The obvious choice is to combine the water wheel with a cooling tower. If the tower has enough holding capacity it should be able to sustain constant gravity fed flow through the block and act as a buffer to remove the pulses of flow from the water wheel pump. Make it so. :)
you could also have several parallell spirals combining into one with the inflow ends evenly spread around the wheel. Also have the wheel fully submerged to avoid air.
Coil pumps were designed to pump up from a stream to a hill or water tower to a reservoir, then use gravity feed water pressure to water crops, etc. It would be really cool to replicate that type of system to cool the cpu, as it should eliminate the air bubbles.
An inventor's greatest aspiration is to have their invention be used for a purpose they couldn't possibly have imagined. It's cool to see such an old design work as well as this does for PC cooling.
V2 should include another pump setup exactly like this. But, have the pumps setup 180 degrees out, meaning have one of the inlet tubes under water, and the other pump have the tube out of the water. Then have them connected via T Flange. See if that would keep a constant flow of water. Probably not, because you still have to account for the air in the system. But i'd be curious to see what happens.
That air could be gravity bleeded out. You could add long vertical tube before junctions and bleeding hole on top of that. Bleeding hole should be higher than any point of water. It's janky, but what part of this really isn't :D And since that bleeding hole is above any point of water (and that drpo should also start abve whole system) then it should work only with gravity pushing all the water trough. Pump just lifts the water to highest point and gravity does rest.
a alcohol makers burp valve set a half foot over the pump wheel would probably do it. if you want jank, a t-valve to catch the air, and a red solo cup with a water trap using a plastic shot glass would be effective enough. make the riser tube 2x the size of the water pipe to lower resistance. make the whole pc downhill would let any air flow back up to the trap.
Two suggestions that should improve this system: If you add a funnel at the inlet of the pump, you should scoup up mor water/less air. And if you add something like a catchtank you could get rid of the air entirely. So basically add a tank at a very high position, where the pump dumps its water in. Have an opening in the top to release the air and an outlet for the water in the bottom. If the tank is positioned high enough above the wateblock gravity should create mor than enough waterpressure to ensure a good flow.
I thought a standpipe to bleed out the air would be a good idea as well, then remembered what these pumps are famous for, pumping up hill!! The standpipe could have to be taller than his house to become a harder path than the micro channels in the core! Would be a good exercise to find out how high it needs to be?
You can get rid of the air bubbles putting a tall tube open at the top in a T junction before the block. The tube just must be taller than the pump pressure
@@manitoba-op4jxNo need for a mechanical cover, just slap a u-turn on the top so the end of the vent pipe faces downwards. Let gravity protect you from dust.
You also need to make sure that the top section is wide enough for the air to escape and let the water flow freely around the bubbles otherwise they'll push the water out too.
That's going to need 6-8 ft of stand pipe/tubing...unless you use a much wider diameter pipe. Hmm. Add a check valve before the T and we now have a reserve/overflow tank as well. 🤓 okay, I'm sold.
Maybe use the pump to fill a reservoir then use that to flow through the block removing the air from the system. If your pump can keep up with the flow... but if the flow is causing a change in water heights it should balance out and become a flow that speeds up and down.
Two thoughts: 1) instead of a radiator, have the coil pump dump water into the top of a cooling tower with a second reservoir at the bottom, and just have it elevated to provide the necessary pressure into the water block. That would help avoid the water becoming thermally saturated *and* would result in a continuous water flow into block, likely resulting in better performance. 2) Can you cool a PC with an archimedes screw? :p
You probably could if you combined the screw idea with a water tower to provide constant flow through the cpu block. You would need to rely on gravity to go from the tower to the CPU to the big water tank.
As soon as I saw the thumbnail I knew exactly where the inspiration for this came from. Very well executed! Hooking it up to a radiator would be neat, I'd like to see how much pressure it can handle.
I kind of want to figure out a way to implement some sort of air catch in the loop that could occasionally release extra air so just water flows through the actual cooling side of things. Cool concept and that it actually provides a useful level of cooling!
Near the motor needs an upward shaft with a ball float, it would regulate air and allow water separation. Plus double hosing equals double pressure, and the motor should be on a band not gears so no energy lost as it tries to compress water between gears.
a manifold with check valves, multiple inlet tubes staggered around the drum so that the moment one tube is up, the next one is in the water. I think that would increase the flow a good bit. I'm not sure how much effect the length of the coil has on the line pressure. It may work or it may not if you had 2 shorter coils with their inlets 180 degrees apart.
Multiple feeds would work and create less flow spikes (more like an alternators ac ripple when passing through the rectifier to become dc... The more loops per inlet hose the higher the head pressure at pump outlet the more inlets the smoother the pump flow
It's a very cool idea, could you try exactly the same set up but print it along the length of the tank instead of the width and make the centre cyclinder smaller so the inlet pipe spends less time in the air ? that should help push more water throught the loop and then compare the 2 results please ? And then maybe if you have time experiment with different rotation speeds ?
@@davidconner-shover51 I didn't mean he should submerg it just rotate the pump 90 deg on the x-axis and elongate it while also making the radius smaller.
A longer cylinder won't do much I don't think... Except increase "cold startup" time. As for the smaller cylinder, causing a higher frequency of water/air pulses (whilst keeping whatever ratio of water/air we have now) would make for more consistent cooling... not sure if it's any better but worth the Science!
This is an interesting build! Only suggestion I have is to put a few more hoses with offset openings, with the other end going to a common collector, so that the intervals between pumps are shorter.
@sinisterthoughts2896 that's true in the screw portion of the 'pump' tubing, but once the water exits the pump, the air is no longer necessary. It's almost an inverted version of a pneumatic screw compressor which has and air/oil separator after the screw.
I was thinking this too. Like pumping an old Super Soaker water gun, but with a bleed valve on top of the tank to keep a constant pressure and give some place for the air to go.
