Bench PS mains switch arcs when switched on...

[jorgen]

I cobbled together a variable bench PS using ideas from Elecronoobs and Great Scott youtube channels and using an old 135W computer PS.

The build uses an old aluminum chassis I had laying around so allowed me to run everything chassis-ground and eliminate a lot of black ground wires, which was pretty necessary because the case was just big enough for the PS board, buck-boost board and a fan (had to accommodate a choke that was part of the original PS too).

Anyway, it all works as expected, the dummy-loads don't get hot, the voltages are stable, BUT when I turn it on, there's a noticeable arc (visible & audible) in the mains power switch... and when I power it down, it takes a full 4-5 seconds for the LEDs and Volt/Amp readout window to go out (caps bleed-off?).

So, right now I'm making/breaking the hot side of the 120vac to power it on, the original power supply didn't have a power switch, you just plugged it in and relied on the green-to-ground to power it up (I still do that as a secondary 'enable'), so I'm thinking that it was originally intended to be 'powered up' at the mains level all the time because it has a heck of an inrush when using a switch to power it from mains.

Does anyone have any ideas that would let me use this with a power switch without frying the contacts? How about bleeding off the filter caps? Right now, I leave the switch on and just plug and unplug the line cord.

The PS board was removed from an old Lenovo, it's an AcBel PS. I upgraded the fan to a 12vdc-40cfm unit, the only difference I can see in the installation is that I didn't include two 2200pf 250vac ceramic caps that were wired from both the hot and neutral sides of incoming power to ground since I used a new three-prong socket on the chassis.

That's all I've got, inrush arcing on power-up and slow bleed at power-down.

 

[Bluejets]

Mains switches will arc if any noticeable load is present.
It really depends on just what switch you are using for the mains as some are not really suitable.
Some may be open to seeing the arc that is always present anyhow.

As far as power down, it is more than likely due to the charged caps in the power supply and quite normal especially if there is not a lot of secondary loading.

 

[kellys_eye]

Do you have the specifications for the switch? They actually have a 'rating' and if you don't have the right type then it will arc 'uncontrollably' - potentially leading to damage and/or fire.

Sound to me like you're using a small DC-rated switch where an AC-rated one should be present and/or you're using one of those tiny 'toggle' switches that are rated at 125V and 1A.......

The actual rating of switches needs to take into account 'in-rush' current which, in a PC-type power supply, can be many amperes at switch-on.

Fit the correctly rated switch - and a double-pole one at that - it would be the proper safe route (rather than a direct connection to the mains) and could save your life one day.

 

[jorgen]

Yeah, thanks Bluejets.

On the power-down, there wasn't any secondary loading, just a couple of dummy loads on the 12 and 5 volt rails and maybe the fan, but that's pretty insignificant. Maybe about 2w per rail.

As for the arcing, I used a couple of rockers that are supposedly rated for 10A @125VAC, but they were left over from a project I did about three years ago and were just cheap Chinese rockers that you get for five bucks a bag on ebay, OK for signal voltage stuff but probably pushing it for mains.

Because I'm already committed to the front-panel layout, I'll probably leave the existing mains switch alone and just run a better quality switch to the back of the chassis, don't really have the real-estate on the front for any further modification.

@kellys_eye: Appreciate your sound advice but as you see in the pics, I've made sure to use a properly rated (if not good quality) switch. Also, here in the US, there's only one hot leg on domestic power, so a DPST isn't necessary (neutral and ground are at the same potential at the breaker box).


Anyhow, thanks again.

 

[dave9]

You could experiment with putting an NTC thermistor in series between the switch and bridge rectifier.

There may already be a bleed resistor to drain the capacitor(s) but you could make it a lower ohm value, taking care to choose a suitably larger (wattage) size if needed, or just not care that it takes a few seconds to drain the caps.

 

[jorgen]

Yeah Dave, I think I'm just going to leave it alone and just go for a switch I have a little more confidence in, these no-name Chinese 'bag-o-switches' don't really give me the 'warm and fuzzys', especially after arcing out from around the bezel with the same 'snap' of authority as lighting up a 7018 rod on rusty steel (maybe just a little exaggeration there...).