Looking for Pulse Induction Ideas

[fixstuff]

Hi folks,

I am thinking of a new project. It will be a pulse induction system. I want to build something based around pulse induction, Raspberry Pi 3 and a solar backpack.

I am going to build the heaviest portions of the unit into a solar backpack, to handle the battery and main PC board and electronics. I found a backpack with a 12000 mah battery with solar array.

As of now, I'm looking at a distributed wireless type system, using ESP8266 node with Teensy 3.2, and the Pi 3 with Codesys 3.5 sp8 as the main computer.

i.e.
wireless coil with ESP8266 and Teensy or XMOS
wireless wrist interface with ESP8266 and Teensy
wireless headphones with ESP8266 and Teensy

I could go with a more powerful SBC, but I don't need it. I need the RTOS and networking of the Codesys more. Pi3 uses wifi set up as AP for ESP nodes.
I could go with more powerful microcontrollers than the Teensy, but I'm using multiples, so the load on each is less.
All wireless is using 2.4ghz Wifi.
Interface communications between Raspberry Pi 3 and ESP 8266 is Modbus TCP master/slave.
Interface communications between ESP8266 and Teensy is I2C.

RPi3 uses a Adafruit Ultimate GPS hat, and a Teensy Hat for position and datalogging.

I want to really go all out on trying some different Pulse induction tech, and I want some ideas if anyone has any.

I'm thinking of tech like:
bipolar Transmit coil
24bit ADC sampling
Separate Transmit/Receive coils
multi-frequency (any advantage?)
would like to try ESP32 in Mesh if possible

Any thoughts, or contributions are welcome.

[ODM]

PI power draw is mostly in the search head transmit pulse, and after that analog processing, so a power cable would be needed from that backpack to the search head. If you want to keep them separate, you need to figure out power transfer. Easily swapped cells from the backpack? Supercap and momentary charge wire connections? A lightweight coiled cable would be relatively foolproof for power and data, in comparison to wireless power and data transfer. (We're looking at something around a few watts continuous power for a simple PI.)

With 24bit processing you might next want to look at speed. If your converter isn't doing MSa/s you could get around by sample-and-holding at your points of interest on the decay signal and then processing at leisure before next transmit pulse.

Bipolar gets you around earthfield sampling and some oddball remanence effects etc. and it's often a requirement for demining use to have a null net TX field to avoid (m)any magnetic field triggers.

There are some advantages to variable transmit pulse timing regarding difficult soil, and pulse repeat rate regarding external disturbances. The patent database and some manufacturers articles will be a good starting point for reading up on the topic - but read with a grain of salt

[Carl-NC]

I doubt a wireless coil is very practical with PI. You'll need to put the entire TX and most of the RX in the coil, and unless you develop a super-low-power TX circuit you'll either have a very heavy coil or a very short run time.

One option to consider is half-sine TX. It's more efficient and offers full target ID, drawback is the PI response isn't as good. I think Barringer had a couple of patents on it from the 1960's.

[Skippy]

I would suggest having a 'regular' size battery pack in the detector (under the elbow cup, or even further back on a 25cm boom for balance), and spare pack(s) in the backpack, on charge from the solar panel. By 'regular' size, I was thinking 3 or 4 18650 Li-Ion cells, about 3400 mAh capacity for the best ones (Panasonic, + others)
How big do you anticipate the panel being? I was thinking 20cm x 20cm would be a starting point.

[fixstuff]

I like the half-sine TX suggestion. It's something different, that I can tinker around with, and might end up useful.

What I really like is the possibility of using this on the rover - I don't have power limitations on that, so a big beefy cable would be fine.

[Tim]

I would suggest having a 'regular' size battery pack in the detector (under the elbow cup, or even further back on a 25cm boom for balance), and spare pack(s) in the backpack, on charge from the solar panel. By 'regular' size, I was thinking 3 or 4 18650 Li-Ion cells, about 3400 mAh capacity for the best ones (Panasonic, + others)
How big do you anticipate the panel being? I was thinking 20cm x 20cm would be a starting point.