I think you might see an increase in temps on the upper cores IF the cpu was in a vertical position as I don't believe the water level would reach the top of the cpu. I believe the cpu being flat on your test bench was a good idea.
If the cooler was vertical, it could actually work better, assuming the outlet was on the upper side, rather than a left or right. As it would self purge the air I believe.
You could try putting a couple of them with their outputs joined using one way valves so you can maybe get better preassure and alleviate the irregular flow
The RUclips algorithm works well sometimes. Not only did I see the pump video I also contemplated the same things. So I'm glad you made this lol, I did think the air pockets would cause more problems.
@MajorHardware I think that you could get better flow with two or more tubes on the pump. If you offset the opennings of the tubesby 90 deg or 120 for three, then you will have pretty steady flow. Make a plenum with inlets and check valves to mix the flow from all the tubes. I think that would give you plenty of flow. If you use the suggestion I saw of a catchtank, with my idea you would have a pump that could supply a steady stream of air-free water to the cooling block.
Re-do this process with three water loops combining into one later in the loop. Each one has a check-valve to prevent back-flow. Then off-set each loop by 120 degrees. And you can showcase the typical 3-phase AC function of power generation, with water. And why it is beneficial in removing the 'air gap'.
It is possible to remove all air in the system if the pump is delivering water into a closed tank with only an outlet for a hose that has 50% of the carrying capacity of the hose on the pump. My assumption is that the pump has a 50% duty cycle, so if we account for that by having double the hose capacity on the pump, then we can accumulate the combined pressure in a closed tank. Then the pressure in the closed tank will deliver a relatively stable pressure on the side with the reduced hose size. Inlet should of course be on top and outlet on bottom, to keep max pressure on the outlet hose and reduce movement of the water. If possible, you can in this configuration add an infinite amount of pumps (the law of diminishing return may apply here) to raise the pressure on the secondary side of the closed tank (for stability of flow's sake, they can be run with an offset of 180degrees).
Though it would not improve the issue of pumped air, it might be interesting to put 2 hoses in parallel where each dips into the water on opposite side of the wheel and both output into the hub connector. One-way valves would surely be needed where each hose connects to the hub (or wherever the combine) to avoid back-pumping, however, which may make this approach impractical. I don't know what's available or can be made, but a t-shaped valve where each hose pushes a ball or valve to close off the other hose with the output in the middle might work if independent 1-ways are too bulky, pricey, or impractical. The extra restriction might be a deal-killer anyway. The idea still interests me, though, as double action should greatly reduce the "off" time without increasing motor speed and increase total volume if the same hose were used. A slightly smaller hose might be more practical for compactness as you'd want to keep a similar number of turns, but you wouldn't want to reduce the hose diameter too much due to the exponential volume loss and increase in surface tension; the smaller the hose radius, the greater the losses and vice-versa.
Have it pump into a reservoir with an overflow. That does double duty, smoothing out the flow and eliminating air. Unfortunately, you lose your pressure unless you take full advantage of your head.
Very cool! These videos are the exact reason why I keep coming back to this channel, always some interesting stuff, that I haven't seen anyone try before, keep up the great work!
No ideas or recommendations. Just wanna say I appreciate your curiosity and skill to test theese weird cooling ideas 😊 Thank you and see you in the next one ✌️
You might be able to get the air out by T-ing the return line back in the system just past the pump and having some sort of pressure release valve on the return as well to manage the pressure. Bonus points for finding an adjustable pressure valve to find the pressure limit of the system.
Have 2 or more coils on the same drum with evenly spaced inlets. With some clever valving you could eliminate the air gap, or use it to fill a water tower so you can have a constant flow.
With the fishtank... set up in a way with no air and actually usable for a daily driver, the goal being daily use shouldn't be enough to heat soak it so each day you can run w/o fans. The truly silent PC XD. so big, no fans required. but it would require a like 100% water cooled pc like mobo etc. (Not to be confused with an actual fish tank pc with is being submerged. although that could be a super cool project too. Specially if you 3d print all the brackets etc, so community could create their own with yoru plans. It would make for an awesome series. ) After that tanget, will just say thank you for awesome content and keep it up. :)
A long flat/horizontal section of tubing with foam/sponge diffusers inside it would allow you to merge the air into the water as fine bubbles which might actually provide some cooling improvements just from turbulence
Might want to add a part to separate the water in the air I would suggest a cup that lets the water air mixture in has a tube draining out the bottom and lets the air out the top as long as you keep it above the rest of the components it should flow properly thanks to gravity
One idea is if you put a static tank that captured the air and water downstream of the swivel, and released only water through one exit that leads to the CPU, and have an air throttle as a second exit. If the throttle is correctly sized, it will regulate a constant higher than atmosphere pressure. Think a tire with a slow leak in it. Another idea is to have multiple intakes to even out the air/water cycles.
You can just use a normal reservoir for it with an automatic vent screwed into the top of it. I have one in my loop on the 80ml reservoir that serves as the air trap. The valve works by means of a membrane that can flex both ways, it lets air in or out when the pressure difference on the membrane exceeds about 400mbar. Aquacomputer sells them for like 7 euro.
I am honestly surprised you did not use the noctua to power this. You could try printing an archimedes screw, for a more stable flow. It's a really easy, if a bit loud design.
The noctua motor wouldn't have enough torque. I'm not sure, but I think that even gearing it down wouldn't work because it would end up turning too slowly. Not to mention the hassle of acquiring or creating a gear box that is efficient enough.
@@SilvaDreams yes, but too slow and you don't move the great away from the CPU block fast enough. The system needs to be balanced for both the needs of the pump and the needs of the CPU block, and I don't think the noctua fan motor is capable of that balance.
@@ChadHadsell True it wouldn't work mostly because it would lack the torque. You need a bigger wheel and then drop the water into a holding tank above the PC and the water exit out to the pumping tank on a lower level to ensure no air bubbles and a consistent flow.
