|
|
Post by stormchaser on Jul 20, 2020 14:05:19 GMT -5
Still waiting on the pump, in line filter and temp sensor with LCD screen.
The LCD will most likely be mounted in one of the spare 5.25" drive bays, if I like the look of it.
So as much as I want to build this thing right now, I have to be patient and wait for these parts before I can begin.
For now I am working on drilling new mounting holes to seat the radiator so it sits flush with the case. So I am definitely moving the second, pulling fan, to the outside of the case.
|
|
|
|
Post by Vinster on Jul 20, 2020 14:50:42 GMT -5
No you want the air in the tubes so it goes to your reservoir and bleeds out. Otherwise you get the air trapped in the rad and air-in-rad = less heat transfer and worse performance. Its only the other way around for AiO units as they have no reservoir and the air would go into the pump.
Ya, but I'm sure we've all ran our PC's on it's head to let all the air out of the crevices... lol. and Yep, AIO's need their barbs down. I cringe when I see them installed up.. and even big youtube guys have done this...
With the barbs down would make flushing the loop easier. A 'T' off the RAD outlet and away you go. If it were me, that's what I would do.
it's easier to fill the reservoir, close up the loop and flip the case to get the air out than having the barbs up and dealing with liquid in the RAD/loop trying to get it out without spilling on the board/HW.
but that's just me.
Vin
|
|
|
|
Post by stormchaser on Jul 20, 2020 16:48:58 GMT -5
|
|
|
|
Post by stormchaser on Jul 20, 2020 16:53:00 GMT -5
If I do run into thermal problems, I am toying with the notion of adding a very small radiator to suppliment the loop. Probably 80mm x80mm would be ideal. I would also have to cut a hole in the bottom of the case and put feet on it for proper airflow. I really like the idea and perhaps I will do this just for kicks, even if the thick rad is able to hold temps to a reasonable level.
|
|
|
|
Post by MainFramR on Jul 20, 2020 18:41:25 GMT -5
That makes sense too. I'm used to doing a rad shake to get air out. Tipping the chassis and rotating it while jumping the loop. Never had an issue with performance with that method.
Seems easier to put them on top though if the air comes out that easily.
|
|
|
|
Post by osmiumoc on Jul 21, 2020 8:45:46 GMT -5
Also of note, I moved my bay reservoir to the very top of the case. So the level of water in the top bay reservoir will be at least a few inches above that of the pump/res combo. Is this going to cause any problems? Is there any performance impact when running two reservoirs? What's the best way to do this? As for the placement of the second res, I'd put it in a spot where its not blocking anything or making access to e.g. HDD/SSDs difficult. Having the reservoir as the highest point is actually the best spot for filling the top. Water can flow down into your loop while you fill, requiring less pump power-cycles to fill the whole system. There will be no issues with performance / flowrate as long as you keep it inside the case. Problematic would be a very high reservoir, like if you put it outside the case on a shelf 2 meters above the pump. Even then a D5 might still work fine. I have no experience with two reservoirs in the same loop. When filling the loop you will leave the fill-port open to bleed air out on the main res. The secondary has to be plugged so it will probably fill up completly with water? Depends on how the internal flow of that res is setup it will either fill up completly or it might trap some air. Either way its gonna be fine and no issue. Reservoirs add very small flow restriction. |
|
|
|
Post by stormchaser on Aug 2, 2020 2:30:08 GMT -5
|
|
|
|
Post by ShrimpBrime on Aug 2, 2020 12:44:30 GMT -5
That's a damn nice waterblock. Wish I had a couple  |
|
|
|
Post by MainFramR on Aug 5, 2020 9:21:16 GMT -5
I love the look and performance of those Heatkiller blocks. I haven't read up on them lately, but they were one of the best when the first/second version was released. I came across this browsing TPU this morning www.techpowerup.com/forums/threads/120mm-liquid-cooling-system-upgrade-for-the-9600kf-5ghz.270089/post-4322161The reason why your pump appears not to be working is: A. The molex isn't plugged in. As far as I know, the molex is for power and the 3/4-pin is for controlling the pump speed. Check the manual to confirm. B. From the looks of some of those pictures, you plumbed the pump wrong. The top of the pump should be coming from the res and is an intake to the pump (res before pump is always the rule so the res feeds the pump) while the side is the outlet and should be going to your blocks (not the res). Again, check the manual to confirm. But yes, assuming that I could interpret what I saw in the pictures, you need to replumb and plug in the pump properly. |
|
|
|
Post by stormchaser on Aug 6, 2020 16:25:48 GMT -5
I love the look and performance of those Heatkiller blocks. I haven't read up on them lately, but they were one of the best when the first/second version was released. I came across this browsing TPU this morning www.techpowerup.com/forums/threads/120mm-liquid-cooling-system-upgrade-for-the-9600kf-5ghz.270089/post-4322161The reason why your pump appears not to be working is: A. The molex isn't plugged in. As far as I know, the molex is for power and the 3/4-pin is for controlling the pump speed. Check the manual to confirm. B. From the looks of some of those pictures, you plumbed the pump wrong. The top of the pump should be coming from the res and is an intake to the pump (res before pump is always the rule so the res feeds the pump) while the side is the outlet and should be going to your blocks (not the res). Again, check the manual to confirm. But yes, assuming that I could interpret what I saw in the pictures, you need to replumb and plug in the pump properly. Okay I will change the order of the loop if you think that's necessary. other's have been saying the routing is okay? as far as I can tell? The loop order I selected is (again first custom loop so I appreciate your help): Pump ----> Reservoir ----> Reservoir ----> Rad----> Waterblock ----> Back to pumpSo you are saying it should be: Reservoir ----> Reservoir ----> Pump ----> waterblock -----> Rad ----> back to reserviorOkay, I was under the impression that loop order was not a significant factor in cooling performance? According to Jay on youtube, loop order does not matter? The reason I chose this particular combination is because (on that side) the pump is located higher in the case than most of the other components of the loop, so it requires less pressure to push water through that side of the loop as opposed dipping down and then all the way around the loop, and then pushing water up through the radiator as well. On the side I have configured, the water is gravity fed through rad because the rad sits lower than the high point of my 5.25" res, thus aiding the pump and allowing it to function with less resistance on the motor. That was my thinking, at least. I selected this order because it offered the least amount of resistance to push a volume of water through the loop, due to in my limited knowledge of water cooling system? I will be happy to change the order, however. More pictures inbound will show how the tower looks with both reservoirs filled looks from the front angle. I am very happy with the aesthetics of this project so far. Including the two matching reservoirs filled with coolant. Mistakenly getting the wrong reservoir was actually a blessing in disguise. Because the lower bay res is about half the length of the res in front of it, making it an ideal place to stick the pump, while still appearing from the outside to match the other bay res perfectly. Cant go any lower in the case with the pump because the massive 5700 XT is on the way.           |
|
|
|
Post by ShrimpBrime on Aug 6, 2020 16:39:24 GMT -5
If you move the pump to the out side of the cage, you can significantly shorten the waterblock supply hose and you won't get a kink from it. A trick to kind get the lines to form better is to run them under hot water. quickly dry with a towel and install. Once it cools to room temp, it'll take memory of the shape and won't give you any fuss.
Looks good though! Something like this awaits my bucket list.
|
|
|
|
Post by stormchaser on Aug 6, 2020 17:00:47 GMT -5
The navy blue coolant is great very happy with the color. Really brings out the best in this build, in addition to matching the blue writing on both bay reservoirs. When I go out in a couple weeks I will be purchasing a tiny 75 x 75 mm fan, and adding it to this existing loop, also manufactured by Alphacool (will be mounted just in front of the 120mm fan in the lower right (front) corner of the case). So at that point it will have two matching reservoirs and two matching radiators, all made by Alphacool (+ 5700 XT Alphacool water block). Aesthetically, it should be very pleasing on the eyes. The little rad is going to be added to the loop as I need/want something slightly more robust than a single 120mm thick radiator for both a 5 GHz six core CPU and a 210w GPU. Love the navy blue color and ease of access to get to the fill port on the res.  The case is of a tool less design that allows the end user to very easily swap hardware in the 5.25 bays. Very helpful for this build as the res has come out about a hundred times up to this point, and also makes it much easier to fill (I simply lift the rear of the case up about 2", that way when I'm filling it's impossible for coolant to go in the direction of my motherboard.  I fill the reservoir from the threaded insert on the rad, as this picture shows:  Thinking of cutting a hole in the back of the case to exhaust the soon to be ordered 75mm radiator. This way it would not expel superheated air back into the case.       Mad max theme, lol just for kicks. I actually think it looks really good but I will most likely be painting over it. Both fans are noctua and the bottom fan has an insane 3.6w power draw. Also goes up to 3000 rpm via PWM. So If I want to dump fresh air on the CPU and memory, these two fans should be up to the task, plus the other 120mm noctua in the lower right front of the case.     |
|
|
|
Post by MainFramR on Aug 6, 2020 19:48:56 GMT -5
Loop order doesn't matter, right, so long as the res is before the pump. After that it doesnt matter (i recall this from skinee/Martin's labs guide among others). Surely the group here remembers that site! Temperatures inside a properly configured, enough rad and flow, loop don't vary much at all. A couple of degrees Celsius at most if I remember correctly, so yes. Loop order doesnt matter after the res->pump for cooling properties. Pressure is also not terribly relevant. It's a closed loop so for all intents and purposes, it's the same. Here is what ekwb has to say about things. www.ekwb.com/blog/does-loop-order-matter/So, yes, replumb it properly and get the power situation corrected and you should be good to go. The loop is looking good so far! But does this forum have spoiler tags? My finger hurts from scrolling through so many images to get to the words! Haha! |
|
|
|
Post by stormchaser on Aug 7, 2020 1:52:51 GMT -5
Loop order doesn't matter, right, so long as the res is before the pump. After that it doesnt matter (i recall this from skinee/Martin's labs guide among others). Surely the group here remembers that site! Temperatures inside a properly configured, enough rad and flow, loop don't vary much at all. A couple of degrees Celsius at most if I remember correctly, so yes. Loop order doesnt matter after the res->pump for cooling properties. Pressure is also not terribly relevant. It's a closed loop so for all intents and purposes, it's the same. Here is what ekwb has to say about things. www.ekwb.com/blog/does-loop-order-matter/So, yes, replumb it properly and get the power situation corrected and you should be good to go. The loop is looking good so far! But does this forum have spoiler tags? My finger hurts from scrolling through so many images to get to the words! Haha! Forgive me for writing a book here. I have too much free time on my hands and I'd like to stay productive. The supposition your are putting forward can easily be refuted by taking a look at current tech and current prebuild liquid cooling systems (AIOs). The elephant in the room, the fact that you can have high performance liquid cooling system without a reservoir at all, has now turned into a "hot air balloon" which is about get popped, for good reason. I mean just look at this AIO! and it's not even using a reservoir!? how "crucial" reservoir placement in the loop really is. lol. And remember, once you prime your custom loop, there is no difference between it and an AIO. You flip the computer upside down, they are still closed systems at the end of the day. My point? Reservoir placement should be considered for aesthetics as well or even favored over performance implications. Because there are none. lol  I get your point, but this borders on misconception. Many people out in the wild have run this same loop/configuration and its as good as any other combo performance wise. Besides, with my reservoir up top, at the highest point, filling the loop is a breeze due to tool free locking mechanism, I can easily slide out the top reservoir for a refill or inspection. With the res being the lowest point in rotation, it doubles as an excellent method for draining the system when the time comes. So the top reservoir is positioned exactly where it needs to be, as well as the pump being positioned exactly where I want it. I have no concerns with placing the res first. Because I built this loop based on a straightforward concept. This is my theory so take with a grain of salt. You place the pump at the lowest part of the loop and this will assure some level of protection against running dry because that area will be the last part of the loop to drain, affording you a level of protection over the standard margin. You also have the two massive reservoirs situated directly above pump, pressing down, so we can apply the simple concept of gravity here. An ideal loop is one not measured by ordering of equipment within it, but by altitude of equipment relative to the confines of the case and other hardware combined with the aspirations and dreams of the owner or creator. A huge mass of computers builders prefer aesthetics over performance. doesn't always have to be maximum performance (although in this case, it is all about performance) I didn't get a HEATKILLER CPU cooler for nothin!Also note, if you look closely, I do not have enough room in the 5.25 area to rotate the pump vertically (as in a straight shot from res to pump)... because it's too tall to fit in with the bay res above, so it must be laid down on its side. So this entire discussion is a moot point but I wanted to break it down a little further... In other words, due to the confines of the case, your recommendation simply doesn't work in this build. In order to straighten it out, I would have to relocate the pump away from the 5.25" drive cage and find another area within the case to use. and add another 90* fitting... And with the massive RX 5700 XT GPU in place, there is not a whole lot of room to work with. there would be no "straight shot" and even if I could make that work by pitching it over, the aesthetics would be horrific and I would be using two 90* fittings. Right now, I only have one in the entire loop and I'd like to keep it that way. Also - perhaps this was overlooked, but the HEATKILLER water block is essentially the lowest part of the loop, with well over 1/2 gallon of coolant sitting directly above it in the form of two alphacool bay reservoirs and large capacity 120mm radiator that is 80mm thick... So barring a catastrophic failure and meltdown of the loop, the pump will always have pressure from a substantial amount of water upon both sides, intake and exit. Running the loop your way may be more convenient, but, its not a silver bullet. it simply aids in loop priming, but beyond that, offers no real value or improvement over my combination. Matter of fact I would argue my arrangement is used quite often and with the large bay reservoirs I will likely have a slight edge for better dealing with load/heat spikes. A larger volume cooling system is going to be able to absorb more punishment to some extent. I could see you making a point that if I was running toothpaste as thermal grease then you could recommend I apply a new layer of some MX4 or something, and based on thermals, make a change. Except your suggestion here regarding loop order does not improve performance and offer a very limited reward, and worst of all not based on improvements of thermals or any performance or reliability change. as you can see, there is not much of an incentive for me to execute your recommendations at this time. That leaves us with the matter of aesthetics and how the creator, me, wants to have his rig set up. And that's exactly what I've done here. I like the neat clean finish and the hidden pump (after all this is a sleeper build)… You gotta remember, in many, many watercooling situations, aesthetics actually favored over performance. Because the performance is already there. Here, I think I'll get three more MHz out of my chip and that will win me the competition if I only shuffle the order of my custom water cooling system. 3 MHz is not going to make or break you. Please dont take that as a personal attack, I'm just saying there are plenty of people out in the wild who have used the exact same loop configuration as me with good results. Also note, all you can do is offer advice and recommendations -- and I will hear you out. I respect you and your opinions and you must also respect the creator who has the final say in how the rig gets built, and that's me. I think that last part gets lost in the shuffle too often and people helping start to feel demeaned because they are attempting to take control of choices that should be better made by the guy who build the PC. That or they go into a tirade about how they 'know best'. Fact is it's not your project its mine. Not sure why this concept flies over the heads of some intelligent thinkers and excellent computer builders. Point is we have different color cars for a reason. Some people like, some people like black. I think you can figure it out from here. This is my blog for this build, hence the reason for all the pictures, and yes, I tend to go a little overboard on picture documentation.... it's essentially my blog where I can take my time and fill it with meaningful assessments (I typically create a new blog, and data and as a digital history or record for the computers progress throughout it's metamorphosis. You might recall, this rig was originally cooled by an MSI core Frozr 120 mm air cooler. At that point in time my goal was 5.0 Ghz across all six cores of my 9600KF. Many, many people said this wouldn't work, and that I would never be able to pull it off. Well, they all got proven wrong (this peak clocks.  I sincerely appreciate your advice, but I am actually going to leave the loop as it is. I actually prefer how it's set up right now. And also how it looks, the aesthetics are pretty much dialed in the way I want. Dont have time to read the book I wrote? Go down to the bottom line IN BOTH RESPECTS and that decision should be made by the owner/user/creator etc. I’ve got my mind made up: I am actually going to leave the loop as it is. With all components in the same order. I actually prefer how it's set up now. And also, how it looks, the aesthetics are pretty much dialed in the way I wanted. Regarding order of my loop components? I know that in most cases, your order of components (including the pump) within the loop “circuit” does not matter, except in rare instances, and for the life of me, I can’t see a reason why I would want to change my routing setup simply ‘because you said so’ for no performance or reliability gains. Please don’t take it personally if I ignore your advice, I am looking at all viable options from all members and weighing that information against my goals for this build. If temps can hold within the margin of error, despite the “musical” chairs element would likely take your advice straightaway to flip the order and go RESERVOIR TO PUMP to increase the safety margin. However, I dont have a D5 and my pump's motor is not cooled by water. That's not to say I want to run it dry, just that the system may be more forgiving to my specific loop and type of water pump. But even that is a little over the top, because the bottom line is the reservoir before the pump will make it slightly easier to prime the system (this is just my theory - not confirmed yet as I still waiting on a couple parts, nothing more. Once both systems have been successfully primed, there with be absolutely zero difference between the two, otherwise identical loops. So when people recommend putting the res right before the pump, that's good and all, but they should know that it's really only an aid to help you bleed the system a little bit quicker than other loops in a different order. In most cases, order of components in a loop does not matter (and that is applicable here to my loop as well), except in rare instances, or perhaps if you are running a D5 and are overly afraid or paranoid of running it dry, then I would likely take your advice straightaway to flip the order and go reservoir to pump to increase the safety margin. However, I dont have a D5 and my pump's motor is not cooled by water. That's not to say I want to run it dry, just that the system may be more forgiving to my specific loop and type of water pump. But even that is a little over the top, because having the reservoir before the pump will make it slightly easier to prime the system, nothing more. Once both systems have been successfully primed, there with be absolutely zero difference between the two, otherwise identical loops. So when people recommend putting the res right before the pump, that's good and all, but they should know that it's really only an aid to help you bleed the system a little bit quicker than other loops in a different order. Once you close the plug, they are both closed systems. Bottom line: In this case I valued aesthetics over convenience. |
|
|
|
Post by MainFramR on Aug 7, 2020 7:38:44 GMT -5
I see you wall of texts and raise you!! AIO rads have reservoirs in them/are reservoirs. I believe some even state as much on their webpages. I also did not mention performance implications, but it is simply there to prime the pump. Remember, I ran across your thread at another site where this loop was not working and you asked why. Outside of the obvious of plugging the pump in the right way, the rest after the order didn't make sense according to all the guides over the years and the ekwb link I sent. Will it work? Likely. Is it optimal, seems like not. But how will that affect it? No idea. I haven't seen someone do this before. Not at XS, Raged3d, TPU, OCN or other sites. Nope. Not at all. You insist many have done this, but again, I haven't run into it nor any testing. I'm sure they do. It looks like you are taking the performance route over aesthetics as well. Function over form FTW! As far as the pump and pressure, it's a closed loop, gravity has nothing to do with it. Your res's are NOT pressing down on anything, they are getting water pumped into them and are of zero help. Let's humor your theory for a minute, wouldn't all that water in the res press AGAINST the flow since it is sitting right at the end of your outlet? How does that work? Again, not saying this is true, just some food for thought. There are plenty of places that pump could have been located to make it happen. It can work in this build. Think outside the box both literally (what shrimpbrime said) or figuratively. Why couldn't it have went on the floor, for example? Why dual res? You took another theory of yours (more water = more cooling capability) and had that work against you. The cooling will be the same once that water reaches an equilibrium. It just takes longer to reach that point. So you could have a couple of more bays to play with while still achieving the same performance. I only stated that it is there for priming the pump. ekwb says it affects flow rates. How much exactly? No idea. Repeating myself, I just saw your system not working and you asked for help. You can argue that point all you want, I've never seen it and been around several forums for nearly 20 years doing this. Surely people have done it, I can't see every build on the net, but commonly held advice over the years from those in the know is res first. You will not have an edge over heat spikes with more water. The issues with CPUs and especially the latest Intel like you have is the tiny die and the density of transistors in that die. It would literally be no different than having cooler ambient temps for an air cooler. You'll still get spikes. Again, I did not say anything about performance. That was you (and ekwb said it?). lol. I assure you, I am not taking what you are saying as a personal attack. This is your first loop and I simply tried to share advice to someone who asked for it, at TPU. This is about the 3rd or fourth time in this wall of text you've mentioned so many others do it this way, but, feels a lot less common than you think it is. All the advice from incredibly reputable websites including Martin's Liquid Lab who was THE watercooling authority at the time. Can't say I feel demeaned at all. Your talking points are your theory. I gave sound advice based on facts and studies. I couldn't care less if you follow it or not. You asked for help, got it, and responded with your personal 'theories' and straw man arguments (performance that I never mentioned). Don't take this personally, but that is your problem. I'll simply walk away no different than when I started. I saw your thread and a couple people had reservations, sure. You really like to go on the defensive quite easily (see wall of text above). Relax a bit man. Nobody is after you in the forums and we are all trying to help. Congrats on your CPUz screenshot. I agree and never said otherwise. Advice was given, it is your choice to follow it or not. Enough of the martyrdom. Jesus man, it was just advice and advice backed up by links. The choice was always yours. How many times does this need to be said (as many times as it is mentioned, I guess!). It's for helping to prime the pump (D5 or not, pumps die QUICKLY if they are run dry, trust me). The res HIGHER than the pump can help bleed as the air bubbles will naturally go there. Perhaps it helps with bleeding too. I leave the top of mine open at the start so I can fill the res and not let the pump dry. My res is the highest point and the drain at the bottom. One man's trash is another man's treasure! Your build isn't trash, for the record. But I just found it curious that a hacked up CM 690(?) case and enough tubing in the case to make a noose is valuing aesthetics. I believe you mentioned performance was the goal earlier and it was a sleeper. I agree with that last two, for sure, and how a sleeper is defined (doesn't look good, performs well). This is a great sleeper build! In the end, of course, do what you want! I'm not bothered at all. Now, get your pump power connected right and enjoy!! |
|
|
|
Post by Mr.Scott on Aug 7, 2020 15:28:08 GMT -5
Res before pump always.
If for no other reason than guaranteeing your pump never running dry.
No other order matters, as long as you have this.
Joe is 100% correct. Nearly 2 decades of WC experience between us tells us this.
I can link posts and sites to back this, but I see it's fairly pointless so...........
|
|
|
|
Post by Vinster on Aug 7, 2020 21:42:02 GMT -5
I typically go Res > Pump > Block > Rad then Back to Res. You want to cool the fluid before you store it again to re-use, if you don't it'll raise the fluid delta from ambient (from my experience, but I'm talking a degree or two). if you have temp issues and want to add another rad, add a rad between the pump and block. So then you'd be Res > Pump > RAD1 > Block > RAD2 then Back to Res
play around with it to get your feel. but this is why there are reservoirs with pump mounts built in. You can never starve a pump unless the fluid is low.
issue with starving a pump is, the fluid cools the pump, and depending on the pump, the fluid is also the lubricant for the pump bearing. Starve the fluid, you will cause air bubbles in the loop and rad and also cause the pump to fail and possibly take out other components (CPU or GPU if you burn them up)
Vin
|
|
|
|
Post by stormchaser on Aug 8, 2020 0:34:11 GMT -5
I see you wall of texts and raise you!! AIO rads have reservoirs in them/are reservoirs. I believe some even state as much on their webpages. I also did not mention performance implications, but it is simply there to prime the pump. Remember, I ran across your thread at another site where this loop was not working and you asked why. Outside of the obvious of plugging the pump in the right way, the rest after the order didn't make sense according to all the guides over the years and the ekwb link I sent. Will it work? Likely. Is it optimal, seems like not. But how will that affect it? No idea. I haven't seen someone do this before. Not at XS, Raged3d, TPU, OCN or other sites. Nope. Not at all. You insist many have done this, but again, I haven't run into it nor any testing. I'm sure they do. It looks like you are taking the performance route over aesthetics as well. Function over form FTW! As far as the pump and pressure, it's a closed loop, gravity has nothing to do with it. Your res's are NOT pressing down on anything, they are getting water pumped into them and are of zero help. Let's humor your theory for a minute, wouldn't all that water in the res press AGAINST the flow since it is sitting right at the end of your outlet? How does that work? Again, not saying this is true, just some food for thought. There are plenty of places that pump could have been located to make it happen. It can work in this build. Think outside the box both literally (what shrimpbrime said) or figuratively. Why couldn't it have went on the floor, for example? Why dual res? You took another theory of yours (more water = more cooling capability) and had that work against you. The cooling will be the same once that water reaches an equilibrium. It just takes longer to reach that point. So you could have a couple of more bays to play with while still achieving the same performance. I only stated that it is there for priming the pump. ekwb says it affects flow rates. How much exactly? No idea. Repeating myself, I just saw your system not working and you asked for help. You can argue that point all you want, I've never seen it and been around several forums for nearly 20 years doing this. Surely people have done it, I can't see every build on the net, but commonly held advice over the years from those in the know is res first. You will not have an edge over heat spikes with more water. The issues with CPUs and especially the latest Intel like you have is the tiny die and the density of transistors in that die. It would literally be no different than having cooler ambient temps for an air cooler. You'll still get spikes. Again, I did not say anything about performance. That was you (and ekwb said it?). lol. I assure you, I am not taking what you are saying as a personal attack. This is your first loop and I simply tried to share advice to someone who asked for it, at TPU. This is about the 3rd or fourth time in this wall of text you've mentioned so many others do it this way, but, feels a lot less common than you think it is. All the advice from incredibly reputable websites including Martin's Liquid Lab who was THE watercooling authority at the time. Can't say I feel demeaned at all. Your talking points are your theory. I gave sound advice based on facts and studies. I couldn't care less if you follow it or not. You asked for help, got it, and responded with your personal 'theories' and straw man arguments (performance that I never mentioned). Don't take this personally, but that is your problem. I'll simply walk away no different than when I started. I saw your thread and a couple people had reservations, sure. You really like to go on the defensive quite easily (see wall of text above). Relax a bit man. Nobody is after you in the forums and we are all trying to help. Congrats on your CPUz screenshot. I agree and never said otherwise. Advice was given, it is your choice to follow it or not. Enough of the martyrdom. Jesus man, it was just advice and advice backed up by links. The choice was always yours. How many times does this need to be said (as many times as it is mentioned, I guess!). It's for helping to prime the pump (D5 or not, pumps die QUICKLY if they are run dry, trust me). The res HIGHER than the pump can help bleed as the air bubbles will naturally go there. Perhaps it helps with bleeding too. I leave the top of mine open at the start so I can fill the res and not let the pump dry. My res is the highest point and the drain at the bottom. One man's trash is another man's treasure! Your build isn't trash, for the record. But I just found it curious that a hacked up CM 690(?) case and enough tubing in the case to make a noose is valuing aesthetics. I believe you mentioned performance was the goal earlier and it was a sleeper. I agree with that last two, for sure, and how a sleeper is defined (doesn't look good, performs well). This is a great sleeper build! In the end, of course, do what you want! I'm not bothered at all. Now, get your pump power connected right and enjoy!! Yes, I brought up the performance aspect on my own because your recommendation did nothing to improve performance. Perhaps that's partly the reason I wont make the switch? Get it now? You basically just admitted that the primary function of a reservoir is to assist in priming the pump, so why are you trying to push this angle that AIOs have them? Do you just not want to admit that in most watercooling systems, the res is chosen for aesthetics and not performance? . You've gone so far as to try to make me believe reservoirs are hiding inside of radiators ... this is news to me lol. And definitely, if you can get some evidence from product documentation, or the manufacturer themselves that the rad doubles as reservoir in their particular AIO. I'd be very interested to see it. Of course you will be searching for the next couple hundred years, because AIOs DO NOT have reservoirs... lol So before the fallout continues, I will ask one time... would you like a Mulligan? If not, you better hope at least one manufacturer talks about their AIO having a reservoir inside the radiator... integrated in to one unit... the actual radiator itself. Because that's what you are wanting me to believe. Yeah it is my build and I will have final say, of course. So please dont be shocked or saddened if I don't take your advice... or take it as a personal attack. But this is nowhere near the first time something like this has happened so I cant say I'm surprised by whats happened here or your misguided perception of the project as a whole. The "cooling gods" have spoken! Almost always, it's like "we know better" Might want to look up the definition of "reserve" and see how that correlates to our discussion here about reservoirs. Yes, function over form + sleeper case is what it's all about. But contrary to popular belief aesthetics do in fact play a large role in creating the perfect sleeper. Like they say, beauty is in the eye of the beholder. This is another issue some of the experts have regarding advice - they seem to think if they dont get through with their recommendations they can always knock the build. Would be the first time! lol. I don't care what you think of the build , that's why its MY build and not YOUR build. And I sure as hell didn't come here to enter a PC beauty contest! Again I am not sure if you are trolling here or what, but do you honestly believe that a closed loop such as this is impervious to the law of gravity? Please explain. Okay, so tell me where I should put the reservoir, giving the fact that I am running a huge GPU that sticks out well beyond that area. You must also keep in mind, I am keeping that GPU area free of clutter for the time being because in the not so distant future I will be adding another 75mm baby alphacool to assist the single 120. All other things being equal, I am hoping to hit 5.2 GHz on this loop. Yes, I could brainstorm for hours as to where to put the pump. Or, here is a novel approach, I can put it where I want it and watch you try to come to grips with that. lol According to Jay, the computer guy has addressed this issue multiple times in multiple videos and he is adamant against any doubt: Placement of components within the loop does not effect overall performance of said loop. Or if it does, it's in like a 1-2% margin of error. Perhaps follow-up with Jay and he can explain to you why it's not mandatory? And I know you are only trying to help, it's just that I've already hammered this out (at least the loop portion) and I've already got the soft tubing in place and I like where the pump goes. But if you have a water cooled rig, I might suggest you find a place to mount the pump OUTSIDE the box and report back to us with results on how that's working out for you. RE: Performance Right, again, it was me who brought up performance. Your recommendation seemed to make it clear that the holy grail of water loops is when the pump is fed by the res and everything else is just second best. In other words, if your suggestion had actually done something to improve loop performance, I might have given it the time of day. But it doesn't. So I'm not. I would prefer not to have my pump located elsewhere, on the floor, or on the ceiling. You keep pushing this angle, even though you know its just hot air. Because you've already admitted the function of a radiator is to assist in priming the pump, primarily. Besides, I am running very low on tubing. If I had to make some changes and cuts, I would be needing more clear tubing and that's just not happening. You are contradicting yourself again. You've already mentioned the major purpose of reservoir is merely to help prime the pump. And now you are saying it is of the utmost importance to use the specific order you recommend... why? If all it does is improve convenience for bleeding air out of the system, then you haven't listed a valid reason for why I should prescribe to your order... It's not like I have to prime the pump every day, once I set I can forget it. Again, this is all very typical stuff I'd expect to see from the overly ambitious types who would prefer to build the rig for you. That's not advice that's something we call pride. My way is better, and if you don't follow it you might lose access to 3Mhz of your overclock ceiling. We will all be waiting with baited breath for you to come back with an AIO that has a reservoir built into the rad. It might just be a simple case of confusing the purpose of a radiator end tank. Which has nothing to do with your reservoir. No, it was more than just a couple people. And beyond that, they thought 5.0GHz was impossible due to thermals. They seemed dead set in their ways and were only snapped back to reality after I posted a 30 minute torture test and passed with flying colors. Wonder who else is going to get snapped back to reality? You wont have to wait very long to find out my friend  Also note, you are explaining that my pump is hooked up wrong and I need to fix the connection. Agreed, but did you honestly think I was going to leave it that way? lol It's the same idea with "gifted" people who criticize the builder for having poor cable management before the build is complete. It's not productive and doesn't serve to further the conversation. Perhaps you should reflect on that? Like I said, call the build trashy all you want. But at the end of the day, it packs a serious punch. You might be just as stunned as those guys from TPU that said 5.0Ghz would never be possible. Might not want to wait and find out, right? lol In any event I think we've both beat the dead horse a number of times back and forth on this. Lets move on. I will have some time to wire the pump tomorrow and hopefully everything kicks over without a hitch. But I do need to RMA the board. This MSI board has been lackluster in terms of quality. First, my primary pcie slot stopped working, then the tiny little code box to show you trouble codes just detached from the rest of it. At the end of the day, you know it's pride when another member has great difficulty accepting your intended vision for the machine. So yeah, its move on, and tomorrow it is my hope to show you a 5.1GHz screenshot. |
|
|
|
Post by stormchaser on Aug 8, 2020 0:48:23 GMT -5
Res before pump always. If for no other reason than guaranteeing your pump never running dry. No other order matters, as long as you have this. Joe is 100% correct. Nearly 2 decades of WC experience between us tells us this. I can link posts and sites to back this, but I see it's fairly pointless so........... So because most AIOs don't have a reservoir, surely they are at a much greater risk of a pump running dry scenario? That's not good! That's why the bay reservoirs I selected have a clear window and gauge to monitor current coolant level. I don't need to play musical chairs with my loop components to figure that out, super simple through one glance at the tower. So Jay must be wrong when he says the order is irrelevant, right? Im not saying he is the holy grail of cooling, but I've heard it from others as well. Wonder how many decades he has in the computer industry? It's not always this cut and dry, guys. Just dont be surprised when my original plan works exactly the way I intended it too, just like with the first iteration of the build and the air cooler. Like I said, lets move on, I will show you my best effort OC tomorrow all things permitting. |
|
|
|
Post by Vinster on Aug 8, 2020 0:51:52 GMT -5
AIO's do have a res. but they are like 30-60ml.. the connection rad bulkhead on an AIO is a touch larger than a normal/convectional rad. Could be marketing, but the Coolermaster H50, H70 and H100i that I have and don't use, where the hosed connect to the rad, that area is larger than my custom cooling rads.
Vin
|
|
|
|
Post by Vinster on Aug 8, 2020 0:54:01 GMT -5
Also , AIO's have next to zero air in them. I believe they are near vacuum. so there is no way you could starve or cavitate the pump because there is no air in the manufactured loop.
Vin
|
|
|
|
Post by stormchaser on Aug 8, 2020 1:42:31 GMT -5
AIO's do have a res. but they are like 30-60ml.. the connection rad bulkhead on an AIO is a touch larger than a normal/convectional rad. Could be marketing, but the Coolermaster H50, H70 and H100i that I have and don't use, where the hosed connect to the rad, that area is larger than my custom cooling rads. Vin Yeah I've noticed larger plenums on my alphacool, which is not an AIO radiator, and built for use in a custom loop, so it goes both ways and is nothing specific to AIOs.  In the grand scheme of things, an extra 30 mL of coolant isn't going to do all that much for you or your loop. Or your safety margin. I will admit that my setup, with res not feeding directly into pump does have it's disadvantages, and in a perfect world, I would want to stand the reservoir up so there is a column of water pressing directly on the intake port of the pump as mr scott and mainframe have said. However, if the system is bled correctly, there should be no major safety concerns doing it this way. Because at that point, assuming you have successfully removed the air from the system, it's essentially operating identical to how an AIO is filled and capped and utilized. I initially intended to put the pump in the bottom front of the case, at the lowest point in the loop. But the idea of installing a mini alphacool 75mm down there was too much to say no too, so I had to find another place for the res. The second bay res originally had provisions for up to two D5 pumps off the back of it. So I took them off, disassembled the res and used a block off plate to convert it too res only. This had a silver lining in effect, because it allowed me to mount the external loop pump in the drive bay below/behind it. It is impossible to stand the pump up for a direct shot from res to pump, so it was the only concievable way of mounting the pump cleanly and not have it interfere with the GPU or extra 75mm rad I will be adding later. |
|
|
|
Post by MainFramR on Aug 8, 2020 2:03:47 GMT -5
I feel like I need to reiterate the point that I'm not trying to force you to do anything to your loop. All that is going on is arming you with facts so you can make whatever choices you want based on the theories you are conjuring up from 'jay'. If I am guilty of anything that would be my insistence on sharing the info. I'll go through this exercise (in futility) and respond one more time. Why does gravity not matter in a loop? Because it's a loop. What goes down, comes back up! This Jay vid? Just watched it. What a waste of time. Glad I'm a couple drinks in on some Woodford Reserve. That video only supports what I was saying, you are aware of that, right? The point of contention is only with the res before the pump. Loop order outside of that mantra, doesn't matter. You can go swap any rad/gpu/cpu/rad in any order and not have much difference. No shit. I was abundantly clear that was my stance. This vid did nothing for that point at all. You really seem to have trouble hanging on to this train.  Again, stick with me. I said you could move the pump, not the res. Looks like it will fit in the front on the bottom where there isn't any soundproofing. Set the pump vertically, intake up to the res (or two in your case), then the output goes to the CPU or rad or whatever because loop order doesn't matter (so long as the res is before the pump). That's how I would do it. Or, again, don't go dual res (since there is little reason to have them) one res up top, pump where it is but vertically. Remember, nobody is forcing your hand. I'm only saying it is possible as you believed it wasn't. I mount a MCP655 (D5 vario) to the outside of an Antec 900. There are grommets on that case to route tubing in and out. Clearly not ideal, but, did it myself. Yes. I do. 100%. Absolutely. Did you just pull that guys chain (fry?) in your reply and your seeming moment of enlightenment about how the pump 3/4-pin works? Must have been AFTER you asked them (TPU) why your loop isn't working and before he (and I here) told you how, lol... on top of that, you misunderstood him thinking he was talking about the block tubing being backwards - he also said your res before the pump for the same reasons. My guy, I'm not going to be stunned by you showing a 5.1 GHz screenshot. I can do that on air, likely higher. It's just a screenshot. I'm sure 5Ghz is in the cards with some worthwhile stress testing (P95 Small FFT, or AIDA64 at least 45 mins to saturate the loop)... likely with no AVX offset even. Cheers for accomplishing what is to be expected of what cooling you have?? Not sure what to say there. Go get em, tiger! In closing, I feel like I need to reiterate the point that I'm not trying to force you to do anything to your loop. All that is going on is arming you with facts so you can make whatever choices you want based on the theories you are conjuring up in your head supported by a video that doesn't support res->pump discussion at all. If I am guilty of anything that would be my insistence on sharing them. Yep, I said it twice (with tweaks here!) in hopes you'll put the hackles down and have a less defensive conversation about things. Cards are out on the table. Do with them as you may as you always could and would. I agree, let's move on. Let's see this thing kick over and pound out a lengthy stress test. |
|
|
|
Post by stormchaser on Aug 8, 2020 2:26:32 GMT -5
I'd also clean up those cables if possible. The chassis looks a bit basic/old, not sure if there is space between the chassis and motherboard panel, but that is where they should all go. This will give you more airflow in front and clean up the rats nest of cables from a non modular PSU. Masters of the obvious here lol Oh yes, no build thread would be complete without the one guy critiquing your PSU wiring before the rig is even half way built... little over ambitious, are we? lolLike the guy who says to the painter, you missed a spot! And you actually need to sit him down and explain to him in no uncertain terms that you have not even started on that side of the house yet. Or the guy who says WHAT?? you haven't spackled or finished the drywall yet!? And out of his other eye he can clearly see half of the wall and most of the room is still gutted, just bare 2 x 4 framing without wiring, light switches and outlets. And that's teh question that pops into his head lol You gotta wonder about the true motives of some people in this life, that's for sure... Okay, okay, just one more Or perhaps, lets say you are building a house. A bystander walks by to inform you that you should put a roof on the house to prevent rainwater from getting into the interior. Then you ask him if he sees the scaffold with all the shingles on it and you say "what do you think those are for?" lol It's fixed, btw, I mean the wiring p.s. try to avoid putting the cart before the horse. You are just muddying the waters. |
|
|
|
Post by stormchaser on Aug 8, 2020 2:29:34 GMT -5
I feel like I need to reiterate the point that I'm not trying to force you to do anything to your loop. All that is going on is arming you with facts so you can make whatever choices you want based on the theories you are conjuring up from 'jay'. If I am guilty of anything that would be my insistence on sharing the info. I'll go through this exercise (in futility) and respond one more time. Why does gravity not matter in a loop? Because it's a loop. What goes down, comes back up! This Jay vid? Just watched it. What a waste of time. Glad I'm a couple drinks in on some Woodford Reserve. That video only supports what I was saying, you are aware of that, right? The point of contention is only with the res before the pump. Loop order outside of that mantra, doesn't matter. You can go swap any rad/gpu/cpu/rad in any order and not have much difference. No shit. I was abundantly clear that was my stance. This vid did nothing for that point at all. You really seem to have trouble hanging on to this train. Again, stick with me. I said you could move the pump, not the res. Looks like it will fit in the front on the bottom where there isn't any soundproofing. Set the pump vertically, intake up to the res (or two in your case), then the output goes to the CPU or rad or whatever because loop order doesn't matter (so long as the res is before the pump). That's how I would do it. Or, again, don't go dual res (since there is little reason to have them) one res up top, pump where it is but vertically. Remember, nobody is forcing your hand. I'm only saying it is possible as you believed it wasn't. I mount a MCP655 (D5 vario) to the outside of an Antec 900. There are grommets on that case to route tubing in and out. Clearly not ideal, but, did it myself. Yes. I do. 100%. Absolutely. Did you just pull that guys chain (fry?) in your reply and your seeming moment of enlightenment about how the pump 3/4-pin works? Must have been AFTER you asked them (TPU) why your loop isn't working and before he (and I here) told you how, lol... on top of that, you misunderstood him thinking he was talking about the block tubing being backwards - he also said your res before the pump for the same reasons. My guy, I'm not going to be stunned by you showing a 5.1 GHz screenshot. I can do that on air, likely higher. It's just a screenshot. I'm sure 5Ghz is in the cards with some worthwhile stress testing (P95 Small FFT, or AIDA64 at least 45 mins to saturate the loop)... likely with no AVX offset even. Cheers for accomplishing what is to be expected of what cooling you have?? Not sure what to say there. Go get em, tiger! In closing, I feel like I need to reiterate the point that I'm not trying to force you to do anything to your loop. All that is going on is arming you with facts so you can make whatever choices you want based on the theories you are conjuring up in your head supported by a video that doesn't support res->pump discussion at all. If I am guilty of anything that would be my insistence on sharing them. Yep, I said it twice (with tweaks here!) in hopes you'll put the hackles down and have a less defensive conversation about things. Cards are out on the table. Do with them as you may as you always could and would. I agree, let's move on. Let's see this thing kick over and pound out a lengthy stress test. Wow. You are really getting your panties in a bunch over this nonsense, aren't you? |
|
|
|
Post by MainFramR on Aug 8, 2020 2:58:20 GMT -5
Woodford Reserve, a smart phone, and batwings im hanging so loose!!
GNight!
|
|
|
|
Post by stormchaser on Aug 8, 2020 3:44:43 GMT -5
If you move the pump to the out side of the cage, you can significantly shorten the waterblock supply hose and you won't get a kink from it. A trick to kind get the lines to form better is to run them under hot water. quickly dry with a towel and install. Once it cools to room temp, it'll take memory of the shape and won't give you any fuss. