Moodz' Awesome Gold Pulse Induction Version 3 - MAGPI V3 Project

[Gunghouk]

So should this be MAGPIV4 thread now?

[Detectorist#1]

Hi pito,
The aim of this construction is to be sensitive for small nugets. The Tx coils are 33uH and only 5tr. The Tx pulse is 12us and peek Tx current is 10A . As a result, the peek voltage on Tx coils is 1600V. The distance for 25c USA is appr. 30cm. This distance is not big - the reason is that the distance between two RX coils is only 12cm. Other advantage of this design is reliable Fe recognition - in "Fe" mode two Rx coils work as magnetometer and Tx pulses are not used.

[moodz]

So should this be MAGPIV4 thread now?

not yet I only made slight modifications to the V3 and code to achieve ferrite rejection. The iron rejection I may have to leave for V4 but i am testing now.

[Gunghouk]

Are the mods easily transferrable to the through hole version?
I've been holding off building the V3 waiting for a definitive list of changes to bring it up to a final/latest spec. Unfortunately a ready built, surface mount board is the only way I can get something newer in the future as I can't construct anything other than through hole with my myopic vision

[moodz]

Are the mods easily transferrable to the through hole version?
I've been holding off building the V3 waiting for a definitive list of changes to bring it up to a final/latest spec. Unfortunately a ready built, surface mount board is the only way I can get something newer in the future as I can't construct anything other than through hole with my myopic vision

You are in luck !
The the holes are already on the V3 thruhole board. The mod is just to remove some components and add some components to the original design. No actual mods ( track cutting etc ) to the PCB.

[Gunghouk]

Can't wait

[moodz]

I was a bit doubtful but I have increased the preamp gain from 100 to 1000 ( 40 db to 60 db ) and had to back off the sampling points.

The reason for this is to hit the 2CDCCD specification.
The 2CDCCD ( or 2 COIL DIAMETERS COKE CAN DETECTION ) test is a indication if your detector is actually working or maybe you should take up building props for science fiction movies.

Tested the MAGPI detector with a 40 cm ( 17 inch coil ) and it detects a coke can at 80 cm .. which is two coil diameters.

The GEB ground balance has been adjusted to move the detection "hole" to the far left so instead of creating a gap in detection range just the end of it has been clipped. Very small targets ( eg smaller that 0.1g ) have reduced sensitivity because of this.
The GEB is running all the time and if you rest the coil will balance out small targets BUT when you move the coil off the target .. will have to balance the ground again ( approx 5 seconds ). Its all done for you and theres no pumping of the coil etc.

Even with the large coil it easily detects 0.1 gram targets and does not respond to ferrites.
Noise has been reduced and there are actually less components now than in the original V3.

I will be out this weekend testing and update you then.

[moodz]

I am waiting on a parts order but the good news is ...
Thanks to a breakthrough in preamp design ... the magpi can now hit larger deep targets and small sub microsecond tau targets much better.

The changes to the V3 circuit are minimal. Few firmware changes also.

Here is a video of the MAGPI detecting a sub microsecond tau target and also a larger target at 1 meter in the lab.

https://youtu.be/2F6O45bi3pA?si=dkC4dogkrjrqWbWO

[Gunghouk]

Thanks for the update. JLC pcbs are on the way here also

I've programmed a Pic but can't get the Lcd to work is there something needs enabling in the software? It's V1. 4 version. I've a 3.3v LCD2004A version on order so I'm hoping it may be that my current display is 5v only.

Here in the UK there's a shortage of diggible gold so another question is can it be modded for artifact hunting?
Thanks in anticipation.

[moodz]

Thanks for the update. JLC pcbs are on the way here also

I've programmed a Pic but can't get the Lcd to work is there something needs enabling in the software? It's V1. 4 version. I've a 3.3v LCD2004A version on order so I'm hoping it may be that my current display is 5v only.

Here in the UK there's a shortage of diggible gold so another question is can it be modded for artifact hunting?
Thanks in anticipation.

Try jumping a wire over to the 5 volt TX supply .. instead of the 3.3 on the header.

Unfortunately I wont have iron reject ... that will be V4. ( so it will be dig everything ?? ).
I never really understood discrim detectors in artifact hunting .. aren't roman spear / arrow heads made of iron ?

[Gunghouk]

If it works on everything then that'll do me.
I tried supplying the lcd with 5v but no joy. Just wondering if I've missed something that needs enabling in the software.

[moodz]

If it works on everything then that'll do me.
I tried supplying the lcd with 5v but no joy. Just wondering if I've missed something that needs enabling in the software.

Should just work ...provided connections are good. Can you post photo of back of display ?

[Gunghouk]

Here's my display and I2C adapter.
The display works fine over I2C on an Arduino Nano project @ 5v.

[Altra]

Every time I have problems with I2C it's always the device address. I can see three empty jumpers on the black pcb, marked A0,A1 A2. Maybe check Moodz's code for the device address and then short the correct jumper?

[Altra]

I found that there are two versions of the PCF8574 IC. One is PCF8574A and the other is PCF8574T. If moodz wrote the code for the "A" version, your "T" may not work without changing the address in code.

See this

https://alselectro.wordpress.com/201...odule-pcf8574/