SMPS low Vout

[madan1]

Re: SMPS low Vout

Bought another top247 and this time it worked.
The odd thing is that the faulty IC reacts normally to all resistance tests...
Thank you for the assistance, I really appreciate it!

[R_J]

Re: SMPS low Vout

So when the primary is being switched there are voltages on the secondaries, which seems to point to a bad top247?
All the secondary voltages are roughly 1/3

[madan1]

Re: SMPS low Vout

Everything was on the board ( just soldered wires on the primary winding and connected them to the power supply ).
I added above and the readings for c663 ( ~7V ).
On the secondary side there are about 4v on the optocoupler.

[R_J]

Re: SMPS low Vout

Was the top247 out of circuit? or the transformer primary isolated from the top247? You should have had D661 removed also. Its hard to isolate the switcher with the pfc circuit connected. but it does look like something was getting through the transformer, What about the voltage on C663? if this winding has 0 volts it might be the problem

[madan1]

Re: SMPS low Vout

Ok, so 14V@0.6A through the daylights module connected straight to the transformer ( still on the board ). See the attached image for the input waveform.

So, first of all - lot's of hi-pitch noise coming from the board ( I guess the transformer? ).
Second - the following voltages on the secondary sides:
c709 - 0.85v
c708 - 1.76v
c707 - 4.27v
c620 - 6.3v
c663 - 6.9v

[R_J]

Re: SMPS low Vout

You could disconnect the primary of the transformer and drive it with the pulse generator. You can first try this with just the primary disconnected (DON'T plug in the power supply) Connect the generator across pins 6 & 7 of T651. Vary the generator frequency close to 132khz. This setup will simulate the switching of the top247 and you might, see some voltage on the secondaries, or you might not due to the low output of the generator. The second option would be to completely remove the transformer and do the same and check the secondaries with the scope.
You cant drive the transformer with 60hz ac, its ment for high frequency.

[madan1]

Re: SMPS low Vout

Well there is not much left to check, I assume you already checked D6651 (should read open both ways) and D652. The top247 does'nt require much to work, so it leaves one thing, T651. If there are no shorts on any of the secondary outputs. I suspect either the primary winding or one of the secondary windings has a shorted turn. it only takes one turn to be shorted to kill the transformer, and you won't find that with a meter.

I will check again the diodes and will also see if there is a way to temporary convert the board to manual on/off smps ( schematics for this top247 application could be found in its datasheet ).

By the way, how to test the transformer?
Here is what I have
low voltage pulse generator 1hz-200khz
several DMMs
DIY scope
DC power supply
probably can scrap a 12v linear power supply ( remove the rectifier and use it as an AC power supply )
DMM which can measure inductance
"daylights module" ( mosfet 12Vin switching power supply with constant frequency and no filtering on the Vout ).

[R_J]

Re: SMPS low Vout

Well there is not much left to check, I assume you already checked D6651 (should read open both ways) and D652. The top247 does'nt require much to work, so it leaves one thing, T651. If there are no shorts on any of the secondary outputs. I suspect either the primary winding or one of the secondary windings has a shorted turn. it only takes one turn to be shorted to kill the transformer, and you won't find that with a meter.

[madan1]

Re: SMPS low Vout

Removing d703 did not make any change
Shorting pins 3 and 4 on the optocoupler also did not change anything.


I just injected ~6 and 8.2 volts on ground and ic651.1. Both times the voltage on pin1 went up to ~6.4 v, but still no secondary side V or 380V.
On the 6V it consumed ~35ma both on and off. With 8.2v the draw was ~100ma on and off.

[R_J]

Re: SMPS low Vout

D662 supplies the run voltage, it should be at least 7-10 volts
If you can, disconnect D703, see if makes a difference.
You could try shorting pins 3 & 4 of the optcoupler and see if the supply starts working, you just won't have any output regulation, butfor a test it should be fine

[madan1]

Re: SMPS low Vout

Thanks for reminding me - just scrapped a cap from an old pci lan card. After I installed it, the V on c661 went down to 0.2v ( was 0.36 ) and on c663 - 0.3V ( was 0.47 ).

