point of making it a quasi-alternative to VLF with not so huge a loss in performance.

[Gwil]

I hope you will assemble some hardware and try out your ideas. The results should be interesting (please post them here!).
One comment - you may find that 80kHz is too high to give useful discrimination.

[Teleno]

I hope you will assemble some hardware and try out your ideas. The results should be interesting (please post them here!).
One comment - you may find that 80kHz is too high to give useful discrimination.

Soon. I'm working on the PCB design, the HW is 100% figured out.

[Funfinder]

Weak Tx power is of no real issues with detectors. Distinguishing Target from Background noise is the Main issue. More Power Produces More Noise. Sorta the same as the Diminished Returns of adding power to a PI detector. It takes a offal lot of power to realize a tiny bit of gain in target signal.

Once I loved PI detectors because of their great voltage (EM-field) power!
BUT: That power gets used often just berserk-style with no real advantage!

What we have to learn is that low power is absolutely not good for metal-detectors!
They are easy to built but ridiculous weak! Add a 1m coil to a pinpointer and be happy! No, surely not!

Meanwhile we have high stable electronic-parts for higher voltages
and this will be the future! There is no more excuse using weak 5 volts detectors!

We have high density LiPo accus and we can use 50 volts together with relative small coils
for maximum depth! Or does someone uses 1 volt of power with a 5 meter diameter coil?

Nexus goes into the right direction by using higher power.
We need some high-voltage-engineers even if just 50v is pretty low!
Later we will use 100 or 200v batteries to crack up the whole treasure-area 5-10 meters deep!

We have high stable electronic parts meanwhile and there is no excuse not using them
for reaching much better penetration depth!

The contrary would be using some weak pinpointer-circuit with 3 meter coil!
Or why don't we use the battery of a wrist-clock for 1m depth detection???
This way you can go treasure-hunting just using your watch!

[homefire]

Maybe you should try some Tesla style Coils and SUCK the targets out of the ground ? Do your Realize there are Laws and Regulations in Regards to Transmitted Power Levels on most if not all usable frequency allocations ?

[Carl-NC]

Existing BFO designs that measure changes in amplitude use crude rectifiers. In the Falcon MD-20, for example, a 0.1mV target response would get buried in the 20V common mode. Using a deadband amplifier instead can reduce the common mode to single figures around the peak. Depth can be vastly improved.

Regarding changes in frequency, a 16 bit counter driven by a 16MHz clock (eg. Arduino Uno) can capture a 1Hz shift of a 80KHz oscillator in only 5ms. Sub-Hertz resolution also means improving the detection depth.

Both effects depend linearly on distance (depth). The ratio of the increments dF/dW would be constant for most of the detector's range and depend on the particular metal.

These technology upgrades could potentially boost a BFO's depth and discrimination capabilities to the point of making it a quasi-alternative to VLF with not so huge a loss in performance.

Energy-theft designs like Garrett's Propointer (and, I assume, the Falcon) aren't really BFO, but they have similar limitations. One is that ferrite has a strong opposite effect of eddy targets and de-biases sensitivity. With a BFO, a little variation in coil height can often give strong responses.

Try it and see what you get. I'm a Big Fan of trying out ideas, sometimes the unexpected comes out of it.

[Dave J.]

Two big problems.

1. Frequency shift depends on the reactive component of the target eddy current. Sorry, the reactive component is the one that iron minerals in the ground screw up.

2. Eddy current loss (the resistive component) is a tiny fraction of the transmitter circuit loss, which is usually mostly copper losses. Those have a tempco of 0.39%/C. Hope the searchcoil isn't being used outdoors.

Back in the bad old days of BFO's, a lot of ingenuity was applied in attempts to get more depth. It managed to coexist with high frequency induction balance ("TR") technology because it suffered from similar problems although for somewhat different reasons. At least a good TR might hit pretty good in an air test thanks to induction balance, but in the ground....... different story.

Ground balanced VLF induction balance killed both those technologies quickly and thoroughly. This is not to say they don't have uses, but they're never coming back as a mainstream technology.

[Teleno]

Thanks to you guys for pointing out potential pitfalls and encouraging experimentation.

Dave J. I'm aware of the effects you point out and measures are in place to counter them. I have found that while ferrite changes the baseline frequency, the shift of a metal target relative to the baseline is the same with and without ferrite. See my experiments in this post: http://www.geotech1.com/forums/showt...ctance-Sensors

Temperature changes are slow compared to those caused by targets, so the usual motion detecting approach of using a slow integrator as the baseline would work here too.

You might be right that BFO technology is not competitive. Anyway, the old days used old solutions, so I thought I'd give it a fresh try with new techniques. I've tried my ideas in the lab and the results were promising. PCB's are expected to arrive in 10-15 days ( http://dirtypcbs.com ). I'm ooking forward for the moment of truth!

[Skippy]

Teleno: Have you studied the ETI561 project?

http://www.geotech1.com/cgi-bin/page...i561/index.dat

That seems to be one of the most advanced BFO variants I've seen published. In addition to being technically interesting, it may be a good reference point for establishing what a BFO can achieve, when comparing it to your designs.