Help to diagnose SMPS issue

[ultrasmurf]

Hi guys,

Is there any easy way to test newer design SMPS IC that has all kind of over voltage, over current limit ? e.g. : M51995AFP.

I have a power supply using this IC with no output. Fuse is good, secondary diode are good, switchers are good. Supply will trip ELCB if powered on with full AC input (>100VAC), ramping input up shows that the PSU suddenly draw high current at about 80VAC (about 2-3A) from mains, and there is trkk trkkk sound for half a second or less and after that it does not pull current anymore. Increasing VAC beyond 100V is fine this way and will not trip ELCB as there is no large current pull after this. Main Electrolytic charge up to about 385V (there is PFC IC in this supply)

All I can think now is the main M51995AFP IC not working correctly. There is no pulse output driving the switcher from this IC. I don't have a new parts for this, and replacing this IC from a salvaged board does not solve the problem (it could also be that the salvaged board has problem with the IC).

Will update with pictures and model number once I get back to lab...

[sam_sam_sam]

Re: Help to diagnose SMPS issue

If were me

I would replace the IC chip with a new one and see what happens

I would also see if you have a short in the power supply some where or what it is powering

I would also check all of the caps with ESR meter and you can compare to a new one

Please post some pictures of the front and back of the power supply board and what is powering

[ultrasmurf]

Re: Help to diagnose SMPS issue

Attached front and back picture of the PSU. It's a Densei-Lamda JWS100-3/RA. 3.3V 20A power supply. The grey wire is just connecting the V+ to Sense+ and V- to Sense-.


Main switcher IC : M51995AFP (Mitsubishi/Renesas)
PFC IC : 5331M (Fuji Electric)

BTW confirmed that the switcher IC is dead, no pulse output. Testing method : Connect VC (Output collector) to Vcc and VCC to + Supply , GND to gnd. Based on test of the working unit, if you monitor the Vout, it should start pulsing once the VCC hits slightly more than 16V, and it should still pulsing when VCC is reduced down below 10V (as stated in the datasheet). Same testing method can be used for the PFC IC, all testing done with chip mounted in the PCB so I assume the OVP/OCP/and all other shutdown signal is properly pulled to ground or VCC.

Still can't figure out why the elcb trips... the current surge I notice when it hits 80VAC using variac is due to the PFIC turns on and pumping the main cap to 385V.

[sam_sam_sam]

Re: Help to diagnose SMPS issue

Do you get any voltage at the output terminals at all if you do check the optic sensor to see if it getting voltage in the input side and dose it climb and then fall or dose it stay steady

[ultrasmurf]

Re: Help to diagnose SMPS issue

No voltage at all at the output terminal. I did check the secondary portion and they seem to be doing ok. Injecting voltage at the output terminal light up the led and feedback through the optocoupler is going up and down according to the injected voltage.

Now I'm waiting for ebay replacement IC... this IC is EOL and only ebay has some... anyone knows of any direct replacement for M51995AFP ?

[Longbow]

Re: Help to diagnose SMPS issue

Hi guys,

Is there any easy way to test newer design SMPS IC that has all kind of over voltage, over current limit ? e.g. : M51995AFP.

Yes. Use the same troubleshooting procedure as for other SMPS. The data sheet for the M51995 is extensive, and shows a number of example circuits.

Supply will trip ELCB if powered on with full AC input (>100VAC),

Where is the ground fault current coming from? That might be the best place to start. The usual power supply failure will not produce ground fault current. You might have an insulation problem.

ramping input up shows that the PSU suddenly draw high current at about 80VAC (about 2-3A) from mains,

Usually, you cannot test a SMPS by ramping up the input voltage with a variac. However, you can often apply a reduced voltage to check startup.

there is trkk trkkk sound for half a second or less and after that it does not pull current anymore.

That noise is the supply trying to start up but failing.

Increasing VAC beyond 100V is fine this way and will not trip ELCB as there is no large current pull after this.

The supply fails the attempt to start up, then it stops trying.

[ultrasmurf]

Re: Help to diagnose SMPS issue

Thanks for the suggestion. I couldn't trace the ground fault current. As said, ramping up with variac does not trip the ELCB. Another info : I disconnected the ground cable to the main (from PSU to variac), and it still trigger ELCB whenever I start it with >80VAC. Slowly ramping up doesn't trigger ELCB, The trk trk sound I believe is coming from the PFC circuitry charging up the main cap to 385V. Sounds right ? Or is it more of the M51995 trying to start up but bogged down by some OVP/OCP etc protection ? I probed the OVP, OCP (CLM pin) and got a 0V while trying to power up this M51995 with 16VDC on VCC &VC (No VAC input) and Vout shows no pulse, which convinced me that this shows the IC is dead.

Having said that, I couldn't really wrap my head around the OCP using the CLM pin as explained in the datasheet of M51995. My testing method could be wrong...

Yes. Use the same troubleshooting procedure as for other SMPS. The data sheet for the M51995 is extensive, and shows a number of example circuits.

Where is the ground fault current coming from? That might be the best place to start. The usual power supply failure will not produce ground fault current. You might have an insulation problem.Usually, you cannot test a SMPS by ramping up the input voltage with a variac. However, you can often apply a reduced voltage to check startup.That noise is the supply trying to start up but failing.The supply fails the attempt to start up, then it stops trying.

[sam_sam_sam]

Re: Help to diagnose SMPS issue

No voltage at all at the output terminal. I did check the secondary portion and they seem to be doing ok. Injecting voltage at the output terminal light up the led and feedback through the optocoupler is going up and down according to the injected voltage.

Now I'm waiting for ebay replacement IC... this IC is EOL and only ebay has some... anyone knows of any direct replacement for M51995AFP ?

