KNX Switching power supply [solved]

[ElectricGuy]

Hi everyone;

I'm trying to fix a Switching por supply of a KNX system.
The power supply is this;
http://www.gds-eshop.com/MTN684016
http://descargas.futurasmus-knxgroup.org/DOC/GB/MERTEN/1938/P683329_GB.pdf

Sympton:
- When i power up the supply, the voltage at the output slowly increases up to 8~9 volts and stops there. The voltage of this supply should be 29V +-1V. To reach the 8~9 Volts it takes ~10 seconds.

What i've done:
- Measured all the electrolytic capacitors, both capacity and ESR. They are normal.
- Measured all the transistores and diodes. They are normal.
- Changed the Switching IC Viper50A. Still same behavior. ( http://www.st.com/web/en/resource/technical/document/datasheet/CD00001219.pdf )

The output stage is very close to the one on the datasheet of the IC - Figure 14. The Opto is a SHF6286 and is biased with a TL431 ( U2 on the datasheet )

The IC SO14 on the picture is LM399. I think is responsible for supervise the outupt, like current and voltage!

Can somebody point me out what can i do more? I can find the schematic also,if anyone has it, please provide me!
I can provide more signals and data.

Thank you!

Pictures of the Power Supply;





Signal on the Oscilator PIN of the IC:


Signal at the DRAIN Pin of the IC:

[SeanB]

Have you checked the 100uF 16V capacitor by the switcher primary, your SMPS looks like it is stuck in startup, or the optocoupler is very leaky and pulling the compensation pin low. Simplest to replace the capacitor and the optocoupler to check them.

Also what is the jumper by the Viper chip for, is there a resistor missing or a jumper to enable it there?

[ElectricGuy]

Have you checked the 100uF 16V capacitor by the switcher primary, your SMPS looks like it is stuck in startup, or the optocoupler is very leaky and pulling the compensation pin low. Simplest to replace the capacitor and the optocoupler to check them.

Also what is the jumper by the Viper chip for, is there a resistor missing or a jumper to enable it there?

Hi SeanB;

That is not a jumper, is just 2pins to hold a protection for not to touch the mains capaciotr, but i remove it so measure better the signals.
I don't have a opto, but i can change the capacitor. I will try and then i will give feedback. Just one moment.

[ElectricGuy]

Hi SeanB;

Sorry, but i was really stupid. I did say that i checked all the electrolytics, but i didn't check the smd ones.......   

I changed the capacitor and its working just fine now. Its under my electronic dc load and is working great.

Thank you SeanB and sorry for my mistake.

[SeanB]

Never trust those little can ones, they are pretty common failures, and in this one the most likely candidate. I have done a few just by changing that startup capacitor, which can at worst blow up the controller IC.

[ElectricGuy]

Never trust those little can ones, they are pretty common failures, and in this one the most likely candidate. I have done a few just by changing that startup capacitor, which can at worst blow up the controller IC.

Yep.. you are right. Thank you for the tip!

[tautech]

Never trust those little can ones, they are pretty common failures, and in this one the most likely candidate. I have done a few just by changing that startup capacitor, which can at worst blow up the controller IC.

@ SeanB
Many of us have seen only the startup cap failure prevent SMPS's from running and as you say they are common failures. It's one of the first point of failure items I look at in a SMPS.
As many are supplied via a dropper resistor from rectified mains, are they subjected to higher voltage when the controller is idleing along at light secondary loadings?

Poor cap quality?
Underrated for application?
Exceeded MTBF?

Interested in your POV.

[SeanB]

Mostly the failure causes increased leakage, and the charging current is very small. The power supply startup voltage is pretty high so it will get at least a half dozen switching cycles in to start the auxiliary rail to actually power it before it reaches undervoltage lockout, and the IC typically in UVLO is in a very low power state. Thus the current is typically under 100uA, and this, along with the drop in capacitance, or the increased ESR dropping terminal voltage under the load of driving the SMPS power device, is what causes it to hiccup and not operate.

Thus you can have the PSU run in startup mode all the time, with a few cycles of operation every few seconds as the cap degrades, with it possibly starting after some time as the capacitor heats up from the current pulses and the heat from the startup resistors, till it eventually is not running at all or just pulsing every few seconds.