5char

[Thompson11k]

5char

[Ian.M]

A bad neutral to your breaker panel or any other fault that could cause the supply to go more than 10% high could cause the transformer core to saturate, and if sustained it would rapidly overheat and blow its internal thermal fuse.  A high voltage mains transient could be sufficient to puncture the insulation between two turns of the primary, that are physically touching but electrically have a lot of turns between them, again leading to overheating and fuse failure.  Sustained overload + restricted ventilation could also direcly blow the thermal fuse due to overheating.   A failed diode in the bridge rectifier or failed amplifier output stage is another possibility.   Just about any good quality 9V transformer rated 15VA to 25VA could probably replace it, (Although the circuit only really needs 9.5VA, you need 5% more voltage than 9V so cant afford the droop that you would get if you run the new transformer at its full rated VA)  but I'd be a bit wary of such substitutions till I'd researched the chips used and reverse engineered enough of the schematic to determine its abs. max. supply voltage limit.

[Ian.M]

It sounds like the diode bridge is totally FUBARed.  Is there any chance at all of reading the part numbers?  You can confirm the polarity by tracing through to the +ve and -ve termials of the reservoir cap.

I wouldn't recommend trying to chisel off any heatsinks. If they are epoxy bonded, the IC will probably break first. 

However the system is so cheap and nasty (4W @ 10% THD !!!  ) that you may decide to simply patch in a cheap 9V 10VA transformer and take a gamble!

Treating it as if the speaker system was actually valuable, you need to make sure your fix never overvoltages it by any significant amount.   Take a very close look at the original transformer.  If you can identify the terminals of the thermal fuse that's in series with the primary and bypass it you may be able to get it to run long enough to take some measurements (not connected to the amp).  Start by measuring the primary DC resistance in case further tests fail destructively.  Then power it up on a dim bulb tester with a 40W or 60W incandescent bulb, with the secondary disconnected.  The bulb should not be glowing.  Get the exact primary and secondary voltages then soak test it for an hour to reduce the risk of an undetected intermittent shorted turn.  Then try it without the dim bulb tester, but with T1A fuses in series with the primary and secondary a 10R 10W resistive load on the secondary to let you get the voltage drop when close to fully loaded.  That gives you enough data to pick a compatible replacement, but you could go ahead and replace the bridge rectifier and try it on the amp, so you can measure rail voltages when muted and at full volume.  Just don't box it up and put it back into service as without a thermal fuse, the transformer is a fire risk.

Once you have a good idea of the rail voltages you can more easily find a compatible replacement.  Add a pair of antiparallel diode in series with the new transformer's secondary if you need to drop its output by about 0.5V or replace the ordinary diodes of the bridge with 3A Schottky diodes if you need to bring the rail voltages up a bit.   If the new transformer doesn't have a thermal fuse, leave the primary and secondary fuses mentioned earlier in circuit.

If you cant get the old transformer running well enough to get the unloaded voltage, the only thing you can do is treat that 9.5V on the label as a hard limit for the unloaded voltage of the new transformer.   A toridial + adding an overwind so you can boost or buck the secondary voltage a bit may be the best option.   

if you were paying a competent tech. labour on the above, you'd already be at several time the new replacement cost of the system so regard it as a learning adventure, and try not to get too far down the rabbit hole buying replacement parts!

[mariush]

You can find loads of transfomers at reputable distributors like digikey.com , mouser, newark.com

Digikey : 120 v to 9..10v AC 10-20VA : http://www.digikey.com/short/3905f8

Newark : http://goo.gl/VzWoMi

You have transformers there with two 115v primary windings, you just parallel them for 110v.
Most of them have two secondary windings... you can just parallel the windings for more current , lower voltage  or you can wire the secondary windings in series for higher voltage less current

If you want closest voltage this is probably the closest, with 9.3v AC @ 1.5A max : http://www.digikey.com/product-detail/en/tamura/PLT20-32-130B/MT3143-ND/182888 
Connect the primary windings in parallel, it has 2 x 115v windings. It has an additional secondary winding for ~ 27.5v at 0.22A  which you could just leave unconnected, just insulte the pins.

[Ian.M]

It looks like you are SOL digging for the fuse.

From the silkscreen, there appear to be two caps - a small nonpolarised one close to the bridge rectifier, and a larger bulk reservoir electrolytic cap above it.  Electrolytic caps aren't much good at RF and most bridge rectifiers are broad-band noise generators due to the time it takes to clear the charge carriers from the junction depletion layer resulting in a high dI/dt pulse when there is already significant reverse voltage across the junction.  The non-polar cap is to filter this noise.  You will quite commonly see a small cap filled across each individual diode in a bridge rectifier to filter the reverse recovery pulse at its source, which this design doesn't have.  As I said earlier: "cheap and nasty".

Although theoretically you could run a 220V transformer on 120V, its fairly crazy to do so - you'd need a 20VA transformer for 10VA output.  Going from 50Hz to 60Hz isn't a problem, it just means the transformer will have approx 20% iron in its core than an equivalent 60Hz transformer because it needs it at 50Hz to avoid core saturation. TLDR: doing that will need a bigger, heaver and more expensive transformer than a 60Hz 120V one.

You should replace all the diodes - anything in the lower half of the 1N400x series is suitable (i.e from 1N4001 to 1N4004). As the silkscreen shows the individual diode polarities you have no excuse if you get it wrong . . . .

I would expect about 9V DC across the reservoir cap at full load, maybe a little less of the original transformer is really crappy.  With the system in standby or with no input this could rise to about 12V.  If you've got a floating output bench supply that can do 9V 1A DC, you could connect it across the cap and power the speaker up so you can test to see if it has any other problems before spending money on a transformer and diodes.
 

[Ian.M]

That one's rated at 2.7A.  1A is only 37% of its rated current so assuming 10% regulation and that it can actually produce 9V @ 2.7A,  it will probably deliver about 10.4V.  If it was a 1A-1.2A transformer it would probably be ideal but you don't need both a 110V primary and more than double the VA.

However a couple of pairs of anti-parallel silicon diodes in series with its secondary could be used to drop the voltage a bit so its not a bad choice as long as you test its output voltage @1A load before hooking it up.