**Experimenting with Temperature Differentials in Water Cooling Loop**

As I sat down to experiment with my water cooling loop, I was eager to see how different pump speeds would affect the temperatures of my graphics cards. With three Titan X's running in a parallel configuration, each card only getting 1/3 of the flow rate entering through this tube, I knew that the flow rate would be divided among all three cards and then recollected as it leaves all three cards and goes back to the pump. This meant that my pumps were being run at max speed, but the cards were only getting a third of that max speed.

The reason for the high temperatures when I had the pump set to zero or its lowest setting was not due to one thing, but rather because of how my system was configured. With multiple radiators and three GPUs, there was a lot of resistance for the D5 pump to overcome, resulting in a lower flow rate than expected. In a single GPU or double/SLI configuration, the difference between the lowest setting and highest setting would likely be more minimal.

However, I wanted to test how the pump speed would affect the temperatures of my cards. To do this, I ran multiple tests with different pump speeds, trying to see if there was any measurable difference in temperature. The results were surprising - even at a relatively high pump speed, I only saw seven differences between the hottest and coldest temperatures on the graph.

**The Importance of Flow Rate**

One thing that became clear during my experiments was the importance of flow rate in water cooling loops. When the fluid didn't move fast enough to cool off the graphics cards effectively or efficiently, the temperatures would rise. This is partially true, but it's also important to consider the specific configuration of the system.

In my case, with three GPUs running in a parallel configuration, I was able to mitigate some of the issues that come with flow rate. The multiple radiators and fittings helped to reduce the pressure drop across the loop, allowing for a more efficient coolant flow.

**Pump Speed vs. Temperature**

While pump speed did affect the temperatures of my cards, it wasn't as significant an impact as I had expected. When running at max speed, the pumps became quite loud - in fact, some pumps, like the DBC pump, are notorious for their noise level even when running at low speeds.

In contrast, pumps that can run at lower speeds without losing too much performance can be a more appealing option. These types of pumps offer manual control over flow rate, allowing you to set and forget it - which is often ideal for systems where temperature stability is crucial.

**Room Ventilation and Ambient Temperature**

Another important factor to consider when experimenting with water cooling loops is ambient temperature. In rooms with good ventilation, the loop will be more efficient and cooler, while in poorly ventilated spaces, the temperatures will rise.

This highlights the importance of considering all factors that affect a system's performance - not just the components themselves, but also the environment in which they are operating.

**Conclusion**

In conclusion, experimenting with pump speed can provide valuable insights into the temperature differentials of water cooling loops. While it may not be a game-changer for every system, it can help you fine-tune your configuration and optimize performance. As always, ambient temperatures and airflow will play a crucial role in determining how well your loop performs - so don't forget to consider these factors when experimenting with different configurations.

The results of my experiments have provided some interesting insights into the relationship between pump speed and temperature differentials in water cooling loops. While there were some surprises along the way, I hope that this article has given you a better understanding of how to experiment with your own system and find the sweet spot for optimal performance.

If you guys are struggling with higher temps and want to try adjusting your loop, finding a balance between pump speed and noise levels is key. On setups like mine, where you have multiple components to consider, mileage may vary - so be sure to play around and see what works best for your system. And don't forget to take note of the ambient temperature and airflow in your room, as these will also impact how well your loop performs.

Thanks for watching today's video! I hope you've enjoyed learning about my experiments with water cooling loops and that it has given you some valuable insights into optimizing performance. If you have any questions or comments, be sure to sound off in the comments below - and don't forget to like and subscribe for more content like this!

