Analyzing TX/RX Circuit

[sled]

Hi,

I have some questions about an TX/RX circuit (X-Terra MD), I'll start off with the TX circuit in this post:

The TX signal is generated by a DAC which has an integrated headphone amplifier (50mW @ 16 Ohm), the TX signal is also tapped and fed to an ADC to get a TX reference signal.

What I'm not sure about the purpose of the small inductors/chokes?

What I could think of is:

• L1/C1/L2 form a T-section LC filter to filter out high frequency noise generated by the DAC / reduce the "stairs" in the DAC output?
• L4 filters out high frequency noise picked up by the coil / coil wire?
• Why are there clamping diodes on the DAC output but none on the ADC input?

(I chose 0.47Ohms as series resistance for the small 15uH inductors. The coil has a series resistance of 13.9Ohms - sums up to ~16Ohms with R1)


Thanks!

[Davor]

Most probably they are here to make a device pass RFI examination. Unlike a search coil, these small chokes retain RF blocking capability waaaaay into VHF.

[sled]

thanks, so removing the chokes should not affect the performance of the detecting right?

I got some more questions about the RX side I'm trying to understand:

• What's the advantage/disadvantage between a parallel LC circuit versus a series resonant LC circuit on the RX side?
• What's the voltage range induced by targets in the RX coil?
• Is there a rule of thumb for the bandwidth of the RX amplifier? Most schemas I've seen use some sort of bandpass filter with gain.

Thanks!

[Silver Dollar]

Is this the circuit we are discussing?


I have some bare boards for sale I had made that you can do a similar circuit with;
http://www.geotech1.com/forums/showt...dec-Bare-PCB-s

I've been busy with other projects so haven't started developing it yet.

[Davor]

thanks, so removing the chokes should not affect the performance of the detecting right?

I got some more questions about the RX side I'm trying to understand:

• What's the advantage/disadvantage between a parallel LC circuit versus a series resonant LC circuit on the RX side?
• What's the voltage range induced by targets in the RX coil?
• Is there a rule of thumb for the bandwidth of the RX amplifier? Most schemas I've seen use some sort of bandpass filter with gain.

Thanks!

With exception of L1 that may prevent current spikes at voltage steps on the circuit input, the rest seem pretty much irrelevant for detecting.

There is no such thing as a parallel LC circuit in Rx front end. On this bombshell just think of a coil as a voltage source (induction) in series with a coil. A classic parallel LC tank would have a bandpass response, while this common detector front end is an L-configuration low pass filter. The advantage of this configuration is impedance up-transformation that results in improved S/N if done right. Downside is a screwed up phase response, so you can't use it at resonance where transformation works best because of ~90° phase shift accompanied with steep phase gradient. Because of that, the tank is tuned to ~1.2x Tx frequency, and is a commonly accepted compromise.

Series configuration has voltage source, coil and capacitor all in series, and has 0° at resonance, and also phase response reasonably blunt and depending on input impedance of a preamp. The only downside is that there is no S/N benefit as there is no up-transformation of impedance. Noise source of a coil is its resistance, say 50ohm, resulting in ~1nV/sqrt(Hz), compared with 4nV/sqrt(Hz) of a typical low noise op amp, which in case of series configuration is ~12dB loss. With a seriously low noise preamp at 1nV/sqrt(Hz) input noise, there is no downside of a series LC configuration.

Rx front end bandwidth is largely irrelevant for a target response, because filtering after a switching mixer reduces the resulting bandwidth to below 50Hz. Purpose of such bandpass filters is mainly to reduce EMI from cars and mains-powered devices. There is no useful information in Rx signal ± 30Hz from a Tx frequency but making extreme Q filters is even more problematic because of tuning and sharp gradient phase response.

[sled]

Thanks for your input @Davor

Here is what I came up with for the TX stage, it's an adapted version of the Kpot2-XM TX stage.

Do you have any advice how to improve the design further? I'm going to put it on a breadboard in the evening and do some basic tests.

The TX voltage swing on the tank circuit seems a bit low (~12Vpp) also the current is limited to about 70mA when powered from a 3.3V source, maybe I should go for a 5V source?

[Davor]

In this configuration you could only reduce R1 and have more current through a coil (I=U/R at resonance), which in turn would increase coil voltage (U=IωL). Rising voltage would increase current as well, same as reducing R1.
Otherwise it is a good idea to keep a Tx power supply firmly related to a low noise supply rail, because in VLF MD amplitude modulation passes through Rx by means of "air signal", and any AM hum will be detected as chatters.

[sled]

In this configuration you could only reduce R1 and have more current through a coil (I=U/R at resonance), which in turn would increase coil voltage (U=IωL). Rising voltage would increase current as well, same as reducing R1.
Otherwise it is a good idea to keep a Tx power supply firmly related to a low noise supply rail, because in VLF MD amplitude modulation passes through Rx by means of "air signal", and any AM hum will be detected as chatters.

The audio amplifier is only capable of driving 8+ Ohms, so R1 is already the smallest value possible (4.3 Ohm coil resistance + 3.9 Ohm R1 = 8.2 Ohm).

What is the industry standard/average used for the coil current in a typical VLF detector? Could you tell me some numbers based on experience?

I could increase the supply voltage to 5V but the ADC/DAC are only 3.3V tolerant and I'm not sure how to protect the ADC inputs from overvoltage if the pre-amp goes into saturation and outputs 5V accidentally, are clamping diodes the way to go? I've read that the clamping diodes may introduce noise due to leakage current...

Thanks

[Davor]

Clamping diodes ... no. If protecting an ADC is what you need, a simple resistor divider is the way to go. Most probably some micro behind the ADC is looking for a zero crossing, but it may also seek a symmetry point of a signal peak, and you'd ruin it with a clamp.
You seem to have reached the very end of this approach using TPA301.