Ivconic Magnetometer/Gradiometer

[ivconic]

[ivconic]

Before the final WS43 project, I did several "lighter" and simpler projects.
One is particularly interesting because it only uses 1 sensor and a gyroscope and serves as a pinpointer (up to several meters depth though).

[ivconic]

[ivconic]

This is a good example to compare the power of ESP32-P4 and ESP32-S3.
On this project with 1 sensor and a gyroscope, the main processor is ESP32-S3.
Considerably weaker than ESP32-P4.
The grid is a fixed size of 5x5 and isometric 3D rendering is obviously slower and of lower quality.
But it is good enough for mapping smaller areas.
It has no file system and no connection with Android, it is designed as an "instant" instrument for quick checks of soil conditions (or some other, perhaps industrial application).

[ivconic]

The simplest project is with a 1.8" (non-touch) display.
"Null" is a physical button and "Sens" is a physical potentiometer.
It is designed as a small manual pointer for very fast tests.

[ivconic]

One of the "transitional" projects is the WS2, which has a 2" TFT Touch screen and an ESP32-S3 processor.
It was the work from which I added an external ADC.
It was the "key work", which showed me the direction in which I had to move further in development.

[ivconic]

So even though my posts are chronologically disorganized; the true chronological list of development to arrive at the final WS43 project is as follows:

1) 1.8", ESP32-S, no "Touch", no external ADC, 1 sensor.
2) 2.8", ESP32-S, "Touch", no external ADC, 1 sensor + gyroscope, 5x5 grid mapping with rudimentary 3D isometry.
3) 2", ESP32-S3, "Touch", external ADC, 1 or 2 sensors, BT streaming, some other features.
4) 4.3", ES32-P4, "Touch", external ADC, 1 or 2 sensors, BT streaming, full scale 2D/3D mapping, file system, micro SD card, etc...

I drew a couple of important conclusions from these works:
- Without an external ADC and a good DSP, the projects are "lighter category", very usable for simple instant tests and checks on soils.
But they provide low-end performance. Usable in various applications that do not require high precision and resolution.
- ESP32-S and ESP32-S3 are of respectable performance but for real multitasking that includes serious DSP and 2D/3D mapping; are not enough.
- ESP32-P4 is a "beast" and can easily multitask, no sweat, very complicated DSP algorithms + do serious 2D/3D rendering at the same time, easily supports file system and BLE communication with little and in most of the cases no effort at all.

I have noticed an increased interest in various "stick" magnetometers and the appearance of several products from small independent manufacturers is noticeable.
I had the opportunity to hold most of them in my hands and test them thoroughly.
Despite the often high prices, none of these products can be classified as better than "low end" category, I'm sorry, but that's a fact.
Serious performance requires much more than mere enthusiasm and putting high prices on beautifully decorated toys.
...
Therefore, from all similar devices that I have tested so far as well as from my personal projects; only and exclusively 4.3", the ES32-P4 project can be classified in the "mid-range" category, with some specifications that even exceed that category.

[ivconic]

And now a few words about the sensors.
I took the famous FLC-100 sensor as the main reference for working on projects.
However, I was fortunate to have a collaborator in China, a very smart and educated young man, who dedicated himself to developing quality sensors.
And after 3 years of hard development work, many expenses along the way, he made a remarkable success by achieving top performance with his "HDC100" sensor and typically the same performance as original FLC-100 with his "FLCH-101" sensor.
He provided me with several copies of those sensors and that's how I got into a situation where I could change the sensors and compare the performances.
Real progress and peak performance can only be seen when using an external ADC and HDC100 sensor.
The differences are ... to put it mildly: HUGE!
The colleague from China did not stop there, in the meantime he is developing an even better sensor.
He had to temporarily stop his work due to some unforeseen personal problems.
But I sincerely hope that he will overcome these problems and continue his development soon.
But until that happens, the HDC100 remains one of the best sensors I've worked with to date.

[Qiaozhi]

Half of them are back... the other half, I don't have. George, did you add these?

I did add a few extras, but not that many.
I don't seem to have them either, but will try and find some replacements.

