After strong interest in the Facebook group, here is a clear breakdown of how to improve WiFi range on the Anthbot Genie mower.

This is a real hardware modification based on teardown and antenna replacement.

The issue

The mower worked well overall, but connectivity was the weak point.

• App connection relies mainly on Bluetooth and short-range WiFi
• Stable control only within ~7 meters from the house doors
• Even with a UniFi setup (4 APs), coverage was not sufficient
• Relocating access points did not significantly improve stability

The key problem was not the network, but the mower’s internal antenna design.

The Solution

ChatGPT : Radiation efficiency & gain

Patch (PCB/internal) antenna
Typical gain: ~1–3 dBi
Efficiency drops if near battery, ground planes, or plastic housing

External antenna (dipole/whip)
Typical gain: ~2–5 dBi (consumer), higher with directional
Better impedance matching → more consistent performance

👉 External antenna often gives +2–4 dB link budget → noticeable range increase

Expectations:
- PCB patch inside enclosure baseline
- External dipole (2–3 dBi) ~1.5–2× range

Step 1: Opening the housing

Luckily there is a video of housing removal - thanks to No Waffling for the videoTop housing removal is straightforward, and this video helps:

*Assembly tip:

• Reinstall upside down for easier alignment
• Press wheel assembly into housing carefully ( instead of putting pressure on wheels )
• A small amount of soapy water on rubber bushings helps reduce friction and risk of damage

Step 2: Locating the WiFi module

Initial disassembly shows:

• There are no WiFi components in the wheel assembly section
• Main electronics are located in the upper camera module (grey housing)

This is effectively the main control unit.

Step 3–5: Full access

• Remove charging plate by unscrewing these two screws

• Remove protective front guard

• Remove camera module.

At this stage, the magnesium camera module should be free from top housing and can be removed as a single assembly.

Step 6: Antenna access

Inside the camera module:

• Next open camera module itself

• The original antenna is a small patch type - glued directly to the plastic housing. Partially open module, find the patch antenna and carefully peel it off to expose. This is how patch antenna is looking.

Step 7: Identifying the WiFi chip

The module used is - Fi-Link 6223A-SRD

1. Antenna cable is secured by black tape
2. U.FL connector

Stock antenna is a low-profile internal patch design - the key limitation point.

Step 8: External antenna modification

Upgrade approach:

• Use an SMA to U.FL extension cable.
  • For example this :(https://cpc.farnell.com/siretta/asmg015x113s17/150mm-1-13-cable-ipex-to-sma-f/dp/RF00763)
• Drill a 6 mm hole in the housing (placement depends on your routing; I used a position near the WiFi module)

• Route the cable externally or to a better RF position inside the housing

Step 9: Reassembly

• Install SMA connector and secure with nut

• Seal the entry point thoroughly with silicone to prevent moisture ingress
• Reconnect U.FL connector

• Reassemble all parts in reverse order

Result

Before modification:

• Unstable connection beyond ~7 meters from doors

After modification:

• Stable connection up to ~40 meters from doors in open yard conditions

The improvement is significant and immediately noticeable in real use.

Additional observations

• Battery expansion: there is physical space, but no supporting power management circuitry, so parallel battery modification would require additional BMS integration
• 4G upgrade: not possible on Genie 600, as the module is not present
• Overall hardware design: surprisingly solid and serviceable once opened

Summary

The limitation was not the WiFi infrastructure, but the internal antenna design of the mower. Replacing the internal patch antenna with an external SMA antenna dramatically improves range and stability.

Additional teardown photos follow.

The problem - some books are not available in audiobook formats. 

The solution - I asked an AI agent to help create a small pipeline

• PDF → extract text
• Clean page numbers and captions
• Send clean text to a text-to-speech model
• Export audio

Three prompts later - I had working code.

First run on my laptop: 14 hour ETA. Overnight run.

Morning check: 8 hours left.

Evening check: Out-of-memory crash.

Classic.

Back to the agent.

It suggested running the job on GPU:

• SageMaker
• Kaggle

SageMaker refused to cooperate that day, so I switched to Kaggle.

Update. Upload. Run.