Archimedes screw doesn't provide a constant flow - from an overall design point of view, this actually *is* an archimedes screw, in terms of how it is providing the pressure and pumping, just optimised and simplified for modern construction.
I'd like to see you try to make the water flow constant, through a bottom feed resivoir + pressure release for excess air? I think it'd be interesting physics lesson too
Big "Gravity based" water flow. Just a big ol reseviour (fishtank) somewhere 5 feet above where the PC is. See how often you need to re-fill it. Could also make some sort of interesting pump to refill it, like a 5 foot long Archimedes screw.
It needs a small funnel for the intake... and for higher pressure and a constant flow it would help to have a reverse "p" trap, or simply force fill the last loop on the drum and pump the water up to the ceiling before it flows down into the computer turning the hose into a sort of reservoir.
I always love the random experiment videos, you tried to make a water pump but you also made a trompe, now you need to make a separator with a check valve that will release some of the air as it builds up, and use it to cool a rad to help pull heat from the system.
You could do a triple pump, or at least a double, separated by 120 or 180 degrees, but with check valves between the units.... that would give you more consistent flow. Or, you could put the water into a small tank, and the exit of that tank could then feed the cooling block. if you left an air bleed port on the top, sort of an overflow back to the fish tank, you could get constant flow, without the air. All told, very neat project you built!
5:57 others have suggested it, but I was gonna say have a second reservoir to pump the water into and then you'll have constant flow if you position the outlet/inlet at the right spots. The Romans did something like it to pump water over hills/valleys
What about one more intake witch is shifted 180° so it can put water into the system without the waiting to get more water this way you can join the two pipe together before it is going into the rotating section of the tube, and maybe after this you can add a closed water tank to get just the water and remove some air, i think a litlebit air is fine because this way the pump can generate preasure in the tank and improve water preasure like that. Anyway nice video! Keep up!😄
Aquarium filter powerheads are simple centrifugal pumps which drive a considerable volume of water through a similarly sized tube without the air gaps you get from anything resembling Archimedes' screw. As you noted, a radiator on the outflow side of the CPU may be helpful in extending the cooling capacity of the input reservoir.
i have seen one hat was 10ft..he had a i gal container with the bottom cut out attached to the end of the tube on the wheel..it worked good..and take the air out of the hose on the wheel so water flows better with less surge in the water. 👍👍
I have generally found that people use much higher flowrates than are actually required for most watercooling setups. Water has massive heat capacity, and the difference between running your pump at full compared to its lowest speed is often just a couple degrees; temperatures is primarily controlled by radiator size and fan speed. Running your pump at minimum is much quieter, your pump will last longer, and the pump will dump less excess waste heat into the loop to boot. Definitely some kind of vertical stand-pipe to release air would be a good addition here to purge the air from the system (the air is blocking a lot of heat transfer) and provide smoother flow. Since the tubing sees variable back-pressure as it goes along the coils, it also might be helpful to jump tubing sizes a couple times across the drum. Start with a smaller inlet pipe and then use couplers to jump up in size every turn or two, which will approximate a widening pipe and provide more weight to keep the air slugs from climbing over the loop near the end at higher rotation speeds.
I've seen it surest a time or two in at least as many ways. Putting a water tower after the rotary union would give you a place to discharge air to. But I think having a bike pump or something similar to pull the water though the hose (end of the hose in the water obviously) would likely fully prime the pump. The second problem would be keeping the inlet in the water 100% of the time. It should work if you get a second rotary union for the inlet side and shove the hose into the tank then it'll work more like an auger which is basically what this is.
I think this needs a showdown sequel: an archimedes pump, this coil pump, maybe a rope pump, all have a risk of sucking in air, and need a pressure reservoir that should continually be allowed to leak through a valve from the top back into the input tank, to separate air before the CPU... then a peristaltic pump (plenty of printable designs for silicone tubing are freely available), and an off the shelf water cooler to bring it home. I believe all of the above (except the "control" ordinary water pump) can be easily made with a brushed motor as was done here, but knocking it out of the park would be using an (e.g.) NF-A12x25 motor and 3D printed parts, rather than an already-geared-down brushed motor to accomplish the task.
One could argue the coil pump is even older. Because it's basically an Archimedes Screw (aka Egyptian screw). And that dates back to ancient times. If you do an internet search for Archimedes screw, you'll see lots of pics and animations of an auger inside a tube. But there were several variants, including coiled tubes. The wikipedia page for the Archimedes Screw also has an equation that can be used to calculate the flow rate. So for example, if you want a certain flow rate, you can use the equation to design your pump. 😎
Try using an air bubbler in a tube to lift the water, like it used in aquariums and under water dredging. If the thermodynamic balance is just right the evaporative cooling by the air bubbles will remove the excess heat ensuring the reservoir stays cool.
The key to eliminating bubbles is simpler than you think, make the length of the coil longer and the diameter smaller. And then tilt it 45 degrees so that the inlet is completely submerged.
The lower the pump sits in the tank, the better it will work. Elevating the tank will make better outlet pressure too. Did one of these for a science fair years ago. A funnel or larger opening is not going to increase efficiency because there still is air present from when the inlet exits the water to when it re-enters. A funnel or cone on the end would introduce more water into the system but also trap the equal amount of air. You're fighting line loss, just like with a vacuum. The only way to improve the design as it sits is to lower the wheel further into the tank, and raise the tank above where you want it to pump to. Use the weight of the water to compress the air inside the tube as much as possible and then increase the height of the actual pump so that anything downhill sees more flow. you can maybe get 50% of the air out of the system that way, the rest is there as a byproduct unfortunately.
Add a large intake paddle to scoop the water into the intake hose. Also, check the power draw on the hub motor and see if the CPU fan output on the MOBO can run the pump and regulate the speed of the wheel. Using a check valve as well on the intake will stop the air from entering when the intake scoop no longer has water pressure to open the valve. Make a small tank just large enough for the wheel, throw a couple of heatpipe fin coolers around the tank to passively cool and keep the power draw to a minimum.
Modern boiler systems for radiant heat often have an air release valve added into the system. You might be able buy one but could probably make one. It's basically a chamber with a fin or two and a floating check valve. Air gets out but water is supposed to stay in and works on pressurized systems. I would really like to know if something like that would work with something like this.
The setup you just made looks pretty good. I doubt doing the extra stuff the comments are saying will do much but experimenting is the fun part. The only thing that concerns me is the water evaporating. Maybe you can make your pump float on top of the water to stay a constant level?
i call that a win! pumping vertical into a tower where it pools again, and then gravity feeds into the block (eliminating air gaps) it would probably be dead stable and might handle a GPU added. then do a heat exchange radiator off the back of the tank to keep it from heat soaking.
Hey, can you do a video on the rotary woofer? (It's a subwoofer that uses a fan and blade pitch to create sound.) It looks like a very powerful and interesting, yet niche and underreported technology. You can do sub-20Hz sounds at high volumes without much power intake. Also according to those few reports on RUclips you can easily shake your house + your neighbor's house with "simple" DIY solutions. So hooking it up to 120mm fan that is more powerful than a Noctua sounds like a pretty exciting idea!
Two things to try maybe 1. A water tower like others have said 2. Some kind of ball valve like some of the cheap snorkels have but it keeps the water in and air out. Also using a larger tube might help
0:28 I know exactly where that clip is from, that's from a Habitual Linecrosser video lol dude is hilarious. Interesting to see such old tech used in this way, surprising to see it actually work not too bad.
Use the pump to feed a siphon reservoir above the CPU. That will eliminate the air gaps. Then just tune the return rate from the siphon to the tank with an adjustable nozzle to match the flow of the pump. Or you could have a secondary catch-up pump that comes on when the siphon reservoir gets low and cuts off once it gets to a certain level.
Building a "water tower" to act as a capacitor for the flow should allow you to keep the air out and flow more continuous. Plus it would look cool, too.
If you want to get rid of the air, you can build a separater tank. You need an airtight tank between the pump and the water block. It needs to have a water inlet, and outlet, _and_ a float actuated valve to let air out. Air will build up in the tank with water until the water level drops enough to let the air out. The _flow_ will stop while that happens, but no air will run past it. Here's kind of an illustration for you: ---[ ^ ]--- Just make sure the water outlet line is lower than the float.
Some sort of air separator system would be neat. Probably could make out of of just a bottle as reservoir where the air can bubble up to the top and some float valve to vent when there is enough air in it.
a bit of standing pipe out from the cpu block could be good to maybe make sure the block is always full. so it has to fill the pipe to escape the block.
Make a second drum , and hook both to a Y fitting with check valves, if you offset their phasing by 90 degrees you should be able to eliminate most of the air pockets and potentially use smaller drums overall
an idea to get rid of the air gaps is to hold the whole thing under the water and rotate it till all the airs out. also have the entire wheel lower under the water the whole time and have maybe a belt/chain drive from the motor
Try using bigger tubing on the drum and having it reduce down to PC cooling tube at the swivel fitting. Should increase the flow rate and probably pressure though the system
The obvious choice is to combine the water wheel with a cooling tower. If the tower has enough holding capacity it should be able to sustain constant gravity fed flow through the block and act as a buffer to remove the pulses of flow from the water wheel pump.
Make it so. :)
you could also have several parallell spirals combining into one with the inflow ends evenly spread around the wheel.
Also have the wheel fully submerged to avoid air.
@@nalissolus9213 unfortunately the air is what actually allows this pump to work, pretty neat concept
This is exactly what I had in mind: Just put a reservoir high enough that the downstream flow continously cools the CPU without any bubbles.
ya, water reservoirs and tanks, solving water pressure problems for centuries.
This was my thought on this too.
Coil pumps were designed to pump up from a stream to a hill or water tower to a reservoir, then use gravity feed water pressure to water crops, etc. It would be really cool to replicate that type of system to cool the cpu, as it should eliminate the air bubbles.
An inventor's greatest aspiration is to have their invention be used for a purpose they couldn't possibly have imagined. It's cool to see such an old design work as well as this does for PC cooling.
A DIY Peristaltic pump would be cool and probably pretty easy to do. No air bubble issues either.
was about to make the same comment. A piston pump would be funny too.
A peristaltic pump going fast enough would also have more pressure behind it than this does.
V2 should include another pump setup exactly like this. But, have the pumps setup 180 degrees out, meaning have one of the inlet tubes under water, and the other pump have the tube out of the water. Then have them connected via T Flange. See if that would keep a constant flow of water. Probably not, because you still have to account for the air in the system. But i'd be curious to see what happens.
That air could be gravity bleeded out. You could add long vertical tube before junctions and bleeding hole on top of that. Bleeding hole should be higher than any point of water. It's janky, but what part of this really isn't :D
And since that bleeding hole is above any point of water (and that drpo should also start abve whole system) then it should work only with gravity pushing all the water trough. Pump just lifts the water to highest point and gravity does rest.
Maybe adding a mini air release valve at the highest point should help bleed the air out?
I was thinking add a second loop on the inside of the wheel.
a alcohol makers burp valve set a half foot over the pump wheel would probably do it. if you want jank, a t-valve to catch the air, and a red solo cup with a water trap using a plastic shot glass would be effective enough. make the riser tube 2x the size of the water pipe to lower resistance. make the whole pc downhill would let any air flow back up to the trap.
