Switch mode PA Amp PSU fault

[Yorkie]

At our woodturning club, of which I'm secretary, we use a PA system - a Chiayo 'Focus 500' with a remote UHF radio mic. Quite a commonplace system and not a cheap affair at about £500. (It was paid for by Awards for All as part of a £3,500 successful bid I submitted for lottery funding five years ago so is well out of warranty). It has a built in 12V SLA battery to make it truly portable, and an external PSU/Charger. We only use it once a month so the battery is permanently run down, but when we plug in the PSU it works happily from that. (The Power supply is 240V AC,1.5A / 15V DC,4.3A output).

http://www.aandcaudio.co.uk/chiayo-focus...em-p131490


A new battery is £48 inc VAT – not that we need one, and if we did, no doubt a cheaper source could be found.

http://www.aandcaudio.co.uk/battery-for-...em-p271541

Last night, there was a pop when I switched it on, and on checking the switched mode PSU it had no output. I stripped it this morning and found a dry joint on the PCB where a 120uF 400V electrolytic cap was connected. It looks like it's been arcing at the dry joint as there are signs of overheating on the PCB. Unsurprisingly, the el cap has an infinite ESR and reads 47pF on a capacitance meter, so is obviously duff. I've ordered a replacement cap (off e-bay - where else?), which should be here tomorrow.

I know little about switched mode PSUs. I checked for continuity at the mains input socket, and it's open circuit. On a conventional PSU that would suggest that the primary winding on the mains transformer was open circuit. However, on an SMPSU it doesn't have a mains transformer as such - it does have a couple of inductors wound from quite heavy gauge wire - I'd guess about 24 - 26 SWG, which look OK. The question I'm asking is would we expect to see an open circuit at the input of an SMPSU, or an Ohmic reading? It's quite a complex PCB with many components including surface mount, so really, if this cap doesn't sure it, it's beyond economic repair.

I dare say I could source a new PSU, but if I can repair this one, I'd prefer to. I'll soon find out when I replace the cap - it will either work or it won't! Hence, my question is out of curiosity rather than desperation. Any thoughts anyone please?

[Joe]

My lap-top PSU reads 2MΩ across the input.

- Joe

[Radio Fixer]

More than you need to know here David http://en.wikipedia.org/wiki/Switched-mode_power_supply

But I'm not surprised that you dont get an ohmic reading. AC, according to above, goes straight into a half wave rect or voltage double for 120V operation, probably with auto switching. So dont worry 'til the cap doesnt work.

I did work at one time on DC SMPSUs and they were a B when they blew! took out semiconductors all over the place.

Gary

[dennishoy]

David, there are several types of switched mode supplies of varying complexities. The cheap Chinese junk type (laptop/phone/etc.) are honestly not worth repairing and you sometimes have to destroy the case to get it apart.

However, this sounds like a 'middle of the road' item which should have protection circuitry in the event of component failure. Normally, the mains supply is full-wave rectified and applied to the 450V electrolytic capacitor to provide a DC source for the rest of the unit. I suggest you locate the bridge rectifier that feeds that capacitor and check that both arms are still ok as you need full wave for the PSU to work.
Also check for a blown fuse on the pcb.
If you replace the capacitor, it will need to be a high ripple current rated type.
It would also checking the power device on the heatsink for short circuits. They normally go S/C and not O/C.

The dry joint(s) suggests this may not be a high quality item. It would be worth inspecting the rest of the board for similar joints. I've found them where the holes for the wound components are to big for the wires to allow for differences in dimensions and the expansion and contraction of the paxolin board couple with the mass of the wound component eventually leads to dry joints. This is especially true if the pin in exactly in the middle of the hole with an air space around it. They all use jigs these days, so it shouldn't be difficult to make the holes in the board fit the components correctly.

When happy, perhaps you should power it up with a dummy resistive load before taking voltage measurements as these supplies often need a load to regulate and won't work under open circuit conditions. If you take 100mA from the supply (12V bulb ?) that should be fine.

[Yorkie]

Thanks for the rapid respose to my query guys. I'll see what tomorrow brings! If a new cap doesn't sort it, I'll get back to the supplier to ask if they can supply a replacement PSU. The case came apart with just two screws and overall, it does look quite well built and designed, and rather more complex than I might have imagined. I guess that a lot of the 'gubbins' is to do with the battery-charging circuitry.

[Nick]

With Dennis on this - 99% of such SMPS units are of a common, simple, architecture...