Three minutes later I had an audiobook.

Small difference, huh? 14h to get OOM crash vs 3 minute GPU jobby

Earlier that day the audiobook didn’t exist anywhere.

Now it did — because a small tool assembled it.

Preview: (quality is really pristine)

Hw hacker

0:00

/34.092979

1×

Success!

That’s the part of AI that excites me most. Not content generation. Tool generation.

Computer science has always been about building tools that reduce friction in everyday tasks. AI lowers the barrier dramatically.

You describe the tool.

The AI assembles it.

And problems that once required weeks of work - can become an evening experiment.

That’s the snowball.

One tool enables the next.

Here’s my latest 3D printing project: a smart bird feeder with a solar panel, Wi-Fi camera, and motion detection.

Designed from scratch, 3D modeled, printed, and assembled over a few evenings.

Build highlights:

• Cost: ~60 EUR (including filament, solar panel, and off-the-shelf camera)
• Power: ~3 hours of sunlight fully charges it; battery hasn’t dropped below 88% so farCamera:
• Built-in app supports motion notifications, event playback, and low-latency live view — all working smoothly out of the box

It’s a small but satisfying project — the kind that reminds me why I like making things that work.

Hey! Recently I started to document my journey of building an automatic pet feeder — a free-time project inspired by a practical problem(will be shared in the next article) and my curiosity to learn about computer vision.

This "series" will showcase the evolution of the project, covering mechanical design,3D printing and "AI" computer vision, web app development, and more.

Of course this will revolutionize the industry - VCs $ are welcome!

Sneak Peek: V3

Sneak Peek: Webapp

Sneak Peek: Object tracking

Once upon a time, I joined a startup. Our mission? To conquer the world—well, at least the world of Building Energy Monitoring. You know, because saving on energy bills is the new world domination. Despite our small team, we somehow managed to snag a few big clients. Things were looking up!

At the same time, I was knee-deep in electrical engineering studies, because clearly, sleep was optional. We had some early wins with software, but we needed to take the next step: hardware. Why? Because gathering data from buildings was key, and paying for existing solutions wasn’t going to happen on our "budget." So, we decided to roll up our sleeves and build it ourselves. Obviously.

The Dream Device: Because What’s One More Prototype?

Our target audience? Large public buildings with those fancy Building Management Systems (BMS)—places spending a small fortune just to keep everything at a "comfortable" 22°C. From our previous work, we knew it was possible to cut their energy expenses by 15% with very little effort (and no one loves "very little effort" more than a big company). We also knew the key protocols and interfaces to tap into these systems.

So, here’s what our miracle device needed to do:

• 12-24V power support – The standard in automation. We’re not here to reinvent the wheel.
• M-Bus interface – The universal language of metering devices. Like English, but with less grammar.
• Modbus/RS485 interface – The go-to for automation systems. If you don’t speak this, you’re out of the conversation.
• GSM connectivity – Because dealing with corporate IT departments is like asking a sloth to run a marathon.
• Ethernet connector – For the rare moments when internet access isn’t a bureaucratic nightmare.
• Cheap – These devices were meant to charm potential clients, so we needed them to be so affordable, you’d think they were party favors.
• Linux support – Our small team needed to move fast. Linux gave us access to libraries, tools, and, well, sanity.
• Overcurrent protection – Because nothing says “we tried” like an overheated circuit.

The Ingenious Plan: Modular Magic

The solution was simple—modular design. Two parts:

1. The Brain – An Orange Pi Zero, because Raspberry Pi is too mainstream. This little computer, costing about €30, ran Armbian OS with solid updates. Six years later, it’s still around, which is more than can be said for some tech startups. Check it out here.

2. The Interface Board – A custom PCB that connected the brain to everything else: RS485, M-Bus, SIM module, voltage converters, LED indicators, and so on. Essentially, everything to make it look complicated enough to impress someone in a boardroom.

Success! Well, Sort Of…

I designed a few boards, ordered them from JLCPCB, and assembled them. Surprisingly, they worked (mostly). But as any good startup story goes, our budget was tighter than a jar lid your grandmother sealed. So, we didn’t make it past the prototype stage.