Just checked the diodes on the secondary.
d702 and d701 behave as expected.
zD703 starts leaking in the opposite direction after 1V ( 100ma@10V ) which I don't know if it is normal for that zener.


Do you know what is the "normal voltage" from d662?

[R_J]

Re: SMPS low Vout

The ic will do its thing for a few cycles but needs to see the run voltage from D662 to go into proper run mode.
I don't know about R754 but resistors rarely go down in value, the go high/open so it could be a bad ic, this ic is used on a lot of power supplies.
Looking at your scope display, It seems the top ic is in Auto-Restart mode
I guess that scope can measure up to 200khz so it should be ok for this

Re check the diodes in the secondary to make sure none are shorted.
Did you ever install the C653 0.1µf cap?

[madan1]

Re: SMPS low Vout

I see, so even without any feedback from the optocoupler, the ic651 should be doing its thing?

For the c663, for ground I used the negative leg of c661 and pin 4 of the optocoupler, for the transformer, I hooked the probes directly to the transformer pins.
If I remember correctly, the scope is able to go up to 180-200kHz ( of course with signal generator and sine wave ).
Indeed this is the second top247 I bought. The first one was faulty, but later I noticed that probably I killed it, by applying too much solder which had reached through the board and shortened the top of the pins.
The funny thing is that the price of that ICs in that particular shop is like two times cheaper than in the rest ( of course not even close to the price in aliexpress ). So in the end it really could be a faulty IC... ( even though I never had issues with stuff bought from there ).

By the way, I got some strange readings on the diode side of the optocoupler.
R750
r751
r752
^^ those were fine

r753 - code3321 / 1.2k in the schematics - measured 0.81k
r780 - code1201 / 1.2k in the schematics - measured 0.81k ( r780 and 753 are in parallel so I guess 0.81 is ok )

r754 - code1002 / 10k in the schematics - measured 0.89k ( in both directions ).. so is the resistor gone or ic750 is lowering its resistance?
Is there a way to test ic750 without getting it out of the board... the "best" thing is that I'll have to remove 3 other caps + probably a choke before the hot air gun can reach the resistors and the ic.

[R_J]

Re: SMPS low Vout

The optocoupler is there to provide regulation feedback, as the load on the 12volt or 2.5 volt lines change the voltage across the opto's led changes, this changes how the transistor conducts, varying the voltage from D662 to the ic's control pin.
What are you using for ground on that scope board? Like I said before I doubt it can measure 132Khz
The voltage on C663 will likely be around 7-10 volts when its working
Maybe you have a faulty to247 ic?

[madan1]

Re: SMPS low Vout

Just got back to that PSU.

On ic651.1 - 5.27V
ic652.3 - 5.27v
ic652.4 - 0.47
c663 - 0.47
t651 between 9 and 10 - 0vDC and 0vAC(@50hz )
c661 - 0.36


by the way, what exactly is the logic of the optoc.?
When the psu is working fine, there should be constant V on the diode side, correct?
This should make the transistor side conductive, right?
What should be the normal working state of coil 9-10? I guess after the diode there should be some kind of a half-wave.. but should the RMS be higher or lower than the voltage of ic651.1 (~5v )? Also, does the ic651 expect high/low on pins 1 and 3 or reads the peaks of the half-wave?

Here are two shots of the scope.
the 0.2v one is from transformer pins 9 and 10
the 10mv one is from c663

[momaka]

Re: SMPS low Vout

Do you think that this cap might be the cause for the current problems?

Unlikely. But depending on the circuit design, it may be possible for the regulation to become unstable under very high or very low load. So if you can, do add it back in.

I guess a couple of li-ion batteries will do the job, right?

Probably. But I don't suggest using Li-Ion batteries in the circuit, because if something goes wrong, you certainly wouldn't want those to short-circuit or over-discharge.

What is expected to happen with that injection?

Essentially, what this does is it supplies a stable Vcc drive voltage to the TOP247 chip in order to rule out bad aux. primary-side winding on the transformer.

