Direct I/Q demodulation

[sled]

In case of a VLF detector, if I oversample the RX signal with a sampling frequency of exactly four times the transmit frequency I should get a nice I and Q signal, because the time between two samples would be 90 degrees apart, correct?

in this case I could only masure the relative change in amplitude and phase of the RX signal and not relative to the TX signal. would this still work?

[kt315]

time between samples is ~70plus degrees apart

[sled]

why 70 degrees?

[Carl-NC]

In case of a VLF detector, if I oversample the RX signal with a sampling frequency of exactly four times the transmit frequency I should get a nice I and Q signal, because the time between two samples would be 90 degrees apart, correct?

in this case I could only masure the relative change in amplitude and phase of the RX signal and not relative to the TX signal. would this still work?

Yes, this is exactly the method used in the White's Prizm 6T (aka CoinGT). If you design a decent transmitter with predictable characteristics, then you can software-calibrate the phase reference and even implement ground tracking, as was done on the 6T. You can sample at 90° as is the norm, or at a different angle as has been done by Tesoro. A drawback with direct quadrature sampling is that you loose the opportunity to amplify the resistive channel and boost the low-conductive response. The 6T is a little weak on nickels.

[sled]

thanks a lot for the insights, Carl! May I ask why I loose the ability to amplify the resistive channel? As far as I'm aware the direct quadrature sampling should deliver the same result as the traditional way, i.e mixing with a local reference oscillator (cos/sin) followed by a low pass filter.

[Carl-NC]

The R channel demod usually has a gain that's ~5x that of the X channel, then both are digitized. When you direct sample with an ADC, both channels have equal gain and therefore equal bits of resolution. Once digitized, you can't change that.

[Skippy]

Presumably if you have enough (or an excess of) bits in your converter, this becomes less of a problem. I recall ML's X-Terra machines use 24-bit direct sampling ADC's.
For info, the Teknetics T2 / Fisher F75 use a gain of 8.0 on the R channel before demodulation /conversion. I assume this power-of-two multiple helps with the mathematics later on.

[Carl-NC]

As I recall, the 6T has a 24b ADC. The problem is getting 24 bits out of it, more bits on paper don't equal more SNR in the circuit.

You always want to run a max signal into the ADC to get the best SNR; the X channel is usually maxed out, but at the same gain the R channel is not. So most analog demod designs gain up the R channel to improve SNR, regardless of how many bits the ADC has. Ferinstance, the F19 and Goldbug (Pro) use a 24b ADC, but still boost the R channel. Yes, the X-Terras are direct sampling with a 24b ADC. Not sure if they use quadrature sampling or oversampling. Oversampling will help, but doesn't solve the fact that the R channel is run on half its cylinders.

[sled]

Thanks for the info!

Let's assume I would use a classical approach with a numerical local reference oscillator. RX signal is A/D converted, and then I mix the RX signal digitally with the cos/sin reference with a weighting of 1:8, would this work?

My first attempt is to use the internal ADC/DAC of an STM32L4 microcontroller, it has only 12bits at 5MS/s however it supports a novel "hardware oversampling", i.e I can get almost 16bits out of it at ~83kS/s with no CPU overhead. The direct quadrature sampling saves me some RAM and clock cycles.

@Skippy Yep I know the X-Terra, they use a stereo audio codec (WM8731) which has some advantages, i.e coherent A/D and D/A sampling. Not sure how many of the 24bits are noise though, a mole's fart on the other side of the planet would probably be enough to wiggle the lower bits

[Carl-NC]

Sounds like you want to oversample, then digitally mix down to DC. I don't see that this has any advantage over quadrature sampling and binning the results directly to X and R, unless you intend to do the mix-down with a sinusoidal NCO to get bandpass filtering. Also, once you've digitized the signal, you've gotten all the resolution you're gonna get; digital gain doesn't mean anything. Again, oversampling can help because it can be used to improve SNR via digital filtering, which is what ST is doing with their ADC.

If you want a learning experience, start with the 12b ADC, sample at 4x the TX freq, and do a simple X & R binning. Get everything working with this (TX phase reference, GB, target ID, etc), this kick on the 16b option, see what difference it makes. Then consider oversampling and doing digital demodulation. You'll probably want a better ADC for that.

