Falcon030 dead after Centuriontech CT60/CT60PSU incident - investigation begins

[exxos]

A UK owner recently got in touch about a Falcon030 motherboard with a serious fault. He'd been running a CT60e accelerator (bought from willyvmm in 2017) on a picoPSU without issue for several years, then bought a CT60PSU and later a CT63 from Centuriontech. He also has a separate, working Falcon motherboard unrelated to any of this, which he used to confirm the fault by testing the original Atari PSU on it.

The sequence of events, as described to me:

1. March 2017: bought the CT60e from willyvmm, ran it on a picoPSU with no problems for several years.
2. 13th August 2025: bought a CT60PSU from Centuriontech, wanting the convenience of its dual-position switch (030/060 mode selectable without opening the case) over the picoPSU's bulky external brick. Within days of fitting it, one of the Xilinx chips on the CT60e exploded with a visible hole, silently, no pop or smoke, just a loss of 060 boot capability followed by the visible chip damage. Centuriontech were contacted shortly after and were not forthcoming with support.
3. August 2025: ordered a second CT63 from Centuriontech to replace the dead CT60e. It took roughly a year to arrive, reportedly due to component supply problems.
4. August 2026: the CT63 arrived and worked fine for a couple of days. Then, at the exact moment of flicking the CT60PSU's dual-action power switch between positions, there was a loud pop and black smoke around the switch. This took out the CT63, the Falcon motherboard, and the small circuit board on the power switch itself. Centuriontech were contacted the following day and again were not forthcoming with support.

So in total: two Centuriontech products bought over this period showing failures (the CT60PSU itself, and the CT63), the CT60e died alongside them despite being bought years earlier from a different source, and there's one dead Falcon motherboard.

Centuriontech were not willing to offer support or replace any of the broken hardware after either of the two earlier incidents. He's since contacted them a third time, asking about a replacement or refund, and as of writing, several days on, still hasn't had a reply.

I don't normally take on Falcon motherboard repair work these days, for reasons I've gone into elsewhere on here. But this isn't just a pile of dead hardware, it's several thousand pounds worth of destroyed equipment. With no support forthcoming from Centuriontech, I decided it was worth at least taking a proper look, to actually see what's blown rather than just guessing.

To blow a visible hole in a chip like that, it sounds like the 3.3V rail went too high at some point, that's really the only thing I can think of that would destroy a chip that catastrophically. That's only a guess at this stage though, nothing confirmed. The 060 CPU also runs on 3.3V, so there's a good chance that's damaged too, along with the RAM. I'll test everything in my own Falcon once the parts arrive, and hopefully it doesn't take out anything of mine in the process.

The motherboard, the CT63, and the CT60PSU are all being sent over for inspection. I'll post photos of the damage once they arrive, chip level and PSU level both, and report what testing actually shows once it's done. No conclusions yet on cause, that's exactly what the inspection is for, but the sequence as described is worth putting on record so other CT60PSU/CT60/CT63 owners have it to weigh up for themselves.

I'm currently waiting for the package of carnage to arrive, and I'll update this thread as I work through everything. Time is limited at the moment, so updates may come slowly.

[Badwolf]

To blow a visible hole in a chip like that, it sounds like the 3.3V rail went too high at some point

Crumbs!

I'd guess it must have been the 12V rail that got shunted through there. Would 5V have been enough to actually blow out a package? Blimey.

Best of luck!

BW

[exxos]

I'd guess it must have been the 12V rail that got shunted through there. Would 5V have been enough to actually blow out a package? Blimey.

Never tried blowing one upon purpose to find out

My only possible other thought, is maybe a spike on the mains or something has damaged one of the power blocks internally somehow and it's just completely blown to bits inside. it could have even sent mains down for all we know.

I don't know the quality of the power blocks used yet but I have seen really bad power supplies over the years where the more you load them the more the voltage increases and that basically very very bad. I've also seen switching power supplies explode if you do not have a load on them.. Anyone's guess at this point..

Part of the discussion I was having with the owner about trying to fix the accelerators by replacing the PLD chips, but the problem is that we don't know if any of the tracks have vaporized on any of the internal layers, so it's questionable if it's even worth attempting to repair them or not.. Similar sort of thing with a Falcon motherboard as well..

I've had Falcon's here before, where the 12 volt has been applied to the 5 volt rail, and it's just basically obliterated the entire thing. I don't even bother even trying to repair those now. it would just be probably cheaper to buy one of the remake boards at that point... Time will tell...

[PhilC]

The only times that I've seen chips literally blown up has been when mains voltage has somehow made its way onto the exploded chip

So I'd be thinking ground or neutral fault or something similar, like a buck converter gone bad

[exxos]

The only times that I've seen chips literally blown up has been when mains voltage has somehow made its way onto the exploded chip

So I'd be thinking ground or neutral fault or something similar, like a buck converter gone bad

You may be onto something there.. just about to do another post...

[exxos]

While waiting for the parts to arrive, I had a look at Centuriontech's own product page for the CT60PSU, to understand what this switch is actually supposed to do.

Their own description calls it a "dual function AC power switch" that "powers on the system in either 030 or 060 mode". So by their own wording, the same switch handles mains power and the 030/060 mode selection.

That raises an obvious question. The 030/060 mode select is a logic level signal, going off to a PLD somewhere, nowhere near mains voltage. Mains switching and a logic signal have no business being anywhere near each other on the same switch, unless there's a genuinely well isolated interface board between the two sides.

The Falcon owner described a small circuit board at the switch itself, which was also damaged with smoke coming from it during the incident, suggesting there is some kind of small interface board there, but there's nothing on Centuriontech's page explaining how it works or what isolates it.

No conclusions yet, I'm just reading what's publicly available and asking the obvious question it raises. Once the PSU and switch board are here I'll open it up and see what's actually going on inside.

Does anyone else know more about how this switch actually works internally, or have any photos of the switch board itself? There doesn't seem to be any information about it anywhere online, so if anyone's got one apart or has pictures, I'd be glad to see them.

[DoG]

There is a review video here about it. Some interesting close up shots.

[exxos]

There is a review video here about it. Some interesting close up shots.

Thanks..

It does look like the mode select wires run underneath the switch, right where the mains wiring also goes in. From what's shown, it looks like it must be a six pin switch, with the centre position off and left/right both switching mains on while also selecting the mode on a separate pole of the same switch.

Since the PCB was apparently burnt in the incident, it looks like there may have been some kind of flashover from the mains side across to the mode select side, which would fit with everything described so far.

This is still just my own reading of what the video shows, not a confirmed finding, but it's looking more and more like the switch design itself may be at the heart of this.