Dusty PC vs. Clean PC Wind Tunnel & Thermal Benchmark: Years of Dust
video description
Date: 2020-06-27
Related videos
Comments and reviews: 10
Michael
That epoxy blob on the anemometer is PTC thermistor; though all electrical components have different thermal characteristics that negatively affect their stated value, thermistors actually use this temperature dependency to their advantage. For the component's safety, they are epoxy encapsulated. They are [much] more accurate than thermocouples, but have a much slower response time (Tc).
I know your methodology doesn't take into account any smoothing, but it can help with dealing with digital noise that you have present and allow a smaller time interval to be used to determine the average of unobstructed airflow, which was naturally turbulent.
That's a nice bit of kit from AD you have their. Looking at its specs, you're paying for the software more than hardware, but still, it's a solid little guy for what you're using. I've very much interested in a deeper dive into the design of the flow chamber. The mesh flow straighteners you have seem to be placed at determined spacing. I'm curious as to what was used, mathematically, for that determination.
reply
That epoxy blob on the anemometer is PTC thermistor; though all electrical components have different thermal characteristics that negatively affect their stated value, thermistors actually use this temperature dependency to their advantage. For the component's safety, they are epoxy encapsulated. They are [much] more accurate than thermocouples, but have a much slower response time (Tc).
I know your methodology doesn't take into account any smoothing, but it can help with dealing with digital noise that you have present and allow a smaller time interval to be used to determine the average of unobstructed airflow, which was naturally turbulent.
That's a nice bit of kit from AD you have their. Looking at its specs, you're paying for the software more than hardware, but still, it's a solid little guy for what you're using. I've very much interested in a deeper dive into the design of the flow chamber. The mesh flow straighteners you have seem to be placed at determined spacing. I'm curious as to what was used, mathematically, for that determination.
reply
phil21s
First of all, great content and thanks for going far and beyond for testing seemingly minor details such as fans or dirty radiators!
I am a bit concerned with that easy conversion of FPm to CFM. The idea with having a constant area (the tube) and well controlled flow seems right, tough you imply a uniform flow rate across the whole surface (at least at the end of the tunnel where you measure).
Seeing that you noticed that you need to precisely control the anemometer position (kudos for testing that!), tells me that the assumption of a uniform flow density (that the correct term?), or the flow through a given area of your tube, is not valid.
I am not sure how to do this better, as a laminar flow profile would likely have a systematic variation due to boundary layer effects, whereas turbulent flow might introduce more 'noise' in the measurement.
Also, that same setup could probably be adapted to measure the static pressure difference a fan can put up, too. Have you thought about testing static pressure for fans?
Keep up the good work!
reply
First of all, great content and thanks for going far and beyond for testing seemingly minor details such as fans or dirty radiators!
I am a bit concerned with that easy conversion of FPm to CFM. The idea with having a constant area (the tube) and well controlled flow seems right, tough you imply a uniform flow rate across the whole surface (at least at the end of the tunnel where you measure).
Seeing that you noticed that you need to precisely control the anemometer position (kudos for testing that!), tells me that the assumption of a uniform flow density (that the correct term?), or the flow through a given area of your tube, is not valid.
I am not sure how to do this better, as a laminar flow profile would likely have a systematic variation due to boundary layer effects, whereas turbulent flow might introduce more 'noise' in the measurement.
Also, that same setup could probably be adapted to measure the static pressure difference a fan can put up, too. Have you thought about testing static pressure for fans?
Keep up the good work!
reply
PappyMan
Steve, any static created by a vacuum might (gasp) be completely voided by an EARTH GROUNDED case/radiator/PS connection, protecting any static sensitive components. So you show an environmentally horrific 'air' can spraying air FROM THE DUST SIDE of the radiator at an oblique angle TOWARD the CLEAN side of the radiator essentially PUSHING the dirt INTO the radiator and blasting any other dirt into the surrounding air that people breathe. As opposed to using a common household HEPA filtered modern vacuum present in most American homes. Dude, really? All that tech science and measuring and so little 'common' sense, seriously SMFH. The VAST MAJORITY of 'harmful dust' is on filters, heatsinks and radiators. NONE of which is static sensitive. Also, any CO2-based 'air' cans will generate tons of static in the plastic hose that most have installed as well as adding to the CO2 problem in the environment we already have.