[latcakir]

FOR SALE: Complete Project — Handheld Fluxgate Gradiometer with 24-Bit ADC and ESP32-P4
Archaeological, geological and industrial magnetometry. Serious principals only. NDA mandatory.

Overview

I am offering for sale a fully developed, completed and tested project: a high-sensitivity handheld fluxgate gradiometer, designed primarily for archaeological and geological surveying.
With minimal adaptation it is also suitable for industrial applications.
The project was finished and field-tested in July 2026.
It has been held in confidence until now.

Key Features

• Two operating modes.

1. Real-time gradiometer, pinpointer-style, with several selectable displays on the Home screen.
2. 2D/3D grid mapping, up to 40x40 cells (adjustable down to 5x5).

• Fully standalone. The device provides 100% of its functionality on its own, with an internal file system holding up to 100 recordings.
• Android integration. Over Bluetooth (BLE), the device streams data in real time to an Android application. When connected, it switches automatically to slave mode and operation continues in the app.
• Open data format. Recordings are saved as standard .csv files, compatible with common survey and analysis software. Files can be saved, deleted, moved to a micro SD card and transferred to other devices or computers.
• 24-bit ADC. All ADC parameters are adjustable from the menus.
• Advanced DSP. A multi-stage chain of digital filters extracts the useful signal from the raw measurement, fully configurable from the menus.
• Sensors. One or two fluxgate FLC-100 sensors (selectable in the menus). Higher-grade alternatives are available: HDC100 sensors, which reportedly exceed FLC-100 specifications by about 5x, and FLCH-101 sensors, which are on par with the FLC-100.
• Main processor. ESP32-P4.

Market Position

As the author, I cannot be fully objective, and I leave the final judgement to the market.
In my professional opinion, however, the device offers exceptional sensitivity, resolution and stability, and I know of no competing product in the entry and mid-range categories that matches it.
Only top-tier instruments (such as Barrington systems) outperform it, at correspondingly far higher prices.

Economics

• Production cost in Serbia, where materials and duties are comparatively high: ~ €300 per unit.
• Realistic retail price of the finished device: €2,200–2,500, based on existing comparable products that are significantly inferior in quality and performance.

Background

I am the author of Euromag3D ,year 2018 , which began with a difficult market entry and went on to become a highly profitable project. It still generates income today.
I have since drawn clear conclusions about how to do business: only quality, exclusivity and speed of decision-making succeed.
I have no interest in image-building, sales theatre or negotiations with parties who intend to invest nothing.
My manufacturing partner in China is currently unavailable for personal reasons, and my attempts to reach decision-makers locally did not succeed.
I have therefore decided to offer the project publicly.

Terms of Sale

• Starting price: €30,000. Offers below this level will not be considered or answered.
• A signed NDA is mandatory before any technical details are disclosed. Serbia recognises and enforces NDAs, and such agreements are valid in the great majority of jurisdictions.
• This is an offer to investors, manufacturers and companies with real capital, the capacity to produce and sell, and the ability to decide quickly.
• The project will be sold fairly and properly, or it will remain archived and unpublished for good.

If you are a serious buyer, contact me directly. Inquiries without genuine purchasing intent will not receive a reply.

Contact

Email: [email protected] Viber / WhatsApp: +381 65 830 4004

Dear Ivconic
Did you make it using vibe coding?

[ivconic]

Yes, sure.
I don't know how to judge properly, it was a long process, several intense months.
But before that, several years of working without AI and gathering a lot of experience, both theoretical and practical.
I wouldn't be able to estimate how much of this is really a product of AI and how complete my work is without AI.
Things quickly get out of hand.
In 2018 I started with "Euromag3D" which was 100% my work without AI (I don't remember but I don't think AI existed as an available tool then).
Later, AI appeared and most of the time, as it is today, it serves as a tool for me to reduce time around trivial tasks.
But as AI advances; so I leave more and more of the work to it.
For example, imagine that I am an Indian living in India and using an elephant to carry huge logs of wood from one place to another!
However, in order to have the widest possible context, it all started in 200x (2003-2005) or so,
I would be lying if I said exactly what year, when I first made the "EPE Magnetometer" from the EPE magazine.
From then until 2018, I regularly made EPE magnetometers, sold them, worked with them, learned from the course of practical work.
"Euromag3D" was my attempt to modernize the "EPE Magnetometer", quite successful.
Everything that bothered me for years and didn't fit me with the EPE Magnetometer; I tried to supplement and correct with the "Euromag3D".
As you can see, things are not simple, they are not random and they do not happen suddenly. It's all a process.

