How to test DC adapter for reliability and performance?

[Surajg]

Hello Everyone,

I am newbie to electronics. Please forgive me if I have posted it in incorrect forum 
Recently I purchased some 12V-2.0A DC adapters from market.

I did check the voltage on multimeter and its output is OK as mentioned.
I was wondering how to test any adapter for reliability or other issues (like noise etc using oscilloscope or any other instrument) if we are planning to use it for mass production.

Thanks

[Zero999]

Where did you buy it from and what's the brand?

If you're going to buy a large quantity, then you could get a hi-pot. tester to check the insulation between primary and secondary. Such a test is normally destructive, or at least weakens the insulation, so the device mustn't be used afterwards. You could also open it up and visually inspect it. What brand and temperature rating capacitors does it use? Are the creepage distances on the PCB sufficient? Does the transformer have sufficient separation? Does it have EMI suppression i.e. a common mode choke and Y capacitor?

[duckduck]

Depending on the expected use, it might make sense to verify that the DC adapters meet their spec of 12V@2A. This could be done by hooking it up to a DC electronic load (or a very robust 6 Ohm resistor) and then monitoring the temperature of the adapter to make sure that it doesn't get too hot after several hours. The voltage could also be monitored at full (2A) load for several hours to make sure that it stays high enough to meet your specifications. You could even do this testing at depressed or elevated ambient temperatures (e.g. 0° C or 45° C). It's really up to you to define what your requirements are.

EDIT:

You could watch the voltage on an oscilloscope to see how much electrical noise there is, but unless you know the requirements/sensitivity to noise of the device that you are powering this is not very helpful.

Probably the most important thing is that you know the range of voltages (and power consumption) that your device operates reliably within e.g. 11.5V - 13.7V, and then ensure that the DC adapters will provide this output with a nice margin under all circumstances that you expect your device to see in the field.

You can also use a variac to ensure that the DC adapter puts out the rated voltage / amperage at whatever the range of voltages are for the countries that you wish to sell your device in e.g. 240V +/- 5V.

[Gyro]

If you are planning to use the adapters for mass production, then you are effectively taking over approvals and safety responsibility for them as they will be shipping inside your product's retail packaging.

In additional to the destructive testing that Zero999 suggests, you need to obtain full EMC and Safety approvals documentation from the manufacturer and ensure that it is honest and correct for your market (consumer, IT etc). Once you are in mass production, responsibility for them becomes yours and the adapter manufacturer, your subcontractor. You need full approval documentation to demonstrate that you undertook due diligence if the worst happens.

[wraper]

Where did you buy it from and what's the brand?

If you're going to buy a large quantity, then you could get a hi-pot. tester to check the insulation between primary and secondary. Such a test is normally distructive, or at least weakens the insulation, so the device mustn't be used afterwards. You could also open it up and visually inspect it. What brand and temperature rating capacitors does it use? Are the creepage distances on the PCB sufficient? Does the transformer have sufficient separation? Does it have EMI supression i.e. a common mode chock and Y capacitor?

It's only destructive if you test a breakdown voltage. Hi-pot test is not destructive because a limited voltage is applied which insulation must withstand and every device like PSU must be tested at factory if things are done right.

[Zero999]

Where did you buy it from and what's the brand?

If you're going to buy a large quantity, then you could get a hi-pot. tester to check the insulation between primary and secondary. Such a test is normally destructive, or at least weakens the insulation, so the device mustn't be used afterwards. You could also open it up and visually inspect it. What brand and temperature rating capacitors does it use? Are the creepage distances on the PCB sufficient? Does the transformer have sufficient separation? Does it have EMI suppression i.e. a common mode choke and Y capacitor?

It's only destructive if you test a breakdown voltage. Hi-pot test is not destructive because a limited voltage is applied which insulation must withstand and every device like PSU must be tested at factory if things are done right.

It depends on the type of test done. There's the test which is often performed on every device which uses a lower voltage and is non-destructive, then there's another test which uses a much higher voltage to establish the breakdown voltage of the insulation. In this instance, I'd recommend performing the latter on a sample of units. It would be nice to test every single one of them, but if they're manufactured to a high enough standard and the manufacturer is trustworthy, then it probably isn't necessary. I don't remember the voltages, or exact names of the tests, which is an exercise for the original poster to research.

[wraper]

It would be nice to test every single one of them, but if they're manufactured to a high enough standard and the manufacturer is trustworthy, then it probably isn't necessary. I don't remember the voltages, or exact names of the tests, which is an exercise for the original poster to research.

