Is this charger safe?

[oldway]

The back side would be interesting too for checking the clearance between primary and secondary side. I see cheap electrolytics, no X2 filter cap, no MOV, a common mode choke, a SMPSU controller with an opto coupler for feedback, and a LC output filter. It's cheap, but doesn't seem to be nasty.

I find very interesting that some people are concerned to check only the things they can see, as clearance between primary and secondary side (must be >8mm), data sheet of optocoupler,  but forget or even refuse to check the internal clearance/insulation of the transformer.....and the only way to check this is to test the power supply applying the specified voltage (4250Vdc) during 1 minute between primary and secondary sides of the charger.

[Zero999]

Normally a sample needs to be tested, for approval purposes, not every single unit produced. Such a test is also pretty stressful to the insulation, which, unless is ridiculously over-engineered, could be weakened as a result, even if it passes, so the test would actually make the device more hazardous, not safer. I would not feel comfortable using a device which I know has been subjected to this sort of stress test.

That's not what I have seen as practice in industry....!!!! Safety is not a concern that can be controlled only by sampling...It must be 100% tested...The only thin barrier between mains voltage (230V) and the user is the primary/secondary insulation of the ferrite transformer....Primary and secondary are very close....thats very dangerous.

If the test must be repeated, you can only apply a reduced voltage of 80% of the nominal test voltage.

I disagree. In theory no testing is necessary. If the insulation is so thin, that every product needs to be tested, then the design is marginal and isn't safe to begin with. It should be redesigned so it will comfortably exceed the minimum requirements.

Electrical safety can be guaranteed, purely by design. The breakdown voltage of various insulating materials is a well known parameter. All one needs to do is make sure the material is thick enough, sufficiently mechanically robust and large enough creapage and clearances are used throughout the design. In reality it may be necessary to test a small sample of materials/products, to ensure they meet the specifications but 100% testing is not desirable. Any products which have been subjected to the high potential test should be destroyed and not sold to the customer.

[wraper]

The back side would be interesting too for checking the clearance between primary and secondary side. I see cheap electrolytics, no X2 filter cap, no MOV, a common mode choke, a SMPSU controller with an opto coupler for feedback, and a LC output filter. It's cheap, but doesn't seem to be nasty.

I find very interesting that some people are concerned to check only the things they can see, as clearance between primary and secondary side (must be >8mm), data sheet of optocoupler,  but forget or even refuse to check the internal clearance/insulation of the transformer.....and the only way to check this is to test the power supply applying the specified voltage (4250Vdc) during 1 minute between primary and secondary sides of the charger.

People don't forget that (at least me), it's just obvious that it cannot be tested without special equipment. Se mentioning it won't change anything.

[wraper]

Electrical safety can be guaranteed, purely by design. The breakdown voltage of various insulating materials is a well known parameter. All one needs to do is make sure the material is thick enough, sufficiently mechanically robust and large enough creapage and clearances are used throughout the design. In reality it may be necessary to test a small sample of materials/products, to ensure they meet the specifications but 100% testing is not desirable. Any products which have been subjected to the high potential test should be destroyed and not sold to the customer.

Power supplies, optocouplers, transformers are Hi-POT tested individually. You cannot guarantee by design against assembly defects.
Good PSU, deadly short due to manufacturing defect:

Any products which have been subjected to the high potential test should be destroyed and not sold to the customer.

[oldway]

I disagree. In theory no testing is necessary. If the insulation is so thin, that every product needs to be tested, then the design is marginal and isn't safe to begin with. It should be redesigned so it will comfortably exceed the minimum requirements.

Electrical safety can be guaranteed, purely by design. The breakdown voltage of various insulating materials is a well known parameter. All one needs to do is make sure the material is thick enough, sufficiently mechanically robust and large enough creapage and clearances are used throughout the design. In reality it may be necessary to test a small sample of materials/products, to ensure they meet the specifications but 100% testing is not desirable. Any products which have been subjected to the high potential test should be destroyed and not sold to the customer.

