Danger, danger, SMPS

[cosmicray]

Today at the rubbish pickup, someone handed me a non-functional Epson Artisan 50 printer. Confirmed that it was dead, so decided to take it apart and examine the SMPS. Not obviously the cause of death, but the first suspect. The SMPS was buried under the printer assembly. It took forever, to find every last screw to get that printer assembly out (and in the process I probably inflicted damage). The SMPS was under a plastic cover, held by two screws. Lifted it out, flipped it over a few times, went to set it down on the table top and brushed the leads on the X2 cap. I felt that one ! So, even after not having had any power applied for 2 hours, the cap still measured 148V DC. 

That fuse looks visually OK, but I'm not going to continuity test it until I bleed off ever cap in sight.

In the future, I will give SMPS (dead or alive) much more respect.

[tautech]

 
Yep they bite.
Commonly they don't have a bleed resistor across the main cap and if the switching circuit is not working that cap will retain rectified mains for quite a while.
First port of call on SMPS is to discharge the primary side cap.

You've just earnt your Learner plates. 

[IanB]

brushed the leads on the X2 cap. I felt that one !

You main that big brown cap sitting on its side? That's not little, that's going to pack quite a punch!

[cosmicray]

brushed the leads on the X2 cap. I felt that one !

You main that big brown cap sitting on its side? That's not little, that's going to pack quite a punch!

Yeah, the one slightly right of center. 30 mins or so after touching it, it's down to 140V. What a relief. I may leave it sit overnight just to see how slowly it drains on it's own.

ETA: the one I got hit with is, I think 160 uF @ 200V. The value are turned facing the board, but I can see parts of this where the label wrap turns over the end of the cap. The other one, near the output, says 1500 uF @ 50V.

[electr_peter]

One more reason not to touch any unknown metal surfaces with bare hands.

On another forum there was thread about an idiot who tried to measure temperature in grossly overheating (with a burn smell) live SMPS by touching top metal surface of mains input cap. All good so far...

[tautech]

brushed the leads on the X2 cap. I felt that one !

You main that big brown cap sitting on its side? That's not little, that's going to pack quite a punch!

Yeah, the one slightly right of center. 30 mins or so after touching it, it's down to 140V. What a relief. I may leave it sit overnight just to see how slowly it drains on it's own.

WHY
Make it safe(discharge) and move on.
Just leaving it charged on your bench is not smart. Visitors, Kids, or worse...the wife.......

Many just use a screwdriver, I suggest you do that once just to confirm how much energy is in it, expect to get a good spark and a fright.
OR better for the cap, an incandesant mains light bulb with some old test leads attached.

[PeterFW]

In the future, I will give SMPS (dead or alive) much more respect.

You should be carefull with old xenon flash units too, learned that the hard way before i knew what a capacitor even is.
Had a big round 10mm hole/burn mark on my hand after that, my hand seized shut with a firm grip around the cap.
It dit feel like it took ages for it to discharge, the feeling i will never forget...

Just discharge it, do not wait for it, it will make a nice spark
Or use a resistor...

[cosmicray]

I am discharging it, with a pair of alligator clips, and one of those red-green-green lights that tests that the wall socket is wired correctly. The center light was green until it dropped to 60V, then it went out. The voltage is continuing to drop.

[cosmicray]

Main logic board from Epson Artisan 50

7 ribbon cables, plus 7 wires plugs

[hamdi.tn]

My guess is the SMPS switching IC is dead or not powered , i had to fix some SMPS ( laptop charger ) and most of one i saw they don't have a bleed resistor ,if the circuit is still operational the switching ic will continue to operate to some low voltage. i don't know what kind of regulation apply to those charger but even genius charger i cracked open were not that good and you can feel the crapy Chinese quality , in the other hand the HP one are awesome built quality and protection

[Davy]

I always check the main smoothing block on SMPS..... you can discharged them with a 470 Ohm  ~ 1K resistor rated at a couple of watts, they can easily be adapted for this purpose, lately I've been using a 240V 15 Watt sewing machine incandescent lamp - just be sure the filament ain't o/c, 60W incandescent lamp can also be used.

Remember, I'm in the UK so 240V  AC is the norm here. Yep, I felt em' at the elbow too and it hertz!!!!

Dave

[retiredcaps]

I have probably worked on/repaired more than 50 consumer electronic devices in the last 5 years and so far 2 of them kept their the big filter capacitor fully charged over a long period of time. 

On the LG 2234S, it kept the charge overnight (165V).  On the Motorola DCT-3416 PVR, it kept the charge over a 3 week period (165V).  Both devices were disconnected from the AC outlet.

Both SMPS were working and only the DCT-3416 had obviously bloated capacitors on the SMPS.

Now, I always check the voltage across the large filter capacitor before doing any work.

[wagon]

I repair a lot of electric fence energisers, and I always check the main cap.  It's not uncommon to see 600-800V  in them.  I've been bitten a couple of times and I can assure you it's in no way pleasant!  I discharge them with either a lamp, a resistor or a screwdriver (CRACK!).

Please remember also that if the mains supply is 230Vac (or more) you'll have 330Vdc or more on the main capacitors off the bridge rectifier.

[SeanB]

I use some 56k 10W ceramic resistors to discharge caps in high voltage areas, just because i have some and they are otherwise not going to be used. Great for putting across the terminals of any capacitor before using the meter to check it, to make sure that dielectric absorption ( bounce back) has not generated close to 100V across the terminals.  Also use them after the high voltage test, a failing cap might measure close to correct but apply 1kV for 30 seconds and you can hear it crackle inside.

[lightman]

Hi
You where lucky that you have 110V mains, if you had 230 like I have, and touch a charged mains cap with 300V it would hurt like hell.
I disagree about discharging it shorting the pins, use a low value mid power resistor, I usually have a 47 ohms 10W with leads and aligator clips to do just that.

[simplifyiyuk]

almost all Epson SMPS' i've worked with remained charged for a very long time, not unusual for newbies to get a sizeable arc when reconnecting to the main board when reassembling. that’s always fun to watch. If your SMPS is reading ok check the clearances of the legs on through hole components to the shield (seems a popular failure point) and check any black marks on the shield. hell it could be as simple as the button doesn't work, enjoy.. 

[electr_peter]

Another quick and easy method for capacitor discharging - use DMM with LoZ functionality. It will provide both 1-3kOhm resistance and indication of voltage.

[albert22]

Relatively new power supplies with PFC tend to use a 400Vcc power rail, independently of the mains. Which could be from 90 to 230Vac in most of them. I repaired 1200W power supplies that had six 270uF capacitors, in parallel, charged to 400V. And they take a long time to discharge by itselves, specially if the switching transistors are dead.
You can easily calculate how many Joules are stored in a such a bank of capacitors and surely that amount of energy is DEADLY (at that high voltage). I would not dare to discharge that with a screwdriver.
Beside the main bank you can find an auxiliary power supply in the same board with another set  of charged capacitors.
Nevertheless extreme caution is required.

[SArepairman]

be sure to read about dielectric soakage.


My 350V 10000uF capacitors on ebay came charged to 50V. I studied them carefully before so I knew they might have a charge (since they came with no shorting bar) but I was not expecting quite such a spark!

[cosmicray]

be sure to read about dielectric soakage.

I watched a minute amount of that happen on this cap. The discharge of the cap was finally accomplished with a 150K resistor. It took a few minutes, due to the small current flow, but eventually the meter showed effectively zero. I removed the resistor, but kept monitoring the cap anyway. Slowly it drifted up towards 1V. It was almost like watching a car rolling uphill.