Offline 230VAC Flyback SMPS with TO220 FET and legs not bent..Too close to each other

[treez]

Reference The following post, it is illegal to have an offline 230VAC flyback SMPS which uses a TO220 FET in the PCB without legs bent to get the clearance above 1.5mm from drain to source.
https://www.edaboard.com/threads/335099/
…But I have taken apart an offline flyback SMPS, which is sold throughout the EU. It comprises a TO220 FET which is soldered into the PCB with legs not bent. There is 1.09mm clearance between the bare metal of the drain and source pins along the PCB. This is illegal.
It is sold by a huge European corporation, throughout the EU, though the SMPS was designed in China.
How is this allowed? The volume sales of these product are enormous.

This thread also speaks of this anomaly, or rather, direct flouting of the regulations by large corporations, because they have big lawyers that can stop them getting sued for it.....small companies cant do this flouting, so they die off
https://www.smps.us/pcbtracespacing.html

 

[Easy peasy]

There is presumably no statute specifying clearance on power supplies - it will not pass UL, it will not pass NZ/AU standards or EU standards for clearance on pcb's, however if the D-S shorts it will likely blow the mains input fuse and the unit will become safe - thus the working voltage creepage & clearance is too low - but the unit still operates as these C&C's have built in margin for the accumulation of dust etc and moisture over time...

If they have made an EU declaration of conformity - then yes technically illegal - any complaint to the EU standards office with proof would - in time - get the product and possibly the company blacklisted - i.e. imports prohibited - but, unsafe for the user? - likely not...

 

[KlausST]

Hi,

Which regulation do you refer to that says you need more than 1.09mm for functional clearance?
Usually the value depends on a lot of parameters....

Is there some coating on the PCB? What case is used? Is the device for indoor use only?

Klaus

 

[treez]

Thanks,
Coating looks like standard solder resist...defintely no conformal coating
Its for indoor use only.
The case is plastic
There are no vent holes. Enclosure is totally enclosing.
The reg i refer to is an IPC one, but also EN60950 says it.

 

[treez]

Anyway, the standard I refer to is "reflected" in IPC2221A…..where on page 53 you can clearly see that 2.5mm clearance is required for category “B2”.
**broken link removed**
…But as we all know very well, it is simply not possible for any of us to state exactly where this law is stated in whichever standard…the standards have simply become too complex. The standards now have become a “hammer” which the big corporations can use to batter smaller rivals into going out of business, and not competing with them.
Take the offline power supply that I have opened up. I actually used to work for the company that sold it. They started out as a small self-owned business. Their products were very good and sold like hotcakes. Then one day a large corporation rang up and said that the product was illegal due to the TO220 leg spacing. They said that the company must recall all products or else they would report us and put us out of business. Then they gave us an “alternative”……..if the company sold out to them completely….then they would let the small company off. Indeed the company did sell out to them. Now the company is not allowed to put its own name on its own products. They are payed a pittance of the total profits. Their plans to expand to new offices have been banned. It is only a matter of time before they all loose their jobs….since the big corporation will want to move the company nearer to its own head offices.
The interesting thing is, that the huge corporation still use the TO220 FET with straight legs…so its still an “illegal” product….however, the big corporation have big lawyers and big influence…

 

[Easy peasy]

anyone can make a complaint to the relevant authorities if the power supplies do not meet UL (in the USA) or EU standards in Europe / UK - any complaint received will be acted on - regardless of the lawyers of the company in default ... making a declaration of conformity that is false in the EU is a serious matter.

 

[treez]

these C&C's have built in margin for the accumulation of dust etc and moisture over time...

Thanks, of course you realise that this SMPS is in a totally sealed enclosure, and no dust or moisture could ever get in.

- - - Updated - - -

The thing is that a small startup in South Wales , which had a brilliant product, was forced to sell out. With respect to New Zealand, you are lucky, you are not part of the EU…..you are therefore able to make power supplys with TO220s with legs in a straight line and get away with it, because obviously your own country isn’t going to destroy a company which is paying tax to the country. And as you kindly said, there is no actual danger with a TO220 with legs “too close” together. Its perfectly safe due to the fuse etc. -However, it is illegal.

In UK, we are not even allowed to make power supplys with TO220’s with straight legs for ourselves in UK…because we are governed and controlled absolutely by the EU….therefore the EU can even penalise and punish British companies which make products solely for the British internal market.

- - - Updated - - -

Also, machines which automatically bend the legs of transistors are extremely expensive, and due to the fact that they involve manipulation of the MOSFET, there is the accompanying threat of ESD damage to the MOSFET. As such, its extremely inconvenient to introduce leg bending as a manufacturing process. Also, a MOSFET with a bent leg is more prone to break with vibration etc.
As such, the big corporations, (with big lawyers and influence) who can break the rules with respect to TO220 legs, are at a massive advantage to smaller startups, who can be sued for doing it.
Another point is that bending TO220 legs to get more clearance distance on the PCB is a total waste of time…because the clearance on the bit where the legs join onto the body of the TO220 FET cannot be changed, and is in itself, totally illegal!!!

 

[cyberrat]

How many other forums are you asking this on?

Again you are using the term "illegal", what law are you referring to as being broken that makes it illegal?
Conformance to standards is not a legal requirement.

 

[Easy peasy]

using a TO-220 in a "sealed" enclosure ( pollution degree 1) with pads for legs inline would likely break the standards here in NZ - which would make it technically against statute - but not illegal in a criminal sense unless it caused injury or major damage to property.

Coating the legs & pads is a common work around for C&C requirements.

 

[cyberrat]

Step the pads apart all you want, when the leg get to the body of the TO220 they are still very close together.

Yet the manufacturers still rate them for the voltages they are rated at, have done for years and they still work.

Waiting for Treez to confirm what law is being broken, he has made this claim on at least 2 forums I know of (as he seems to do with every thread he starts - spams the rest of the world )

 

[treez]

Thanks, the "law" that is being broken is EN60950.
Its as good as law because if any product does not conform to it in absolution, then that product can be banned from sale, and unpayably huge fines can be imposed on the company who sold the product....taking the company out of business....closing them down.

 

[FvM]

Waiting for Treez to confirm what law is being broken

Easy peasy has explained the legal aspect in detail, there's nothing you need to wait for, I think.

Essentially an unsubstantiated CE manufacturer declaration can be suspected. Declaring conformity with EU low voltage directive includes conformity with a number of applicable technical standards, e.g. EN 61010 (= IEC 1010) or EN 60950 and the clearance and creepage distances specified therein.

Placing a product on the market without complying with the low voltage directive risks a fine and trade prohibition.

Nevertheless I don't agree with the moaning attitude. The standards represent state-of-the-art and can be accomplished.

 

[KlausST]

Hi,

Mind that the distance between the legs of an scr or triac is always functional isolation, it is not safety isolation.
Functional isolation distance is less than safety isolation distance.
But i don't have the values in mind.

Klaus

 

[treez]

Thanks, there is a British company just down the road from us, who have "muscled in" on what we believe is our market, and they are now selling products to "our" British customers, and they have offline SMPS's with TO220 FETs in them which do not have bent legs to increase spacing. We are therefore going to report them and get them to do a total product recall, and we also want them to receive a very large fine. Can we expect to be sucessful in this action?

 

[Kajunbee]

Search " testing by competitors in enforcement of product standards".

 

[marce]

Anyway, the standard I refer to is "reflected" in IPC2221A…..where on page 53 you can clearly see that 2.5mm clearance is required for category “B2”.
**broken link removed**
…But as we all know very well, it is simply not possible for any of us to state exactly where this law is stated in whichever standard…the standards have simply become too complex. The standards now have become a “hammer” which the big corporations can use to batter smaller rivals into going out of business, and not competing with them.
Take the offline power supply that I have opened up. I actually used to work for the company that sold it. They started out as a small self-owned business. Their products were very good and sold like hotcakes. Then one day a large corporation rang up and said that the product was illegal due to the TO220 leg spacing. They said that the company must recall all products or else they would report us and put us out of business. Then they gave us an “alternative”……..if the company sold out to them completely….then they would let the small company off. Indeed the company did sell out to them. Now the company is not allowed to put its own name on its own products. They are payed a pittance of the total profits. Their plans to expand to new offices have been banned. It is only a matter of time before they all loose their jobs….since the big corporation will want to move the company nearer to its own head offices.
The interesting thing is, that the huge corporation still use the TO220 FET with straight legs…so its still an “illegal” product….however, the big corporation have big lawyers and big influence…

