AMX Digital

[Willy Bayot]

Getting a sample rate of 1us may require a dedicated micro just for data accumulation. I think 10us is feasible, and maybe faster but I don't know how fast. I'm sure it will depend on the clock speed of the micro and how much needs to be done in the processing loop. Probably just build it and see. I do think you want the ADC samples synched with the TX timing, even at 1us. That should be quite easy.

Yes, the first phase of capture is made using DMA and feeding the FIFO, it is continous without any synch with the pulses.
The second capture phase (made in parallel by the main program loop) is made by synchronizing the reading in the FIFO with the start of each period and accumulating the FIFO values into the slots of another array. The end result is an array made of as many slots as necessary to cover the whole pulse period. Example : period of 200µsec captured @ 1Msps generates 200 slots.
Each slot is the integration of a number of consecutive periods (e.g. integration of 20 consecutive periods gives one net series of accumulated slots every 4msec).
Next, the real windowing is defined by selecting the starting slot and width of each digital integration window and they are accumulated in as many virtual demodulators.
Finally, the demodulators are used for the DSP and reporting.

[Olly]

Yes, the first phase of capture is made using DMA and feeding the FIFO, it is continous without any synch with the pulses.
The second capture phase (made in parallel by the main program loop) is made by synchronizing the reading in the FIFO with the start of each period and accumulating the FIFO values into the slots of another array. The end result is an array made of as many slots as necessary to cover the whole pulse period. Example : period of 200µsec captured @ 1Msps generates 200 slots.
Each slot is the integration of a number of consecutive periods (e.g. integration of 20 consecutive periods gives one net series of accumulated slots every 4msec).
Next, the real windowing is defined by selecting the starting slot and width of each digital integration window and they are accumulated in as many virtual demodulators.
Finally, the demodulators are used for the DSP and reporting.

Because I have one laying around I intend to use one of the CMOD-A7 FPGA modules to do my experimental work with. To this end I've started writing some Verilog code to drive the ADC and capture the 1uSec samples into Fifos. I have a question regarding the accumulation of data for the 4mSec processing loop. Do you envisage this merely being the accumulation of each slot over the 4mSec period which is then reset at the start of every period or would a rolling average of each slot be a more desirable solution?

Regards,
Olly

[Olly]

Ivica, you're asking all the right questions. I'm a Big Fan of illustrations so let's dive in (again, all this belongs in the Concept discussion, but we'll go with the flow). Here is a multi-pulse TX using 2 rates:


I used 25kHz because Tony suggested it, and 5kHz because it gives me a familiar 100us TX width. If you transmit only 25kHz then you are stuck with trying to squeeze a GB sample in that 20us window which is tough. Plus, your GB has a dreaded target hole. In this example I used (5) pulses of 25k per (1) pulse of 5k; in reality we might instead use 4:1 or 8:1 just to make scaling in DSP a little easier. You can also change the ratio to implement variable frequency weighting.

Let's suppose we use analog demods. Here is my first guess as to what to try:



Boy, I love being able to just <paste> these images! So I have 4 RX channels: (1) for 25kHz and (3) for 5kHz. Maybe we could squeeze in another sample for the 25kHz pulse but I doubt it would do any good. GB is achieved with a combination of RX25, RX5a, and probably RX5b. Discrimination might be achievable with RX5a,RX5b,and RX5c.

The clocks above are demod clocks. The demods are integrating the results so the ADC does not need to run this fast. You can sample the demods every 400us (2.5kHz) or even slower if you like. Let's say we sample at 2.5kHz. The ADC creates a DataReady pulse which triggers a micro interrupt. The micro reads the ADC and DMA's the result. So far this takes nearly zero overhead. But you don't want to run a processing loop at 400us unless you are using a really fast micro. So instead, you pile up the samples (either in a circular buffer or by just summing the data) and say, every 5ms, you send all the data to the DSP task. (Most metal detectors run a processing loop of 5ms.) A processing loop of 5ms on a micro running at 32MHz gives you 160,000 single-cycle instructions to work with.

So now you have a hard example of an analog demod solution: 4 demod channels, a 16-18 bit 4-channel simultaneous sampling ADC running at 2.5kHz, and not-too-aggressive micro.

