Positive or Negative Enclosure Pressure Tests

[martin-w 3,234]

It seems that the debate regarding positive or negative enclosure pressure has raged for years but still opinion is equally divided. Well, after reading yet another article where the author [claiming to be a physicist] cited Bernoulli's principle and Newton's second law in order to justify his choice of positive enclosure pressure, I decided to run some tests.

 

The proof of the pudding is in the eating as we say in the UK. You can babble on about the Bernoulli principle, conservation of energy and fluid velocity as much as you like but there's no substitute for experimentation.

 

Because I had nothing better to do, I spent a couple of hours testing, comparing extreme negative pressure and extreme positive pressure, running stress tests for the CPU, GPU and monitoring temperature, and what did I find?

 

Absolutely no difference whatsoever in regard to CPU temperature, GPU temperature, or motherboard temperature. Actually, I tell a lie, motherboard temperature was down one degree but given that one degree is well within the margin of error you would expect it's meaningless.

 

My tests certainly aren't ultra scientific but I find it interesting that there was no difference at all. Given that I was formerly a negative enclosure pressure advocate, perhaps I should be less scathing of those that choose the opposite approach.

 

All of this is system specific of course, but the conclusion I draw from this is we shouldn't be too concerned about positive, negative or balanced enclosure pressure, it's pretty much irrelevant.

 

I'd be interested if anyone else has discovered the same, or has anything else to add.

 

Edit: The caveat of course is that I was measuring temperature with a limited number of sensors, CPU, GPU and wherever Asus position the temp sensor on my board. Therefore it could well be that with a positive enclosure pressure, that there were pockets of stagnant warm air present. This wouldn't be the case with negative pressure.

[Ted Striker 338]

I agree with the physicist that in theory the positive pressure case with higher density air flowing through your CPU and GPU coolers should result in better cooling.

 

However,

1. By what percentage did you change the pressure of the case air? Normal air pressure is 14.7 psia. All else being the same I think you would have to increase the pressure quite a bit to notice a difference.

 

2. How did you change the case air pressure? If you turn off your exhaust fans to achieve this it will increase your case pressure slightly but it will also reduce the airflow through your case. The reduced airflow will increase your average case air temperature which will have a much more detrimental effect on your component cooling than the benefit from increased air density.

 

I could discuss theory all day long, but you did the right thing which was to actually perform some tests. To accurately perform the test you are discussing I think you would need inside and outside case pressure measurements as well as inside and outside case air temperature measurements.

 

Ted

 

I have 3 speed case fans and get a 3 degree C CPU temperature drop during load testing with the case fans on high vs. low speed. I haven't bothered to measure the air temperatures but I expect the low speed settings is allowing the case air temperature to rise thus providing hotter cooling air to the CPU heat sink.

 

Ted

[martin-w 3,234]

I agree with the physicist that in theory the positive pressure case with higher density air flowing through your CPU and GPU coolers should result in better cooling.

Well he claimed he was a physicist. Some of his comments were contradictory though, and despite citing how Bernoulli's principle tells us that lower pressure results in increased air velocity, and thus more rapid cooling, he contradicted himself with his final conclusion. I'll try and find the link later. We would need a fluid dynamicist with custom PC experience I reckon.

 

But essentially, given that we're talking about a miniscule pressure deferential in a PC, probably a fraction of a PSI, positive or negative enclosure pressure arguments are probably mute. [Probably why I found no difference] But for sake of argument, I'll put the notion to one side for a moment...

 

 

Higher density air and thus lower component temperature is no help unless it's evacuated from the enclosure rapidly of course. And by definition, higher enclosure pressure necessities that air intake is greater than exhaust and thus the rate at which a given volume of warm air is replaced with cool air is slower. But that's only initially, I suspect very briefly...

