Case fan thrust vs flow?

[TeraSeraph]

I'm considering coming back from my hiatus of case fan testing to see if I can't get better comparative results using thrust rather than flow. This would shift the comparison to mass flow rather than air velocity and allow greater ease of comparison between different sizes of fans. Rather than being a calculated value, the force applied to the scale/load cell is a direct measurement. 

 

My main questions before I make a test apparatus are:

• Is thrust a good measurement for fan comparison? Is it relevant to cooling potential or performance?
• Should there be flow correction to ensure linear flow exiting the fan? Will this effect the thrust?
• Should I isolate the intake and exhaust sides of the fan to prevent byflow and avoid frame aero dynamics from effecting results? 

[Stahlmann]

Just throwing my opinion out there: I don't care how good a fan performs in a perfect test environment. I care about how well it performs on a cooler, with relateable numbers to draw comparisons. For example normalize both different fans for noise and then keep track of the temperature of the CPU they're supposed to cool. Most consumers will not understand thrust, pressure or flow and how it'll impact their cooling potential. But most consumers will understand that a lower temperature at a given noise level means it's better.

 

I don't know how exactly you can adapt this concept to case fans, but you should still try to get relateable results. An average consumer should understand your charts without reading up on your exact testing methodology. Most enthusiasts don't even go in that deep.

 

43 minutes ago, TeraSeraph said:

• Should I isolate the intake and exhaust sides of the fan to prevent byflow and avoid frame aero dynamics from effecting results? 

This would only mean that your results are even less representative of real-world use.

[TeraSeraph]

6 minutes ago, Stahlmann said:

Just throwing my opinion out there: I don't care how good a fan performs in a perfect test environment. I care about how well it performs on a cooler, with relateable numbers to draw comparisons. For example normalize both different fans for noise and then keep track of the temperature of the CPU they're supposed to cool. Most consumers will not understand thrust, pressure or flow and how it'll impact their cooling potential. But most consumers will understand that a lower temperature at a given noise level means it's better.

 

 

There's far too many variables involved with live testing on a computer, run to run consistency, precisely the same fan placement, software updates compromising past or future results, degradation of thermal paste over time, etc. There's also environmental factors to consider on top of all these other variables which will effect the run to run consistency. All of these inconsistencies that live testing introduces makes it a poor way to measure true performance from the fan. 

My data base isn't for average consumers as most of them care more about RGB that they do about performance. I really have no interest in making it palatable for people who aren't smart enough to interpret the data. I care far more about having fair comparison and the ability to compare all types of fans on the same test apparatus.

 

6 minutes ago, Stahlmann said:

 

This would only mean that your results are even less representative of real-world use.

This is the same situation you see when you have a fan installed in a PC case, how is this not representative? 

[GuiltySpark_]

13 minutes ago, Stahlmann said:

Not even enthusiasts go in that deep.

This is basically how I feel. I don't really look for or care about deep metrics, just if the price is right, it looks good and performs well enough that others have no concerns with it. 

 

A case fan isn't that complicated in my experience, though some like to make it complicated. As if a 2CFM difference is going to make or break their build. 

[manikyath]

i dont think the amount of force a fan applies is necessarily a useful factor. while in theory it is directly relative to the air it is pushing, and how much resistance it can work against.. why not just test those variables directly?

 

as for comparing different size fans, i dont necessarily see much use in that past 80/92 or 120/140, but in these cases we go back to the actual volume and air pressure being what we really are looking for.

 

Just now, TeraSeraph said:

There's far too many variables involved with live testing on a computer,

that's why your test setup should be something that "emulates" the real world conditions as accurately and repeatable as possible.

[TeraSeraph]

1 minute ago, manikyath said:

i dont think the amount of force a fan applies is necessarily a useful factor. while in theory it is directly relative to the air it is pushing, and how much resistance it can work against..

 

The force directly correlates to how much mass of air is flowing through the fan and thus correlates with the cooling potential of that air.

 

1 minute ago, manikyath said:

why not just test those variables directly?

How do you propose one would get that direct measurement? Don't tell me to put it on a live computer and measure the temps because that's far more indirect and inconsistent than measuring the fan itself.

 

1 minute ago, manikyath said:

as for comparing different size fans, i dont necessarily see much use in that past 80/92 or 120/140, but in these cases we go back to the actual volume and air pressure being what we really are looking for.

This still doesn't explain why measuring the thrust wouldn't be effective? How would you measure the volume and pressure?

Thrust combines air density, velocity, and pressure into one force measurement.

 

1 minute ago, manikyath said:

that's why your test setup should be something that "emulates" the real world conditions as accurately and repeatable as possible.

I plan to place an aircooler and a radiator in pull and push configurations to represent real world flow resistance. 

The added complication of temperature doesn't seem necessary and would have to be performed in laboratory conditions to be relevant. At that point I don't think anyone would consider it emulating real world conditions. 

[TeraSeraph]

25 minutes ago, GuiltySpark_ said:

This is basically how I feel. I don't really look for or care about deep metrics, just if the price is right, it looks good and performs well enough that others have no concerns with it. 

This isn't for the average consumer to be referencing, because as you stated they probably don't care. 

25 minutes ago, GuiltySpark_ said:

A case fan isn't that complicated in my experience, though some like to make it complicated. As if a 2CFM difference is going to make or break their build. 

