Solution Finder Get Tech Support

LNK PH failure

Posted by: Osnirok on

We are using LNK420 in 75W design based in DER-291 single section. see schematic. After some days of working, one unit has a catastrophic failure, the IC has burned. The entire driver is polyurethane potted, but after disassembling we see the ic exploded.  We see that many associated components are apparently well, lets say MOV RV1 ( a 275V, 20mm) , C9 470nF x 630V polipropilene, c13 100nF x 630V polipropilene, C14  10uF 450V, diodes D5, D7, VR1, bridge rectifier,  D3, D6, D8, R19, C 15, C16. There are no evidence of PCB short circuits

Many times during bench tests, we had short circuits at output and open circuit and entirely design works very well

A little difference is that we need to put a 100nf connected at V pin and source pin to dont have shift in output current after potting.

We think the layout of PCB is good, based in layout of DER-291

We have many drivers working very well, but we are afraid and would like to know what happened. We are using the driver in a industrial application, and probably the failure happens at night. Next week we will monitor the power line to see if we have some over voltage, but the MOV rated to 275V was intact.

Could you please tell-us something, or some ideas about this failure ?

Comments

Submitted by PI-Dunwell on 09/14/2014

Based on your description, it may be a surge voltage causing the VDS of the device to exceed its rating. A 100nF on V-pin probably inhibits the device to shutdown. Its a good test to do surge  on the unit with the 100nF on the V-pin and check whether max VDS is exceeded.

Submitted by Osnirok on 09/15/2014


Well, I think we need to use the 100nF, I will try to use a 10nf and see what happens with potting results. but I think the unit will again have a inhibit shutdown.Do you think is a good idea to put two inductors at input, one in each power line, to improve differential surge immunity? What value of inductance could I use? If I use a too high value is there any problem? obviously not a too high value to don't have a big drop in power line.  Actually I don't have a surge immunity equipment, then is better to spend some more varistors (metal oxide varistor) and inductors... thankyou ! 

Submitted by PI-Dunwell on 09/15/2014

Yes inductors inline will improve surge immunity. The limit on the max value will depend on your efficiency requirement. Note that 275VAc rated varistors clamping voltage is ~700V. This will definitely fall down in value if you use a higher energy (bigger size) rating device.

The bottom line is you need to know the VDS during surge event either by measurement or approximation. You can approximate it by knowing the maximum dc input voltage that will bring the VDS to its maximum. The next step is to limit the max dc input voltage so that the device will not reach its max value.

Submitted by Osnirok on 05/05/2015

Hello, we still have problems with some drivers. Definitely the problem is the power line of this particular installation but I would like to do some comments and ask some more help.

In the lasts installed units we use the following:

Not using C 100nf at Vpin, use 4 parallel MOVs rated to 275V, 20mm, at the Ac input, and a TVS rated to 440V (1.5KE440A ) located in parallel with C8 – 470nFx630V,  C3 is 100nF.  The other things are similar to RDR290.  But, again some units fail, the ic explode probably during a big surge event.

In some lab tests, during surge event, the voltage in C8 (with has a 440V TVS in paralell) increase to a peak of 500 or 550V, and the IC burns.  The IC has a 725V rating.  But, I think the voltage VDS is higher than the voltage developed at C8 because we have some interaction with transformer.

The best place to put a 440V TVS is in parallel with C8 ?

Now, in lab tests we see that a good way to improve the surge immunity is to increase the C2 to 40uF, and change the common mode filter inductor, to a 2 separated units of 1,5mH, and use a 440V TVS in paralell wit C8. Lats see what will happen in the next days...

With this relatively large capacitance, can we use D1 as a single DL4007 (We use a US1M, a fast version of 1N4007, rated to 30amp pk during 16ms) diode to charge this caps?

 

Do you have any other recommendation to improve the surge immunity?  

Thanks again,

Submitted by PI-Dunwell on 05/17/2015

The best place to put VR1 is in parallel with C3 not C8. if possible on the layout, cathode is close to pin 1 of transformer and anode to source. Across C3 I guess should be fine. Increasing C2 will definitely help. This is actually the function of this capacitor, to absorb the surge energy, and the larger its value, the better. make sure on the layout that this network will see the surge before going to C3.

Submitted by Osnirok on 05/21/2015

Ok, the cathode of TVS is located near pin 1 of transformer (V+ bus) and the anode conected near source of LNK. these means in paralell with C8. perhaps you meant C8 and not c3.

But, an other litle question: On the v pin, I´m using R top and R bottom. In some designs there is a zener diode from V pin to source. According to my experience, When I was using a 100nf at V pin this zener was usefful because in some situations the driver not start. But now, I do not use a 100nF at V pin because these inhibit the over voltage shutdown. Well, my question is: Is it necessary to place a zener at vpin for protection purposes?  Since these place is a high impedance connection, I´m not using these zenner to prevent a  decrese in the shutdow speed.

Thanks

 

Submitted by PI-Dunwell on 05/22/2015

Hi,

Maybe we are looking at different schematic but I think you know what I'm refering to (V+ bus to source of the IC)

Zener diode is not required on the Vpin.

Regards

Submitted by Osnirok on 05/23/2015

Yes, thank you. have a nice wekend.