I am trying to solve an SMPS that keeps blowing it's MOSFET.
I bought the Blue Ring tester from Anatek. I tried contacting them but they do not answer the phone and the forum is dead. Will not even let me Log in once I have registered.
When I take the transformer out of the board and test the primary I get full green lights.
When in the board I get no lights indicating a short.
When I take out the 2 1000uF 100v Filter caps from the secondary side I get full green lights.
However, these to caps pass ESR, Capacitance, Multimeter charging, and even a 30v Leak test. I believe they are good.
Is it possible they are causing a false reading of short on the Blue Ring tester?
You did not tell us what SMPS it is.
Low voltage, mains powered, ??
I assume mains powered, because of the 100v caps you mention.
A SMPS can be a difficult beast to repair.
I had mosfets fail because of failing parts in the snubber network.
Leo..
I only have an IPOD and I know the pics are way to blurry from trying to use it before. However, I Tested all the caps first thing and they are good. I actually remembered seeing a video on how these things burn up without the snubber network. I thought to myself....Why did I never check that main Diode. It is one of the first things I usually check. I can't believe I forgot that. So I went to get the board and look for that diode. I couldn't see one on the primary side at all. So I started looking at the labels on the board. D5..D4 ...D3...D2.....nothing. WTF? No D1?? Finally, I find D1 labeled on the board right where it should be. Someone put a large value large wattage resistor in it's place. I have no idea what kind of current this diode needs to take so I ordered a 400v Shotkey with the highest amperage I could find that was still in the axial package like the board is meant for. Also knew MOSFETS and a replacement Metal film cap that looked discolored that was part of the network. I hope it works this time.
How did you measure the capacitors.
You can't use a capacitance meter for that.
With SMPSs the internal RESISTANCE of the capacitor is more important than the actual capacitance.
Only an ESR meter, like the "blue ESR", will reveal that.
That snubber network must have a capacitor. usually high voltage ceramic.
Discoloured means trouble.
If that cap (partially)fails, the mosfet blows randomly.
Leo..
You generally cannot test transformers in circuit. It works by firing a pulse of current into the winding, and seeing how long it takes to ring down to about 70% voltage. With external parts on the transformer, it loads it down and it may not ring at all.
This is the circuit for it, it just counts the number of times it rings with a 4015:
Just an update, I had to take the filter caps out of the secondary and then the blue ring tester gave a great result for a good 2ndary. Also My blue ESR passed the burned cap but I'm changing it anyways. Burned is bad IMO.
Also...I could not find the main diode for the SMPS snubber network to test it. Finally found where the board was labeled for it but some one put a large value resistor in there instead so that is getting replaced back to a diode again.
Ooops. I should have said no snubber diode. I'm worried that he thinks that any coil connected to a transistor gets a snubber diode like a relay or solenoid would.
I only have an IPOD and I know the pics are way to blurry from trying to use it before.
Okay Jarrod, pity, pics help a lot.
I find D1 labeled on the board right where it should be. Someone put a large value large wattage resistor in it's place.
Has this SMPS been repaired before?
Where in the circuit does D1 go, a quick reverse engineering circuit diagram around D1 should give some indication, I assume it is in the AC mains/DC bus area.
Yes I wish I had a good camera too. The D1 goes parallel with the primary winding across the + and ground bus but of course is reversed so that it will only junction when the "Switch" is off and the primary needs to discharge the inductive spike through it. Someone put a large value resistor in it's place and I do not see how that could be a good thing. It has been worked on before and I have no idea what parts were changed or if the presets were changed or anything like that. I am mostly just working on it because It wont matter what I do now as it is pretty much junk on a low value amp. I need to practice my SMPS skills because I'm just learning them. I just got the parts in the mail and will let you know weather it works now or not. I would not hold my breath though
A switch mode power supply does -not- "discharge the inductive spike". That inductive spike is how it works! When the primary current is shut off and the magnetic field begins collapsing, a voltage is induced in the secondary and that energy ends up as the output of the power supply.
I ask again - has someone else worked on this before you? Do you have a schematic? What makes you think that should be a diode and not a resistor?
No schematic, Yes someone has been in here before, Missing screws, wrong fuse etc.
I know the field collapses and sends energy to the secondary side but if you don't have a diode you get smoke. There was a really awesome video on YouTube with a breadboard showing what happens without the diode and then with one but of course I can't find it now. I did have this video which shows the diode and explains the basic concept.
BTW, I know one is supposed to go there because the spot on the board is labeled D1 and that is the spot where the diode would normally go as per the basic design of a buck converter etc.
Are you sure you mean "transformer" and not "coil"?
If it is a buck mode SMPS, there is no primary or secondary, just one coil. And that diode does not go across the coil, it goes from one end of the coil to ground.
if it has a transformer, it is a flyback converter and where there are diodes depends on the exact type. Some have diodes on the primary, some do not. All have diodes on the secondary.
So you have determined that it -is- supposed to be a diode, based on the silkscreen.
Do you have a schematic? Model number? As previously requested, photos of both sides of the board? It doesn't sound like it takes AC line voltage, since you mentioned 100V capacitors on the input side.
Been reverse engineering it a bit because I can't find any drawing. Board number is A4824A. or perhaps HYPOWER-JBL-12. Turns out I made a mistake in my original tracing. My eyes are getting old. It was supposed to be a resistor. Is part of a voltage divider that feeds the primary winding on the transformer. It's definitely a transformer. Has multiple taps on the secondary side for various voltages. Most Importantly the + and - 60V I can't coax out of it.