It's necessary because once in a while there may be defective transformer, solder bridge or something like that.

[wraper]

at 8:50 every PSU is Hi-pot tested

[Zero999]

It would be nice to test every single one of them, but if they're manufactured to a high enough standard and the manufacturer is trustworthy, then it probably isn't necessary. I don't remember the voltages, or exact names of the tests, which is an exercise for the original poster to research.

It's necessary because once in a while there may be defective transformer, solder bridge or something like that.

It depends on the exact construction of the transformer and PCB. There are ways of designing something so that can never happen and eliminate the need for such tests, i.e. winding  the transformer on a double section bobbin, using wire with thick insulation, which won't scrape of and having generous creepage and/or slots between the mains and secondary traces on the PCB. Either way, if you test and inspect a few of every batch and they're good, it's highly likely the manufacturer has a decent testing regime in place. In an ideal world, you should be able to trust the manufacturer to get this right and not have to perform any inspection and tests.

[wraper]

i.e. winding  the transformer on a double section bobbin,

Does not eliminate assembly defect where single wire gets into both sections.

using wire with thick insulation

Defective PSU in the video above used exactly that, insulation was mechanically damaged. The result of not Hi-pot testing may be that someone gets electrocuted and then you will be required to recall millions of power supplies even though 99.999% of them are completely fine.

[Zero999]

i.e. winding  the transformer on a double section bobbin,

Does not eliminate assembly defect where single wire gets into both sections.

Yes it does. I should have been more clear. By double section bobbin, I meant two bobbins: one for the primary and a separate one for the secondary, so there's no chance that can happen.

[Gyro]

Unfortunately you don't get double bobbin transformers in smps mains adaptors, the leakage inductance is too high. To make matters worse, the primary and secondary are usually interleaved.

I very much doubt the OP would be sourcing traditional 50/60Hz transformer ones.

[TimNJ]

At my first internship, for a new project, the company sourced around five (somewhat dodgy) power adapters with similar specs. The lead engineer had me open all of the units, apply polyester tape to the transformer winding, output cap(s), output diode, and primary-side switching transistor. (This was to give each component roughly the same emissivity, around 0.95-0.99.) Then, I ran each power supply at the same load for about an hour each. After one hour, I checked the temperatures of each of the aforementioned components with an IR gun. We compared the temperatures of the five units and the chose the power adapter with the best overall temperatures.

If testing at room temperature (25°C) like this, with the power adapter housing cracked open, I'd suggest to ensure that following temperatures are met:

Transformer: 85-90°C (max.)

Output diode, primary-side transistor, bridge rectifier, etc: 85°C (max.)

Output capacitors(s) and primary bulk cap:

Should be 105°C rating on the label. If it's 85°C, don't even bother. Check the manufacturer datasheet to find the endurance rating (as below):

~50°C  (max.) for 2,000 hour rated cap
~60°C  (max.) for 5,000 hour rated cap
~70°C  (max.) for 10,000 hour rated cap

These temperatures would give you about 30,000 - 35,000 hours life at full load and at max ambient of 40°C. Obviously, it's up to you to determine what is "reliable enough".


Those are just ballpark suggestions. And again, the temperatures above are assuming you are measuring component temperatures at room temperature ambient. Safety-wise, the transformer is most critical. Temperature testing won't tell you if the transformer is constructed correctly, but the farther away the temperature rating is from the enamel rating of the wire (usually 130°C), the better. The components will probably run at least 10°C hotter in the enclosure, plus add another 15°C to account for a typical worse case ambient of 40°C.

Also check that the Y-caps (usually blue ceramic disc capacitor) are Y1 or Y2 type. It should be printed with various symbols like VDE and other agencies. If it just says "102 1KV" that's probably not a safety rated capacitor.

[wraper]

i.e. winding  the transformer on a double section bobbin,

Does not eliminate assembly defect where single wire gets into both sections.

Yes it does. I should have been more clear. By double section bobbin, I meant two bobbins: one for the primary and a separate one for the secondary, so there's no chance that can happen.

I know what it is. And there is a chance that something goes awry during winding (like a few turns going into a wrong section) or bobbin being defective and quite likely nobody will notice it visually.

[Zero999]

i.e. winding  the transformer on a double section bobbin,

Does not eliminate assembly defect where single wire gets into both sections.

Yes it does. I should have been more clear. By double section bobbin, I meant two bobbins: one for the primary and a separate one for the secondary, so there's no chance that can happen.

