Why so many ferrite beads in Meanwell AC/DC power brick ?

[BravoV]

Its just a common Meanwell AC to DC 12 Volt 8.5 Amp, and took the cover off to clean the dust recently.

Noticed there are so many ferrite beads, I counted there are 12 pieces, why so many ?

What will happened, say I take out all those beads and leave none ?

Photos attached below.

[senso]

Its to attenuate high frequency radiated noise, if you remove them(why would you do that??? ) it will no longer be able to meet FCC and other standards.

[SeanB]

The ones on the leads of the secondary rectifier are there to slow down switching, without them the diode will have increased reverse conduction loss, and will run hotter at full load. The rest are there to also tame sharp pulse edges, so they do not use the wiring as a convenient radiating antenna.

[JoeO]

Try to get access to the internet.  When you do, type in Google.com and into the search box type "ferrite beads".  Click on the link that takes you to the wikipedia page for ferrite beads.

Googling also works for other topics, not just ferrite beads.

[T3sl4co1l]

The ones on the leads of the secondary rectifier are there to slow down switching, without them the diode will have increased reverse conduction loss, and will run hotter at full load. The rest are there to also tame sharp pulse edges, so they do not use the wiring as a convenient radiating antenna.

Not much effect on switching losses, but all the EMC thingys.

Ferrite beads are mainly used to increase impedance and decrease Q of parasitic resonances in a system.  They aren't useful for much gross attenuation, but generally give a few dB worth of filtering on their own.  When dampening resonances, the advantage could be 10 or even 20dB.

Notice that not all "Y type" capacitor positions are populated.  It's normal to design a board with many options in mind, then optimize the component placement on later testing.  Probably, the ferrite beads were added as a necessary afterthought as well.

Tim

[keymaster]

This can also be the result of a very frustrated engineer. how give up finding the problem of failing emc levels. so he/she nuke the thing with ferrite beads. problem solved .

[BravoV]

Its to attenuate high frequency radiated noise, if you remove them(why would you do that??? ) it will no longer be able to meet FCC and other standards.

As I understand slightly that these beads must have something to do with those EMI or EMC matters.

Its just I'm curious, as there are many other typical AC/DC bricks that I have seen, like those "decent" ones for laptop, those don't use that many as this, this is the 1st time I noticed them, hence this thread.

No, don't worry, not planning to remove them, this thread is purely to learn. 

[BravoV]

The ones on the leads of the secondary rectifier are there to slow down switching, without them the diode will have increased reverse conduction loss, and will run hotter at full load. The rest are there to also tame sharp pulse edges, so they do not use the wiring as a convenient radiating antenna.

Ok, noted, especially for the secondary side schottky diodes.

What about those at the primary mains silicon rectifier, two are installed at the HV rectified output pos & neg terminals ? To damp the switching noise getting out into the mains line ?

[BravoV]

Try to get access to the internet.  When you do, type in Google.com and into the search box type "ferrite beads".  Click on the link that takes you to the wikipedia page for ferrite beads.

Googling also works for other topics, not just ferrite beads.

Thanks, noted, will do once I have access to internet, and good for you too, did you just discover that ? 

[BravoV]

The ones on the leads of the secondary rectifier are there to slow down switching, without them the diode will have increased reverse conduction loss, and will run hotter at full load. The rest are there to also tame sharp pulse edges, so they do not use the wiring as a convenient radiating antenna.

Not much effect on switching losses, but all the EMC thingys.

Ferrite beads are mainly used to increase impedance and decrease Q of parasitic resonances in a system.  They aren't useful for much gross attenuation, but generally give a few dB worth of filtering on their own.  When dampening resonances, the advantage could be 10 or even 20dB.

Notice that not all "Y type" capacitor positions are populated.  It's normal to design a board with many options in mind, then optimize the component placement on later testing.  Probably, the ferrite beads were added as a necessary afterthought as well.