[scrungy_doolittle]

I am assuming that the R channel is actually the Y channel on the 5900's. When you direct sample with an ADC, you are sampling the signal AFTER the 5x gain, are you not? If so, then the channels do not have "equal gain". One has already been amplified by 5x outside of the ADC.

[scrungy_doolittle]

Yes, the STM32 part does have automatic hardware oversampling. I was going to use that on a STM32F411 or maybe it was F401, don't remember now. And use direct DMA to create a realtime buffer in memory. Additionally, I was going to use the STM32 to generate the frequency, and tag the sampled result with the start of every cycle of the frequency, by storing a 16 bit 0xFFFF in the 12 bit data stream.
I think the F4xx also had a 63ks 16 bit adc as well.

[ivconic]

Yes, the STM32 part does have automatic hardware oversampling. I was going to use that on a STM32F411 or maybe it was F401, don't remember now. And use direct DMA to create a realtime buffer in memory. Additionally, I was going to use the STM32 to generate the frequency, and tag the sampled result with the start of every cycle of the frequency, by storing a 16 bit 0xFFFF in the 12 bit data stream.
I think the F4xx also had a 63ks 16 bit adc as well.

Sorry for interrupting with off topic stuff here; but i just got DSO138, put it together and played with it for couple hours, trying to investigate a bit detailed the real power of STM32F103C8T6's ADC.

I was thinking to add STM32F103C8T6 to my "artillery" along with dozen Arduinos that i already have.
There are pretty affordable STM32F103C8T6 ARM STM32 modules on market:

But there is one question i would like to ask you, before i order it.
I already installed Code Sourcery G++ Lite toolchain and tested few simple steps with it.
But it is still not that comfortable to work with as for example with Arduino IDE is.
I saw there is a way to use Arduino IDE to flash STM32F103C8T6 board through UART.
Not really a problem, i am already using UART to flash Mini Pro's.
But i was wanting to ask if there is any dedicated software similar to Arduino IDE for more comfortable work with STM32F103C8T6?
Something more "user-friendly" than invoking a toolchain functions through command prompt?
Is there a such thing? And what are you using in your work?

[Altair]

Sorry for interrupting with off topic stuff here; but i just got DSO138, put it together and played with it for couple hours, trying to investigate a bit detailed the real power of STM32F103C8T6's ADC.
[ATTACH]38423[/ATTACH]
I was thinking to add STM32F103C8T6 to my "artillery" along with dozen Arduinos that i already have.
There are pretty affordable STM32F103C8T6 ARM STM32 modules on market:

But there is one question i would like to ask you, before i order it.
I already installed Code Sourcery G++ Lite toolchain and tested few simple steps with it.
But it is still not that comfortable to work with as for example with Arduino IDE is.
I saw there is a way to use Arduino IDE to flash STM32F103C8T6 board through UART.
Not really a problem, i am already using UART to flash Mini Pro's.
But i was wanting to ask if there is any dedicated software similar to Arduino IDE for more comfortable work with STM32F103C8T6?
Something more "user-friendly" than invoking a toolchain functions through command prompt?
Is there a such thing? And what are you using in your work?

I have buy the same STM32F103C8T6 ARM STM32 module you asking for; the module is arrived but I am not able to play with it due to the serial connection: on this module the usb port is suitable only for power supply and NOT for serial communication. I am waiting for a serial converter 232 -> 3V level to complete the kit.

On this address: http://grauonline.de/wordpress/?page_id=1004
are there some info in order to use Arduino IDE on it.

[ivconic]

I have buy the same STM32F103C8T6 ARM STM32 module you asking for; the module is arrived but I am not able to play with it due to the serial connection: on this module the usb port is suitable only for power supply and NOT for serial communication. I am waiting for a serial converter 232 -> 3V level to complete the kit.

On this address: http://grauonline.de/wordpress/?page_id=1004
are there some info in order to use Arduino IDE on it.

Yes, uart dongle:


I am having few, using them.
I saw that page the other night.
But i was hoping that there is dedicated IDE software already...