This is my power option on all my homebuilds now. 3x18650s, with the $3 charge controller (alongside battery holder) and a $6 1amp solar panel. The bracket is a phone holder for a bike $6 You can also get a 3amp panel not much bigger.You can charge from any external 12v adaptor if you ever used it in the dark for hours.
Also now I am making my brackets and armrests from 4inch pvc shaped with a heat gun. The extra bracket is for the hipstick/bungy, for the big coil. The last coil I made, 6 inch, 3 rounds of Cat5 twisted. 380uH, (on Newhobby NH-5-last picture) can see 1.8g of flour gold at 4 inches on the SurfPro.
If not for the broomstick lower shaft, it would be super-light.
9v LM386 speaker/amp box-just visible in last picture.

[fixstuff]

I like the idea of using Cat5 for a coil. How does it compare to a normal coil build?

[Tim]

Super easy, and dare to say as good as any basket weave or similar design. Amazingly with all my trawling here I have only found one article on it, imported from another site. Crazy-tech(.com) uses it for his 1Msq. coils, but says nothing of the twisted smaller designs either. I found by end joining all strands, then measuring inductance and tapping at the suitable join, any remaining rounds seem to act as a shield(?). Alternatively one or more remaining strands can be disconnected to further vary inductance if needed. Also I would presume that doubling the number of winds and only joining the positive/negative wires would further decrease self-capacitance (??)
At $9AUD for 15M it's good value for experimenting too.. Perhaps it would be suitable for non-PI's with centre-tapped coils etc...?

[KingJL]

One option to consider is half-sine TX. It's more efficient and offers full target ID, drawback is the PI response isn't as good.

Carl, are there any reference designs for a half sine bipolar TX that you are aware of for study?

[Carl-NC]

A few times I've mentioned that the Fisher Impulse TX circuit is a good candidate for people to study and play with. It will support bipolar half-sine with nothing more than a judicious choice of capacitors.

[ODM]

With a class-G styled coil switch and a capacitor snubber to power it, it makes for a fast rising coil current enabling short pulses with high excitation. Granted, works unipolar as well, but bipolar gets rid of earthfield sampling.

[kt315]

I like the idea of using Cat5 for a coil. How does it compare to a normal coil build?

try
- basket coil design
- litz wire

https://en.wikipedia.org/wiki/Skin_effect

[ODM]

Snubbing the coil current with a relatively small, high voltage capacitor to give a high voltage "start" would give benefits relative to constant-current switch-on. The bipolar implementation is more elegant in doing with just diodes. Notice that the switch BJTs can be driven by a current sink/source at battery voltage level. Energy efficiency is not as sweet as Impulse. That is one circuit that could do with more homebuilding and basis for experimenting.

Playing with capacitor sizes and pulse lengths gives optimized responses to different target TC, and capacitors are something relatively easily switched "in and out" of the circuit. I see at least one PI patent that deals with constant-current driving a coil, with familiar names under inventors... Class G switching driver doesn't give equal benefits, but I don't see any current patent on it, or haven't had luck searching. However I don't believe the circuit is novel in any manner. It's also missing voltage shunt regulation or clever sampling for consistency.

[KingJL]

half-sine anybody?

One option to consider is half-sine TX. It's more efficient and offers full target ID, drawback is the PI response isn't as good.

Carl, are there any reference designs for a half sine bipolar TX that you are aware of for study?

A few times I've mentioned that the Fisher Impulse TX circuit is a good candidate for people to study and play with. It will support bipolar half-sine with nothing more than a judicious choice of capacitors.

WOW... totally by a fat fingered mistake!
bi-polar_sine(1).pdf
bi-polar_sine(2).pdf
I was playing around with a simulation based on a variation of Dave Johnson's latest patent design, when I fat fingered a value change and boom... this appeared on the subsequent simulation run! I will not be posting a schematic. However, I will post a link that will lead to public information about Dave's design.
http://www.geotech1.com/forums/showt...nson-PI-Patent

I will tell you that the coil I was using, was 1000uH. I have been experimenting with larger coils as a result of Eric Foster's discoveries/discussions about the Vallon mine detector, that stands on it's head the notion that large slow coils are not suitable for detecting small low conductivity targets. With innovative TX designs and DSP RX processing, almost anything is possible.