"WEBVTTKind: captionsLanguage: enwhat's going on YouTube J's two cents here one point out it's J's two cents not J two cent or JZ two cents it's just James two cents I mean why is that so hard but we are fast approaching the warm months of summer and some climates get super freakin hot and that's when I notice a lot of questions come my way about very specific things when it comes to water cooling so this summer we're going to try and touch on all of that we're going to rehash some old topics and we're going to do something that I've never actually done on this channel we are going to talk about pump speed and does it matter today we're just going to perform a test here on skunk works and we're gonna see what happens in my scenario when it comes to pump speed I thought it'd be fun but before we start playing with our pumps let's go ahead and talk about today's sponsor and that is squarespace.com / J's two cents now last year I use Squarespace is online e-commerce tools to sell some custom writing jerseys and if you guys bought one kudos to you if you didn't sorry they were a limited run but I had no idea where to start and how to actually run any sort of an e-commerce site and Squarespace made it freakishly easy with online inventory tracking order tracking you can even export to spreadsheets and keep track of everything you need to know and even get inventory alerts when things start to get low and sell out so if you're interested in getting your own website then head on over to squarespace.com slash Jays two cents and use offer code J's two cents to get ten percent off your order and if you sign up for a year you get a domain absolutely free now inevitably what happens every time we talk about radiators water pumps and water cooling and computers is people want to compare it to things like automotive cooling when it comes to radiators and thermostats and water pumps and the thing is you can't directly compare that you're talking about very low temperatures here very low pressures whereas in a car you're dealing with huge delta temperatures between ambient temp and the coolant temp reaching you know over two hundred and twenty degrees fahrenheit in many conditions obviously our coolant here isn't getting anywhere near that hot so you really can't compare the two it just isn't directly comparable so inevitably happens is you you start to hear things on forms along the lines of well if you run your pump too slow the coolant is going to move too slow through the loop and the coolant is going to touch the block too long and it's going to get really hot so if you really about that the flipside would be so the coolant is going to stay in the radiator longer and cool off more so on the flip side though the other side of the argument is well if your pumps moving too fast then the coolant is moving through the radiator too quickly and it's not going to cool down fast enough now pretty much both of those scenarios are true in the sense of when they talk about the timing of fluid flow where yes it's going to sit in the blocks a little bit longer it's going to hopefully pick up some heat but it's going to sit in the radiator longer as well if you run it really fast it's going to make more passes per minutes a crummy Oh what if you were to say be able to track this one molecule right inside the loop in time how many times it went around the entire loop per minute yeah there would be more times per minute which is going to mean it's going to run through the radiator more and in the end of the day running it slower or running it faster tends to give you about the same result and that's what I'm going to hopefully show you guys here today hopefully physics aren't going to change suddenly and I'm going to be making out to look like a liar and even if that's true it will stay in the video because this again is all about facts here I like to deal with facts they're they're fun to deal with actually you can't really change those so here's what we're going to do right now I'm going to load up Heaven benchmark on my 3-way Titan X water loop which gets really hot it's running a d5 it's independent from the CPU so there's actually no contamination of temperature between anything happening with the CPU and the V RMS and the GPU so this actually can be a pretty control to test here now it's three tight nexus with a 560 radiator and a 280 radiator running a 250 millimeter reservoir and then again as I said a variable speed d5 that's a manual control not a PWM so I'm actually going to control it I'm going to do a control here where I let the test run for about 10 minutes in heaven let the heat get but nice and built up in the system and then we're going to log what the max temperature was and then I'm going to let the system completely idle for about a half hour I'm going to cool off the room open the windows let the hot air get out of here and start all over from the same ambient temp and then we're going to pump the move the speed of the pump to a hundred percent and we're going to compare the results and see exactly what happened so I'm gonna go ahead and do that now and then we'll come back we'll talk about the results well for me it'll be a while but for you it'll be like right now just so you know what you guys are looking at this is what we care about right here these are the three tight necks as you can see 35 32 and 35 they're running a bit warmer at idle because I do have the three panels hooked up so they do go up a little bit higher on their base clock as you can see right there encore clock they're running at a thousand megahertz right now so they're not ramped down to idle that's why the temperature a little warmer so we don't really care about the minimum we care about the maximum which is right now 35 C so I'm gonna go ahead and run heav'n benchmark and then we're going to see like I said after the system gets full fully heated up and the settings on here we're going to push it we've got it running a 34 40 by 1440p panel tessellation I'm going to go ahead and put that on extreme and then anti-aliasing I'm leading at 2x before I start the test one other thing I'm going to do here is I'm going to take fan speed out of the equation I don't want it speeding up and slowing down so I've set my fans static to 70% inside a speed fan so that's our constant right now the only thing we want changing is going to be the actual speed of the pump we don't want the load to change on the software and we don't want the fan speed to change because that can obviously impact our results so here's the first result right here 63c was the hottest GPU temp as you can see the loads were pretty consistent temperatures had stabilized they weren't still going up so that was our max temperature 63 which might lead you to go wow that's really hot for water-cooled card yeah it actually really is and obviously the pump is running really really slow so what we're going to do now is I'm going to go ahead let the system completely cool off as you can see it's taken a little while to come down water-cooled systems take time for the temperature to drop we're going to pump that we're going to rent that pump up to 100% and then we're going to see exactly what we get after that alright so there's the new result 56c that came down from what 63 so that's 7c difference by just changing the pump speed yeah obviously there's a reason why that happens so let's go ahead and talk about that because that's the thing a lot of people leave out is there's a lot of variables that lead to these temperature fluctuations all right so let's talk here about what just happened some people might look at this conclusion and go oh that's it that's a set conclusion across the board higher pump speed equals lower temp done ship it doesn't really work that way guys you have to understand in my system it's a bit of an extreme situation here where I am trying to get as far on one end of the spectrum as I can and as far as I can on the can on the other end of the spectrum to see what the measurable difference is and then and between that white of a gap I only saw seven differences I'm not going to say seven is a bad number seven is actually a very very good number when it comes to temperature difference between just changing the pump speed but remember in my system I've got three Titan X's that are running in a parallel config which means each one of those cards is only getting 1/3 of the flow rate entering through this tube right here so it's going to divide between all three cards and then it's going to recollect as it leaves all three cards and go back to the pump so really my pumps being at max speed right here means that my cards are only getting a third of that max speed if that makes sense so the reason why the temperatures were so high when I had the pump set to zero or and zero actually being the lowest setting it doesn't mean off it's because one third of the lowest flow rate of this d5 with a 560 radiator and three blocks and another 280 radiator on the back and then of course all of the fittings and bends meant that the flow rate going through these three GPUs was very very low probably just a trickle level of low so what was happening is the fluid wasn't moving fast enough to actually cool off the graphics cards effectively or efficiently so when I said in the beginning was one of the arguments is if the pump is moving too slow then the coolant sits there too long and gets too hot it doesn't pull the heat out of a GPU fast enough is partially true now again remember this situation I'm talking about here specifically because of my 3-way parallel config now most of the time you're not going to see a huge prints in temperatures between lowest setting and highest setting in a single GPU and even a double or sli GPU configuration or crossfire depending what you're running because of the fact that you're not going to see as much of a flow drop-off again my loop here having multiple radiators and three GPUs is a lot of resistance for d5 to have to overcome so if I had only had one or two cards in here I would have expected the results to be a little bit more narrow and not as big of a difference there now one thing to keep in mind though is when you run water cooling pumps at their max speed they tend to get pretty noisy so there's going to be a trade-off there where you're going to really have to think about you know is the noise factor of running a pump at max speed D fives aren't too bad believe it or not and mines actually isolated with rubber mounts so it's not really transferring vibration to the case or the desk can't feel the case with my hand right next to it right here some pumps actually will create vibrations that just resonate and sound terrible DBC pumps are perfect example a DDC pump at max speed is incredibly loud so you're going to want to keep that in mind as well but if you guys are struggling with higher temps and you would like in your loop and you've got your pump running a lower speed try finding a sweet spot where the noise and the performance is going to help you but ultimately I wouldn't look at pump speed as being the kind of thing that is going to make a huge difference when it comes to the overall effectiveness of your loop that's one of the reasons why i don't run pwm pumps i like to have manual control over those and i tend to set it and forget it and i usually don't have temperature issues now if you're running a config with a multi speed pump or variable speed pump once you sound off in the comments and tell people what your results have been that way you can kind of get a collection of data down there so people can learn from not just my results but your results as well remember on setups like this and yours and other people's mileage is going to vary not your mileage may vary your mileage will vary anyway thanks for watching today's video here again this is not a conclusive like this can be applied to all systems you're going to have to play with yours you don't have to play with yours and see exactly what feels good if you know what I'm saying but you're going to have to determine for yourself what's best for your system again ambient temperatures or in fact airflow in the room was a factor as the room heats up the loops going to get hotter it doesn't just stay it's going to be dependent on the air cooling the radiators a lot of people forget room ventilation is as important as you know your case ventilation anyway time to go guys thanks for watching today's video I hope you've enjoyed hope you've learned something and as always I will see you in the next onewhat's going on YouTube J's two cents here one point out it's J's two cents not J two cent or JZ two cents it's just James two cents I mean why is that so hard but we are fast approaching the warm months of summer and some climates get super freakin hot and that's when I notice a lot of questions come my way about very specific things when it comes to water cooling so this summer we're going to try and touch on all of that we're going to rehash some old topics and we're going to do something that I've never actually done on this channel we are going to talk about pump speed and does it matter today we're just going to perform a test here on skunk works and we're gonna see what happens in my scenario when it comes to pump speed I thought it'd be fun but before we start playing with our pumps let's go ahead and talk about today's sponsor and that is squarespace.com / J's two cents now last year I use Squarespace is online e-commerce tools to sell some custom writing jerseys and if you guys bought one kudos to you if you didn't sorry they were a limited run but I had no idea where to start and how to actually run any sort of an e-commerce site and Squarespace made it freakishly easy with online inventory tracking order tracking you can even export to spreadsheets and keep track of everything you need to know and even get inventory alerts when things start to get low and sell out so if you're interested in getting your own website then head on over to squarespace.com slash Jays two cents and use offer code J's two cents to get ten percent off your order and if you sign up for a year you get a domain absolutely free now inevitably what happens every time we talk about radiators water pumps and water cooling and computers is people want to compare it to things like automotive cooling when it comes to radiators and thermostats and water pumps and the thing is you can't directly compare that you're talking about very low temperatures here very low pressures whereas in a car you're dealing with huge delta temperatures between ambient temp and the coolant temp reaching you know over two hundred and twenty degrees fahrenheit in many conditions obviously our coolant here isn't getting anywhere near that hot so you really can't compare the two it just isn't directly comparable so inevitably happens is you you start to hear things on forms along the lines of well if you run your pump too slow the coolant is going to move too slow through the loop and the coolant is going to touch the block too long and it's going to get really hot so if you really about that the flipside would be so the coolant is going to stay in the radiator longer and cool off more so on the flip side though the other side of the argument is well if your pumps moving too fast then the coolant is moving through the radiator too quickly and it's not going to cool down fast enough now pretty much both of those scenarios are true in the sense of when they talk about the timing of fluid flow where yes it's going to sit in the blocks a little bit longer it's going to hopefully pick up some heat but it's going to sit in the radiator longer as well if you run it really fast it's going to make more passes per minutes a crummy Oh what if you were to say be able to track this one molecule right inside the loop in time how many times it went around the entire loop per minute yeah there would be more times per minute which is going to mean it's going to run through the radiator more and in the end of the day running it slower or running it faster tends to give you about the same result and that's what I'm going to hopefully show you guys here today hopefully physics aren't going to change suddenly and I'm going to be making out to look like a liar and even if that's true it will stay in the video because this again is all about facts here I like to deal with facts they're they're fun to deal with actually you can't really change those so here's what we're going to do right now I'm going to load up Heaven benchmark on my 3-way Titan X water loop which gets really hot it's running a d5 it's independent from the CPU so there's actually no contamination of temperature between anything happening with the CPU and the V RMS and the GPU so this actually can be a pretty control to test here now it's three tight nexus with a 560 radiator and a 280 radiator running a 250 millimeter reservoir and then again as I said a variable speed d5 that's a manual control not a PWM so I'm actually going to control it I'm going to do a control here where I let the test run for about 10 minutes in heaven let the heat get but nice and built up in the system and then we're going to log what the max temperature was and then I'm going to let the system completely idle for about a half hour I'm going to cool off the room open the windows let the hot air get out of here and start all over from the same ambient temp and then we're going to pump the move the speed of the pump to a hundred percent and we're going to compare the results and see exactly what happened so I'm gonna go ahead and do that now and then we'll come back we'll talk about the results well for me it'll be a while but for you it'll be like right now just so you know what you guys are looking at this is what we care about right here these are the three tight necks as you can see 35 32 and 35 they're running a bit warmer at idle because I do have the three panels hooked up so they do go up a little bit higher on their base clock as you can see right there encore clock they're running at a thousand megahertz right now so they're not ramped down to idle that's why the temperature a little warmer so we don't really care about the minimum we care about the maximum which is right now 35 C so I'm gonna go ahead and run heav'n benchmark and then we're going to see like I said after the system gets full fully heated up and the settings on here we're going to push it we've got it running a 34 40 by 1440p panel tessellation I'm going to go ahead and put that on extreme and then anti-aliasing I'm leading at 2x before I start the test one other thing I'm going to do here is I'm going to take fan speed out of the equation I don't want it speeding up and slowing down so I've set my fans static to 70% inside a speed fan so that's