You do know that you have 3 optic sensors on that power supply you should check all three of them

[Relayer]

Re: Help to diagnose SMPS issue

Hello ultrasmurf,
Please see your modified picture below:



The component to the right inside the box looks like its been damaged. Or is it flux or some other type of contamination on it?
The component to the left/bottom also seems to look a bit unhealthy as well.

Please check another modified picture below:



Can you please tell me what component/s are sitting underneath the heat-sink shown by the red box?
Also, you need to check the fusible (safety) resistors as pointed out by the green arrows. You may have to lift one leg to get a more accurate reading.

I have to admit I'm a tad skeptical that the M51995 IC is at fault.
Did you check to see if there was any voltage on Pin 16 at all?
Make sure you measure between Pin 16 and Pin 13.
If there is no voltage and the main filter cap has 350+ volts on it, then the problem lies elsewhere.
Please let us know how you get on.
Good Luck!!!
Regards,
Relayer

[ultrasmurf]

Re: Help to diagnose SMPS issue

Have a little bit more time to do more troubleshooting

Component circled red is good... its a diode, just got coated with heavy browning flux..

Fusable resistor are all good..

Under the heatsink is NTC. 2x 5Ohm for a total 10Ohm NTC. Seems a bit low for my liking to limit inrush current...

So.. testing the M51995AFP again off board shows they are good with pulse out from the Vout pin. Placed on board and it goes quiet. I have a similar lambda power supply (24V instead of 3.3V output) laying around on the table and upon closer inspection there is a jumper that seems to be missing (although there are a bunch of other things that are different, but this jumper was connected to the on off pin and will short it to ground).

Thanks to sam reminder, eventhough all 3 optocoupler tested good, it makes me wonder why need to have 3 optocoupler. The application note only have the third opto for the on off switch (one for feedback and one other for OVP protection). After hunting around a little bit, and re reading the datasheet I realize one stupid mistake. On/Off pin supposed to be low to turn on the IC(typically should be below 2.4V, or at worst condition below 1.9V). And mine measured 3.7V. Somehow my brain is wired with high voltage to turn on and low voltage to turn off. After a little bit more investigation, more stupidity resurface... This is a remote controlled PSU, there is supposed to be a 5V input at remote pin to turn on PSU (that third optocoupler is really an on off switch... just like the application note says.. ).

So the missing jumper is the one that determine whether this PSU is a remote controlled turn on or immediate turn on type. The connector for remote turn on is the white connector at the bottom right corner (top side board where you have your main cap) picture. Putting a jumper in (shorting the on off pin of 51995 to ground) will now make 51995 produce pulse output on board.

So what I have all along is a working PSU.. no issue at all.

As for tripping ELCB, my take is that based on the NTC rating will draw 24A of inrush current from my 240V outlet that only have a 10A ELCB...

Moral of the story : RTFM... properly.. could have avoided troubleshooting a fully working SMPS if I had known that there is a switch to turn this thing on and off...

And my lesson learned for anyone coming to this thread looking at new way to test SMPS IC :

1. Test IC on board by connecting DC power directly to the IC (less dangerous than your main line and it lets you use your oscilloscope). Monitor your current pull from the external DC power supply. It does not take that much current to turn on just the SMPS IC. If it shows anything more than 100mA, be very wary, you are turning on some other function of the board, it might or might not be causing problem. Remember you are powering it up using external power supply, there could be some circuitry that try to pulls VCC down to ground if it sense a voltage without main connected. In doubt turn your power off if it start pulling high current and make sure you know what is pulling the current before turning it back on.

2. If it does not show life at the Vout (the pin that control external power transistor) then look at the OVP and OCP pin, and alson on off pin. Read the datasheet, find out what is the expected voltage at these pin for normal operation.

3. Take the IC offboard, connect ground to all pin looking for low signal and VCC to all pin expecting high signal (check datasheet, make sure all pin can tolerate VCC turn on threshold), or give those pin required voltage (troublesome...)

4. Read datasheet with patience. And read every single word.. skipping word will get you in trouble

5. Like everyone has already said in the internet... Most designer will follow the application note... try to fit your PSU to the application note, and it will make much more sense, and you have your "schematic".

Thanks everyone... I learned a great deal of SMPS trying to troubleshoot this one...

Hello ultrasmurf,
Please see your modified picture below:



The component to the right inside the box looks like its been damaged. Or is it flux or some other type of contamination on it?
The component to the left/bottom also seems to look a bit unhealthy as well.

Please check another modified picture below:



Can you please tell me what component/s are sitting underneath the heat-sink shown by the red box?
Also, you need to check the fusible (safety) resistors as pointed out by the green arrows. You may have to lift one leg to get a more accurate reading.

I have to admit I'm a tad skeptical that the M51995 IC is at fault.
Did you check to see if there was any voltage on Pin 16 at all?
Make sure you measure between Pin 16 and Pin 13.
If there is no voltage and the main filter cap has 350+ volts on it, then the problem lies elsewhere.
Please let us know how you get on.
Good Luck!!!
Regards,
Relayer

[smartfox]

Re: Help to diagnose SMPS issue

Here is Test circuit for this IC
When i wrote to Renesas to give me test circuit - they
answer : we can send to you datasheet....
Datasheet is very bad explained everiting is too
difficult to be done simple tasks as calculating of frequency
NO OTHER IC WITH MORE STUPID DATASHEET
User not worry how IC works...
He wants to use IC with several simple steps
to be explained adjustment of modes and frequency...
In my case SMPS wont to start stable because optocoupler
in F/B is bad. Transistor leaks current from FB pin to GND.
This is without signal on LED.
MY ADVICE IS :
CHECK ALL ELECTROLYTE CAPS
CHECK ALL OPTOCOUPLERS
CHECK ALL SEMICONDUCTORS