@@TheSanpletextthe whole point of the pump design was head pressure. The vent pipe would need to be super tall
Two suggestions that should improve this system: If you add a funnel at the inlet of the pump, you should scoup up mor water/less air. And if you add something like a catchtank you could get rid of the air entirely. So basically add a tank at a very high position, where the pump dumps its water in. Have an opening in the top to release the air and an outlet for the water in the bottom. If the tank is positioned high enough above the wateblock gravity should create mor than enough waterpressure to ensure a good flow.
yup was typing about funnel but read this.. so yeah =)
I thought a standpipe to bleed out the air would be a good idea as well, then remembered what these pumps are famous for, pumping up hill!! The standpipe could have to be taller than his house to become a harder path than the micro channels in the core! Would be a good exercise to find out how high it needs to be?
the air is required for this type of pump to work. if you eliminate the air it will cease to flow.
@@sinisterthoughts2896you can eliminate it after it leaves the pump though
@@colinmetzger6755 no you cant because you wouldnt be able to maintain the flow rate unless your delivery far outweighed your consumption.
You can get rid of the air bubbles putting a tall tube open at the top in a T junction before the block. The tube just must be taller than the pump pressure
Nice. This is the way.
@@manitoba-op4jxNo need for a mechanical cover, just slap a u-turn on the top so the end of the vent pipe faces downwards. Let gravity protect you from dust.
You also need to make sure that the top section is wide enough for the air to escape and let the water flow freely around the bubbles otherwise they'll push the water out too.
That's going to need 6-8 ft of stand pipe/tubing...unless you use a much wider diameter pipe. Hmm. Add a check valve before the T and we now have a reserve/overflow tank as well.
🤓 okay, I'm sold.
Maybe use the pump to fill a reservoir then use that to flow through the block removing the air from the system. If your pump can keep up with the flow... but if the flow is causing a change in water heights it should balance out and become a flow that speeds up and down.
Two thoughts: 1) instead of a radiator, have the coil pump dump water into the top of a cooling tower with a second reservoir at the bottom, and just have it elevated to provide the necessary pressure into the water block. That would help avoid the water becoming thermally saturated *and* would result in a continuous water flow into block, likely resulting in better performance.
2) Can you cool a PC with an archimedes screw? :p
You probably could if you combined the screw idea with a water tower to provide constant flow through the cpu block. You would need to rely on gravity to go from the tower to the CPU to the big water tank.
Steven Mould would be proud of your set up! Such a cool use for the Wirtz pump.
As soon as I saw the thumbnail I knew exactly where the inspiration for this came from. Very well executed! Hooking it up to a radiator would be neat, I'd like to see how much pressure it can handle.
I was also like "well, I know where this came from" and I'm certainly here for it. I love when my favorite RUclipsrs follow my other favorites lol
A Habitual Line Crosser reference is the last thing I expected to hear lol
This performed infinitely better than I expected. Love it! Cheers 🍻
I kind of want to figure out a way to implement some sort of air catch in the loop that could occasionally release extra air so just water flows through the actual cooling side of things. Cool concept and that it actually provides a useful level of cooling!
Near the motor needs an upward shaft with a ball float, it would regulate air and allow water separation. Plus double hosing equals double pressure, and the motor should be on a band not gears so no energy lost as it tries to compress water between gears.
Missed opportunity to print this in wood filament 😁
😂 I did not expect to see one of my clips used on a CPU but I'm here for it 😂
a manifold with check valves, multiple inlet tubes staggered around the drum so that the moment one tube is up, the next one is in the water. I think that would increase the flow a good bit. I'm not sure how much effect the length of the coil has on the line pressure. It may work or it may not if you had 2 shorter coils with their inlets 180 degrees apart.
that was my first thought was to have a double system with staggered loops so you are always getting a constant flow of water into the system
Multiple feeds would work and create less flow spikes (more like an alternators ac ripple when passing through the rectifier to become dc... The more loops per inlet hose the higher the head pressure at pump outlet the more inlets the smoother the pump flow
@@daxconnell7661it still wouldn't be constant flow but more like a ripple instead of a pulse
It's a very cool idea, could you try exactly the same set up but print it along the length of the tank instead of the width and make the centre cyclinder smaller so the inlet pipe spends less time in the air ? that should help push more water throught the loop and then compare the 2 results please ? And then maybe if you have time experiment with different rotation speeds ?
I wouldn't go too far with submerging it, maybe 3/4 of the way in. the air gap is what makes this work.
@@davidconner-shover51 I didn't mean he should submerg it just rotate the pump 90 deg on the x-axis and elongate it while also making the radius smaller.
it functions because of the air. his design is spot on.
@@sinisterthoughts2896 I said less time in the air not no air =D
A longer cylinder won't do much I don't think... Except increase "cold startup" time. As for the smaller cylinder, causing a higher frequency of water/air pulses (whilst keeping whatever ratio of water/air we have now) would make for more consistent cooling... not sure if it's any better but worth the Science!
Find a way to cool a CPU with one of those drinking bird toys or a tipping bamboo fountain. Now those are real challenges! Hahaha.
Man I'm in love with your channel! awesome work and awesome personality as a host you have!
This is an interesting build! Only suggestion I have is to put a few more hoses with offset openings, with the other end going to a common collector, so that the intervals between pumps are shorter.
really cool idea! probably should have put some type of accumulator to separate the air/water in the line before passing through the water block
the water only moves because the air is pushing it.
@sinisterthoughts2896 that's true in the screw portion of the 'pump' tubing, but once the water exits the pump, the air is no longer necessary. It's almost an inverted version of a pneumatic screw compressor which has and air/oil separator after the screw.
Really cool concept. I would add a smaller fill tank to bleed the air and act as a capacitor to keep the flow constant. Great video as always.
I was thinking that a vertical tube right after the pump would allow the air to bleed off, but yeah a reservoir would accomplish the same thing.
I was thinking this too. Like pumping an old Super Soaker water gun, but with a bleed valve on top of the tank to keep a constant pressure and give some place for the air to go.
I think you might see an increase in temps on the upper cores IF the cpu was in a vertical position as I don't believe the water level would reach the top of the cpu.