 

 

I'm just hypothesising here, but I guess, given that in a positive enclosure scenario, [more air entering than escaping] pressure doesn't keep building until the enclosure bizarrely explodes, there must obviously be a peek pressure reached. At that point, air intake and air exhausted must be balanced. And that peek pressure is probably miniscule, a fraction of a PSI, due to the feeble capabilities of the PC fans we install.

 

But I see a possible issue. Regions of the enclosure without adjacent vents or fans may form stagnant warm air pockets. the question is whether that scenario would be more likely in a positive pressure environment than a negative pressure environment.

 

Complex stuff, and given my results not worth worrying about, but interesting all the same.

 

Incidentally, I read another article yesterday that demonstrated a 5 degree drop in temperature in favour of the negative case pressure scenario. Not huge, no big deal.

 

The positive enclosure advocates claim less dust ingestion of course, which makes a lot of sense. However, what I've found with my enclosure, which is normally configured to be a very negative pressure environment, is that fine dust still passes through the front filters. And given that dust is electrostatically charged, it soon clumps together. It would seem to me that even in a positive pressure scenario, that fine dust would still pass through the filters and from larger clumps.

 

 

 

2. How did you change the case air pressure? If you turn off your exhaust fans to achieve this it will increase your case pressure slightly but it will also reduce the airflow through your case. The reduced airflow will increase your average case air temperature which will have a much more detrimental effect on your component cooling than the benefit from increased air density.

 

 

No fans were turned off or removed Ted. I have two 140 fans on the top sucking out, one 120 at the rear sucking out, and one 120 at the front blowing in. To configure for a positive case pressure environment, I simply inverted the top two 140 fans, so two 140's blowing in, one 120 blowing in , and only one 120 sucking out.

 

 

 

To accurately perform the test you are discussing I think you would need inside and outside case pressure measurements as well as inside and outside case air temperature measurements.

 

 

Don't think so, all variables were kept the same, just the two top fans inverted. So clearly with so many fans blowing in and only one exhausting, it has to be a positive pressure.

[martin-w 3,234]

Linus seems to be talking common sense to me. Balanced, with perhaps a slight positive bias.

 

Not forgetting of course that a PSU orientated up and our GPU's are exhausts too.

 

 

[stans 702]

I am not a fan of negative air pressure cooling simply because with negative air pressure you suck air and dust into the case through every little opening.  There is no practical way to filter the air drawn into the case.

 

My Antec 900 case has two intake fans on the front, set to medium speed, and covered with used clothes dryer fabric softener sheets and fabric softener sheets over the side vent.  The back and top exhaust fans are set to low speed.  I have good cooling and minimal dust build up.  Best cooling and least dust buildup would come from liquid cooling and I may go that route in the future.

[martin-w 3,234]

I am not a fan of negative air pressure cooling simply because with negative air pressure you suck air and dust into the case through every little opening.

Yes that's certainly a plus point for positive pressure, "less" dust ingress. But as I said, I would say dust still enters through the filters, all be it fine dust. And as it's electrostatically charged it does form clumps. So I guess it's not possible to totally eliminate dust build-up, even with positive pressure.

 

 

My Antec 900 case has two intake fans on the front, set to medium speed, and covered with used clothes dryer fabric softener sheets and fabric softener sheets over the side vent. The back and top exhaust fans are set to low speed. I have good cooling and minimal dust build up.

 

 

We also have to take into consideration that when we place filters in front of our intake fans, or in your case "clothes dryer fabric softener sheets" that we are restricting airflow considerably. So we may think our front fans are pushing in a certain CFM, when in reality they are underperforming. Hence, it may be that we're running with balanced or negative pressure when we assume it's positive. Or of course still positive but much closer to balanced than we assumed. Couple that with PSU's exhausting and graphics cards exhausting and negative pressure is even more likely.

 

At the end of the day, it's pretty much imposable to state definitively that we have a negative or positive enclosure pressure, "all the time". PSU fan speeds vary with temp, CPU fan speeds vary with temp, graphics cards vary with temperature too. Couple that with our case fans controlled by PWM, plus different CFM ratings for our fans and it's an impossibility, too many variables. 