reply
Steve, any static created by a vacuum might (gasp) be completely voided by an EARTH GROUNDED case/radiator/PS connection, protecting any static sensitive components. So you show an environmentally horrific 'air' can spraying air FROM THE DUST SIDE of the radiator at an oblique angle TOWARD the CLEAN side of the radiator essentially PUSHING the dirt INTO the radiator and blasting any other dirt into the surrounding air that people breathe. As opposed to using a common household HEPA filtered modern vacuum present in most American homes. Dude, really? All that tech science and measuring and so little 'common' sense, seriously SMFH. The VAST MAJORITY of 'harmful dust' is on filters, heatsinks and radiators. NONE of which is static sensitive. Also, any CO2-based 'air' cans will generate tons of static in the plastic hose that most have installed as well as adding to the CO2 problem in the environment we already have.
reply
gatordontplay417
I clean my AIOs with water once a year. just the rad. The pump and coldplate get detailed with cleaners not water. NEVR-DULL works wonders on coldplates. I use a round felt for a dremel on its side soaked in NEVR-DULL (not with the dremel with my hands) I rub it tell shiny. It is corrosive to plastic so be sure not to get it on the front of the pump. Plastic around the coldplate is fine as long as you wipe it off after. Then when it is fully polished I wipe it off with a cloth rag soaked in Iso. I have done this to coldplates with literal scorching and they still come out looking like new. I am waiting tell next time I do it and I will take pics and put it on the Discord.
reply
I clean my AIOs with water once a year. just the rad. The pump and coldplate get detailed with cleaners not water. NEVR-DULL works wonders on coldplates. I use a round felt for a dremel on its side soaked in NEVR-DULL (not with the dremel with my hands) I rub it tell shiny. It is corrosive to plastic so be sure not to get it on the front of the pump. Plastic around the coldplate is fine as long as you wipe it off after. Then when it is fully polished I wipe it off with a cloth rag soaked in Iso. I have done this to coldplates with literal scorching and they still come out looking like new. I am waiting tell next time I do it and I will take pics and put it on the Discord.
reply
dosduros
Dood,
I had a 360mm rad (an old one from thermaltake, came with the Kandalf LCS, brass core, very thin alu fins that were not air flow optimized) that had over 10 years of dust build up. It was so clogged up, I had to use water to get it out (regular compressed air was no longer viable).
The thing is, it was STILL cooling the CPU and the GPU on that radiator rather well (after a full clean up the GPU temps droped by 5 max C, CPU 7).
How in the hell did I allow that to happen? Well, everytime I would use the compressed air, some dust WOULD get out from the surface so I assumed it was clean and the temperatures seemed to be kinda ok. (This is one of those Im dumb moments).
reply
Dood,
I had a 360mm rad (an old one from thermaltake, came with the Kandalf LCS, brass core, very thin alu fins that were not air flow optimized) that had over 10 years of dust build up. It was so clogged up, I had to use water to get it out (regular compressed air was no longer viable).
The thing is, it was STILL cooling the CPU and the GPU on that radiator rather well (after a full clean up the GPU temps droped by 5 max C, CPU 7).
How in the hell did I allow that to happen? Well, everytime I would use the compressed air, some dust WOULD get out from the surface so I assumed it was clean and the temperatures seemed to be kinda ok. (This is one of those Im dumb moments).
reply
Utsuro
Interestingly enough, I recently just cleaned my old corsair h75 (120mm AIO) and had it mounted as rear exhaust in my case. I NEVER cleaned it during the 6 years I've owned it and when I finally did... the radiator wasn't even visible due to the sheer amount of dust that had accumulated into it (my computer was kept in a pretty dusty basement with cobwebs and a drop ceiling). However, my cpu temperatures (4770k 4.4 ghz, 1.25v) were never out of the ordinary. In fact, they only went down by about 3 to 4 degrees Celsius on average after dusting off the radiator. Watching this video suggests that this isn't normal, so I wonder why I didn't notice an improvement in temps.