Actually B4 when solder resist is present...

- - - Updated - - -

Thanks, there is a British company just down the road from us, who have "muscled in" on what we believe is our market, and they are now selling products to "our" British customers, and they have offline SMPS's with TO220 FETs in them which do not have bent legs to increase spacing. We are therefore going to report them and get them to do a total product recall, and we also want them to receive a very large fine. Can we expect to be sucessful in this action?

Read post #13 functional isolation.

 

[betwixt]

Putting official requirements aside - what difference does it make to the reliability or safety of a product if the tracks/pads are far enough apart that no risk is evident?

I fully understand the reasons for spacing in situations where safety is an issue, for example between 'in' and 'out' sides of a mains PSU. Isolation is required to prevent danger to the device or user on the output side. However, within the boundary of the input circuit, who cares? Presumably to have a 'CE' mark, the FETs source and drain must have a gap of 2.5mm on the silicon itself! Given the need for their intimate contact on the die, the wiring to them is irrelevant.

Brian.

 

[treez]

Actually B4 when solder resist is present...

Thanks, though i am speaking of a TO220 FET soldered into a PCB with straight legs. The actual soldered pads , as you know, do not have solder resist over them...There is solder resist between the PTH pads on the PCB, but i dont think that makes it category B4...to get category B4, we would have to spray a insulative coating over the pads, and we have not done this, but as Betwixt says, the D and S are close together n and inside the TO220 body anyway.

 

[Easy peasy]

I feel like an Irishman stating the obvious flat out, but, inside the package (i.e. covered in epoxy) there is no way dust, moisture or other pollution can gain ingress to short out the high voltage and the V/mm rating of the epoxy is very high, on the surface of the package the CTI (current tracking index - see standards) is very high too - the epoxy was designed for this - quite a bit higher that plain pcb (FR4). However the spacings at the surface (of a TO-220) often do not meet standards (230 mains operation) and thus would not gain UL in the USA (and many other places). Hence conformal coating (using quality coatings) will solve these issues for meeting standards and for giving the product longevity in the field. PCB pads and device legs coated similarly. [ for 500V rms AC, or DC, 1.3mm creepage is the min allowed, long term working stress ]

- - - Updated - - -

Without these spacings, there is the possibility of arcing, with the arc plasma affecting the nearby insulation and eating away the copper tracks, causing unknown failures - but possibly fire - which UL is all about (fire prevention).

We see lot of 3 phase (400/230) chargers (6kW) with a rectifier on the mains, without very effective conformal coating, the ~565VDC manages to track around the edge of the pcb 5mm both sides in some cases to cause faults, at the fets which switch at 100kHz, tracking occurs easily on the fet bodies (TO-247) and on the pcb, cutting into the fet legs and creating extra paths on the pcb (until they stop) - thus even the mandated creepages and clearances are not enough when the environment is damp and pollution is allowed to enter ...

 

[FvM]

We see lot of 3 phase (400/230) chargers (6kW) with a rectifier on the mains, without very effective conformal coating, the ~565VDC manages to track around the edge of the pcb 5mm both sides in some cases to cause faults, at the fets which switch at 100kHz, tracking occurs easily on the fet bodies (TO-247) and on the pcb, cutting into the fet legs and creating extra paths on the pcb (until they stop) - thus even the mandated creepages and clearances are not enough when the environment is damp and pollution is allowed to enter ...

I wonder if the observed long term failure is substantially caused by partial discharge, starting at the sharp terminal edges?