For sure, you have not the least experience in quality control of electricity and power electronics gears to write such a bullshit.
Should you be quality control manager, the first customer who will be electrocuted will send you to jail , and for a very long time.

For exemple, technical specifications of a SMPS...High pot tested at 4242Vdc 1m between primary and secondary.

[fourtytwo42]

And if you saw the way transformers are mass produced (thrown together) in Asia you would understand that the designers intention is completely sacrificed to cut corners so specifying certain types of insulation is useless.
As we now live in a deregulated world to encourage trade nothing is tested anymore hence we have corner cutting manufacturing directly in contact with the consumer, it is only a matter of luck that we do not have more serious accidents and of course much goes unreported as the units are simply thrown away if deemed to be faulty.
As I said before never trust your life to this kind of mass produced rubbish

[madires]

I find very interesting that some people are concerned to check only the things they can see, as clearance between primary and secondary side (must be >8mm), data sheet of optocoupler,  but forget or even refuse to check the internal clearance/insulation of the transformer.....and the only way to check this is to test the power supply applying the specified voltage (4250Vdc) during 1 minute between primary and secondary sides of the charger.

Please feel free to send the OP a proper isolation tester And please see Table 2K of IEC-60950-1 for the clearance. IIRC, it should be 4mm in this case.

[oldway]

Please feel free to send the OP a proper isolation tester And please see Table 2K of IEC-60950-1 for the clearance. IIRC, it should be 4mm in this case.

Not at all. For class II appliances, you must use the reinforced insulation values (R of the table) for peak working voltage of 420V and the clearance is 6.4mm.
But this apply only for free air clearance and it can't be apply to circuit board.
Here are the PCB's recommanded clearance....
To withstand a 4.242Vdv test, clearance should be 12.96mm.
That's huge.....this is the reason why air gap's are use instead.

For safety, there would be an air gap between primary and secondary circuits.
Dave remember this in almost all his teardowns

[madires]

Based on that we would have to trash 95% of all SMPSUs.

[oldway]

Based on that we would have to trash 95% of all SMPSUs.

Yes, you would have to trash 95% of all SMPS you buy on Ebay for 10 bucks from Chinese sellers....yes, indeed...

[madires]

... and also nearly all wall warts which come with gadgets, mobile phones, SOHO routers and what have you.

[oldway]

... and also nearly all wall warts which come with gadgets, mobile phones, SOHO routers and what have you.

Not all manufacturers are only (qualified ?) hobbyist , some are true and serious professionals....you don't seems to believe this.

I worked by Cherokee Europe and even after a repair, the used SMPS had to be High pot tested successfully again before to be delivered to the customer.

[james_s]

Overall it looks reasonable, but you'd have to disassemble the transformer to really tell, that's the most likely place for a breakdown between primary and secondary to occur.

For what it's worth, I've got a bunch of various cheap power bricks like this and they have all been fine. I wouldn't use it to power a medical device or something I was going to use while sitting in a pool or something but for general use I see no reason to be alarmed.

[fourtytwo42]

So you would be happy with this product then https://youtu.be/3Hdn0MuCK_0

Glib platitudes are not good enough when it comes to personal safety, there are to many people advising the unwary that everything will be ok and just ignore those that try to warn them!

[Zero999]

I disagree. In theory no testing is necessary. If the insulation is so thin, that every product needs to be tested, then the design is marginal and isn't safe to begin with. It should be redesigned so it will comfortably exceed the minimum requirements.

Electrical safety can be guaranteed, purely by design. The breakdown voltage of various insulating materials is a well known parameter. All one needs to do is make sure the material is thick enough, sufficiently mechanically robust and large enough creapage and clearances are used throughout the design. In reality it may be necessary to test a small sample of materials/products, to ensure they meet the specifications but 100% testing is not desirable. Any products which have been subjected to the high potential test should be destroyed and not sold to the customer.