Tim

This can also be the result of a very frustrated engineer. how give up finding the problem of failing emc levels. so he/she nuke the thing with ferrite beads. problem solved .

Now, this is actually made me curious from the 1st place.

Do all this beads mean they're not doing it right from the beginning ? As I noticed similar AC/DC brick, like those used in quality laptops don't have this many beads used in it.

[BravoV]

Btw, this is genuine product, not those Meanwell clones.

Also attached the safety compliance from the datasheet.

[MagicSmoker]

This can also be the result of a very frustrated engineer. how give up finding the problem of failing emc levels. so he/she nuke the thing with ferrite beads. problem solved .

Now, this is actually made me curious from the 1st place.

Do all this beads mean they're not doing it right from the beginning ? As I noticed similar AC/DC brick, like those used in quality laptops don't have this many beads used in it.

I suspect keymaster is on to something here. Ferreting (or should I say ferriting?) out the true source(s) of EMI/RFI requires a spectrum analyzer, a set of H & E field probes, and the knowledge of how to use them... or a sufficiently large budget to pay a compliance testing lab to find them for you and then make recommendations on how to solve them that may or may not require additional rounds of testing (all billed hourly, of course).

I can easily imagine a harried, and perhaps not quite experienced, engineer just throwing a bunch of ferrite beads on any through-hole component showing some leg (!) and calling it a day. After all, a single hour at a compliance testing lab - even one in China - probably costs as much as 10,000 ferrite beads...

[T3sl4co1l]

Do all this beads mean they're not doing it right from the beginning ? As I noticed similar AC/DC brick, like those used in quality laptops don't have this many beads used in it.

Unclear.

Likely, a dV/dt and dI/dt snubber around the switching transistor and/or output diode would help.  And improved transformer windup can do wonders: lower LL, moving the "switching" end of the primary away from the secondary, add a shield, etc.

Ferrite beads on Y caps is kind of peculiar.  That's usually a situation you want minimum impedance, not added impedance.  It may've been a "shotgun" approach, close enough for their purposes, and not fully optimized.

Ideally, when testing different parts, you want to try many things, then undo each change to see which ones are the most effective.

Tim

[tszaboo]

This can also be the result of a very frustrated engineer. how give up finding the problem of failing emc levels. so he/she nuke the thing with ferrite beads. problem solved .

Now, this is actually made me curious from the 1st place.

Do all this beads mean they're not doing it right from the beginning ? As I noticed similar AC/DC brick, like those used in quality laptops don't have this many beads used in it.

I suspect keymaster is on to something here. Ferreting (or should I say ferriting?) out the true source(s) of EMI/RFI requires a spectrum analyzer, a set of H & E field probes, and the knowledge of how to use them... or a sufficiently large budget to pay a compliance testing lab to find them for you and then make recommendations on how to solve them that may or may not require additional rounds of testing (all billed hourly, of course).

I can easily imagine a harried, and perhaps not quite experienced, engineer just throwing a bunch of ferrite beads on any through-hole component showing some leg (!) and calling it a day. After all, a single hour at a compliance testing lab - even one in China - probably costs as much as 10,000 ferrite beads...

Yes. Probably they have the design for chinese market, and they wanted to introduce it to the west, failed EMC testing, maybe also a second time, so just throw it wherever. Sometimes even I get angry at those power rails, and I just thow a dozen capacitor at it to solve the problem. Or just start with that in the first place. 100nF capacitor costs 0.1 cent each? Get some, get some more.

[sokoloff]

I can easily imagine a harried, and perhaps not quite experienced, engineer just throwing a bunch of ferrite beads on any through-hole component showing some leg (!) and calling it a day. After all, a single hour at a compliance testing lab - even one in China - probably costs as much as 10,000 ferrite beads...

The electrical engineering equivalent of boating's "If you can't tie a knot, tie a lot..."

[Kilrah]

Does the "ferrite over longer lead" have that much benefit over just keeping the leads short in the first place?