our constant right now the only thing we want changing is going to be the actual speed of the pump we don't want the load to change on the software and we don't want the fan speed to change because that can obviously impact our results so here's the first result right here 63c was the hottest GPU temp as you can see the loads were pretty consistent temperatures had stabilized they weren't still going up so that was our max temperature 63 which might lead you to go wow that's really hot for water-cooled card yeah it actually really is and obviously the pump is running really really slow so what we're going to do now is I'm going to go ahead let the system completely cool off as you can see it's taken a little while to come down water-cooled systems take time for the temperature to drop we're going to pump that we're going to rent that pump up to 100% and then we're going to see exactly what we get after that alright so there's the new result 56c that came down from what 63 so that's 7c difference by just changing the pump speed yeah obviously there's a reason why that happens so let's go ahead and talk about that because that's the thing a lot of people leave out is there's a lot of variables that lead to these temperature fluctuations all right so let's talk here about what just happened some people might look at this conclusion and go oh that's it that's a set conclusion across the board higher pump speed equals lower temp done ship it doesn't really work that way guys you have to understand in my system it's a bit of an extreme situation here where I am trying to get as far on one end of the spectrum as I can and as far as I can on the can on the other end of the spectrum to see what the measurable difference is and then and between that white of a gap I only saw seven differences I'm not going to say seven is a bad number seven is actually a very very good number when it comes to temperature difference between just changing the pump speed but remember in my system I've got three Titan X's that are running in a parallel config which means each one of those cards is only getting 1/3 of the flow rate entering through this tube right here so it's going to divide between all three cards and then it's going to recollect as it leaves all three cards and go back to the pump so really my pumps being at max speed right here means that my cards are only getting a third of that max speed if that makes sense so the reason why the temperatures were so high when I had the pump set to zero or and zero actually being the lowest setting it doesn't mean off it's because one third of the lowest flow rate of this d5 with a 560 radiator and three blocks and another 280 radiator on the back and then of course all of the fittings and bends meant that the flow rate going through these three GPUs was very very low probably just a trickle level of low so what was happening is the fluid wasn't moving fast enough to actually cool off the graphics cards effectively or efficiently so when I said in the beginning was one of the arguments is if the pump is moving too slow then the coolant sits there too long and gets too hot it doesn't pull the heat out of a GPU fast enough is partially true now again remember this situation I'm talking about here specifically because of my 3-way parallel config now most of the time you're not going to see a huge prints in temperatures between lowest setting and highest setting in a single GPU and even a double or sli GPU configuration or crossfire depending what you're running because of the fact that you're not going to see as much of a flow drop-off again my loop here having multiple radiators and three GPUs is a lot of resistance for d5 to have to overcome so if I had only had one or two cards in here I would have expected the results to be a little bit more narrow and not as big of a difference there now one thing to keep in mind though is when you run water cooling pumps at their max speed they tend to get pretty noisy so there's going to be a trade-off there where you're going to really have to think about you know is the noise factor of running a pump at max speed D fives aren't too bad believe it or not and mines actually isolated with rubber mounts so it's not really transferring vibration to the case or the desk can't feel the case with my hand right next to it right here some pumps actually will create vibrations that just resonate and sound terrible DBC pumps are perfect example a DDC pump at max speed is incredibly loud so you're going to want to keep that in mind as well but if you guys are struggling with higher temps and you would like in your loop and you've got your pump running a lower speed try finding a sweet spot where the noise and the performance is going to help you but ultimately I wouldn't look at pump speed as being the kind of thing that is going to make a huge difference when it comes to the overall effectiveness of your loop that's one of the reasons why i don't run pwm pumps i like to have manual control over those and i tend to set it and forget it and i usually don't have temperature issues now if you're running a config with a multi speed pump or variable speed pump once you sound off in the comments and tell people what your results have been that way you can kind of get a collection of data down there so people can learn from not just my results but your results as well remember on setups like this and yours and other people's mileage is going to vary not your mileage may vary your mileage will vary anyway thanks for watching today's video here again this is not a conclusive like this can be applied to all systems you're going to have to play with yours you don't have to play with yours and see exactly what feels good if you know what I'm saying but you're going to have to determine for yourself what's best for your system again ambient temperatures or in fact airflow in the room was a factor as the room heats up the loops going to get hotter it doesn't just stay it's going to be dependent on the air cooling the radiators a lot of people forget room ventilation is as important as you know your case ventilation anyway time to go guys thanks for watching today's video I hope you've enjoyed hope you've learned something and as always I will see you in the next one\n"