I know what it is. And there is a chance that something goes awry during winding (like a few turns going into a wrong section) or bobbin being defective and quite likely nobody will notice it visually.

You clearly don't know what it is. The primary and secondary coils are wound separately, on different formers, so it isn't possible for any of the turns to go onto the wrong section. They are then placed, either side by side, or inside of one another, on the same core. I have disassembled transformers before and have been able to hold the secondary in one hand and and the primary in the other, without unwinding them. I have a small ignition transformer on my desk at the moment, which is wound using this method. The primary is wound in the middle, on a small former, then a larger former, containing the secondary is placed on top and the whole thing potted in transparent resin.

[wraper]

You clearly don't know what it is. The primary and secondary coils are wound separately, on different formers, so it isn't possible for any of the turns to go onto the wrong section. They are then placed, either side by side, or inside of one another, on the same core. I have disassembled transformers before and have been able to hold the secondary in one hand and and the primary in the other, without unwinding them. I have a small ignition transformer on my desk at the moment, which is wound using this method. The primary is wound in the middle, on a small former, then a larger former, containing the secondary is placed on top and the whole thing potted in transparent resin.

You talked about bobbins with multiple sections then about separate bobbins. Nonetheless even in that case screw up is still possible during transformer assembly. Wires still come out of bobbins and still need to be connected to terminals.

[Zero999]

You clearly don't know what it is. The primary and secondary coils are wound separately, on different formers, so it isn't possible for any of the turns to go onto the wrong section. They are then placed, either side by side, or inside of one another, on the same core. I have disassembled transformers before and have been able to hold the secondary in one hand and and the primary in the other, without unwinding them. I have a small ignition transformer on my desk at the moment, which is wound using this method. The primary is wound in the middle, on a small former, then a larger former, containing the secondary is placed on top and the whole thing potted in transparent resin.

You talked about bobbins with multiple sections then about separate bobbins. Nonetheless even in that case screw up is still possible during transformer assembly. Wires still come out of bobbins and still need to be connected to terminals.

And there can be no electrical connection between the primary and secondary, if the terminals are on oppisite sides of the transformer. I'll have to dig out one of the transformers I've salvaged from something to show you want I mean.

[wraper]

You clearly don't know what it is. The primary and secondary coils are wound separately, on different formers, so it isn't possible for any of the turns to go onto the wrong section. They are then placed, either side by side, or inside of one another, on the same core. I have disassembled transformers before and have been able to hold the secondary in one hand and and the primary in the other, without unwinding them. I have a small ignition transformer on my desk at the moment, which is wound using this method. The primary is wound in the middle, on a small former, then a larger former, containing the secondary is placed on top and the whole thing potted in transparent resin.

You talked about bobbins with multiple sections then about separate bobbins. Nonetheless even in that case screw up is still possible during transformer assembly. Wires still come out of bobbins and still need to be connected to terminals.

And there can be no electrical connection between the primary and secondary, if the terminals are on oppisite sides of the transformer. I'll have to dig out one of the transformers I've salvaged from something to show you want I mean.

There can be some manufacturing defect you cannot possibly imagine. So relying on "defect cannot possibly happen" in safety critical situation is flawed logic. For example there may be a monkey brain personnel who can assemble wrong something that shouldn't be possibly assembled wrong. For example Proton rocket with 3 out of 6 acceleration sensors mounted upside down, even though there were mounting guides which should make doing so impossible, not to say it even passed inspection.

[Zero999]

You are funny the way you like to argue for no reason. Is it a hobby? Anyway, that video is completely irrelevant and off-topic, since it's a about a rocket launch, rather than a mass produced item.

Of course, there's always a possibility something unforseen might happen and no amount of testing can be done to stop it.

There are ways to design things to be intrinsically safe, with minimal testing required. Indeed, there are instances where, a high pot test won't gaurantee safety, for example if the manufacturer of the wire skimped on the insulation material, so it melts under normal load conditions. The device will pass the high pot test in the factory, but fail dangerously in the field.

Anyway, here are a couple of pictures of transformers which are designed to be intrinsicly safe, without any need for a high pot test. Note how the primary and secondary windings are on completely separate formers and the terminals are opposite one another. There is no way, a wire from the primary can touch the secondary. I've also dismantled transformer which appeared from the outset to have two coils wound on one former, but they were really separate formers, as I was able to completely remove the secondary and rewind it. They cost more to make, but money is saved with less testing.

[TimNJ]

Regarding this whole split-bobbin/multi-section bobbin discussion, probably >99.9% of SMPS adapters <100W do not use this kind of transformer construction. Flyback has been the topology of choice for many years (in various incarnations) and splitting the bobbin into primary and secondary halves is a great way to reduce the efficiency by 10-20%. Of course, for high voltage transformers, as was alluded to, the multi-section bobbins are necessary to prevent turn-to-turn breakdown.

Split bobbin is useful in some resonant topologies where leakage inductance is purposefully added to create an 'L' in the resonant tank circuit. Otherwise, it's not particularly a good ideas.

[Zero999]

Regarding this whole split-bobbin/multi-section bobbin discussion, probably >99.9% of SMPS adapters <100W do not use this kind of transformer construction. Flyback has been the topology of choice for many years (in various incarnations) and splitting the bobbin into primary and secondary halves is a great way to reduce the efficiency by 10-20%. Of course, for high voltage transformers, as was alluded to, the multi-section bobbins are necessary to prevent turn-to-turn breakdown.

Split bobbin is useful in some resonant topologies where leakage inductance is purposefully added to create an 'L' in the resonant tank circuit. Otherwise, it's not particularly a good ideas.

That must explain why I normally see such a construction used for mains frequency and large SMPS transformers.

I have seen a much smaller transformer using an intrinsically safe constuction, with separate bobbins for the primary and secondary, except they're inside one another. It was used inside an SMPS in an old TV. I'll see if I can find it, so I can take and post a picture.

Intrinsically safe transformers are also common in electronic halogen lamp transformers, except a toroidal transformer is used, with the primary encased between two, thick, double insulated plastic cups.
https://320volt.com/en/electronic-transformer-circuit-schematics-12v-halogen-lamp/

I generally prefer to design things which are intrinsically safe. If people need to perform tests, they can, but I certainly don't like relying on them.

[wraper]

Anyway, here are a couple of pictures of transformers which are designed to be intrinsicly safe, without any need for a high pot test. Note how the primary and secondary windings are on completely separate formers and the terminals are opposite one another. There is no way, a wire from the primary can touch the secondary. I've also dismantled transformer which appeared from the outset to have two coils wound on one former, but they were really separate formers, as I was able to completely remove the secondary and rewind it. They cost more to make, but money is saved with less testing.

Those pictures have nothing to do with transformers used in SMPS. If you try to make them like that, performance will suck. Not to say they are not inherently safe. Crack in the bobbin, and you may have mains voltage on the core.

[wraper]

I generally prefer to design things which are intrinsically safe. If people need to perform tests, they can, but I certainly don't like relying on them.

You are checking for possible insulation defects as safety test. Not relying on it while ignoring the rest of production.

[Zero999]

Anyway, here are a couple of pictures of transformers which are designed to be intrinsicly safe, without any need for a high pot test. Note how the primary and secondary windings are on completely separate formers and the terminals are opposite one another. There is no way, a wire from the primary can touch the secondary. I've also dismantled transformer which appeared from the outset to have two coils wound on one former, but they were really separate formers, as I was able to completely remove the secondary and rewind it. They cost more to make, but money is saved with less testing.

Those pictures have nothing to do with transformers used in SMPS. If you try to make them like that, performance will suck. Not to say they are not inherently safe. Crack in the bobbin, and you may have mains voltage on the core.

The core is normally earthed, or in a double insulated enclosure. Either way, choose a material which isn't brittle and it won't crack and it's a non issue.

I generally prefer to design things which are intrinsically safe. If people need to perform tests, they can, but I certainly don't like relying on them.

You are checking for possible insulation defects as safety test. Not relying on it while ignoring the rest of production.

True, I jut was pointing out that is isn't always necessary to perform a hi pot test. I can think of plenty of electrical items which meet, or exceed the safety standards and haven't been subject to such a test. It's definitely mandatory in something like a transformer, with the primary and secondary on the same former, where there's the potential for a wire to be in the wrong place, or the tape to fail, but in other instances it's debatable where's there's any point, or not.

Anyway, it's off-topic. If it needs to be hi pot tested, it's the manufacture's job. I suggested the original poster test a small sample, as due diligence.

And here's a picture of the transformer I was talking about before. It came from an SMPS. The primary is on the outside and secondary in the middle. The two coils are on separately on a material which isn't brittle and won't crack.

[Surajg]

Hello Everyone,

Thanks all for replies 
I have this DC electronic load tester. Can this help for basic test during sampling?
Like I keep it set for 12V and 2A and keep adapter ON and see if heats or any abnormal thing happens to adapter?