Baffles in a PC power supply

[Shrek]

I have noticed that a lot of PC power supplies have baffles next to the fan to direct flow, but this seems a bad solution; any units that avoid this?

This didn't used to be for end mounted fans but seems the case for the larger top mounted fans.

 

[dirtyferret]

I've noticed it being more common as well also units themselves becoming smaller so more OEMs are moving back to 120mm fans as opposed to 140mm. You would need to look up the build of each unit separately to avoid such a unit but why do you believe it to be a bad solution?

 

[P4-630]

I don't care one bit, if they give me a 10~12 year warranty.....

 

[Shrek]

... but why do you believe it to be a bad solution?

I suppose it reduces the fan efficiency and makes for more noise.

 

[#22]

I suppose it reduces the fan efficiency and makes for more noise.

Assume that PSU engineers know better than you E.g. lack of directing the airflow resulting in need of fan spinning faster.

 

[dirtyferret]

I suppose it reduces the fan efficiency and makes for more noise.

On the flip side it "can" direct cooler air to key components to lower temps and prolong component life. My guess is the fan & PSU have test evidence of that otherwise why bring in additional part & labor costs but I doubt they will share those tests results.

*It actually increases airflow in the PSU therefore making it more efficient.

 

[Bobaganoosh]

That looks more like a guard than a baffle to me. Are there wires underneath that area they can't guarantee will never touch the fan blades? As they try to cram more stuff into smaller boxes, it might just be an added precaution to reduce returns or complaints of things like wires getting too close to the blades.

 

[dirtyferret]

That looks more like a guard than a baffle to me. Are there wires underneath that area they can't guarantee will never touch the fan blades? As they try to cram more stuff into smaller boxes, it might just be an added precaution to reduce returns or complaints of things like wires getting too close to the blades.

it's a baffle same as the one on the Leadex VII Pro which has no wires that high near the fan (nor should any unit if properly designed)

 

[Bill_Bright]

but this seems a bad solution;

What would be a better solution?

I agree that it could result in more fan noise - never good, IMO.

But undoubtedly, its purpose is, as dirtyferret noted, to direct or concentrate additional air flow over sensitive areas that need additional cooling. And that might actually result in less fan noise if it keeps those sensitive devices cooler, thus minimizing the need to ramp up fan speeds.

That looks more like a guard than a baffle to me.

I don't see it as a guard. A guard for what? To keep tiny fingers out? They why only cover only 1/3 of the fan? And besides, PSU cases are not intended to be opened, except by a qualified technician. So the "grill" as seen on the backside of that fan is there to keep exploring fingers out.

 

[dirtyferret]

Baffle on the PSU fan from 15 years ago (easy give away for the PSU OEM). Only possible wire that could come close to the fan is in the filtering stage (but it's too tight) the rest are zip tied

 

[Shrek]

The problem, of course, is to avoid the flow going straight out the back, by-passing the circuit, which was not an issue with the old end placement; I just dislike the idea of a partially blocked fan.

I only got as far as the problem... can't say I have a solution, unless directed slats is an option.

 

[dirtyferret]

The problem, of course, is to avoid the flow going straight out the back, by-passing the circuit,

it would be counter productive if the airflow by passed components that need airflow

which was not an issue with the old end placement.

Not following, old end placement?

 

[Shrek]

Not following, old end placement?

 

[TumbleGeorge]

I have noticed that a lot of PC power supplies have baffles next to the fan to direct flow, but this seems a bad solution; any units that avoid this?

This didn't used to be for end mounted fans but seems the case for the larger top mounted fans.

View attachment 353840

Well, well, it's a movable part. Once you don't like it, take a screwdriver, unscrew the two screws and remove it. "The problem" is solved.

 

[Shrek]

And the flow will be improved as well as the noise... except now the air will by-pass the board.

 

[TimatPSUTest.com]

On designs I worked on 15 years ago the issue was to direct the airflow to lower the temperature rise on transformers and inductors due to safety agency requirements. For example, when a power supply is rated to work at 100 volts AC (on the label) the input filter inductor needs to pass temperature rise at 90 volts AC at full load.

 

[dirtyferret]

And the flow will be improved as well as the noise... except now the air will by-pass the board.

Don't get me wrong, I like to try and figure out why things are the way they are.

those fans tend to be 80mm or 92mm so smaller fan will have to spin faster making more noise

 

[Shrek]

I don't care one bit, if they give me a 10~12 year warranty.....

I can understand that, but I find it entertaining to try and figure things out; why should Engineers have all the fun?

 

[Bobaganoosh]

I don't see it as a guard. A guard for what? To keep tiny fingers out? They why only cover only 1/3 of the fan? And besides, PSU cases are not intended to be opened, except by a qualified technician. So the "grill" as seen on the backside of that fan is there to keep exploring fingers out.