I believe the cpu being flat on your test bench was a good idea.
for sure, it seems to keep the cooling fins submerged
If the cooler was vertical, it could actually work better, assuming the outlet was on the upper side, rather than a left or right. As it would self purge the air I believe.
@@klampykixx Good point you have. Thanks for sharing.
You could try putting a couple of them with their outputs joined using one way valves so you can maybe get better preassure and alleviate the irregular flow
three phased pumps! apply the same principles as electricity and this should work halfway decent
You could also use three separate pieces of tubing on the same rotor, phasing each tube inlet 120* apart
@@thejo6331 Something like that would be really nice! I was thinking something more like inline motors, but the electricity analogy works even better!
The RUclips algorithm works well sometimes. Not only did I see the pump video I also contemplated the same things.
So I'm glad you made this lol, I did think the air pockets would cause more problems.
@MajorHardware I think that you could get better flow with two or more tubes on the pump. If you offset the opennings of the tubesby 90 deg or 120 for three, then you will have pretty steady flow. Make a plenum with inlets and check valves to mix the flow from all the tubes. I think that would give you plenty of flow. If you use the suggestion I saw of a catchtank, with my idea you would have a pump that could supply a steady stream of air-free water to the cooling block.
Re-do this process with three water loops combining into one later in the loop. Each one has a check-valve to prevent back-flow. Then off-set each loop by 120 degrees. And you can showcase the typical 3-phase AC function of power generation, with water. And why it is beneficial in removing the 'air gap'.
I love the little habitual linecrosser bit.
It is possible to remove all air in the system if the pump is delivering water into a closed tank with only an outlet for a hose that has 50% of the carrying capacity of the hose on the pump.
My assumption is that the pump has a 50% duty cycle, so if we account for that by having double the hose capacity on the pump, then we can accumulate the combined pressure in a closed tank. Then the pressure in the closed tank will deliver a relatively stable pressure on the side with the reduced hose size. Inlet should of course be on top and outlet on bottom, to keep max pressure on the outlet hose and reduce movement of the water.
If possible, you can in this configuration add an infinite amount of pumps (the law of diminishing return may apply here) to raise the pressure on the secondary side of the closed tank (for stability of flow's sake, they can be run with an offset of 180degrees).
habitual line crosser? great fella!!
nice video man, love everything to do with thermals!!
Though it would not improve the issue of pumped air, it might be interesting to put 2 hoses in parallel where each dips into the water on opposite side of the wheel and both output into the hub connector.
One-way valves would surely be needed where each hose connects to the hub (or wherever the combine) to avoid back-pumping, however, which may make this approach impractical. I don't know what's available or can be made, but a t-shaped valve where each hose pushes a ball or valve to close off the other hose with the output in the middle might work if independent 1-ways are too bulky, pricey, or impractical. The extra restriction might be a deal-killer anyway.
The idea still interests me, though, as double action should greatly reduce the "off" time without increasing motor speed and increase total volume if the same hose were used. A slightly smaller hose might be more practical for compactness as you'd want to keep a similar number of turns, but you wouldn't want to reduce the hose diameter too much due to the exponential volume loss and increase in surface tension; the smaller the hose radius, the greater the losses and vice-versa.
Ok this was super cool. These kinds of creative videos are my favorite.
Men, every video from you is a surprise. THX 🥰
Bwt: Sink that thing deeper into the tank, so maybe you can transport more water and less air.
Make it spin faster! In a closed fish tank, use magnets to make it spin! Also remove all the air from the tubes! This was sooo cool!
Have it pump into a reservoir with an overflow. That does double duty, smoothing out the flow and eliminating air.
Unfortunately, you lose your pressure unless you take full advantage of your head.
Very cool! These videos are the exact reason why I keep coming back to this channel, always some interesting stuff, that I haven't seen anyone try before, keep up the great work!
No ideas or recommendations. Just wanna say I appreciate your curiosity and skill to test theese weird cooling ideas 😊 Thank you and see you in the next one ✌️
You might be able to get the air out by T-ing the return line back in the system just past the pump and having some sort of pressure release valve on the return as well to manage the pressure. Bonus points for finding an adjustable pressure valve to find the pressure limit of the system.
Woo Steve Mould ! I'm sure this pump would fit in some theme for a PC
This is an interesting consept. Whit some modifications to original structure this can be scaled for more higher pressure and flow rate.
These are the best videos, love this kind of content dude
Have 2 or more coils on the same drum with evenly spaced inlets. With some clever valving you could eliminate the air gap, or use it to fill a water tower so you can have a constant flow.
a Major Hardware video, inspired by Steve Mould, with a HabitualLineCrosser reference in the intro. Yeah, i like this.
With the fishtank... set up in a way with no air and actually usable for a daily driver, the goal being daily use shouldn't be enough to heat soak it so each day you can run w/o fans.
The truly silent PC XD. so big, no fans required. but it would require a like 100% water cooled pc like mobo etc. (Not to be confused with an actual fish tank pc with is being submerged. although that could be a super cool project too. Specially if you 3d print all the brackets etc, so community could create their own with yoru plans. It would make for an awesome series. )
After that tanget, will just say thank you for awesome content and keep it up. :)
I can't get past the habitual linecrosser clip at the beginning. I nearly died lmfao
A long flat/horizontal section of tubing with foam/sponge diffusers inside it would allow you to merge the air into the water as fine bubbles which might actually provide some cooling improvements just from turbulence
Might want to add a part to separate the water in the air I would suggest a cup that lets the water air mixture in has a tube draining out the bottom and lets the air out the top as long as you keep it above the rest of the components it should flow properly thanks to gravity
One idea is if you put a static tank that captured the air and water downstream of the swivel, and released only water through one exit that leads to the CPU, and have an air throttle as a second exit. If the throttle is correctly sized, it will regulate a constant higher than atmosphere pressure. Think a tire with a slow leak in it.
Another idea is to have multiple intakes to even out the air/water cycles.