For sure, you have not the least experience in quality control of electricity and power electronics gears to write such a bullshit.
Should you be quality control manager, the first customer who will be electrocuted will send you to jail , and for a very long time.

For exemple, technical specifications of a SMPS...High pot tested at 4242Vdc 1m between primary and secondary.

You're talking nonsense. For a start, there's nothing in that data sheet to say that every single device has been tested, just that it can withstand 4242V, from primary to secondary and that the design (not necessarily every single model) has been tested to comply with various safety standards.

Safety can be guaranteed by design. There are far more potentially dangerous things, than a mains powered switching regulator, which are engineered to be safe and never tested to the design limits. Take a house for example, the supporting structures in the building, will be designed to hold the maximum nominal load, multiplied by a large safety factor (typically two), but it will never be tested to confirm this. The engineer who designed the building, ensured that it will be strong enough and conform to the relevant building regulations. The same is true for cars, aeroplanes, bridges, etcetera.

Testing can be useful to verify a design meets certain requirements (safety being only one of them) but once that's been achieved, it only needs to be done sporadically, to confirm the suppliers and manufacturing process are adequate. It's not necessary or desirable repeat tests, which are potentially dangerous. I certainly wouldn't want to live in a house, which has been tested right up to the design stress limits, that could cause microscopic cracks to form. I'd rather live in a house which has been designed to be safe and not subjected to unnecessary stress.

[fourtytwo42]

I disagree. In theory no testing is necessary. If the insulation is so thin, that every product needs to be tested, then the design is marginal and isn't safe to begin with. It should be redesigned so it will comfortably exceed the minimum requirements.

Electrical safety can be guaranteed, purely by design. The breakdown voltage of various insulating materials is a well known parameter. All one needs to do is make sure the material is thick enough, sufficiently mechanically robust and large enough creapage and clearances are used throughout the design. In reality it may be necessary to test a small sample of materials/products, to ensure they meet the specifications but 100% testing is not desirable. Any products which have been subjected to the high potential test should be destroyed and not sold to the customer.

For sure, you have not the least experience in quality control of electricity and power electronics gears to write such a bullshit.
Should you be quality control manager, the first customer who will be electrocuted will send you to jail , and for a very long time.

For exemple, technical specifications of a SMPS...High pot tested at 4242Vdc 1m between primary and secondary.

You're talking nonsense. For a start, there's nothing in that data sheet to say that every single device has been tested, just that it can withstand 4242V, from primary to secondary and that the design (not necessarily every single model) has been tested to comply with various safety standards.

Safety can be guaranteed by design. There are far more potentially dangerous things, than a mains powered switching regulator, which are engineered to be safe and never tested to the design limits. Take a house for example, the supporting structures in the building, will be designed to hold the maximum nominal load, multiplied by a large safety factor (typically two), but it will never be tested to confirm this. The engineer who designed the building, ensured that it will be strong enough and conform to the relevant building regulations. The same is true for cars, aeroplanes, bridges, etcetera.

Testing can be useful to verify a design meets certain requirements (safety being only one of them) but once that's been achieved, it only needs to be done sporadically, to confirm the suppliers and manufacturing process are adequate. It's not necessary or desirable repeat tests, which are potentially dangerous. I certainly wouldn't want to live in a house, which has been tested right up to the design stress limits, that could cause microscopic cracks to form. I'd rather live in a house which has been designed to be safe and not subjected to unnecessary stress.

Well IMOP your both right but viewing the matter from opposite ends without much knowledge of what goes on in the middle! In my experience specifying something to an Asian manufacturer is rather different to how things might work in a euro/usa concept, where as in the latter it is assumed the other party to the contract will adhere to the specification throughout the contract in Asia it is seen as an initial requirement that must be met for a few samples but thereafter the manufacturer may at will relax constraints to meet lead time demands or improve profit margins or component supply difficulties. Quite simply eastern and western minds do not think alike So as a designer with western manufacturing it is indeed only necessary to test a few examples rigorously and rely upon the manufacturer afterwards for eastern fabs rigorous quality control is necessary to ensure continuing compliance with the contract. Add to this a commodity product and a whole new ballpark opens up that of copying or stealing IP, this is remarkably common and identical designs but badged under different names rapidly appear of course any adherence to the original design requirements are completely ignored as often all that is stolen is a pcb layout and parts list, the transformer being reconstructed (but cost reduced) by reverse engineering (stripping and turn counting). This is the real world of what actually happens, many products like smps landing on consumer laps in the west are 3 or more times removed (copied) from the original design. To imagine any original designer specifications remain in there construction is pure wishful thinking...sorry!!