Let's say we want to use direct sampling on the above TX waveform. I will take a guess and say I want 4 samples for every 20us pulse. This now sets my ADC sample rate at 200kHz, and it needs to be 24 bits. You will be hard-pressed to grab data that fast in the same micro as the one doing the DSP. So maybe you need a data capture micro that simply sums up the data and eventually feeds it to the DSP processor at the 5ms loop rate. Or maybe you need a really fast M7 micro. Or maybe I only get data at 100kHz or, worse, 50kHz. Note that you probably don't need 20 samples per 5kHz pulse, but it's way easier to run the ADC at a constant rate and just toss unwanted samples.

Before I hit the Post button, notice that the TX waveform has a positive 100us pulse every 400us. The TDI/Goldscan has a 100us pulse every 320us. So just the response from the single 100us positive pulse should be somewhat comparable to the TDI. But we also get a second (negative) 100us pulse, and 10 (!) 20us pulses in the same time frame. So this thing should hands-down beat a TDI on small gold.

Hi Carl

I've been re-visiting this post to get a better understanding of the various virtual channels proposed and how to go about realising them with direct sampling.

I'm happy with the locations of the RX25 & RX5a + and - slots, but am not so sure about the RX5b & c slots, let me try explain: -

X5b+ and RX5c+ occur in a period where the previous positive going transition of the TX clock occurred after a 20uSec active period, whereas RX5b- and RX5c- occur in a period where the previous negative going transition occurred after a 100uSec active period. Is it possible in this case that the + and - samples may not be equal and opposite to each other (barring EFE effects) ?

Regards,
Olly

[Carl-NC]

The 5kHz pulse serves 2 purposes. One is for better detection of medium-high conductors and in that respect you are right, RX5B+ and RX5b- maybe have different responses. Same with RX5c+/-. However, for target detection these are simply added so it's not a big deal. The other purpose is for GB and this is where balanced responses are more important. In the case of viscous responses I think whether you sample after a previous short or long pulse matters little. I may be wrong and this is something that needs verification.

[Olly]

The 5kHz pulse serves 2 purposes. One is for better detection of medium-high conductors and in that respect you are right, RX5B+ and RX5b- maybe have different responses. Same with RX5c+/-. However, for target detection these are simply added so it's not a big deal. The other purpose is for GB and this is where balanced responses are more important. In the case of viscous responses I think whether you sample after a previous short or long pulse matters little. I may be wrong and this is something that needs verification.

Thanks for explaining, makes more sense now.

Regards

[Olly]

Complementary TX signals generated with the CMOD-A7 FPGA board.
The dead band is 100nSec, but can be altered.

[Olly]

Hi all, in preparation for Carl's TX and analogue front end stage, I've been beavering away writing Verilog code for the little CMOD A7 FPGA module. I know that most of you will want to use some other processing module but for those interested this is what I've got so far: -

• Complementary TX pulse outputs with one 5kHz cycle followed by 5 x 25kHz cycles as per Carl's diagram.
• ADC Interface to the LTC2338-18 for capturing RX data at 1uSec intervals.
• Capturing of the + and - slots for RX5 a, b & c and RX25 as per Carl's diagram.
• Getting the difference between the + & - values for each slot and accumulating the 5 x RX25 slots.
• This gives us 16 x RX5a slots, 32 x RX5b slots, 48 x RX5c slots and 8 x RX25 slots. (Can be changed as required)
• Passing all of this data (104 slots) through rolling average (boxcar) filters/accumulators 16 deep (6.4mSec)
• Making this data available via a Dual Port Ram for processing every 400uSec.
  Note: The time taken to actually filter all 104 slots is less than 4uSec which leaves over 396uSec (39,600 clock cycles)
  for processing if such a fast processing loop is required - we will probably process much less frequently.

The development has been done using Xilinx Vivado tools and I can make the source code available if anyone is interested.

Regards

[ivconic]

[ivconic]

Finally arrived!
23 days from NY/USA to Sokobanja/Serbia.
Now I have real "stage fright"!
I have to study these seriously for several days, maybe weeks...
Before I start producing white smoke with them!

