Apple power supply A1098

[Porto]

Hello,

Got this faulty Apple Cinema display 23" power supply for a few bucks and can't seem to find the main problem with it; hell, am still in the learning phase!

It's a SMPS with 100-240v, 50/60Hz input and 24,5v 6,1A (150w) output.


I found a shorted mosfet on the secundary site, the PSMN015.
Nothing else seems out of the ordinary by look and smell.

What I've done so far:

- Changed the shorted mosfet for a similar one --> no change.
- Checked the main fuse --> still good.
- Checked the main rectifier --> still good.
- Checked the voltage on the main filter cap --> 320v.
- Desoldered and checked the two 11N60C3 mosfets on the primary site for shorts --> still good.
- Checked the voltage on the primary transformer pins --> 320v on 3/5 pins.
- Checked all the primary diodes --> still good.
- Checked the two optocouplers --> still good.
- Desoldered and checked all electrolytics for capacitance and ESR --> still good.
- Changed two small startup electrolytics --> no change in Vcc voltage to the TEA1552t IC.


The primary has two mosfet driver IC's, the 14 pin TEA1552t and the 8 pin L6561D.
The secundary only has the 8 pin N4300 op-amp / voltage regulator.

The TEA1552t has 7.5v on the Vcc pin 8 (too low, has to be at least 11v), 320v on the DRAIN pin 7 and a very fluctuating 0-5v on 5v Vcc output pin 11 (obviously hangs in some kind of startup loop).

The L6561D has no voltages present on any of the pins.

The N4300 op-amp / voltage regulator also has no voltages present on any of the pins, because the whole secundary circuit is not getting any voltage.

My suspicion is that the TEA1552t IC is malfunctioning, hence the low Vcc input of 7.5v. As I understand from the datasheet, especially the block diagram and the text just beneath it, the IC is self sustaining and draws his Vcc from the mains on the DRAIN pin 7; this voltage is internally lowered to the correct voltage by the means of an internal cap (just suspect this, don't know for sure). I'm awaiting a new IC I've ordered to replace it.

I can't understand why only the mosfet on the secundary site has shorted out and not the mosfets on the primary site; I would guess these guys will catch the first blows, or am I wrong?

If anyone have any more ideas / suggestions what to check, I'll be happy to hear.

If better pictures are needed, I'll be happy to post them!


Some pictures:

[stj]

Re: Apple power supply A1098

check the bleed resistor(s) that supply power rfom the main cap to the startup cap

[Porto]

Re: Apple power supply A1098

I don't know exactly which are the bleeder resistors so I measured alot of (SMD) resistors in between the two components; some I had to desolder because of weird measurements but off-circuit they were good.

Any more things to check?

[momaka]

Re: Apple power supply A1098

If you have a spare, ungrounded/non-grounded wall wart or power adapter, you can supply the necessary Vcc voltage to the TEA1552t to see if the standby section will start or not.

To do this, though, you must connect the (-) of the power adapter to the (-) of the main (primary-side) capacitor, and (+) of the power adapter to Vcc. Also, I encourage you to use the series light bulb trick if you perform such tests to avoid blowing up parts.

Series incandescent light bulb trick:
https://www.badcaps.net/forum/showpo...4&postcount=70

[Porto]

Re: Apple power supply A1098

I presume doing this with the board live on 230v also?

Never done such trick because I'm always afraid to toast things.

I don't have an ungrounded power supply but I can uncouple the ground from the wall outlet because I use a specially installed fused outlet for testing mains equipment, without blowing the fuse in the main house fusebox when something is shorting to the mains input; this way I can use my lab bench as power supply with just the local negative line used.

[momaka]

Re: Apple power supply A1098

I presume doing this with the board live on 230v also?

Correct.

Never done such trick because I'm always afraid to toast things.

That's why I suggest the series incandescent bulb. Even if you have a special wall outlet with fuse and whatnot, that won't save the board if something goes bad. On the other hand, the series incandescent bulb will because it will limit the power to whatever it is rated for (i.e. 60W for a 60W bulb). No explosions or blowing up fuses.

I don't have an ungrounded power supply but I can uncouple the ground from the wall outlet...

That will work too.

[saularciga]

Re: Apple power supply A1098

Hi, change the ic L6561 and... ready!!!

[saularciga]

Re: Apple power supply A1098

Diagram here:
https://cdn.badcaps-static.com/pdfs/...eb6fd947eb.pdf

[Porto]

Re: Apple power supply A1098

Hi Saularciga,

Thanks alot for the schematics, I searched all over Internet for it but thought that it was never was published at all!

Do you also have the schematics of the A1096 and A1097 power supplies?


I have already ordered those L6561 IC's, so was already planning on changing it for a new one.

Again, thanks alot for the schematics... it's of a BIG help!

[Porto]

Re: Apple power supply A1098

Well, in the meanwhile, I have replaced the MOSfet for the original PSMN015-110P and the L6561d Driver IC.