[Zero999]

I disagree. In theory no testing is necessary. If the insulation is so thin, that every product needs to be tested, then the design is marginal and isn't safe to begin with. It should be redesigned so it will comfortably exceed the minimum requirements.

Electrical safety can be guaranteed, purely by design. The breakdown voltage of various insulating materials is a well known parameter. All one needs to do is make sure the material is thick enough, sufficiently mechanically robust and large enough creapage and clearances are used throughout the design. In reality it may be necessary to test a small sample of materials/products, to ensure they meet the specifications but 100% testing is not desirable. Any products which have been subjected to the high potential test should be destroyed and not sold to the customer.

For sure, you have not the least experience in quality control of electricity and power electronics gears to write such a bullshit.
Should you be quality control manager, the first customer who will be electrocuted will send you to jail , and for a very long time.

For exemple, technical specifications of a SMPS...High pot tested at 4242Vdc 1m between primary and secondary.

You're talking nonsense. For a start, there's nothing in that data sheet to say that every single device has been tested, just that it can withstand 4242V, from primary to secondary and that the design (not necessarily every single model) has been tested to comply with various safety standards.

Safety can be guaranteed by design. There are far more potentially dangerous things, than a mains powered switching regulator, which are engineered to be safe and never tested to the design limits. Take a house for example, the supporting structures in the building, will be designed to hold the maximum nominal load, multiplied by a large safety factor (typically two), but it will never be tested to confirm this. The engineer who designed the building, ensured that it will be strong enough and conform to the relevant building regulations. The same is true for cars, aeroplanes, bridges, etcetera.

Testing can be useful to verify a design meets certain requirements (safety being only one of them) but once that's been achieved, it only needs to be done sporadically, to confirm the suppliers and manufacturing process are adequate. It's not necessary or desirable repeat tests, which are potentially dangerous. I certainly wouldn't want to live in a house, which has been tested right up to the design stress limits, that could cause microscopic cracks to form. I'd rather live in a house which has been designed to be safe and not subjected to unnecessary stress.

Well IMOP your both right but viewing the matter from opposite ends without much knowledge of what goes on in the middle! In my experience specifying something to an Asian manufacturer is rather different to how things might work in a euro/usa concept, where as in the latter it is assumed the other party to the contract will adhere to the specification throughout the contract in Asia it is seen as an initial requirement that must be met for a few samples but thereafter the manufacturer may at will relax constraints to meet lead time demands or improve profit margins or component supply difficulties. Quite simply eastern and western minds do not think alike So as a designer with western manufacturing it is indeed only necessary to test a few examples rigorously and rely upon the manufacturer afterwards for eastern fabs rigorous quality control is necessary to ensure continuing compliance with the contract. Add to this a commodity product and a whole new ballpark opens up that of copying or stealing IP, this is remarkably common and identical designs but badged under different names rapidly appear of course any adherence to the original design requirements are completely ignored as often all that is stolen is a pcb layout and parts list, the transformer being reconstructed (but cost reduced) by reverse engineering (stripping and turn counting). This is the real world of what actually happens, many products like smps landing on consumer laps in the west are 3 or more times removed (copied) from the original design. To imagine any original designer specifications remain in there construction is pure wishful thinking...sorry!!

I see your point. Yes you need to supervise foreign manufactures. Still 100% testing is not required. Spot testing, and inspection, including destructive, should catch contractors cutting corners.