For fun, I installed one of the prototypes in my parents’ house as a Zigbee hub, for monitoring temperature. That was four years ago. I recently found it, still running. It felt a bit like finding an old pet rock—nobody needed it anymore, but there it was, quietly doing its job.

The Here and Now: Same Spirit, New Problem

Fast forward to today. My latest problem? I can’t control my pellet boiler remotely. Naturally, instead of buying something off the shelf, I’ve decided to hack a solution together. Because if I can build an energy monitoring device on a shoestring budget, surely fixing this boiler will be a piece of cake.

Stay tuned.

Hey there! 👋 Check this out!

Pretty cool, right? And guess what? This wasn't created by some data scientist in a Google data center. I made this, and I don't even have a background in computer vision. Not even writing code professionally, but hey, I made it happen.

So, chances are, you can too!

I'm planning to dive deeper into this tech in the coming weeks. Let's see what else we can build 😉💻

Just a few things I wanted to note to myself about the book The Mom Test. by Rob Fitzpatrick.

As a technical guy deeply entrenched in the world of product development, "The Mom Test" by Rob Fitzpatrick provided invaluable insights into the art of customer validation. One of the most impactful lessons from the book is the importance of avoiding questions that inadvertently validate our assumptions about features. Instead, Fitzpatrick emphasizes the need to focus on understanding customer problems rather than jumping to solutions or features. This resonated deeply with me, as it highlighted the danger of biasing our conversations with customers and missing out on genuine insights.

Moreover, "The Mom Test" stresses the importance of keeping conversations casual and avoiding framed questions that lead customers to provide answers we want to hear. This approach encourages authentic dialogue and prevents us from falling into the trap of confirmation bias. I found this particularly useful in my own interactions with customers, as it allowed for more open and honest communication.

One of the key strategies outlined in the book is the concept of asking questions about how customers currently solve their problems. By delving into their existing processes and workflows, we not only gain deeper insights into their pain points but also evaluate the potential value of our solution. This approach allows us to quickly validate whether there is a viable business opportunity around this problem or if it's merely an annoyance that customers are unwilling to pay to solve.

"The Mom Test" also provides practical advice on how to politely navigate conversations and escape from compliments that don't provide meaningful insights. Instead of getting caught up in praise, Fitzpatrick encourages us to steer the conversation towards exploring pain points and uncovering valuable insights.

Lastly, the book introduces the concept of the "advance" – a small commitment or action that indicates genuine interest from the customer. Types of advances can include signing up for a beta trial, providing contact information for follow-up, or even making a small purchase. By asking for advances, we can gauge customer interest and commitment, helping us identify potential early adopters and refine our go-to-market strategy.

Overall, "The Mom Test" offers practical guidance and actionable strategies for technical professionals looking to validate their product ideas effectively. By avoiding biased questions, focusing on customer problems, keeping conversations casual, navigating compliments gracefully, and leveraging advances, we can gain deeper insights into customer needs and build products that truly resonate with our target audience.

Remember the great MCU chip shortage of 2020?
Those were the days when even getting your hands on a small batch of microcontrollers felt like winning the lottery.

Flashback to that time: armed with my freshly minted engineering degree, in my free time I set out to conquer the world with my grand vision of creating the most mind-blowing long-range temperature sensor system. But fate had other plans. The chip shortage hit, leaving me with only 3D designs and some rough PCB designs. Temperature sensor chips were as elusive as a unicorn in those dark days.

Fast forward a few years, turns out, that shortage was a good thing after all. It forced me to pivot, adapt, and ultimately land a gig where I'm leading a team of eight professionals, tackling exciting products with real-world impact. From developing cutting-edge video telematics solutions to spearheading integrations that streamline operations, every day brings new challenges and opportunities to make a difference.

Inspired by Tony Fadell's #BUILD, I'm now contemplating starting some cool side projects during my free time. After all, who knows where the next great idea might come from?
Let's embrace the lessons learned from scarcity and turn them into opportunities for growth and innovation!

This is PROTO BLOG, a brand new site by M.G. that's just getting started. Things will be up and running here shortly.