Popular Discussion
Welcome, 匿名
Legend
Community Member
Power Integrations Engineer Power Integrations ModeratorTNY266PN failing 61000-4-5
I have an interesting problem I hope someone can help me with.
I have a mature 5W universal input power supply circuit that has been manufacturered in their thousands. I have just submitted this same design in a new product to an EMC test house, and it is failing on 61000-4-5 surge tests. It can survive 500V positive and negative pulses. It can survive 1kV positive pulse - but blows the protection fuse on 1kV negative pulses. I suspect that I have a duff TNY266PN as this is the same device that has been used in exactly the same circuit as other products that have sailed through these tests. My back up test unit is on its way to the test house, but in case this is not the answer, has anyone seen this before? Or can offer me any suggestions?
4 replies | Created on Sep 6, 2010 07:24 AM by shug
shug
Many thanks for your feedback.
I hope to survive a 2kV surge waveform as per AN-20 figure 1, for industrial immunity rating. You are indeed correct the configuration is differential mode, as I do not have an earth connection. The TNY266 does still work after replacing the fuse. What is confusing me is that it will pass the 1kV positive surge but not the 1kV negative surge and is repeatable? This to me impiles that there is a component that is failing due to the negative surge rather than the fuse I2t rating - otherwise the fuse would blow on the positive surge as well? The actual fuse I am using a Siba 189140 1A anti-surge with a breaking capacity of 1500A, and an I2t rating of 1.5A. How does one select a fuse to survive this test?
My input circuit is already as per figure 13 in AN-20.However I am unsure why the varistor has been positioned after the fuse, I would have thought that the varistor should have been put in front of the fuse to absorb any surge before it got to the fuse? I am contemplating adding a Transil to clamp the surge. This will have to be positioned directly after the fuse, as when these devices fail, they fail short circuit.
shug | 4 posts It may just be that the test house tests multiple positive strikes first, which weaken the fuse, which then subsequently fails during the negative strike tests. Fuses do have a wear-out mechanism - or it may simply be that there is insufficient time between strikes to cool the fuse down, and its temperature gradually climbs during the testing. Higher temps mean lower I^2*t ratings.
Check that the test house is doing the standard wait time between strikes.
MOV's can fail short circuit, and the fuse prevents a very loud bang which some customers do not accept. It may also be a Safety violation.
If the 1^2*t rating of the fuse is indeed lacking, you may simply select on with a higher rating.
Another technique is to place series resistive elements in series with the fuse in order to reduce peak currents during the surge, such as any differential EMI suppression inductors, or flameproof resistors (which in some designs, act as the fuse, negating the need for an actual fuse). Note that you cannot place a common-mode choke before the fuse because those can fail shorted L-N, and the fuse would not offer protection against it.
PI-Tucker | 290 posts Many thanks for you continued feedback.
OK I understand about the fuse weakening. I once worked for a company that made an electric compressor and you could see the fuse sagging during the piston stroke. I have test gear arriving later today so I can try out some of these theories.
The test house usually does a 20 second dwell time, but this was increased to 60 seconds and the fuse still failed.
I will seek advice about the safety impilcations of the position of the varistor.
shug | 4 posts Is this common mode mode or differential mode surge?
I'll assume differential mode. After the fuse blew, did you try replacing it to check if the TNY266 was still functional? If so the problem is lack of current limiting for the surge, and/or the fuse has too low an I^2*t rating.
If the test is common-mode, the setup makes a big difference (how the DUT is grounded, where the generator's output is connected, etc).
Pls. study some of the basics in this appnote:
http://www.powerint.com/sites/default/files/product-docs/an20.pdf
PI-Tucker | 290 posts 