Electrical safety can be guaranteed, purely by design. The breakdown voltage of various insulating materials is a well known parameter. All one needs to do is make sure the material is thick enough, sufficiently mechanically robust and large enough creapage and clearances are used throughout the design. In reality it may be necessary to test a small sample of materials/products, to ensure they meet the specifications but 100% testing is not desirable. Any products which have been subjected to the high potential test should be destroyed and not sold to the customer.

Power supplies, optocouplers, transformers are Hi-POT tested individually. You cannot guarantee by design against assembly defects.

Yes, you can. By design, I mean the whole process, not just the components and laying out the PCB, but the materials and how they're physically put together. The video you've linked to demonstrates this perfectly. Proper design could have avoided this fault. Testing could have made this worse by weakening some marginal components, causing them to fail in the field, which might even be the case here!

[fourtytwo42]

Sadly I beg to differ, It seems you have limited or no experience of manufacturing!

[Mr. Scram]

And if you saw the way transformers are mass produced (thrown together) in Asia you would understand that the designers intention is completely sacrificed to cut corners so specifying certain types of insulation is useless.
As we now live in a deregulated world to encourage trade nothing is tested anymore hence we have corner cutting manufacturing directly in contact with the consumer, it is only a matter of luck that we do not have more serious accidents and of course much goes unreported as the units are simply thrown away if deemed to be faulty.
As I said before never trust your life to this kind of mass produced rubbish

You could argue that if you don't fully test all the electrical equipment yourself, you are trusting your life to it. A faulty USB charger can quite easily burn your house down, with you in it. You trust the equipment around you every day.

[Zero999]

And if you saw the way transformers are mass produced (thrown together) in Asia you would understand that the designers intention is completely sacrificed to cut corners so specifying certain types of insulation is useless.
As we now live in a deregulated world to encourage trade nothing is tested anymore hence we have corner cutting manufacturing directly in contact with the consumer, it is only a matter of luck that we do not have more serious accidents and of course much goes unreported as the units are simply thrown away if deemed to be faulty.
As I said before never trust your life to this kind of mass produced rubbish

You could argue that if you don't fully test all the electrical equipment yourself, you are trusting your life to it. A faulty USB charger can quite easily burn your house down, with you in it. You trust the equipment around you every day.

Indeed, the same thing could be said about anything. There's such thing as due diligence. If you can prove, you've designed it properly and you've got paperwork from the manufacturer, stating they've followed your instructions to the letter, to meet the specification, then you've gone as far as you can. If there's a manufacturing defect, which causes a safety issue, then you've done all you can. The blame then falls on the manufacturer for not building it to specification.

In practise, most companies don't even make their own power supplies any more. They use subcontractors. I doubt Apple or any other company, tests their PSUs made by a third party.

Why is it that as soon as the issue of electrical safety is raised some people seem to lose all perspective? They'll happily live in buildings, drive cars and take flights in planes, which haven't been 100% stress tested, to the design requirements, yet they mandate the same for a mundane, comparatively low risk, switched mode power supply?

[wraper]

Yes, you can. By design, I mean the whole process, not just the components and laying out the PCB, but the materials and how they're physically put together. The video you've linked to demonstrates this perfectly. Proper design could have avoided this fault. Testing could have made this worse by weakening some marginal components, causing them to fail in the field, which might even be the case here!

Hi pot testing does not weaken insulation if it was good in the first place. While under high voltage, leakage current is measured, if it exceeds (tiny) specified value, device does not pass. If the insulation was weakened during test, it would not pass test in the first place. I don't know for sure if that PSU was Hi-POT tested, likely it was and during test insulation was good enough to withstand it. But then it could drop, and wire could move a little bit over damaged insulation area. There was no fault in the design itself, reinforced multiplayer insulation was mechanically damaged during production. BTW I was myself in the factory where they Hi-POT test their devices where high voltage is present. Haven't you ever seen hi-pot stickers?

