Custom ESP32-S3 development board — professionally manufactured by JLCPCB. A far cry from where it all started. It Started in a School Science Lab — Around 1998 Most people who get into electronics start with a kit, a tutorial, maybe a breadboard and some LEDs. I started by sneaking ferric chloride out of a school science lab to etch my first PCB. That was around 1998. I was living in the Maldives — a small island nation in the Indian Ocean — where there was no electronics supply chain, no maker community, no local PCB fab. Just a chemistry cabinet at school, a copper-clad board from somewhere, and a lot of curiosity. This post is about what the next 25+ years of PCB prototyping looked like from there. The early wins with proper chemicals, the years of improvisation when those chemicals disappeared, the real injuries, the failed boards, and finally — the moment JLCPCB changed ever...
 |
| Custom ESP32-S3 development board — professionally manufactured by JLCPCB. A far cry from where it all started. |
It Started in a School Science Lab — Around 1998
Most people who get into electronics start with a kit, a tutorial, maybe a breadboard and some LEDs. I started by sneaking ferric chloride out of a school science lab to etch my first PCB.
That was around 1998. I was living in the Maldives — a small island nation in the Indian Ocean — where there was no electronics supply chain, no maker community, no local PCB fab. Just a chemistry cabinet at school, a copper-clad board from somewhere, and a lot of curiosity.
This post is about what the next 25+ years of PCB prototyping looked like from there. The early wins with proper chemicals, the years of improvisation when those chemicals disappeared, the real injuries, the failed boards, and finally — the moment JLCPCB changed everything.
Era One — Ferric Chloride and First Boards (1998 – Early 2000s)
In those early years, ferric chloride (FeCl₃) — the standard PCB etchant — was actually obtainable. School science labs stocked it as a chemistry reagent, and with a bit of resourcefulness you could get small quantities from a few other informal sources.
The process was entirely manual. PCB layouts were either hand-drawn or printed from early CAD tools onto glossy paper, then ironed onto a copper-clad board — the classic toner transfer method. The ferric chloride did the rest, dissolving the exposed copper and leaving behind the traces. Drilling was done by hand.
The boards were rough by any standard. Inconsistent trace widths, generous pad sizes, single-sided layouts forced by the limits of the process. But they worked. Simple timers, LED controllers, sensor circuits — each one a proof of concept built from nothing. I don't have photos from this period, but every mistake made back then built the instinct that carried everything that came later.
📌 No Photos From the Early Era
The 1998–early 2000s period predates smartphones and casual photography for me. All images in this post are from 2015 onwards — but the techniques and mindset started much earlier.
Era Two — When the Right Chemical Disappeared
At some point, ferric chloride became impossible to obtain in the Maldives. The exact regulatory reason was never made fully clear, but the result was simple: the standard tool for the job was gone. Restricted. Unavailable.
So I did what any determined engineer with no other option does — I went looking for an alternative. I spent days visiting local hardware stores and pharmacies, reading labels, researching reactions. The goal was to find something that could oxidise copper the same way ferric chloride does, using only what was sitting on local shelves.
What I found was a mixture of pool maintenance chemicals — oxidising agents used to treat swimming pools — combined with compounds from local pharmacies. It worked. It etched copper. But it was nowhere near as safe or predictable as ferric chloride.
⚠ SAFETY WARNING — Do Not Try This
The improvised chemical mixture described in this article is genuinely dangerous. It produces toxic fumes and causes severe chemical burns on contact with skin. I have personally experienced burns from this process — this is not a theoretical risk.
This method is documented here for historical record only. Do not attempt to replicate it. JLCPCB and other professional PCB services exist, are affordable, and give you far better results with zero chemical hazard.
The Process — UV Photoresist, Caustic Soda, and Improvised Etchant
By this period I had also moved from toner transfer to the UV photoresist method — it produced significantly finer and more reliable traces, which was essential as the designs got more complex. The full process looked like this:
01
Design the PCB
CAD software, exported as a mirrored image
02
Print the transparency mask
Laser-printed onto acetate / transparency film — the photomask
03
UV exposure
Photosensitive copper-clad board under the mask in a homemade UV box — burns the circuit image into the photoresist
04
Developer bath — caustic soda (NaOH)
Agitate for a few minutes — dissolves UV-exposed photoresist, reveals copper pattern
05
Etch in the improvised bath
Pool chemical mixture dissolves exposed copper — protected traces remain
06
Strip, drill, populate
Remove remaining photoresist, drill holes, solder components
The Night Everything Went Wrong — And Many Other Nights
The failure rate was high. A single board could represent two to three hours of careful work — designing, printing, exposing, developing, etching. If the chemical mixture was slightly off, the entire copper layer could dissolve in seconds, taking every trace with it. Two hours of work. Gone.