You can just use a normal reservoir for it with an automatic vent screwed into the top of it. I have one in my loop on the 80ml reservoir that serves as the air trap. The valve works by means of a membrane that can flex both ways, it lets air in or out when the pressure difference on the membrane exceeds about 400mbar. Aquacomputer sells them for like 7 euro.
I am honestly surprised you did not use the noctua to power this.
You could try printing an archimedes screw, for a more stable flow. It's a really easy, if a bit loud design.
The noctua motor wouldn't have enough torque. I'm not sure, but I think that even gearing it down wouldn't work because it would end up turning too slowly. Not to mention the hassle of acquiring or creating a gear box that is efficient enough.
@@ChadHadsellThese pumps actually work at a low speed, too fast and it doesn't function
@@SilvaDreams yes, but too slow and you don't move the great away from the CPU block fast enough. The system needs to be balanced for both the needs of the pump and the needs of the CPU block, and I don't think the noctua fan motor is capable of that balance.
@@ChadHadsell True it wouldn't work mostly because it would lack the torque. You need a bigger wheel and then drop the water into a holding tank above the PC and the water exit out to the pumping tank on a lower level to ensure no air bubbles and a consistent flow.
Archimedes screw doesn't provide a constant flow - from an overall design point of view, this actually *is* an archimedes screw, in terms of how it is providing the pressure and pumping, just optimised and simplified for modern construction.
I'd like to see you try to make the water flow constant, through a bottom feed resivoir + pressure release for excess air? I think it'd be interesting physics lesson too
Big "Gravity based" water flow. Just a big ol reseviour (fishtank) somewhere 5 feet above where the PC is. See how often you need to re-fill it.
Could also make some sort of interesting pump to refill it, like a 5 foot long Archimedes screw.
It needs a small funnel for the intake... and for higher pressure and a constant flow it would help to have a reverse "p" trap, or simply force fill the last loop on the drum and pump the water up to the ceiling before it flows down into the computer turning the hose into a sort of reservoir.
I always love the random experiment videos, you tried to make a water pump but you also made a trompe, now you need to make a separator with a check valve that will release some of the air as it builds up, and use it to cool a rad to help pull heat from the system.
You could do a triple pump, or at least a double, separated by 120 or 180 degrees, but with check valves between the units.... that would give you more consistent flow. Or, you could put the water into a small tank, and the exit of that tank could then feed the cooling block. if you left an air bleed port on the top, sort of an overflow back to the fish tank, you could get constant flow, without the air.
All told, very neat project you built!
5:57 others have suggested it, but I was gonna say have a second reservoir to pump the water into and then you'll have constant flow if you position the outlet/inlet at the right spots. The Romans did something like it to pump water over hills/valleys
Amazing idea, please keep up the good content😂i really enjoy watching it
Loved this! very creative method of cooling a PC :D
What about one more intake witch is shifted 180° so it can put water into the system without the waiting to get more water this way you can join the two pipe together before it is going into the rotating section of the tube, and maybe after this you can add a closed water tank to get just the water and remove some air, i think a litlebit air is fine because this way the pump can generate preasure in the tank and improve water preasure like that. Anyway nice video! Keep up!😄
Saw that video as well, his videos are always interesting.
Aquarium filter powerheads are simple centrifugal pumps which drive a considerable volume of water through a similarly sized tube without the air gaps you get from anything resembling Archimedes' screw. As you noted, a radiator on the outflow side of the CPU may be helpful in extending the cooling capacity of the input reservoir.
i have seen one hat was 10ft..he had a i gal container with the bottom cut out attached to the end of the tube on the wheel..it worked good..and take the air out of the hose on the wheel so water flows better with less surge in the water. 👍👍
cool experiment
you'd wear out the o ring in the rotary fitting in a day or two but it works as a proof of concept
I have generally found that people use much higher flowrates than are actually required for most watercooling setups. Water has massive heat capacity, and the difference between running your pump at full compared to its lowest speed is often just a couple degrees; temperatures is primarily controlled by radiator size and fan speed. Running your pump at minimum is much quieter, your pump will last longer, and the pump will dump less excess waste heat into the loop to boot.
Definitely some kind of vertical stand-pipe to release air would be a good addition here to purge the air from the system (the air is blocking a lot of heat transfer) and provide smoother flow. Since the tubing sees variable back-pressure as it goes along the coils, it also might be helpful to jump tubing sizes a couple times across the drum. Start with a smaller inlet pipe and then use couplers to jump up in size every turn or two, which will approximate a widening pipe and provide more weight to keep the air slugs from climbing over the loop near the end at higher rotation speeds.
I've seen it surest a time or two in at least as many ways. Putting a water tower after the rotary union would give you a place to discharge air to. But I think having a bike pump or something similar to pull the water though the hose (end of the hose in the water obviously) would likely fully prime the pump. The second problem would be keeping the inlet in the water 100% of the time. It should work if you get a second rotary union for the inlet side and shove the hose into the tank then it'll work more like an auger which is basically what this is.
I think this needs a showdown sequel: an archimedes pump, this coil pump, maybe a rope pump, all have a risk of sucking in air, and need a pressure reservoir that should continually be allowed to leak through a valve from the top back into the input tank, to separate air before the CPU... then a peristaltic pump (plenty of printable designs for silicone tubing are freely available), and an off the shelf water cooler to bring it home.
I believe all of the above (except the "control" ordinary water pump) can be easily made with a brushed motor as was done here, but knocking it out of the park would be using an (e.g.) NF-A12x25 motor and 3D printed parts, rather than an already-geared-down brushed motor to accomplish the task.