 

 

Best cooling and least dust buildup would come from liquid cooling and I may go that route in the future.

Do you mean for your CPU and GPU? Of course even with a full blown water cooling loop enclose fans are still required. And water cooling radiators still accumulate a very nice dust build-up.

[stans 702]

Probably best to cool both CPU and GPU with water.  Yes, the fans and radiator will collect dust, so they still have to be cleaned.  It would also be best to have the fans and radiator somewhere other than mounted in the fan bay of the case, keeping it cooler and easier to clean, but most kits today mount the radiator to the case.  A really great cooling system might be a drag racer's fuel line cool can.  The radiator is actually a coiled pipe contained inside of a metal can, the can is filled with ice and alcohol.  I guess to keep things a little easier to clean up, for computers you could use plastic freezer packs.  The best cooling would be liquid nitrogen, but most of us do not have access to such things and you have to be very careful with the stuff.  Just my thoughts, but back to the topic; Slightly positive case pressure is really good for air cooling, but proper flow accross the components must be maintained.

[martin-w 3,234]

Slightly positive case pressure is really good for air cooling, but proper flow accross the components must be maintained.

 

Here's a weird story for you...I was once INSIDE an enclosure.

 

I'm not kidding. I used to be a pro photographer.  One of my duties was to operate an Agfa Repromaster camera. [The tech doesn't exist now]. This guy was a camera as high as your shoulder with two huge banks of halogen lighting. The heat generated in a small darkroom was phenomenal.

 

Anyway, I finally managed to get my company to install an extractor fan. The fan could be set so that it blew cool air in from outside [positive pressure] or sucked air and heat out of the room [negative pressure].

 

Others that used to used the darkroom would set the fan to blow cool air in from outside. The result was barely any change in the room tempreture. There would be just a trickle of air seeping around the door and any other gap, out of the darkroom and into the adjacent room.

 

When I used the darkroom I would immediately set the fan to suck the warm air out of the darkroom, so negative pressure. The result was phenomenal, within seconds the hot air was gone and the room was cool. The gust of cool  air that was sucked into the room around vents and the door frame was phenomenal. The huge amount of atmospheric pressure pressing down on all of us did the trick. Nature abhors a vacuum.

 

The most rapid way to cool any enclosure is with negative pressure. However, we PC enthusiasts aren't dealing with a room that requires general cooling. We are dealing with multiple individual components that require cooling priority. Thus, we like to spot cool components which necessitates fans blowing in. We also have dust to contend with. Therefore balanced air flow is optimal, equal intake and equal exhaust. This way we limit dust ingress and still get to spot cool our precious components.

 

Unfortunately though it's very difficult to configure for a precisely balanced airflow, so a slight positive pressure in order to mitigate dust ingress is probably okay.

[martin-w 3,234]

Probably best to cool both CPU and GPU with water.

In terms of dust ingress I'm not sure if water cooling is better than air cooling or not to be honest. I'm not prepared to comment definitively as I've never had a rig with a full blown loop.

 

However, In terms of general cooling efficiency I still prefer high end air coolers. No pump to fail, zero chance of leaks. In addition, my NH-D14 does an amazing job of cooling my overclocked CPU and is ultra quiet. In fact I'm so confident, that my new Skylake rig will be cooled with an NH-D15S.

 

I also like the way the D15 has a centrally mounted fan that's deliberately larger than the heat sink, so a considerable gust of air across the VRM's.

 

When closed loop coolers are guaranteed not to leak, have pumps that are more reliable than they are now and quieter, and don't have nosier fans than my D14/15... then I might consider one.

 

 

 

A really great cooling system might be a drag racer's fuel line cool can. The radiator is actually a coiled pipe contained inside of a metal can, the can is filled with ice and alcohol.

 

Something similar already exists, radiators can be mounted outside of cases and water cooled systems can be equipped with chillers.

 

http://koolance.com/exc-800-portable-800W-recirculating-chiller