FRONTPAGE Dust Prevention: Positive Pressure, Negative Pressure, Does It Really Matter?

For years we've all heard the claims that positive case pressure can help reduce the amount of dust entering your PC case. There have been countless opinions stating why positive case pressure is beneficial, but until now there have been no scientific data to prove these theories that we've run across. Being proponents of empirical testing and the scientific method in general, Overclockers has teamed up with be quiet! to conduct what could be the first physical testing as to whether positive case pressure really does reduce dust in your PC.

Click to expand...

Click here to view the article.

Very nice article. I still can't get my head around parts of it but I appreciate the test.

I also appreciate the time taken to show this data. Well written and thought out. I will continue to use positive pressure when possible anyway.

Good article. All Hail Roomba!

Thanks everyone. I had a lot of fun doing this one. Like Dejo I've always been a positive pressure advocate. I was shocked when I opened up the dust chamber for the first time. It took me a long while to figure out why.

I'm sure Earthdog is happy it's completed now too. I must have PM'd him every other week asking him to poke different companies looking for sponsors. I was super stoked when be quiet! replied back with a confirmation. Thanks Joe for your patience with me.

Now to clean up the mess.

I always rocked negative pressure because I don't care about dust, and never thought it made much if any difference myself, and testing has shown that negative has better cooling ability in most pc case configurations. ... ok its just a couple degrees but still.

Dust is gonna happen no matter what. Clean your sh1t once and a while. Negative pressure for me. I'll take the extra couple degrees.

Mr.Scott said:

Dust is gonna happen no matter what. Clean your sh1t once and a while. Negative pressure for me. I'll take the extra couple degrees.

Click to expand...

the only time mine gets cleaned is the every 4 or 5 years i open it up to do hardware changes lmao IF i even do it then. Albeit my place has very little dust, there's little to no accumulation in that time frame.

Very interesting article, thanks!

Wrong testing method.

You run positive pressure to control where the air flows (dust), not for less dust in the case but for less dust in the cracks rather than the filter screens! How is this overlooked?

If you create a vacuum, the fans will be sucking in the air from every crack on your PC case. This will introduce pet hair in around your 5.25 covers, where you seal the case covers, the front fascia panel, etc.
Instead, run a positive pressure by positioning your intake fans where the case manufacturer designed the air to flow. Typically there will be filters here and allow for easy cleaning. Instead of trying to figure out if it's ok to blow air in your Blu-ray burner because it's full of hair and dust.

I'm not sure how sand was helpful here.

Hopefully, this helps people. The review above is mostly useless.

I appreciate your input NGL. Some of what you said is true though I question if you even read the full article.

The cases used for this test did not have front 5.25 bays which is more common in today's chassis' than those with them. Secondly, the intake fans were located where the manufacturer intended them, as you said, in the front of the chassis where the filters are located.

The whole purpose of this testing was to determine if the vacuum created from negative pressure was enough to significantly increase the amount of dust collected. The math shows that the +1.56 H2O vs. The -1.56 H20 difference is not significant enough to skew the results. As a point of fact it's actually the cracks that prevents a positive pressure or vacuum from ever occurring in the first place.

If you've ever lived in the south west you would understand the importance of the sand in the test. That shtuff gets everywhere and is far more troublesome than dust bunnies.

As a student of science I appreciate your comments and views. I was completely surprised by these test results. If you believe these results are "useless" prove it. Please feel free to conduct your own tests and post the results.

Good article. I've always run negative pressure for the couple degrees because I'm currently sporting an 8350 and every degree matters. I will give some more thought to a positive pressure system on the ryzen build I have in the works.

Repeat the tests with dryer lint now [emoji14]

Something wrong with the setup

I don't know what you did wrong, but I've made positive pressure cases (for a living, built touchscreen kiosks) with filters on all intake fans and they lasted much much longer with far less dust.

Also you'll see that is the experience with others like here (https://www.scholarlygamers.com/tech-talks/2017/11/14/building-pc-positive-vs-negative-air-pressure/):
"Previously I had purposely created a negative pressure environment, as I believed this would give me the lowest temperatures possible. Sure enough the system ran nice and cool, but I found myself having to clean my PC quite often due to the accumulation of dust. After changing out a few components (replacing the huge fan on the top of the case with a window), installing filters on all of the intake fans, and configuring my fans so I had more air entering the case than exiting; my system still has not required any dusting since. The ambient temperature within the case has only risen by a degree or two, and the end result is much more optimal for my configuration."

Also here he says why tests fail ... it's because you didn't have good enough filters on all the intakes:
In fact what happens is that the well filtered intakes clog up the filters and then all your intake is coming from poorly filtered intakes, which is why you get tons of dust even with a positive pressure system: it wasn't filtered properly.