They run the mains through a common & differential RF filter (mainly to prevent cr*p from the SMPS getting onto the mains, not the other way round!) which is then rectified by a full wave bridge. This is then smoothed, chopped at high frequency by a PWM control chip, fed into a transformer (normally via a FET or two), the output of which is your desired voltage that tends to be half-wave rectified, then filtered (to remove RF as much as to smooth) and connected to the output lead. A second secondary winding is used to provide feedback to the SMPS PWM chip via an optocoupler - if you look at the underside of the PCB you'll see distinct HV & LV sections, often with a routed cut between them to increase safety (stops condensation bridging the gap) - the optocoupler and the transformer should be the only things bridging the HV/LV divide.

The PWM chip has a number of safety features that shut off the supply in out-of-bound conditions, such as the output getting too high.

Replace the obviously dead capacitor, but ask yourself: "Why did it go"? Before powering up, check the bridge with some low voltage AC input and see if its giving full-wave rectified DC - if not, check the bridge (or the discrete diodes if that's what's being used).

Other things to check - the single diode on the LV DC side and the caps round that - some of the better small SMPSs have a crowbar protection system that shuts off the output in emergencies - check the HV side, then the LV side. There can be input and output normal or polyswitch fuses too..


The transformer primary should be centre-tapped - check it and check the output windings too - sometimes (rarely) the xfrmr dies... if the PWM/SMPS control chip, the transformer or the optocoupler has gone AWOL, chances of fixing the supply are slim.

HTH

[Yorkie]

Thanks for the info Nick. I think the dry joint on the PCB has cased arcing between the track and the lead-out wire of the cap. The cap itself isn't swollen and has no physical signs of any problems, but over time, (five years), as is all too common, it has developed a high ESR, which I think SMPSUs are particularly intolerant to. In another thread I mentioned that I had a TV antenna amp which had developed a fault. It had a simple conventional PSU and a separate amp unit in the one case. The electrolytic cap on the PSU had a high ESR and when I fitted a new one, the amp was fine and is now back in use.

I'll report progress, and I hope it will be good news, but I'm not going to spend too much time on this - the amp is still in production, so I'm assuming that a replacement PSU will be available.

[AlanBeckett]

David,

Might I respectfully say 'Pillock'.

Many moons ago the Company I was working for had supplied some gear to another Company who we didn't normally deal with. The gear included a NiCad battery which was float charged from the supply. After a couple of Months they complained that all the Batteries were dead. At the time we were having some problems with supplies of Cells, so we sent new Batteries. After another couple of Months they complained again. It turned out that they ran the gear for a few days then turned the power off and went away for a couple of weeks to do some more design. Put simply that was not, and never was, what it was designed to do.

I would suggest that you are/were in a similar situation.

Alan

[dennishoy]

I forgot to mention that some untis have fusible resistors which have caught me out in the past. These were colour coded as ordinary resistors, so I replaced them as such only to find out later (when I had the circuit) that they were safety devices. These were low value items, highest was 4.7 Ohms, lowest was 1 Ohm.
Don't get caught out if you find an O/C 'resistor'; it may not be quite what you thought.

[Yorkie]

David,

Might I respectfully say 'Pillock'.

Many moons ago the Company I was working for had supplied some gear to another Company who we didn't normally deal with. The gear included a NiCad battery which was float charged from the supply. After a couple of Months they complained that all the Batteries were dead. At the time we were having some problems with supplies of Cells, so we sent new Batteries. After another couple of Months they complained again. It turned out that they ran the gear for a few days then turned the power off and went away for a couple of weeks to do some more design. Put simply that was not, and never was, what it was designed to do.

I would suggest that you are/were in a similar situation.

Alan

None of the above has any relevance. The instructions clearly state: "While charging, the unit may be operated simultaneously". There are no cautionary notes about only using the unit when the battery is fully charged. All of the equipment - projectors, lathe, PA system, projectors, screens etc, is stored in a village hall. It's used once a month for two hours. We would have preferred a small PA system which is simply mains powered, but at the time of purchase (when we consulted a national audio visual equipment sales and hire company), no suitable alternative equipment was available. The SLA battery isn't fully discharged when we set the equipment up, but only lasts for about half an hour, so we plug in the PSU at the outset. Short of someone (ie me,) taking it home to store and re-charge it each month is readiness for use, we use it in accordance with the instructions - 'simultaneously using and charging'. (The PSU is rated at 4 AMPS).

(It is jiggered by the way, so I'll be sourcing a replacement).

Thanks guys for all the advice and guidance, but having checked the bridge and replaced the el cap, there is still no output, and as both sides of the PCB, with a high component density, is peppered with SMDs, I have neither the time, expertise nor inclination to furtle about with it any further. IMHO, five years for a PSU from China sounds about par for the course, albeit it looks well constructed, and I don't label everything that comes out of China as tat).