Arduino Nano PI Main Discussion

[green]

Hi Green,
I have allot of questions about your front end but best done In the proper thread.
To Hit A little ? Switchable Unipolar Bipolar? Dual Channel?
Delay time to first sample 3.5us? (I have several coils that damp to 3us but never been able to sample that low.)
Gb type ? Dual Integrator?


I like where your going with your detector project is there a thread I've missed please point me that way so I can catch up?
Trying to recreate your coil is still on my list as well. went coil crazy for a bit had to take a break.

I don't know if micro or at least arduino would be capable of 3.5 us first sample doing all the load but wondering if maybe it was just used for all the sampling and audio.

On your ground balance sample Im thinking the Gb sample should be adjustable aside the main sample so you can play where you want the hole. That's where I left off on my Mpp Gb Board.
The way the TDI Gb is Set up really is a great design to take a look at.

I can sample my 133mm fig8 with 3.5 us delay unipolar, 4us bipolar. Maybe 3us would be to low even if I could. Lower the delay, lower the GB hole. Not sure how small a nugget I should look for. I never learned programing micro's and didn't know what was needed to make a good PI detector when I started. So I hard wired logic chips instead of making a PCB. Allows making changes. I would suggest not everyone try it. I started a thread https://www.geotech1.com/forums/show...hp?22777-PI-MD few years back. What I've learned since then is scattered thru other threads. Still think defining what is desired should be first, then decide how best to do it.

Changing target delay and target sample changes the hole. Not sure a 100us ground sample is best but so far I'm not seeing a better option.

[godigit1]

One problem we all face around the world when out in the field detecting is mains electricity interference which is either 50 or 60 Hz,
if we make the PI Tx rate 1200 pps this number is divisible by both 50/60 Hz .

By both using 1200 pps and having the ability to change the Tx rate + or - slightly in code can help tune out and reduce the mains interference .
It usually manifests in a PI detector as a "beat" frequency in the receive audio by giving a constant rise and fall in detector threshold audio
the surf pi is prone to it in a 50 Hz country.
Just something to consider when writing code.

Thanks 6666,
I was wondering if the Arduino platform might be more susceptible to outside interference rather than just using a atmega chip?
Changing the Tx was one of the ideas behind the SD card boot. If I started having problems somewhere in the field I cold try another frequency.

[godigit1]

What I'm doing today, maybe different tomorrow.
Tx time, 160us
Tx rate, 1000pps unipolar, 2000pps bipolar
target delay time, adjustable front panel(3.5us to 20us)
target sample time, adjustable front panel(5us to 20us)also used to adjust GB, needs fine adjust
GEB, front panel on-off
ground delay time, 4us haven't played much with different times
ground sample time, 100us
Open to suggestions to make it better. Never learned micro's so can't help with programing. Still need to define control loop whether it's done with logic or micro.

unipolar needs an EF sample also

Hi Green,
I have allot of questions about your front end but best done In the proper thread.
To Hit A little ? Switchable Unipolar Bipolar? Dual Channel?
Delay time to first sample 3.5us? (I have several coils that damp to 3us but never been able to sample that low.)
Gb type ? Dual Integrator?


I like where your going with your detector project is there a thread I've missed please point me that way so I can catch up?
Trying to recreate your coil is still on my list as well. went coil crazy for a bit had to take a break.

I don't know if micro or at least arduino would be capable of 3.5 us first sample doing all the load but wondering if maybe it was just used for all the sampling and audio.

On your ground balance sample Im thinking the Gb sample should be adjustable aside the main sample so you can play where you want the hole. That's where I left off on my Mpp Gb Board.
The way the TDI Gb is Set up really is a great design to take a look at.

[Carl-NC]

One problem we all face around the world when out in the field detecting is mains electricity interference which is either 50 or 60 Hz,
if we make the PI Tx rate 1200 pps this number is divisible by both 50/60 Hz .

Ideally you would want to exactly lock to a multiple of the mains frequency, as you suggest. This is usually not practical because the mains frequency isn't exact, it varies +/- perhaps 1/2 percent or so. Over the course of an hour or so it tends to average very accurately to 50/60Hz, hence AC clocks keep pretty good time. The only time I try to exactly lock to the mains is in a walk-through design which is powered by AC, so I have access to the exact frequency and can PLL-multiply it to the pulse rate.

In a handheld detector the alternative is to make sure the mains frequency aliases outside the target filter bandwidth. This is usually set by the demod and the Nano has a BW = 1.5Hz (which is pretty low). Assuming 50Hz mains, if you set the detector to 1200Hz then the alias is synchronous and becomes a DC offset. But if the mains drifts by 0.1% then the alias is at 1201.2Hz and you hear a 1.2Hz beat frequency. However, if you set the detector to 1225 Hz then the 50Hz alias is at 25Hz which is suppressed by the target filter BW. And any minor mains variation remains suppressed. BTW, the demod is often followed by additional filtering that roughly matches the demod BW to offer better mains suppression. In the Nano, this would be R26-C15... setting them to 1.5Hz will help (they are currently 16Hz).

[godigit1]

There would probably be even more delay introduced by using the millis() function.

If I wanted to create a shorter main sample delay, I would combine state 1 and 2 of the interrupt state machine and create the delay by invoking the NOP command a few times via an assembler call.
For example: "__asm__ __volatile__ ("nop\n\t");".



