Fractal SP8 Rev-02 1S AIO Flight Controller

How Did We Get There?

1-Preface

Let me start with this; I am NOT an electrical designer. My trade is in industrial product design.

The frames are at the heart of what makes Fractal; they are the foundation of the flight experiences I want to create and share with you.

All the electronics I’ve been dealing with were specifically meant to support my frames, as spare parts and even more so in the BNF.

Needless to say, by now we’ve built hundreds of them. They enable folks to skip right to flying and simply enjoy the frames as I intended.

They also constitute a major part of the revenue that makes it even possible for me to keep developing new concepts for the hobby

So far most components have been fairly steady to procure; cameras, motors, props…(let’s not talk about VTX these days thought lol)

Flight Controllers though, oh man.

We had a long streak with happymodel supplying all our AIOs for 2 years; a mostly positive partnership thanks to their flawless support, but not without hiccups with rather high DOA rate,

Of course those DOA units would be detected in BNF assembly, but as a small and growing business we couldn’t order large amounts, and often times ended up running short of FCs because of it, on BNF orders already purchased by customers.

This part really put me through lots of stress as I always want to offer the best possible service to our supporting customers, and asking them to wait for it really wasn’t it.

2-The arms race

I had to do something about it.

I had been mindlessly watching the likes of EEVBlog, Louis Rossman, GreatScott for a while by now

So I thought, well when I return those FCs for RMA, they will perform on-board repair and send them back anyway, how about I just skip this step and perform the repairs anyway? how hard could it be?

And I got some equipment; a hot air gun, electronic microscope, some shoddy flux. Great.

After a few attempts I was able to replace burned mosfets. Cool.

Then I came across a board that kept frying the fets even after replacing; something else was off, couldn’t figure it out.

I realized repairing the gear was only a part of the work; most of it was actually diagnosing what to repair exactly…

So I got more equipment; lab power supply, oscilloscope, IR thermometer…

They did help a lot getting more repairs done, keeping the inventory up through the DOA cases.

And before long I realized I had all the equipment to start developing my own electronics, so came out the boostyboi and boostybaby later on

However, then the FCs I got so familiar of using/repairing, got discontinued.

So again, no FCs, no BNF, I went out searching for new ones, with mixed results…

Until finally it occurred to me that the FC market got so absolutely bad in all that mean time, that betafpv was now considered the best around?!?

That was around February this year 2024.

That’s when I snapped.

3-The SP8

I hit up my good friend Hanfer, head dev of Quicksilver, and offered him to collab on an AIO truly made for us.

We decided on a chip and feature set, and started the schematics, leaving him most of the FC part, where he assigned resources to the MCU in the most optimal way possible, while I mostly took care of the 1S optimized ESC.

What followed -on my side at least- was 2 weeks of pure, uninterrupted, spite-fueled blind rage laying out the AIO in KiCad.

Mind you, I had never worked on a project this size (the boostybaby is 9 components, the SP8 over 240) let alone on 4 layers.

Yet, somehow all that spite accumulated from years of dealing with all the FC issues managed to drive me through that process, as all the pieces came together in that critical mass moment.

I’m telling you for these 2 weeks I barely slept a couple hours a day, as the moment I’d lay down in bed and close my eyes I’d start seeing ratslines and still layout the board in my head 😂

It didn’t help that we practically worked shifts with Hanfer, he would finetune the schematics during his european hours and I’d take over on my asian time.

What a bonkers time that was.

But after just 2 weeks we had it, a fully resolved, laid out AIO, exactly to our specs, ready for prototype.

Hell yea.

4-The result

You’d think that surely, if all those big dog fpv companies can’t get it right, there’s no way 2 idiots in a shed would right,

Well so did we.

So imagine our surprise when I assembled that initial prototype, and after a few bodges found most of the features ended up working. first time. surprise pikachu face.

Not exactly a flying FC yet, but all of the FC and most of the ESCs were fully functional.

So we fixed the little goofs here and there; some pin strapping, layout reliabilizing…

And with our 2nd board we had it. A fully working, FLYING, hand assembled prototype board.

It took us a few months due to mostly idle time dealing with other urgencies and a loop around home manufacturing,

but essentially, the 2 idiots in a shed made it.

But made what exactly?