[ivconic]

If anyone is interested, and I think this is an interesting trivia, the biggest problem wasn't writing the code.
Writing the code is the easy part.
The hardest part of the job is mastering the diverse hardware.
In theory, everything can be connected to everything.
In practice, you discover that you have just opened a can full of worms... and the smelliest ones at that!
Ha,ha,ha! 100% TRUE!
And that is the main point of dispute between theory and practice!
It's easy to use methods that someone has developed before you.
Github is full of ready-made solutions.
But it's the hardest when you don't have the support of the open source community, when Github is empty on a certain topic.
Then you start a real pioneer trench warfare! You dig to exhaustion!
I can say that I had the most trouble with the WS43 project!
Because its hardware is very poorly supported by the manufacturer itself.
Because it is a rather new product.
Support... ZERO!
I had to "dig" step by step, in order to enable each piece of hardware on the module.
I wrote a lot of it myself from the beginning.
The general recommendation of the manufacturer is that all delicate things CAN ONLY BE DONE in ESP-IDF and no other way.
I'm old school, a slave to habits, I'm reluctant, almost impossible, to move on to new things easily.
Out of 5-6 months of hard work (10-12 hours a day), I spent 80% of my time tuning the hardware and only 20% just writing the code.
Everything could have gone easier and differently.
It may only took a couple of weeks to adapt to ESP-IDF and the whole job probably would be done in a few weeks.
Although, this should be taken with a grain of salt, I don't know ESP-IDF and I'm just guessing.
I have it installed on my computer and every time I run it; after only 10 seconds i get really annoyed and turn it off!
Ha, ha, ha!

[j570]

As the author, I cannot be fully objective, and I leave the final judgement to the market.
In my professional opinion, however, the device offers exceptional sensitivity, resolution and stability, and I know of no competing product in the entry and mid-range categories that matches it.
Only top-tier instruments (such as Barrington systems) outperform it, at correspondingly far higher prices.

Sorry, but this makes no sense. You’re selling a design for €30,000 yet fail to provide a basic parameter: the device's sensitivity. You wrote, "I know of no competing product in the entry and mid-range categories that matches it." The fact that you don't know of any is enough to make that statement true. You claim the HDC100 sensors are "5x better"—that’s just marketing gibberish. Provide their sensitivity specs, the relevant graphs, and field test results.

[ivconic]

Sorry, but this makes no sense. You’re selling a design for €30,000 yet fail to provide a basic parameter: the device's sensitivity. You wrote, "I know of no competing product in the entry and mid-range categories that matches it." The fact that you don't know of any is enough to make that statement true. You claim the HDC100 sensors are "5x better"—that’s just marketing gibberish. Provide their sensitivity specs, the relevant graphs, and field test results.

Constructive criticism is welcome.
Dismissing someone’s work as “marketing gibberish” without offering a substantive technical contribution is not constructive criticism.
Skepticism is easy. Building something worth evaluating is the harder part.
Fair questions deserve technical answers. Dismissive language adds nothing to the discussion.

[ivconic]

Despite the dismissive tone and lack of constructive substance in your remarks, I’ll address the technical points you raised.

The HDC100 is a complete single-axis fluxgate magnetometer designed for measuring weak magnetic fields, with a measurement range of ±100 µT (±70 µT in the gradiometer configuration).
Key specifications:
- Noise density: 30 pT/√Hz at 1 Hz — five times lower than the FLC100
- Zero drift: <2 nT/K
- Accuracy at 20 °C: ±2% ±0.3 µT
- Bandwidth: DC to 1 kHz (−3 dB)
- Output sensitivity: 1 V per 50 µT
- Supply voltage: 5 V ±5%
- Supply current: <1 mA
- Operating temperature: −40 to +85 °C
The sensor provides an analog output proportional to the magnetic field component along its sensing axis and operates without external excitation circuitry.