Case fans are quite complicated, the aero dynamics involved are hardly trivial. I care more about the decibel to flow/trust through the RPM range rather than the min/max stated on paper. I'd hardly call someone a case fan enthusiast if they're only concerned with what the manufacturer is stating. 

[manikyath]

14 minutes ago, TeraSeraph said:

The force directly correlates to how much mass of air is flowing through the fan and thus correlates with the cooling potential of that air.

but why corelate when you can just measure the airflow itself?

 

15 minutes ago, TeraSeraph said:

How do you propose one would get that direct measurement? Don't tell me to put it on a live computer and measure the temps because that's far more indirect and inconsistent than measuring the fan itself.

airflow meter in open air, airflow meter with some known quantity restriction, and a large enclosed space for the fan to blow into with a sensitive pressure meter.

 

this way you get the unrestricted airflow, one or more datapoints for restrictive scenarios, and a worst case measurement.

18 minutes ago, TeraSeraph said:

Thrust combines air density, velocity, and pressure into one force measurement.

explain to me how this combined figure is useful for any real world scenario past moving a vehicle forward? reducing these variables into one measurement essentially removes the nuance of different fan designs being better for different scenarios.

 

19 minutes ago, TeraSeraph said:

The added complication of temperature doesn't seem necessary and would have to be performed in laboratory conditions to be relevant.

i never said temperature needed to be involved.

 

7 minutes ago, TeraSeraph said:

This isn't for the average consumer to be referencing

who is this for then? for companies intending to embed the fans into their products? because those people will prefer a fan manufacturer to have done their own testing in a validated process, to come with numbers (you know.. m3/hr and mmH2O aka flow rate and pressure, like we've been documenting fans for a decade) to base their decisions off of.

 

in a sense, you have more chance for joe average pc builder to be interested in an alternative means of measuring fans, than the people who actually care about these figures.. because they will want the figures that have been aquired in a way they can translate to what matters for them.

 

if anything, i'd say try providing more detailed data, dont try to condense variables down.

[Stahlmann]

1 hour ago, TeraSeraph said:

There's far too many variables involved with live testing on a computer, run to run consistency, precisely the same fan placement, software updates compromising past or future results, degradation of thermal paste over time, etc. There's also environmental factors to consider on top of all these other variables which will effect the run to run consistency. All of these inconsistencies that live testing introduces makes it a poor way to measure true performance from the fan. 

Environmental facturs such as ambient temperature has to be controlled, but if you're in the US chances are your house has AC?

Fan placement shouldn't matter as much as you make it out. Just mount the fan to the same mounting holes as the last one.

You can use a graphite pad instead of paste to get more consistent results over time. The numbers won't be exactly the same as using paste, but also not too much off.

Software updates can be counteracted by taking the test system offline once you have all the software installed.

Run to run variance can be counteracted by making a few runs and averaging them out.

Allmost all of the CPU variables can be set to fixed values in BIOS.

 

I'd say when doing these few things you can get the testing accurate enough that the data is applicable to the average consumer.

 

1 hour ago, TeraSeraph said:

My data base isn't for average consumers as most of them care more about RGB that they do about performance. I really have no interest in making it palatable for people who aren't smart enough to interpret the data. I care far more about having fair comparison and the ability to compare all types of fans on the same test apparatus.

You can do all the comparisons you want. If they don't translate into real world use, why should people care? It doesn't matter to anyone if a fan has a few CFM more if it doesn't translate into lower temperatures or noise.

 

1 hour ago, TeraSeraph said:

This is the same situation you see when you have a fan installed in a PC case, how is this not representative? 

That depends on how your apparatus is designed. Is it supposed to recreate a case or is it something else entirely?

 

 

All in all, what i'm saying is that you shouldn't just collect data. You should collect useful data.

[TPCEA]

On 1/12/2024 at 11:12 AM, manikyath said:

airflow meter in open air

I'm curious about this one. When setting the volume on the anemometer, which is required in order to measure AF, what number should be used?

 

On 1/12/2024 at 11:12 AM, manikyath said:

On 1/12/2024 at 10:41 AM, TeraSeraph said:

The added complication of temperature doesn't seem necessary and would have to be performed in laboratory conditions to be relevant.

i never said temperature needed to be involved

Temperature affects the performance of the bearings. Humidity affects air resistance.

The question I have is are the differences significant enough to warrant tracking them?

On 1/12/2024 at 9:31 AM, TeraSeraph said:

Is thrust a good measurement for fan comparison? Is it relevant to cooling potential or performance?

Since you know a lot about propellers, can you explain how thrust is going to be useful? I think it's an interesting idea, but I want to better understand it. 

On 1/12/2024 at 9:31 AM, TeraSeraph said:

Should there be flow correction to ensure linear flow exiting the fan? Will this effect the thrust?

Why should there be? Do most cases and fans have this feature? I've got over 300 fans and most don't even try to ensure linear flow. Yes, it will affect thrust, although I can't remember in what way. Take a look at Dr Frank Fish's research related to protuberances on fans and props. That said, I encourage you to research the difference between turbulent and linear flow and, if you see value, create a prototype and test with it. 

 

On 1/12/2024 at 9:31 AM, TeraSeraph said:

Should I isolate the intake and exhaust sides of the fan to prevent byflow and avoid frame aero dynamics from effecting results? 

What is the purpose? Isolation will allow you to determine maximal results but how can you do it if the natural flow of the wind is a shape that doesn't work well within those confines? I don't really think that, aside from knowing those theoretical maximums, that this is going to be useful, but I'm open to hearing why you think it might be.