One could argue the coil pump is even older. Because it's basically an Archimedes Screw (aka Egyptian screw). And that dates back to ancient times. If you do an internet search for Archimedes screw, you'll see lots of pics and animations of an auger inside a tube. But there were several variants, including coiled tubes. The wikipedia page for the Archimedes Screw also has an equation that can be used to calculate the flow rate. So for example, if you want a certain flow rate, you can use the equation to design your pump. 😎
実用性はともかく見た目にも遊び心があり素敵ですね。
そして5GHz近い動作速度のCPUを、200年以上も昔の古い構造のポンプで冷却するという試みは純粋に面白いです。
Try using an air bubbler in a tube to lift the water, like it used in aquariums and under water dredging. If the thermodynamic balance is just right the evaporative cooling by the air bubbles will remove the excess heat ensuring the reservoir stays cool.
This was cool, everyone making extremely cooling systems, ai like this one is more chill.
The key to eliminating bubbles is simpler than you think, make the length of the coil longer and the diameter smaller. And then tilt it 45 degrees so that the inlet is completely submerged.
The lower the pump sits in the tank, the better it will work. Elevating the tank will make better outlet pressure too.
Did one of these for a science fair years ago. A funnel or larger opening is not going to increase efficiency because there still is air present from when the inlet exits the water to when it re-enters. A funnel or cone on the end would introduce more water into the system but also trap the equal amount of air. You're fighting line loss, just like with a vacuum. The only way to improve the design as it sits is to lower the wheel further into the tank, and raise the tank above where you want it to pump to. Use the weight of the water to compress the air inside the tube as much as possible and then increase the height of the actual pump so that anything downhill sees more flow. you can maybe get 50% of the air out of the system that way, the rest is there as a byproduct unfortunately.
Add a large intake paddle to scoop the water into the intake hose. Also, check the power draw on the hub motor and see if the CPU fan output on the MOBO can run the pump and regulate the speed of the wheel. Using a check valve as well on the intake will stop the air from entering when the intake scoop no longer has water pressure to open the valve. Make a small tank just large enough for the wheel, throw a couple of heatpipe fin coolers around the tank to passively cool and keep the power draw to a minimum.
Modern boiler systems for radiant heat often have an
air release valve added into the system. You might be able buy one but could probably make one. It's basically a chamber with a fin or two and a floating check valve. Air gets out but water is supposed to stay in and works on pressurized systems.
I would really like to know if something like that would work with something like this.
I want more of this! Scale it up with a higher speed/power motor!
try wrapping multi tubes around and add more inlets to feed water thru, might help with the air bubbles
The setup you just made looks pretty good. I doubt doing the extra stuff the comments are saying will do much but experimenting is the fun part. The only thing that concerns me is the water evaporating. Maybe you can make your pump float on top of the water to stay a constant level?
i call that a win! pumping vertical into a tower where it pools again, and then gravity feeds into the block (eliminating air gaps) it would probably be dead stable and might handle a GPU added. then do a heat exchange radiator off the back of the tank to keep it from heat soaking.
Hey, can you do a video on the rotary woofer? (It's a subwoofer that uses a fan and blade pitch to create sound.)
It looks like a very powerful and interesting, yet niche and underreported technology.
You can do sub-20Hz sounds at high volumes without much power intake. Also according to those few reports on RUclips you can easily shake your house + your neighbor's house with "simple" DIY solutions.
So hooking it up to 120mm fan that is more powerful than a Noctua sounds like a pretty exciting idea!
Two things to try maybe
1. A water tower like others have said
2. Some kind of ball valve like some of the cheap snorkels have but it keeps the water in and air out.
Also using a larger tube might help
0:28 I know exactly where that clip is from, that's from a Habitual Linecrosser video lol dude is hilarious.
Interesting to see such old tech used in this way, surprising to see it actually work not too bad.
Add a 1 way check ball at the inlet or 1 full winding in. this should close on the way down preventing water loss and air coming in.
would be super to add an air release valve after the spooling
Try using larger tubing on the pump, and slow the rotation to allow the water to burp back the air so it doesn't get forced into the water block.
Use the pump to feed a siphon reservoir above the CPU. That will eliminate the air gaps. Then just tune the return rate from the siphon to the tank with an adjustable nozzle to match the flow of the pump. Or you could have a secondary catch-up pump that comes on when the siphon reservoir gets low and cuts off once it gets to a certain level.
Building a "water tower" to act as a capacitor for the flow should allow you to keep the air out and flow more continuous. Plus it would look cool, too.
Run 2 drums side by side w/ 180deg offset on rotation. combine their outputs w/ 1 way valves and put a breather to remove air from the output line.
As you squeezed the outlet, you could see the back pressure in the hoze!
this is the stuff!!!! love it man
If you want to get rid of the air, you can build a separater tank. You need an airtight tank between the pump and the water block. It needs to have a water inlet, and outlet, _and_ a float actuated valve to let air out. Air will build up in the tank with water until the water level drops enough to let the air out. The _flow_ will stop while that happens, but no air will run past it.
Here's kind of an illustration for you: ---[ ^ ]--- Just make sure the water outlet line is lower than the float.
The 7700K at 0:30 was awesome
I hope all is well with you. Normally you have a video or 2 a week. Looking forward to more from you when you can.
Some sort of air separator system would be neat. Probably could make out of of just a bottle as reservoir where the air can bubble up to the top and some float valve to vent when there is enough air in it.
a bit of standing pipe out from the cpu block could be good to maybe make sure the block is always full. so it has to fill the pipe to escape the block.
That masked-face thing.
"Would you overclock me?"
"...I'D overclock me..."
To get rid of the air, submerge the whole coil. The tube will pickup only water. Just a few gears or a belt to transmit the power.
Make a second drum , and hook both to a Y fitting with check valves, if you offset their phasing by 90 degrees you should be able to eliminate most of the air pockets and potentially use smaller drums overall
Thats fun, well done!
an idea to get rid of the air gaps is to hold the whole thing under the water and rotate it till all the airs out. also have the entire wheel lower under the water the whole time and have maybe a belt/chain drive from the motor
Try using bigger tubing on the drum and having it reduce down to PC cooling tube at the swivel fitting. Should increase the flow rate and probably pressure though the system