PSU still don't work.

I still keep 7.5v on pin 8 of the TEA1552t IC.

If I remove a little 25v 47uF electrolyte cap which is directly connected to this pin 8 of the TEA1552 IC, I presume this is the startup cap, I get a voltage which is fluctuating between 8.5 and 9.65v,
so I think this more a filter cap. If I then touch any pin of the TEA1552 IC with my DMM probe, I hear something (the transformer?) making a high pitched sound, with the cap installed the sound is gone and the
voltage goes back to a quite stable 7.5v.

Anyone got any suggestions what to check next?

[Porto]

Re: Apple power supply A1098

Sorry, forgot to mention I changed the TEA1552t IC also for a new one.

[momaka]

Re: Apple power supply A1098

PSU still don't work.

I still keep 7.5v on pin 8 of the TEA1552t IC.

I think that is a bit too low. Datasheet says VCC needs to be 11V (typical). Anything below 9.3V could trigger an undervoltage (UVO) lockout.

There is likely still a problem elsewhere in the PSU. But still, I would say try what I mentioned in post #4 - insert about 12V on pin 8 (VCC) with external ungrounded power adapter. And have a series incandescent light bulb with PSU mains, in case it tries to do something funny (i.e. if something shorts out).

Or perhaps before trying any of that, did you check the output rectifier/diodes(s)? Those would be D8 and D9, and Q11 and Q12 (Q11 and Q12 appear to be used as synchronous rectifiers) in that schematic that saularciga posted above.

[Porto]

Re: Apple power supply A1098

Hi,

I tried to inject a 11.5v voltage to that pin 8 but I think it blew the IC because now I don't get even the 7.5v anymore, just a few milliVolts. But as I mentioned before, I suspect the Drain pin 7 Mains voltage is feeding the Vcc pin its needed supply voltage because when I lift that pin 8 and measure the voltage on the board there is none, but now something weird, when I then measure the voltage on that pin there is also no voltage present; only when I reattached that pin to the board the 7.5v has come back so I guess the voltage is coming from the chip itself (see picture, on top).



Unfortunately, that schematic is not this PSU, it's a similar of one which also uses the TEA1552 and the L6561 IC's together... Pity, was too good to be true because Apple schematics are very rare or you have to pay lots of $$$ for it.

BTW: that bulb trick does work good, I got no short whatsoever, bulb stayed off... tried a 60w and 100w bulb.

[momaka]

Re: Apple power supply A1098

But as I mentioned before, I suspect the Drain pin 7 Mains voltage is feeding the Vcc pin its needed supply voltage because when I lift that pin 8 and measure the voltage on the board there is none, but now something weird, when I then measure the voltage on that pin there is also no voltage present; only when I reattached that pin to the board the 7.5v has come back so I guess the voltage is coming from the chip itself (see picture, on top).

Yes, the way it works is this:
When you plug in the power cord, the main cap charges to ~170 - 380 V DC (depending if the PSU has APFC and/or the main voltage in your country). This voltage is travels through the primary winding on the main transformer and so is present on pin 7, Drain, of the TEA1552 IC. The TEA1552 IC uses this voltage to charge the electrolytic cap attached to pin 8, Vcc. Once the voltage goes above 11 V (typical value), the TEA1552 then uses the power stored in that cap to turn ON the main MOSFET in it very quickly (to minimize turn-ON loss). This makes a current flow through the primary winding of the main transformer, which in turn produces a current flow on the main winding on the secondary side and also the auxiliary winding on the primary side. The current from the auxiliary winding usually gets rectified through a diode and then is used to charge that same cap that the TEA1552 charged through the Drain pin.

So in essence, once the power supply begins working, the Vcc for the TEA1552 is self-sustained through the primary side auxiliary winding of the main transformer. If there is a problem, however, the voltage and currents going to pins 2, 12, 13, and 14 (Isense, LOCK, CTRL, and DEM) are used to shut down the TEA1552 IC.

So the fact that the power supply is not working with an externally-supplied Vcc voltage suggests there is something else wrong with the power supply.

Even if that schematic is not the same as your PSU, see if at least some sections match your PSU somewhat closely. But even without it, the main things to check are if there are any short or open-circuited components on the primary and secondary sides - in particular check that the main output voltage rail of the PSU is not shorted to ground. From there, start checking any diodes and transistors and high-power resistors. If there are any SMD ceramic capacitors (these are small tan-colored rectangular components on the solder-side of the PSU), check them as well. They can go short-circuit for no reason sometimes. If in doubt about a component's condition, remove it and test it outside the PSU. Let me know what you find.

BTW: that bulb trick does work good, I got no short whatsoever, bulb stayed off... tried a 60w and 100w bulb.

That's good. It means at least nothing is badly shorted on the primary side.