If inserting this voltage doesn't produce any voltage on the output of the power supply, there is something else wrong still. In which case, I suggest replacing the optocoupler and LM431 shunt regulator. In addition to that, make sure any resistors for the LM431 and optocoupler are not open-circuited.

Also, should I have the d661 back?

It doesn't really matter. D661 simply rectifies the 20VPFC supply, which is needed for the APFC. But at this point, we just want to make sure the TOP247 circuit works first. So D661 is irrelevant for the moment.

[R_J]

Re: SMPS low Vout

You can reinstall D661 but the purpose of D661 is to provide the Vcc for the pfc circuit, this top247 circuit will operate just fine with the lower B+ voltage of 320 volts, so don't worry about the pfc circuit at this time. The fact that there is no real load on the 12,5,2.5 this supply should operate with the lower B+ supply to the top247

[madan1]

Re: SMPS low Vout

Looks like 0.1 uF to me as well. If it's a tiny SMD ceramic cap, and you want to replace it with same, then get a dead PCI-E video card. The small ceramic caps near the PCI-E slot are always 100 nF (i.e. 0.1 uF), so those should work. Better yet, just get a 104 ceramic disc cap and use that instead. I'd say that cap is somewhat important, as the "C" pin of the TOP247 is used both for feedback and Vcc (and likely compensation?)

I'll see if we have any of those cards in the office, but yes - it's one of the smallest smd ceramic caps.
Either way, I'll get one of those.. just probably will wait to collect a decent list of components to justify the trip to the shop.
Do you think that this cap might be the cause for the current problems?

With D661 removed, what voltage do you get on C663? This is the first filter cap for the TOP247 Vcc supply and feedback voltage.
If nothing, check if diode D662 is good.

will check it next time and report

You can also remove D662 and supply an external DC voltage at the cathode of D662 - and if you do, make sure your external DC supply is NOT ground-referenced, because you will need to connect its negative lead to HOT ground (i.e. connect the external DC supply across cap C663). 6-8V supply should do. Even 5V might work, but not 100% sure.

I guess a couple of li-ion batteries will do the job, right?
What is expected to happen with that injection?
Also, should I have the d661 back?

[momaka]

Re: SMPS low Vout

I will try to find one, but what do you think 0.1UM means ( quoted from the SM )? Is this 0.1uF? The cap is of the smallest smd type and is real pain in the a** to solder it in that overpacked space.

Looks like 0.1 uF to me as well. If it's a tiny SMD ceramic cap, and you want to replace it with same, then get a dead PCI-E video card. The small ceramic caps near the PCI-E slot are always 100 nF (i.e. 0.1 uF), so those should work. Better yet, just get a 104 ceramic disc cap and use that instead. I'd say that cap is somewhat important, as the "C" pin of the TOP247 is used both for feedback and Vcc (and likely compensation?)

With d661 removed:
c661 - 0.07V
ic651.1 & ic652.3 - 5.2V
ic652.4 - 0.35v

With D661 removed, what voltage do you get on C663? This is the first filter cap for the TOP247 Vcc supply and feedback voltage.
If nothing, check if diode D662 is good.

You can also remove D662 and supply an external DC voltage at the cathode of D662 - and if you do, make sure your external DC supply is NOT ground-referenced, because you will need to connect its negative lead to HOT ground (i.e. connect the external DC supply across cap C663). 6-8V supply should do. Even 5V might work, but not 100% sure.

By the way, could a damaged transformer explain the exploded top247?

Absolutely. Although it's more likely the other way around: TOP247 went bad and damaged transformer. But going by your readings, I doubt that transformer is bad. In general, transformer damage is extremely rare. I've personally only seen it on those 2-transistor flyback designs in cheap ATX PSUs. Never anywhere else.

[madan1]

Re: SMPS low Vout

With d661 removed:
c661 - 0.07V
ic651.1 & ic652.3 - 5.2V
ic652.4 - 0.35v

secondary side:
c707 - 0.01v
c708 - 0.05v
c709 - 0.1v

320v on the high V rail

By the way, could a damaged transformer explain the exploded top247?