I also I'm quite doubtful about the relevance of this test.

A positive pressure setup with filtered intake only works as long as the filters are relatively clean so the air can flow through them.

Given the testing environment, I'm ready to bet that all filters where 100% obstructed half-way through the test, likely quite early.
Which mean that half-way through the test, all 3 systems were actually negative pressure builds (since intake fans didn't have any unobstructed intake to take air from), and all air intake was done through the case unfiltered openings.
Looking at the test result images, the main intakes were likely the mesh openings in the PCI slots brackets : there are massive dust deposits in front of them, on the PSU cage and on the GPU backplate.

So in the end, the only conclusion that can drawn from this test is that a dust filtered positive pressure setup doesn't work if you don't clean the filters regularly, which is no big news...

I'm also surprised you did all that work and didn't bother to schedule running temperature benchmarks and logging the results at regular intervals, something like every 15 minutes. This would likely have provided a bit more insight.

Welcome to the forums Got.

As you can plainly see, none of the filters were 100% obstructed as you are betting. (See image below) Was there obstruction? Undoubtedly, but the obstruction would be nearly identical in all three scenarios. If you compare the intake filter on all three, you will notice a similar amount of collected dust, indicating as much.

I do agree with you as far as regular cleaning is concerned. In fact here is a quote from the article.

This test also concludes that there is no substitute for regular cleaning.

Click to expand...

The beauty of empirical testing is it ensures the test can be duplicated by its peers. I would encourage anyone who doubts these results, no, I challenge anyone who doubts these results to prove me wrong. Set up your own rig with 3 identical cases and configure them similarly, One positive, one neutral, and one negative. This is the nature of science. You can doubt science all you like, but until we have results proving this test is wrong, it's what we have to go with.

Logging temperature was considered, but to be honest, others have done this and it is fairly conclusive that negative pressure actually wins out. It's not by much, but the other articles I've read were all in agreeance. Also, I was concerned it would distract from the main meat of this article, which is dust prevention and whether case pressure really matters. In the end, just practice regular cleaning and everything will be fine.

Undoubtedly, but the obstruction would be nearly identical in all three scenarios

Click to expand...

The issue is that filter obstruction can completely turn around the positive pressure build into a negative pressure build. The fact that you get the exact same results (dust buildup in front of the PCI slot unfiltered openings) for your 3 setups is a clear indication of that.
It doesn't take a lot of obstruction to reduce a fan CFM to a tiny fraction of its nominal value. In fact, when designing for positive pressure, it is quite important to account for the flow reduction due to the air filter. A 100µ dust filter put right on front of a fan reduce the nominal CFM by 20-30%, in a best case scenario and when the filter is perfectly clean. Additional obstruction (like the front case mesh) further reduce fans air flow.

All it take to turn your positive pressure setup into a negative pressure setup is for the 3 intake fans to generate less than 33.3% of their nominal CFM, since you have a 3:1 intake:exhaust ratio.

Blaylock said:

As you can plainly see, none of the filters were 100% obstructed as you are betting.

Click to expand...

What this picture is telling me is that there was a ton of dust on the filter, but it fell off when you opened the case.


My take on this is that the dust you're using is relatively uniform, dry and clean and doesn't stick much to the filters.
In a real world usage, dust is quite greasy, sticky and composed of a greater diversity of fibers. From my own experience, "normal" dust buildup on an air filter doesn't fall off easily.
In fact, I could argue that the differences between "real world" dust and what you're using already make this kind of test quite questionable in the first place.
Long exposure to a low concentration, high diversity dust can't reliably translate into short exposure, high concentration, low diversity dust.
First, this change how the filters perform, second this changes how much dust will settle and stick to surfaces.
The fact that you're actively ventilating the test chamber with an air compressor also likely cause quick to settle heavy particules to be supended in the air in "unnatural" concentrations.

But in any case, my point is that by ignoring that "are dust filters obstructed" variable, your test setup can hardly be used to prove anything. Not making my own test doesn't change the fact that your test has an obvious flaw...

You are way too confident in your conclusions. As others have pointed out, dust prevention through positive pressure filtered intakes is an industry standard.
All industrial electrical/electronics cabinets are designed that way, and it's the basic principle used to design cleanrooms, the physics behind that concept are solid and this has been experimentally validated for decades.

If a specific positive pressure design doesn't work as expected as far as dust prevention is concerned, it's because something is wrong either in the design, or in the usage.

As you mentioned, your 50h test simulate a multi-year exposure to an extremely harsh environment. For it to be representative of a real-world usage, you would need to replicate the real-world usage : filters must be cleaned regularly, before they become significantly obstructed.