Use it whenever you are testing a PSU that has had major problems like this one. The only limitation of the light-bulb trick is that you cannot test the power supply with a normal load - only no or very small load on the output. If the load is too large, the input voltage to the PSU may drop too far, and the PSU may go into over-power protection (OPP) or under-voltage protection (UVP).

[Dumah Brazorf]

Re: Apple power supply A1098

Reflow every single solder joint in the primary, even the littlest tinyest smd.

[stj]

Re: Apple power supply A1098

i bet this turns out to be a zener diode.
just a feeling i get.

[Porto]

Re: Apple power supply A1098

i bet this turns out to be a zener diode.
just a feeling i get.

Well, maybe your hunch is right, there are a few Zeners on the backside which I have yet to check.

Some strange thing I've found while measuring the board:

If I follow that 7.5v from the IC to the little elektrolytic cap and further I come across some resistors and a diode. The top side of those parts measure a stable 7.5v, but on the other side the voltage goes berserk between 0v and 7.5v. I guess something on that part of the board isn't working right.

[Porto]

Re: Apple power supply A1098

@Momaka,

I have checked the secundary circuit and I've found some shorts to ground.

For instance, pins 1 and 7 of that NCP4300 IC are connected to eachother. I have yet to find the short is in the IC itself or if they are connected externally.

Also, ALL the pins of the transformer are shorted to secondary ground. 2 pins are originally connected to ground, so the 3 others are shorted. In the picture, the top 2 pins are originally to ground.

I'll begin with removing that NC4300 IC, if the short is gone I know what to replace.

Also, on the primary I found a tiny SMD transistor with a hole in it... I think it says '2F7' with the number '7' in a 90degree angle. I don't know if this a NPN or PNP tor though. I guess this transistor was blown along with the TEA1552 IC when I tried to inject voltage on its Vcc pin; I had to replace that IC because it was shorted afterwards and also blew one of the 11N60C3 mosfets on the primary side, which resides an inch from that blown transistor.

So, this project is getting better and better by the day!

[momaka]

Re: Apple power supply A1098

Sorry for the late reply. I've been in a somewhat remote part of the world and without internet for about a month now. Talk about disconnecting from society .

Anyways, let me still give you a reply here.

For instance, pins 1 and 7 of that NCP4300 IC are connected to eachother. I have yet to find the short is in the IC itself or if they are connected externally.

That is strange indeed. Pins 1 and 7 are the two output pins of the op-amp IC. Generally, you shouldn't find these connected to each other in any circuit. It's a bad practice to directly connect outputs to each other as they can have a disagreement and have one overload the other.

That said, you might have to remove the IC itself to check. Or at the very least trace all of the tracks on the PCB in that area to make sure there isn't something else that is making them appear shorted.

Speaking of which, are you using your multimeter's continuity (aka "beep") function or actually measuring resistance? I recommend you do the latter. Preferably on the lowest resistance scale. Sometimes the multimeter will beep even for resistances less than 50 Ohms. It all depends on the multimeter itself, though. So that's why I always suggest to measure resistance when checking for short circuits vs. using diode/continuity test.

Also, ALL the pins of the transformer are shorted to secondary ground. 2 pins are originally connected to ground, so the 3 others are shorted. In the picture, the top 2 pins are originally to ground.

That is normal.

Switchmode PSU transformers have too few windings to measure any significant resistance with a multimeter, so they regularly appear shorted or nearly-shorted. On the other hand, if you own an inductance meter, that will show a different story. But that is not needed here, as these type of transformers almost never fail.

I'll begin with removing that NC4300 IC, if the short is gone I know what to replace.

Sounds good.

Also, on the primary I found a tiny SMD transistor with a hole in it... I think it says '2F7' with the number '7' in a 90degree angle. I don't know if this a NPN or PNP tor though.

Probably a "2F" transistor. I think the sideways "7" is a Fairchild oddity and you can ignore it. If you punch "2F sot-23" in Google, you should get a MMBT2907A datasheet (60V PNP transistor). Link here:
https://cdn.badcaps-static.com/pdfs/...4e424ee03b.pdf

You may be able to check the transistor on your board with your multimeter's diode test function. Put the multimeter negative (-) lead on the base of the transistor, and see if you get 600-800 mV (or 0.6-0.8 V) on your multimeter's display when you hook the positive (+) lead to either collector or emitter of the transistor.

So, this project is getting better and better by the day!

Well, never loose hope!
I always find things like these are the most rewarding when fixed: you learn a lot in the process.

[Porto]

Re: Apple power supply A1098

Hi Momaka,

Haven't had much time myself also to work on this PSU but I've more time now.

But thanks for the reply, I will test those transistors as suggested, I suspect one or multiple are malfunctioning.

I often use the DMM's 'beep' function but will use the resistance method more often; I read that's the better way to measure for shorts.

I never loose hope but this PSU is a real PITA!