[Zero999]

Yes, you can. By design, I mean the whole process, not just the components and laying out the PCB, but the materials and how they're physically put together. The video you've linked to demonstrates this perfectly. Proper design could have avoided this fault. Testing could have made this worse by weakening some marginal components, causing them to fail in the field, which might even be the case here!

Hi pot testing does not weaken insulation if it was good in the first place. While under high voltage, leakage current is measured, if it exceeds (tiny) specified value, device does not pass. If the insulation was weakened during test, it would not pass test in the first place. I don't know for sure if that PSU was Hi-POT tested, likely it was and during test insulation was good enough to withstand it. But then it could drop, and wire could move a little bit over damaged insulation area. There was no fault in the design itself, reinforced multiplayer insulation was mechanically damaged during production. BTW I was myself in the factory where they Hi-POT test their devices where high voltage is present. Haven't you ever seen hi-pot stickers?

Yes. I've seen hi-pot stickers and have never seen the point. It won't stop marginal devices from slipping through or show up poor manufacturing. Only disassembly and visual inspection can do that. Yes, you're right, it shouldn't damage the insulation, as long as the test voltage isn't too high (much lower than the withstand voltage) and hasn't been applied for too long.

I disagree with you about the construction of the transformer. There isn't enough insulation, where the windings pass near one another and it appears too much force has been used when it was wound. Very shoddy.

[Mr. Scram]

Please guys, remember that we're here to help Charkel. I don't feel he is benefiting much from this discussion. It seems pretty clear some products are tested to their design specifications, but that others are not. The discussion which are tested in what way and why seems more suited for another thread.

[wraper]

Yes, you can. By design, I mean the whole process, not just the components and laying out the PCB, but the materials and how they're physically put together. The video you've linked to demonstrates this perfectly. Proper design could have avoided this fault. Testing could have made this worse by weakening some marginal components, causing them to fail in the Yes. I've seen hi-pot stickers and have never seen the point.

The point is to indicate that test was passed. Therefore no untested device may accidentally slip further.

It won't stop marginal devices from slipping through or show up poor manufacturing. Only disassembly and visual inspection can do that. Yes, you're right, it shouldn't damage the insulation, as long as the test voltage isn't too high (much lower than the withstand voltage) and hasn't been applied for too long.

Visual inspection will happily pass most of the faults  , it's only good for finding the most obvious defects. It's really obvious you never dealt with actual production. And you cannot inspect inside the parts (where most of the faults happen), and you won't be able to see imperfections in insulation anyway. Exactly Hi-POT test is what weeds out marginal devices. Withstand voltage means withstand don't confuse that with breakdown voltage.

I disagree with you about the construction of the transformer. There isn't enough insulation, where the windings pass near one another and it appears too much force has been used when it was wound. Very shoddy.

Wire used in secondary winding had reinforced insulation by itself. Basically you could wind primary winding directly on top of it, and it would be just fine. The issue was that a big piece of that insulation was missing because of mechanical damage.
EDIT: Something like this http://www.totoku.com/products/cables/tiw/post-2.php, additional insulation between windings is not required at all.

[oldway]

"High voltage test (also called hipot or HV test worldwide) has by far the highest priority among all of the electrical safety tests. This test of the dielectric strength is mandatory as it is part of the requirements of all national standards (e.g. VDE and UL) and international standards (e.g. EN and IEC). The often used term insulation test or insulation measurement is wrong in this context as a real, ohmic resistance is measured with the direct voltage in this case. In contrast, the high voltage test is carried out with both direct and alternating current. The result of this test is a leakage current, flowing between live parts and the metallic casing. Devices of protection class I always come with a metal casing, whereas the casing of the devices of protection class II must be replaced by a metallic contour during the test. In reality, the high voltage test is a geometric distance measurement in the DUT (Device Under Test): If the DUT withstands a certain voltage without flashover, an appropriate minimum distance between the metal casing and electrical components in the device is ensured. Thus, the term withstand voltage is also plausible again."

https://www.spselectronic.com/en/products/high-voltage/