reply
Interestingly enough, I recently just cleaned my old corsair h75 (120mm AIO) and had it mounted as rear exhaust in my case. I NEVER cleaned it during the 6 years I've owned it and when I finally did... the radiator wasn't even visible due to the sheer amount of dust that had accumulated into it (my computer was kept in a pretty dusty basement with cobwebs and a drop ceiling). However, my cpu temperatures (4770k 4.4 ghz, 1.25v) were never out of the ordinary. In fact, they only went down by about 3 to 4 degrees Celsius on average after dusting off the radiator. Watching this video suggests that this isn't normal, so I wonder why I didn't notice an improvement in temps.
reply
[SAW]Spitfire
Just some advice on getting consistent readings from the anemometer: All wind tunnels used for aerodynamic testing have the fans in a pull configuration. That's because air pushed by a fan is very turbulent and uneven, however air being pulled through a circular tunnel like yours will be linear and a lot more consistent. You can see that by looking at the No Rad control graph 4:11 , the way the flow fluctuates like that could be avoided by running the fans in a pull configuration. Other than that, loved the video, the effort you put into your testing is greatly appreciated
reply
Just some advice on getting consistent readings from the anemometer: All wind tunnels used for aerodynamic testing have the fans in a pull configuration. That's because air pushed by a fan is very turbulent and uneven, however air being pulled through a circular tunnel like yours will be linear and a lot more consistent. You can see that by looking at the No Rad control graph 4:11 , the way the flow fluctuates like that could be avoided by running the fans in a pull configuration. Other than that, loved the video, the effort you put into your testing is greatly appreciated
reply
Dan
I've noticed a big big thing too is dust on the fan blades itself. It seems dust on the leading edge of a fan blade really creates a lot of problems for them what I assume is cutting the air to travel over the rest of the blade. Also the hot wire anemometer setup is very very similar to how a mass air flow sensor works on a car. Very interesting! When tuning a car any change in the size of the intake tube its in, or its position in it renders the car unable to get accurate air mass readings.
Great work, I look forward to more use of this setup in the future.
reply
I've noticed a big big thing too is dust on the fan blades itself. It seems dust on the leading edge of a fan blade really creates a lot of problems for them what I assume is cutting the air to travel over the rest of the blade. Also the hot wire anemometer setup is very very similar to how a mass air flow sensor works on a car. Very interesting! When tuning a car any change in the size of the intake tube its in, or its position in it renders the car unable to get accurate air mass readings.
Great work, I look forward to more use of this setup in the future.
reply
manitoublack
I see what you did there ;)
Vacuum's have an extremely high potential of carrying static electricity...
Also As a ventilation engineer: your test chamber looks good, but it would be better to see the anemometer placed inside the controlled flow chamber. Having it external does open it up to sundry airflow from within the room. The actual best way to measure airflow in a setup like this is with pitot tube's (static and total) and a water column (in this case you'd lay the water column on a shallow angle to see finer variations.)
reply
I see what you did there ;)
Vacuum's have an extremely high potential of carrying static electricity...
Also As a ventilation engineer: your test chamber looks good, but it would be better to see the anemometer placed inside the controlled flow chamber. Having it external does open it up to sundry airflow from within the room. The actual best way to measure airflow in a setup like this is with pitot tube's (static and total) and a water column (in this case you'd lay the water column on a shallow angle to see finer variations.)
reply
Off-Grid
Living off grid in an RV, with a small dog that sheds. I find myself cleaning out the dust filters every 3-4 days, and see a 6C drop in temps whenever I do. Running an R9 3900x OC'd to 4.425 GHz all core with a 40 mm thick 240mm rad with static fans and one corsair fan as intake. 2 Corsair fans as exhaust. See up to 76C on blender and usually 48C-56C while gaming.
reply
Living off grid in an RV, with a small dog that sheds. I find myself cleaning out the dust filters every 3-4 days, and see a 6C drop in temps whenever I do. Running an R9 3900x OC'd to 4.425 GHz all core with a 40 mm thick 240mm rad with static fans and one corsair fan as intake. 2 Corsair fans as exhaust. See up to 76C on blender and usually 48C-56C while gaming.
reply
Add a review, comment
Other channel videos