I can see your as passionate about these things as I am and I appreciate your taking the time to try to help me. I am unfortunately much better versed in Linear supplies.
Basicaly 120VAC comes in through a Switch and Fuse and some EMI coils and caps. Gets rectified then dropped across 2 4700 uF Caps in series which sets up 3 rails, +, -, and "Hot Ground" The plus side goes through the primary winding of the transformer and then gets switched to Hot Ground by the MOSFET. That MOSFET is controlled by KA3525 SMPS controller IC.
This thing came to me AFU. Here's is what I found so far.
Main power MOSFET was blown, Large Film Capacitor burned, 1 of the 2 electrolytic filter caps for the main rectifier puked it's wax out, Main PWM chip (In this case SMPS controller KA3525) was fried, and 1 of the 4 schottkey diode packages on the secondary was toast. Replaced all of that stuff.
Also I decided to investigate a jumper that was installed on the board when I got it and it went directly from cold ground to hot ground!! That would explain why I was blowing fuses and drawing a lot of current even when there was a tiny load. Removed that....Short is gone.
Now back to the main problem. The MOSFET is not switching yet. KA3525 is not pumping out a pulse yet. The Shutdown pin has 0 volts. However, both the + and - Error input pins have exactly 2.4 volts. Which is fed to a comparitor inside the IC so this would result in no output pulse I believe. Need to reverse that section now.
Thank You everyone who has helped so far. Some people just tell me if you don't have schematics just quit. Never LOL!!! I'm winning and I'm not going to stop until I achieve Supreme Victory
Regarding using a Diode to discharge the primary winding after the charge circuit opens. There isn't one. What is happening is that there are 2 power MOSFET's. They open and close in an alternating pattern. One opens and closes the primary winding. When it opens, the other MOSFET closes and takes the inductive spike to ground thus preventing damage to the components in the primary side. These 2 MOSFET's gates are driven buy to opposite phase output coils of a small transformer which the primary is driven by the tiny daughter board where the SMPS IC provides the PWM from. I already know it wasn't running but just wanted to get a more clear understanding of the circuit in between.
There is a toroid coil in series with a film capacitor and the primary winding. I'm wondering if it's the toroid that stores the energy and then releases it into the primary winding of the transformer. That would be more like what you suggested before. Perhaps it can hold more energy in it's field then the coil of the primary can by itself.
Update: I thought the KA3525 was not outputting it's pwm but I was wrong. It seems to take a little while to start. The soft start cap is not the value advised in the papers for KA3525. I might look into that soon. It's not making it to the Gates of the MOSFETs but it has previously because they keep blowing up on me LOL. I already tested the 4 transistors on the daughter board that I believe drive the MOSFETs. They are very tall and I think I read they were darlingtons. They passed the Peak DCA 75 tester so I think the problem is just 3 presets that I was playing with when I was trying to get the shutdown voltage adjusted down below it's threshold. Much has happened since then so I will have to trace out that part of the circuit now.
Alas....I'm out of time. Back to work in 12 hours. Need to sleep. If I ever solve this son of a gun I'm going to make a movie out of it LOL.
Hi,
Any reason to repair this apart from gaining experience, if need to be put back in circuit then I would forget it and replace it.
The minute you said that the two filter caps had fried as well, in my opinion, if that is all that went ok try and fix, but with what sounds like the entire DC bus circuitry has shorted, then it not worth the stress.
Starting a smps is something of a black art, various manufacturers have different ways of doing it.
Please check that with the MOSFET out of circuit that you can on the PCB measure some resistance, Gate to Source, this keeps the MOSFET turned OFF until the IC turns it ON.
I have returned damaged $800 SMPS as unservicable to customers, because even though all the parts are available, with the number of components directly damaged and possibly indirectly damaged I cannot guarantee the repair.
Those MOSFETs are being used as synchronous rectifiers. It may come as a surprise, but they are actually wired up backwards regarding source and drain current flow. That is because there is an internal inherent PN junction that is normally reverse biased in normal operation. But if you tried to use it as a synchronous rectifier with (conventional) current from Drain to Source, the internal diode would be in conduction when the MOSFET was off.
Fortunately, we take advantage of the fact that a MOSFET is really a voltage controlled resistor. Current can flow either way when it is turned On. So we wire it "backwards", with current flow in opposite the usual direction.
Neither MOSFET is "tak(ing) the inductive spike to ground thus preventing damage to the components in the primary side." It is a centertapped secondary, with the MOSFETs acting as diodes in a full wave centertapped rectifier.
Because a MOSFET in saturation has a resistance in the milliohms region, and therefore much lower voltage drop than even a schottky rectifier.
The only thing absorbing any spikes is the output of the power supply. You want all energy from the transformer to end up in the output. There will always be some that is lost, that is unavoidable.
Here you can see synchronous rectification driving the gates from extra windings on the secondary. Notice that both MOSFETs perform exactly the same function. There is a series RC network to absorb very fast spikes, but everything else ends up in the output capacitors.
It is important to remember that the gate drive is -still- with respect to the Source terminal, and must be positive with respect to the Source for an N Channel MOSFET.
Note the direct resemblance with this full wave centertapped power supply:
Here is a reasonably good explanation of a 250W SMPS using the SG3525.
