Offline flyback smps leakage inductance reduced by having no margin in transformer?

[treez]

Hello,
Page 4, bottom RHS of page states that the flyback winding width should be at least twice the creepage distance in order to reduce leakage inductance.
http://www.powerint.com/sites/default/files/product-docs/an17.pdf

In mains flybacks in led light bulbs, why not do away with the creepage requirement altogether, and thus reduce transformer leakage further?
After all, enamelled copper wire has 1000VAC of insulation on it. (not talking about triple insulated wire).

After all, some led light bulbs are being made with no creepage built in at all..as follows

.....................................................................................................................................
....At time 20:52 of the following, is this the proof that bulb is not isolated?....
http://www.youtube.com/watch?v=YLlpoZaBx-g

..also, if you view from time 6:34 of the above video , you can see that the MCPCB that contains the (non-isolated) led circuit is literally lying right on the heatsink which which appears to be opening out to the external of the bulb
..............................................................................................................................

 

[FvM]

What are you exactly asking? The Powerint AN is dedicated to switchers with reinforced insulation.
For lesser isolation requirements or no isolation at all, transformer design will be easier.

In any case, a manufacturer is responsible for the safety of his product. He don't find the answer on youtube or by studying the competitor's weak points.

 

[treez]

If a manufacturer of a light bulb can find a way to do it, which is more efficient, because it comprises less leakage inductance, then that manufacturer has good reason to approach the regulatory bodies and request a clause in the regs so that they can do it the more efficient way...because lighting is 20% of all energy usage, so if you make it more efficient, you are helping all of us.

The goverenements should be asked if regs on isolation can be bypassed for light bulbs, as we all want them more efficient.
The regs are written for general cases...and cant suit all needs, and its up to the manufacturers to bring forward cases which are for the common good.
I honestly cant see why offline lighting trafos cant be wound with no creepage, as long as enam copper wire is used.

 

[FlapJack]

Enamel wire does NOT have a working voltage rating, meaning it is to be used for low voltage hook up wire only. The higher number you see like 1000 or 2000 volts is dielectric strength and is NOT to be used for a wire working voltage.

The reason that enamel wire works in transformers of hundreds or thousands of volts is because the transformer designer is very careful to watch the volts per turn. Also, between layers in the transformer there are not 1 but 3 layers of insulation to prevent any voltage breakdown.

 

[treez]

1000 volts dielectric strength sounds fine to me......the flyback primary peaks up to about 500 volts...so much less than 1000v..i cant see dielectric breakdown happening...
The above youtube vid shows just the insulation on top of mcpcb being used for dielectric between mains live and metal where human hands touch..theyre not actually using trafo isolation at all...so I reckon im doing well with the 1000vac isolation in a trafo.
The pcb is also totally sealed.

 

[FvM]

The above youtube vid shows just the insulation on top of mcpcb being used for dielectric between mains live and metal where human hands touch.

The video shows the author's conclusions and guesses rather than the manufacturer's isolation concept and safety dispositions. Thus it's just fruitless to discuss the video over and over.

 

[treez]

the video is half an hour and I cannot expect people here to watch all that on my behalf, I agree with that.
Though I have watched it thoroughly and can assure that what I mentioned about the isolation barrier (or lack of) is matter of fact.

 

[FvM]

I also watched the video at times. Firstly you don't know which isolation concept the designer tried to implement, e.g. if the LED PCB still has sufficient creepage for 120 VAC basic isolation (could be) and if reinforced insulation (most likely not achieved) would be required. If you come to the result that the design doesn't comply with safety rules, you secondly shouldn't jump into conclusions regarding other products.

 

[treez]

From time 6:30 to 6:40 of the video it is absolutely clear that standard mains isolation creepage is violated.
I thought with the mains, that it was 5mm for 120VAC and 8mm for 240VAC?
That is it, there is no other standard for mains creepage.

The mains isolation creepage laws (for 120VAC and 240VAC) are based on the fact that the mains can spike up to well above 1KV for short transients.

Ive gotten capacitive mains shocks from touching the outer plastic of 7/0.2mm wire with live mains in it...and that plastic is thicker than the combined mcpcb insulation and white heatsink covering seen in that video.

The only exceptions that I know of for mains creepage is when, like in an electric shower, the circuitry is not mains isolated, however, the on/off button is mechanically fitted such that theres no way you can be shocked from creepage or water ingress.
In the video you can see that the outer white covering of the heatsink is what can be touched, and indeed will be touched, since its the majority of the outer casing.

By the way, ive no idea if the guy who made that video is doing the tear-up legally, if its allowed to do that to a product on a public video, but its certainly helpful and if it leads to extra safety for people, then that's a good thing.

I am sure Feit will be interested that the assemblers didn't bother to use a full covering of heatsink paste.
I worked in an electric drives place once, and a load of drives came back failed, and with no heatsink pasted between heatsink and 3-phase rectifier.
The assembly staff said they didn't think the paste was necessary.
They didn't like using the paste as its sticky and doesn't easily come off clothes

 

[FvM]

I thought with the mains, that it was 5mm for 120VAC and 8mm for 240VAC?
That is it, there is no other standard for mains creepage.

No, there are different values for basic and reinforced insulation, enviromental conditions, material CTI, overvoltage category. The smallest creepage distance for 150 VAC basic insulation, clean enviroment , CAT II is e.g. 0.5 mm. Basic insulation might be sufficient in this particular case under several conditions, e.g. the heat sink powder coating has at least 0.4mm thickness and sufficient dielectrical strength so that it can be considered as supplemental insulation.

 

[schmitt trigger]

There is a reason for safety standards and the fact that regulatory bodies SELL those standards...they are the compilation of many years of sweat and tears. They are there to prevent the rookie engineer from repeating the mistakes that someone has previously done. Ignore them at your own peril.
They are also there for the safety of the consumer and to protect -as much as possible- the manufacturer or seller from litigation.

Having said that, regulations can be amended. But it would cost dearly in time and money to prove that the proposed amendment does not compromise safety or increase liability.

What I have seen is that regulations, when they are amended, actually become stricter.
Case in point...about 20 years ago I built a home. Back then only two smoke detectors for a single story home were required. And they were battery powered.
I just finished a new home, and the new regulations require a smoke detector on each bedroom, plus one in the kitchen and another by the main entrance. Also they require to be AC powered with a battery backup.

Automobiles are another excellent example of stricter regulations as time goes by.