Beachmaster Digital

[waltr]

That sounds like a good plan for the batteries.
Yes, one of the battery sets will discharge before the other but since you have monitoring of each the lowest one sets the time to stop detecting.
A Balance charger is then needed to properly put the juice back.

thanks for the mpc33131 date sheet. Should do the job if the PI_pico can do the SPI fast enough.

I have played with using a PIC32 with a fast ADC to sample the coil decay waveform. Found I needed to sum multiple decay cycles to beat the noise down. This worked well since the Coil pulsing ran at 1.6kHz but only needed to update to the audio/display at 30Hz so lots of time to get multiply samples and do the math.

When you do the PCB be careful of Currents flowing in the ground plane. A single solid plane is not the best way to go. Ensure that the current in the coil switching does not go into the analog circuits. Keep these two plane areas separate then tied together at ONE POINT (near the battery.

[Olly]

I can't understand + 8V, is it from the second section 18650? In this case, it will cause the battery's imbalance.

The intended power supply for the detector is via 5 x 18650 LiIon Cells.
3 cells in series to give a nominal -12V and 2 cells in series to give a nominal +8V.

I'm doing it this way rather than the traditional approach of using a charge pump for two reasons: -
a) The current requirements of the Raspberry Pi Pico and LCD combination approach the max possible from a charge pump.
b) We eliminate the switching noise from the charge pump.

It is possible that current drain through the +8V and -12V segments won't balance and one of them will get flat before the other, which is why I have provided ADC monitoring of both levels so that the CPU can take the necesarry steps to warn the user and protect the batteries.

My intention is to have a separate battery pack mounted under the cuff with a 2-pole switch powering the detector.
I also intend to have a multi-way connector on the battery pack in order to balance charge the cells without having to remove them. (A multi-cell balance charger as used for model aircraft is ideal for this and I have a few kicking around).

Hope this helps,
Regards

[jeep6789]

What I face is that the MCU, HCO5, and the ADC chip use 3.3V power supply, and the op amp is ? 5V power supply, so how to reasonably arrange these levels, how to generate ? 5V

[jeep6789]

I want to divide the detector into the AB section, through the Bluetooth Control A section of the function, so convenient operation, can reduce the volume of the A portion, but I don't know how to design hardware, how to design reasonably, have a love person to help me, thank you very much

[jeep6789]

[jeep6789]

I can't understand + 8V, is it from the second section 18650? In this case, it will cause the battery's imbalance.

[Olly]

Yes, R18 I use in my HH2 and it decreases the coil TC On switching.

I see some possible issue in the schematic.
1- The TX gain op-amp, U9, is +5 to -5V output but the ADC is single ended, +5 to gnd.
Prototype this to ensure you can get the signal you need digitized by the ADC.
I could not find an MPC33131 chip with a google search. Do you have a data sheet link?

2- I recommend a resistor from base to gnd of Q1 to ensure Q1 stays OFF during PI boot-up.

3- Power supplies and grounds look like they should work. But not sure...
When I put a processor (PIC) on the HammerHead, The processor's Ground is the Battery's -12V rail. The processor's + is then 5V above the Bat's -12V. This made it easier to control and transistor switches and measure the analog Voltages.

4- U11 look ok. It is fully differential with the coil center taped to gnd . with an offset of 2.5V the RX signal should be in the +5 to gnd range.

Any reason you use a separate ADC on the TX coil's signal but the PI_pico's internal ADC on the RX coil?

1. You are correct, the ADC is single-ended. Diodes D2 and D3 ensure that there is unity gain for the half cycle that we don't need. In other words, the output from U9 will go up to 5V but only go down to -0.7V. This allows us to have the full resolution of the amplified flyback available at the ADC input. The ADC also has protection diodes and D6 has been included in case the built-in protection diodes have a lower forward voltage than 0.7V.

2. Good idea thanks !

3. I actually find it easier to refer everything to a proper GND and then have a solid plane for it on the PCB, the intention is for this to be a 4-layer board to keep noise to an absolute minimum.

4. Actually the GND reference for J4 is for the cable screen, the intention is not for the coil to be centre-tapped. The offset as you correctly surmise is to put the output signal in the correct range for the ADC.

I'm using a seperate ADC for the PI flyback as the intention is to try and analyse the slope, the Pico ADC is only 12-bit and has a maximum sample rate of 500kHz, whereas the MCP33131-10 is 16-bit and samples at 1MHz. The Pico ADC should be adequate to determine the phase and amplitude of the RX signal so the high performance ADC is not required here.

https://www.microchip.com/en-us/product/MCP33131-10

Thanks for the feedback,
Kind regards

[waltr]

Yes, R18 I use in my HH2 and it decreases the coil TC On switching.

I see some possible issue in the schematic.
1- The TX gain op-amp, U9, is +5 to -5V output but the ADC is single ended, +5 to gnd.
Prototype this to ensure you can get the signal you need digitized by the ADC.
I could not find an MPC33131 chip with a google search. Do you have a data sheet link?

2- I recommend a resistor from base to gnd of Q1 to ensure Q1 stays OFF during PI boot-up.

3- Power supplies and grounds look like they should work. But not sure...
When I put a processor (PIC) on the HammerHead, The processor's Ground is the Battery's -12V rail. The processor's + is then 5V above the Bat's -12V. This made it easier to control and transistor switches and measure the analog Voltages.

4- U11 look ok. It is fully differential with the coil center taped to gnd . with an offset of 2.5V the RX signal should be in the +5 to gnd range.

Any reason you use a separate ADC on the TX coil's signal but the PI_pico's internal ADC on the RX coil?

[jeep6789]

Thank you, I will ask the other end of the R25 to pick up the positive pulse or negative pulses.

[Olly]

Who can explain the role of R18 R30 D7, what role does it play on the entire circuit? Is it improving sensitivity, or stability, thank you

R18 in conjunction with the DC resistance of the TX coil limits the flat-top current through the coil to around 2.4A, the value of R18 can be altered depending on the flat-top current required and the DC resistance of the coil in use.

D7 effectively decouples the Mosfet Coss from the coil allowing a faster decay slope and hence an earlier sampling point.

R30 in conjunction with C27 reduces current transients and therefore noise on the -12V supply. Essentially, the full TX current is provided via C27, with R30 providing a charging path for C27.

Hope this helps...

[jeep6789]

Who can explain the role of R18 R30 D7, what role does it play on the entire circuit? Is it improving sensitivity, or stability, thank you

[MartyJ1963]

Thanks Graham, much appreciated! Looking forward to the layout and reading your next posts on your progress and tests. Regards, Marty.

[Olly]

Nice one Graham! Hope you're keeping well. The design sounds great! Regards, Marty.

Thanks Marty,

I think the schematic design is just about complete, about to start the PCB layout so will probably end up with some pin-swapping still but here is the updated schematic for now.

Beachmaster-Digital.pdf

Thinking of a compact 4-layer board with proper ground and split power planes in order to keep the noise down to a minimum.

Regards

[MartyJ1963]

Nice one Graham! Hope you're keeping well. The design sounds great! Regards, Marty.

[Olly]

It will be interesting to see a PI detector designed around the Raspberry Pi Pico.

It's got a lot of horsepower for ?3.60