My take on a proper test would be to run that test in cycles.
First, you need to determine after how much time the filters are clogged. You could do a test run, checking the filters at regular interval. Also, this is where monitoring system thermals would be useful, as a degradation would be an indication of significant airflow obstruction. A setup with a passively cooled (just a big tower radiator, no fan attached) low-end (~30W TDP) CPU would be extremely sensitive to case internal air flow, that would be a good way to get useful results. A simple test to validate if the setup is still positive pressure would be to put some smoke (incense, electronic cigarette...) in front of the PCI openings to see if it gets sucked in or not.

Once you have a decent estimate of "filter clogging time", run the test for that time, stop it, clean the filters, and repeat. Stop repeating when you have enough dust buildup inside one of the test setups to make a meaningful comparison.

The fact that you get the exact same results (dust buildup in front of the PCI slot unfiltered openings) for your 3 setups is a clear indication of that.

Click to expand...

Is it though? (confirmation bias? though I may have some of my own, lol!)


I'd say it's worth noting is that many users don't clean their chassis at regular intervals (when it gets clogged as is being requested here). Many people leave it and after they 'see' it (or something, lol), they clean it. If we test as you mention, I believe it alienates a lot, perhaps even a majority, of users. In my experience, as someone who always runs negative (and doesn't have dust issues...) 'crack dust' isn't a huge deal.

I think of it this way... akin to GPU testing, you can't please everyone. You test at XX resolution and XX setting(s). But what bout the person who doesn't play at [insert settings and/or resolution here]?? This is similar, to me.

That said, I'd like to see what you said tested again. I get it. 100%. I've worked in data centers for almost 20 years (sans the last 3 now) and we ran positive pressure to minimize dust as well. But a datacenter/clean room doesn't have, what amounts to be, wide open cracks either. The 'pressure' difference is more than negligible in those situations.

Many things to consider here... a clogged filter is one of them. Happy to read some links and such of other testing if you have it.

EarthDog said:

I'd say it's worth noting is that many users don't clean their chassis at regular intervals (when it gets clogged as is being requested here). Many people leave it and after they 'see' it (or something, lol), they clean it. If we test as you mention, I believe it alienates a lot, perhaps even a majority, of users.

Click to expand...

Indeed. Dust filter accessibility is paramount. It's unfortunate that almost every case is designed with hidden filters. Plus moving filters away from the fans also increase fans ability to stay closer to their nominal CFM rating. But aesthetics are usually put before function...

There are many actually relevant and useful things that would be interesting to test experimentally :
- After how much dust buildup on a filter does the build become a negative pressure build ?
- How is that affected by the intake vs exhaust ratio ?
- What is the actual intake vs exhaust ratio needed, filter/mesh obstruction taken into account ?
- Testing with various filter materials : stainless, nylon, plastic, and various filter size (100 µm, 200 µm...)
- How is fan performance affected by the distance between the fan and the filter.
- When a filter is not directly on the fan (pressure buildup chamber), what is the fan performance with various filter surface area...

Those are questions that are worth experimenting for. The "Dust prevention: positive pressure, negative pressure, does it really matter ?" one is not one of them.
But how to make a positive pressure setup that actually work for dust prevention in practice is a very interesting topic.
Not everyone cares, indeed. Enthusiast users that rebuild their PC every year or so don't care about dust prevention, and a negative pressure build is usually better in terms of thermal performance.

I understand what your point is, and it has merit. The dust was dry and did not stick to the filter the way "real" dust sticks. This is an unfortunate result of a simulation. There simply wasn't enough time (or willingness TBH) to run these three rigs nonstop for 5 years to get a real-world result. The dust did stick long enough to take all three chassis from the garage to the downstairs photo booth but fell off when removing the front panel. I was hoping to get a pic of the dust on the filter but alas, it wasn't in the cards. You'll have to take my word that they weren't 100% obstructed. even 30% would be a stretch, but closer to reality for sure.

There is a reason these cases were defined as "Intake Heavy, Neutral, and Exhaust Heavy" rather than "Positive Pressure, Neutral, and Negative Pressure" in the article. Primarily, that is because when measuring the actual pressure inside the case, even with the max 5 in and 1 out vs 5 out and 1 in, the pressure difference was negligible. I wish I had the numbers still. This is actually a very simple test you can do at home with your own PC case. The fact is, cases are very porous and it is nearly impossible to generate a decent amount of pressure difference inside the case.

I think we are actually in agreeance here. Your statement that once the filter becomes clogged it's actually negative pressure, is an explanation as to why the results are similar. I'm not suggesting you are wrong. I'm just suggesting if you aren't cleaning regularly, it doesn't matter if you are intake or exhaust heavy.

Anyone without enough dust to run their own study please PM me for dust. I have plenty and will ship USPS in 10# increments to anyone postage paid. Overnight shipping will be on your dime.