- dangerous Chemical fumes during etching — had to work outdoors, often at night
- dangerous Skin contact caused chemical burns — happened multiple times despite precautions
- inconsistent Etchant behaviour varied batch to batch — what worked Monday might destroy a board on Friday
- inconsistent UV exposure time had to be re-calibrated constantly — overexposure hardened resist that should wash away
- inconsistent Developer concentration too strong → attacked all photoresist. Too weak → incomplete development
Despite all of this, functional boards came out. The satisfaction of holding a working PCB that you designed, etched, and assembled yourself — in a place with none of the right resources — was real. But I won't romanticise it. The process was dangerous, inconsistent, and ultimately a workaround for a regulatory barrier that should not have existed.
The Boards — 2015 to 2017
 |
| UV-exposed photoresist board alongside the printed transparency mask (2015). The dark coating shows the circuit image burned into the photosensitive layer. |
 |
| First populated prototype — QFP microcontroller, crystal oscillator, and passive components hand-soldered onto a home-etched FR4 substrate (2015). |
 |
| A larger, more complex board (2016). Months of process refinement allowed denser routing and finer traces. |
 |
| GSM wireless prototype — SIM-based cellular modules hand-wired onto a custom home-etched PCB (2016). |
 |
| The etching process in progress (2017). PCB submerged in the improvised pool-chemical bath — the most hazardous stage of the entire process. |
 |
| Multiple completed boards after etching (2017). By this point the process had reached a working — though never fully reliable — degree of consistency. |
Era Three — JLCPCB Changes Everything
Submitting a Gerber file and receiving a stack of perfectly manufactured boards a week later felt almost surreal after years of improvised etching. Green solder mask. White silkscreen. HASL-finished pads. Drill-precise vias. The quality difference is not incremental — it is categorical.
$2
5 Boards From
24h
Build Time
0
Chemical Hazard
SMT
Assembly Available
 |
| Custom ESP32-S3 development board — professionally manufactured by JLCPCB. Fine-pitch SMD components, clean ground plane, HASL finish. This kind of quality was impossible with home-etching. |
 |
| The same board in a working prototype system — with power supply, Ethernet, and wireless module. Designed and assembled in the Maldives. Built on a JLCPCB PCB. |
For makers in countries like the Maldives — where local supply chains don't exist, import restrictions are real, and even basic electronics components can be confiscated at customs — JLCPCB is not just convenient. It is a genuine lifeline for independent prototyping. Being able to design something, upload the files, and receive professional boards by post has kept this work alive and advancing.
✅ Support JLCPCB
If you're doing any kind of electronics work — student, hobbyist, or professional — JLCPCB is the service I recommend without hesitation. Fast, reliable, exceptional value, and they genuinely support the global maker community.
Key Takeaways
01
Restriction doesn't stop experimentation — it makes it more dangerous
Banning ferric chloride didn't stop PCB etching in the Maldives. It redirected it toward improvised pool chemicals that were far more hazardous. The safer option was the regulated one.
02
You learn more from a failed board than a successful one
Every dissolved trace, every ruined batch, every chemical burn taught something that reading a tutorial never could. The hard way is slow and painful, but it builds real understanding.
03
The improvised method is not worth replicating — ever
Now that JLCPCB exists, there is no practical reason to use improvised chemical etching. The risk is real, the results are inconsistent, and the professional alternative is cheaper than you think.
04
Access to tools is not equal — and that has a cost
Small island nations like the Maldives face real systemic barriers to independent innovation: import restrictions, customs seizures, no local suppliers. The maker community here is resilient, but it shouldn't have to be this hard. That's a topic for another post.
📄 Full Article
A detailed academic-style write-up of this journey — covering all three eras, the full UV photoresist process, and the chemistry behind the improvised etchant — is also available as a PDF document.
Comments