[ivconic]

Although I am quite skilled with the Adruino IDE and have done a lot with it so far; I feel like now is the time to add another development environment,
a little more "serious" to work with around the Nucleo dev board.
I previously programmed the STM32 Blue Pill with the Arduino IDE.
But these Nucleo 144 (l496zg) dev boards are far more serious and I would like to use them with a software environment that is adequate for them.
Half an hour of searching the net...
I found a lot and nothing is clear to me!
I found on one page a very vague instruction with suggestions to download 3 packages:
STM32Cube IDE, STM32Cube MX and STM32Cube Programmer.
Randomly when I read on those pages, about those packages; I don't understand anything, what is it for?
Do all 3 have to be installed? What is it for, etc.
The short descriptions of the packages are unclear to me, it turns out that you can do everything with each of them?
And I can't find any "for dummies" instructions with explanations about these issues.
Does anyone of you have the will to write a brief instruction on what and how to explain the exact role of those packages?
And in as brief as possible to explain setting up a development environment on a system like Windows.

[Willy Bayot]

Although I am quite skilled with the Adruino IDE and have done a lot with it so far; I feel like now is the time to add another development environment,
a little more "serious" to work with around the Nucleo dev board.
I previously programmed the STM32 Blue Pill with the Arduino IDE.
But these Nucleo 144 (l496zg) dev boards are far more serious and I would like to use them with a software environment that is adequate for them.
Half an hour of searching the net...
I found a lot and nothing is clear to me!
I found on one page a very vague instruction with suggestions to download 3 packages:
STM32Cube IDE, STM32Cube MX and STM32Cube Programmer.
Randomly when I read on those pages, about those packages; I don't understand anything, what is it for?
Do all 3 have to be installed? What is it for, etc.
The short descriptions of the packages are unclear to me, it turns out that you can do everything with each of them?
And I can't find any "for dummies" instructions with explanations about these issues.
Does anyone of you have the will to write a brief instruction on what and how to explain the exact role of those packages?
And in as brief as possible to explain setting up a development environment on a system like Windows.

The 'STM32Cube IDE' should be enough for a start.

[ivconic]

It can be used with Arduino IDE too.
I tried "Blink" and it uploads and runs alright.
Before that I just needed to install board definitions in IDE.
Next (before your post) I installed STM32CubeMX and now I just installed STM32CubeIde.
Will see...
Writting a code is one thing, but setting up the proper software environment is quite another thing.

[Carl-NC]

CubeMX is a design tool for selecting a micro and configuring the pins. I've only played with it a little, never really used it.
CubeProgrammer is a manual programming tool. I only use it for production-level programming, or when I need to un-brick a micro.
CubeIDE is the main tool. It's a GUI development system, similar to MPLAB-X if you are familiar with that. It's all you really need, plus the libraries for the specific processor.

[ivconic]

CubeMX is a design tool for selecting a micro and configuring the pins. I've only played with it a little, never really used it.
CubeProgrammer is a manual programming tool. I only use it for production-level programming, or when I need to un-brick a micro.
CubeIDE is the main tool. It's a GUI development system, similar to MPLAB-X if you are familiar with that. It's all you really need, plus the libraries for the specific processor.

I installed it after I saw Willy's comment.
And spent some time since then playing with it.
Ok, some of the things are intuitive and easy to understand.
But some are not and will need certain learning/adapting curve.
One thing is annoying me right now.
My boards are NUCLEO 144 - L496ZG.
So I started IDE, new project, picked right board and even marked it as "favorite" (I guess that's the puprose of a "star" next to a board name in the list).
But once I delete the project; I have to pick up the board all over again, from the list.
Is there a way to make it always to use the same board as "default" ... or I have to pick the board each time I start new project?
As things goes on, I see; I will have to spend much more time watching the videos and reading the documents.
Jumping from Atmega 328P to ARM-M4 is not simple and easy at all.
It looks scary to me now and is actually a quantum leap.

[ivconic]

Phew! All day search on the net!
The first step to easy learning is to find a good source.
In a lot of mediocre and even bad demos;finally a good course,at least for now it seems good to me.
We'll see how it goes...

https://youtu.be/hYYRfhhvTYo