Falcon030 dead after Centuriontech CT60/CT60PSU incident - investigation begins

[exxos]

When I was a bit younger I worked for a computer repair company, and it was situated in a smaller town, it was amazing how many PSU's/mb/modems I had to replace just because of overcurrent conditions and lightning strikes, it was at leat 3x as much as the computer store in the city..

Sounds about right.. I used to repair power supplies for all sorts of equipment in my old job. Out of every switched mode power supply I had seen for literally years and all the failures. That is why I chose to use a low voltage front end transformer on my own power supplies because that was the only type of design which had never exploded before.

[rubber_jonnie]

I removed the switch from the PCB.


PICT0002.jpg

Looking at that first picture it looks like the two low voltage switch lines (Maybe even only one) have become live and flashed over to earth and killed the CT60e in the process. I'm basing this on the flash around the bottom left switch pin. That definitely looks like a 240v flash to me.

Just my view on it.

[exxos]

Looking at that first picture it looks like the two low voltage switch lines (Maybe even only one) have become live and flashed over to earth and killed the CT60e in the process. I'm basing this on the flash around the bottom left switch pin. That definitely looks like a 240v flash to me.

Yeah but from where as that track is GND.. The 2 thin tracks are CT60 wires 5 volts..

The gap is about 3mm from ground to thin tracks.. The highest voltage potential is 5 volts.. Which leads the only source of 230v via the switch.. The switch would have to flash over internally, that gives mains on the 2 thin tracks. Then as theres only 3mm gap to GND, that's what burnt the PCB. That's my best guess. The switch is supposed to have 1000v insulation internally.

[rubber_jonnie]

Looking at that first picture it looks like the two low voltage switch lines (Maybe even only one) have become live and flashed over to earth and killed the CT60e in the process. I'm basing this on the flash around the bottom left switch pin. That definitely looks like a 240v flash to me.

Yeah but from where as that track is GND.. The 2 thin tracks are CT60 wires 5 volts..

The gap is about 3mm from ground to thin tracks.. The highest voltage potential is 5 volts.. Which leads the only source of 230v via the switch.. The switch would have to flash over internally, that gives mains on the 2 thin tracks. Then as theres only 3mm gap to GND, that's what burnt the PCB. That's my best guess. The switch is supposed to have 1000v insulation internally.

Yeah but the thin tracks go inside the switch along with 240 volts. If the flashover started inside the switch the two thin tracks could have been shorted to live and then subsequently shorted to GND outside the switch.

[exxos]

Yeah but the thin tracks go inside the switch along with 240 volts. If the flashover started inside the switch the two thin tracks could have been shorted to live and then subsequently shorted to GND outside the switch.

Yeah that's what I meant..

[rubber_jonnie]

The switch is supposed to have 1000v insulation internally.

It might do, but is that between the inside and outside of the switch or or just between the internal switch contacts?

I'd like to see the inside of that switch.

[exxos]

Best match I can find..

1491719.pdf

(496.91 KiB) Not downloaded yet

AI extracts info..

Alco A203 switch family: published ratings (for reference)

Several people have asked what the switch in the CT60PSU is rated for. I haven't yet been able to read the exact markings on the switch from the unit, so I can't say which variant it is. For now, this is what the manufacturer's literature says about the A203 family. Nothing here is a conclusion about the cause of the failure.

Source
TE Connectivity / Alco "Gemini Series Switches Quick Reference Guide" (April 2011, Farnell 1491719). This is a short quick reference, not a full datasheet. TE state that dimensions are for reference only and to consult them for current specifications.

Contact ratings (maximum, as listed)

• Gold: 0.4VA @ 20V AC or DC
• Gold over silver: 5A @ 125VAC or 28VDC resistive, or 0.4VA @ 20V, or 2A @ 250VAC resistive
• Silver: 5A @ 125VAC or 28VDC resistive

Other electrical figures

• Electrical life: 6,000 cycles
• Contact resistance: 10 milliohms max (2 to 4VDC, 100mA)
• Insulation resistance: 1,000 megohms
• Dielectric strength: 1,000Vrms at sea level
• Break-before-make contact timing
• Actuator travel: 22 degrees (D and E function), 18 degrees (F and G function)
• Operating and storage temperature: -30 to +85C

Materials (as listed)

• Contacts: B = gold over nickel, G = gold over silver, Q = silver
• Actuator and bushing: brass, nickel plated
• Case: thermoplastic
• Internal actuator: phenolic
• Spring: music wire
• Support bracket: brass, tin plated
• Seal: epoxy
• G and Q contacts are listed as UL recognised

Rocker / paddle ON-OFF-ON examples from the guide

• A203J51ZQ0004 (TE 2-1571987-6)
• A203J61ZQ0004 (TE 2-1571987-7)

Both are double pole, ON-OFF-ON, snap-in with frame, wire lug terminals, silver contacts.

Reading the part number: A = rocker/paddle, 2 = double pole, 03 = ON-OFF-ON, J1/J51/J61 = actuator style, Z = wire lug, Q = silver contacts, 04 = RoHS.

A similar vintage listing
A UK parts seller lists a "Vintage Alco Switch DPDT Rocker Switch On/Off/On" (part 069-231), six terminals, rated 5A @ 125VAC and 2A @ 250VAC. It has no datasheet. An eBay listing with the title "Alco Switch A203 2 amp @ 250VAC 5 amp @ 125VAC Japan" appears to be the same rating.

Things to bear in mind

• All the ratings above are resistive ratings. The guide gives no inductive or capacitive rating, and none for the inrush of a switch-mode supply.
• In this table the silver-contact (Q) version is listed at 125VAC. The 250VAC figure appears only under gold over silver. I'm not yet sure which variant was fitted, so I'm not drawing any conclusion about 230V use.
• The 2011 guide may not match the older or Japanese-made parts that were sold under the same family name.
• I could not confirm the "TE 2267079-5" part number mentioned earlier in the thread. The 2267079-2 I found is a different subminiature toggle (3A / 125VAC).

What is still unknown

• The exact variant and markings on the switch from this unit
• Whether it was a Q (silver) or G (gold over silver) type
• How it behaves with the inrush of the supply it was switching

I'll update this once I've read the markings on the switch itself.

[exxos]

Curious...

PICT0014.jpg (397.61 KiB) Viewed 15 times

PICT0013.jpg (400.56 KiB) Viewed 15 times

PICT0012.jpg (417.32 KiB) Viewed 15 times

PICT0010.jpg (393.03 KiB) Viewed 15 times

PICT0015.jpg (418.13 KiB) Viewed 15 times

[exxos]

...and the money shot...

PICT0017.jpg (611.68 KiB) Viewed 10 times

PICT0016.jpg (597.56 KiB) Viewed 10 times

Can I goto bed now ?

[stephen_usher]

Not much isolation between the two sides of that switch. One spark and the atmosphere inside the cavity becomes a conductive plasma.

One thing that puzzles me is if the ground is earth referenced then assuming that the power cable is also earthed then why didn't the socket or even the house RCD trip as soon as it started to conduct?