Not that kind of guard lol, that's on the outside. this thing is on the inside and it would be like I said, to keep wires out of it (that was the thought behind the comment anyway). Side note: test fingers always creep me out:

In response to the rest of this, I just don't see that being effective with this type of fan. All this will do is create pockets of back-pressure and block off some of the airflow (as well as adding noise). It won't improve airflow in the non-blocked area. At least, that's what all the testing and simulations I've seen of that scenario would suggest....is what I had typed...which is sort of true, but I think I see the purpose now.

The problem, of course, is to avoid the flow going straight out the back, by-passing the circuit, which was not an issue with the old end placement; I just dislike the idea of a partially blocked fan.

I only got as far as the problem... can't say I have a solution, unless directed slats is an option.

This^ is close and really pointed me to a theory, but it's only half the picture.

I initially had a problem with this statement above as it's not going to just make all the air go straight out the back if the baffle isn't there.

Here's my theory...without the baffle, air is pushed down towards the entire pile of heatsinks and components there, creating turbulence and reflections every which way. Being that the back of the box is vented, the air bouncing around on that side has a relief for the pressure that the fan generates constantly and that air can be easily pushed out the back, but the rest of the air (on the opposite side of the vented panel) will tend to just recirculate (like eddies in a river) and have a harder time finding it's way out the back as the fan is pushing against all of it with constant pressure.

If you create a zone of the supply (right in front of that exhaust-vent) where the fan pressure is removed, it would actually create more of a directionality of the airflow in the tight space as the whole area in front of that vent now also has a directionality of the pressure because the fan isn't fighting it. So while it would be inefficient from the fan's perspective, it does cause more of the air in the PSU to actually cycle and leave (replaced by fresh intake air) instead of just circulating in the corner due to all turbulent pressure from all directions in the box.

I was previously thinking more in terms of ducting and had it in my head they were trying to direct the air from the baffled side to the unbaffled side (like take all the fan's pressure and direct it over here ->), but that's not really it. It's blocking some of the fan's airflow so that the pressure causes the air to move from one side of the box to the other more effectively.

Edit: one more thing to add. As dirtyferret said, 120/140mm fans are quieter than 80mm fans, so this design allows them to get airflow with a more pleasant noise, but they also are just throwing away airflow and fan performance to achieve that result in a way that's more pleasant to the ears. It's still going to be worth it though.

 

[Shrek]

I have no doubt that having a baffle is better than none; I just find it an inelegant solution.

I am no way suggesting I could do better; I just like to understand technology, rather than accept it as a black box.

 

[Bobaganoosh]

I have no doubt that having a baffle is better than none; I just find it an inelegant solution.

I am no way suggesting I could do better; I just like to understand technology, rather than accept it as a black box.

As someone who's had to come up with some of those "inelegant solutions" (very kind phrasing by the way), there's always a reason. Sometimes, it's just the best option based on restrictions ("no, you can't make the box bigger", "no, you can't use a more expensive heat-sink", etc. that the marketing and project managers tell me lol). That said, I've worked with plenty of MEs that really need to be dialed back sometimes...they find out they have to blow air at something and next thing you know there's a giant squirrel-cage blower fan that sounds like a jet-engine in the design and their only response is "well I found it on McMaster and it had a good CAD model I could download and use...airflow looked good, what's the problem?" and I'm like "it's 93dBA...".

 

[Bill_Bright]

I just dislike the idea of a partially blocked fan.

I don't really see it as blocked. The fan still has unrestricted flow capacity out the back (other than what the grill imposes). If anything, the entire PSU interior, along with all the components inside the case and in the way, imposes more restriction. The baffle simply directs the flow where needed most.

I think you are overthinking this. Why would the maker spend the extra money in parts, design, and assembly if it was not beneficial?

 

[dirtyferret]

I am no way suggesting I could do better; I just like to understand technology, rather than accept it as a black box.

Think of the AC system in a car where you can direct air up and down by vertical vents so that a short, medium and tall person in height can direct the air towards themselves to feel cool.

 

[Bobaganoosh]

I don't really see it as blocked. The fan still has unrestricted flow capacity out the back (other than what the grill imposes). If anything, the entire PSU interior, along with all the components inside the case and in the way, imposes more restriction. The baffle simply directs the flow where needed most.

I think you are overthinking this. Why would the maker spend the extra money in parts, design, and assembly if it was not beneficial?

It is partially blocked. It is also beneficial to do that in this situation.

 

[Shrek]

Slats might be better, but I think there is no room even for small, directed slats.

For me, it's a puzzle... why are they blocking part of the fan, and then 'could this be done better'.