I think that this single instruction will introduce a delay of 62.5ns, but I would have to test it to be certain. Try adding a for loop around the asm call in your sketch and measure the delay on the scope.

Thanks Q
I'm Going down a long road here maybe bitten off more than I can chew but gonna give it a go.
Still waiting on parts but got the Nanno hooked up by itself with the delay pot hooked up.
A bit more studying and I can start playing with some settings before parts get here.

I'll start looking into the Asm commands, that's the great thing about Arduino there are tones of Sketches and tutorials to look at.

[waltr]

You can easily add an external pot to the Arduino Nano PI to allow adjustment of the TX repetition rate, in the same way as the Delay pot is implemented.

Yep, PIC or Arduino can do this.

Tried a AC to DC wall adapter power supply. Could see alias frequency. Difference between sample frequency and mains harmonic. Haven't had a front panel trim frequency adjust, probably should add one.

Good you can now see the beat and can experiment with nulling.

[green]

I have seen (heard) this when trying to detect near some high Voltage transmission lines also a few times in my basement lab when the heating system's pump runs.

To cause it put a AC mains device near the coil. Something inductive.

My HH2 runs at 1600Hz and has a fine frequency adjustment pot the 'null' the AC mains beat. Pot into ADC on PIC processor which adds/subtracts a small delay on the cycle timer. Due to the PIC running its internal RC clock the Frequency drifts with temperature enough that in my lab verse the field the cycle time changes. A XTAL would be better but will still have a temperature frequency change. So slight adjustment may be needed in the field (like on the TDI detector).

Tried a AC to DC wall adapter power supply. Could see alias frequency. Difference between sample frequency and mains harmonic. Haven't had a front panel trim frequency adjust, probably should add one.

[surfdetector]

The Amazon nano's arrived, tried one, it works with the IDE, more later.

Good to hear. Glad that they are working for you.

[Qiaozhi]

My HH2 runs at 1600Hz and has a fine frequency adjustment pot the 'null' the AC mains beat. Pot into ADC on PIC processor which adds/subtracts a small delay on the cycle timer.

You can easily add an external pot to the Arduino Nano PI to allow adjustment of the TX repetition rate, in the same way as the Delay pot is implemented.

[waltr]

I'm guessing the problem is aliasing on a harmonic of the mains. An exact 1200 Hz should work. 1201 Hz sample should give a 1 Hz signal, maybe worse than sampling at 1000 Hz. What am I missing? I haven't seen the "beat" frequency. Any suggestions what I might try to cause it?

I have seen (heard) this when trying to detect near some high Voltage transmission lines also a few times in my basement lab when the heating system's pump runs.

To cause it put a AC mains device near the coil. Something inductive.

My HH2 runs at 1600Hz and has a fine frequency adjustment pot the 'null' the AC mains beat. Pot into ADC on PIC processor which adds/subtracts a small delay on the cycle timer. Due to the PIC running its internal RC clock the Frequency drifts with temperature enough that in my lab verse the field the cycle time changes. A XTAL would be better but will still have a temperature frequency change. So slight adjustment may be needed in the field (like on the TDI detector).

[green]

One problem we all face around the world when out in the field detecting is mains electricity interference which is either 50 or 60 Hz,
if we make the PI Tx rate 1200 pps this number is divisible by both 50/60 Hz .

By both using 1200 pps and having the ability to change the Tx rate + or - slightly in code can help tune out and reduce the mains interference .
It usually manifests in a PI detector as a "beat" frequency in the receive audio by giving a constant rise and fall in detector threshold audio
the surf pi is prone to it in a 50 Hz country.
Just something to consider when writing code.

I'm guessing the problem is aliasing on a harmonic of the mains. An exact 1200 Hz should work. 1201 Hz sample should give a 1 Hz signal, maybe worse than sampling at 1000 Hz. What am I missing? I haven't seen the "beat" frequency. Any suggestions what I might try to cause it?

[6666]

Got to wait a few weeks for mailing from USA, but fingers x'd.

The Amazon nano's arrived, tried one, it works with the IDE, more later.

[gbrowne]

Thanks for the background, sounds like par for the course eh I've just orderd a Pi zero and a fast AD board. I'll capture some curves from various targets, ground conditions etc and see what can be gleaned. If a Pi-Pico at 133Mhz can sample at up to 500k samples/sec and also fo FFT in real-time, who knows what maths magic can happen with a 1Ghz clocked Zero ? MInd you, my coding's ok, but my maths is pants! Probably just load it all into calc and see what it looks like first. G.

[bernte_one]

Hi Carl,

can you place the complete partlist here as pdf or txt document, would be nice to have this extra

i dont want to add the hooks on the book pages after installing components (yes i do checklist for each detector while assembling)

[6666]

One problem we all face around the world when out in the field detecting is mains electricity interference which is either 50 or 60 Hz,
if we make the PI Tx rate 1200 pps this number is divisible by both 50/60 Hz .

By both using 1200 pps and having the ability to change the Tx rate + or - slightly in code can help tune out and reduce the mains interference .
It usually manifests in a PI detector as a "beat" frequency in the receive audio by giving a constant rise and fall in detector threshold audio
the surf pi is prone to it in a 50 Hz country.
Just something to consider when writing code.