[tooki]

Try to get access to the internet.  When you do, type in Google.com and into the search box type "ferrite beads".  Click on the link that takes you to the wikipedia page for ferrite beads.

Googling also works for other topics, not just ferrite beads.

This is the Beginners forum. Simple questions ARE allowed. 

[T3sl4co1l]

Does the "ferrite over longer lead" have that much benefit over just keeping the leads short in the first place?

Err, the point is to make the leads "long" -- and also lossy.

Tim

[Bud]

Speaking of which, i've seen an application note, i believe one from ADI, saying running separate wires from analog AVDD and digital DVDD terminals on a PCB to a same power supply solves crosstalk problem AVDD to DVDD because of added wire inductance, so no additional board level measures are needed.

Edit: just remembered it was not an applicatioj note but a private email communication with an ADI engineer in regards to spurs attributed to powering a DDS analog and digital terminals from a same power supply.

[Dago]

These power supplies tend to use 1-layer PCBs which is not a great starting point for a design which should pass EMC tests. The ferrite beads are to work around that. They are cheap.

[analogo]

These power supplies tend to use 1-layer PCBs which is not a great starting point for a design which should pass EMC tests.

What EMC problem do 1-layer PCBs create that 1-layer PCBs don't?

[homebrew]

These power supplies tend to use 1-layer PCBs which is not a great starting point for a design which should pass EMC tests.

What EMC problem do 1-layer PCBs create that 1-layer PCBs don't?

You mean in comparison to 2-layer designs? Well, on a single-layer board, you cannot have a ground plane. Hence you will inevitably have larger current loops and hence more radiation.

However, even on a two-layer board it is difficult to control the return paths as you want to also use the second plane as a signal layer and thus it is not continuous anymore.

A friend of mine advised some time ago first to maximize ground plane area on both sides of the board and then to hold my two-layer designs against the window ans see where light shines through the substrate. The less light, the better the coverage of ground-planes. And of course you should use heavy via stitching to create low impedance connections between the planes ...

[Gyro]

These power supplies tend to use 1-layer PCBs which is not a great starting point for a design which should pass EMC tests.

What EMC problem do 1-layer PCBs create that 1-layer PCBs don't?

You mean in comparison to 2-layer designs? Well, on a single-layer board, you cannot have a ground plane. Hence you will inevitably have larger current loops and hence more radiation.

However, even on a two-layer board it is difficult to control the return paths as you want to also use the second plane as a signal layer and thus it is not continuous anymore.

A friend of mine advised some time ago first to maximize ground plane area on both sides of the board and then to hold my two-layer designs against the window ans see where light shines through the substrate. The less light, the better the coverage of ground-planes. And of course you should use heavy via stitching to create low impedance connections between the planes ...

Unfortunately virtually all power bricks and consumer equipment PSUs are built on CEM-1 boards (punched, not drilled). That makes single layer layout more or less inevitable, it's just not economic to use 2 layer FR4 in high volume. There are alwayssome compromises with current loops that require layout ingenuity to minimise. usually with a couple of ferrites, and extra snubbers ending up in the mix. The primary and secondary side of the circuit should look very different, lots of track spacing  and empty area on the primary side (with critical tracks as short as possible), and close to 100% copper coverage, with ground copper surrounding everything, on the secondary.

I agree, that NANDBlog etc. That Meanwell looks like an extreme example of a 'scattergun' approach to getting an existing PSU though an emmissions test with minimum rework and test house time. Different cost and time constraints result in different solutions - this one isn't subtle but it obviously worked for them! It probably wasn't even worth the extra expense to go back and see which ones they could safely remove.

[T3sl4co1l]

It also helps that the power level is low, so not much current is being switched.  The primary side basically doesn't care about stray inductance (the switching impedance is close to the transmission line impedance for traces of that construction), and the secondary side may well benefit from the extra inductance (since, as you can see, diode recovery is slowed down, intentionally, using ferrite beads).  If nothing else, the transformer dominates in LL and Cp.

Tim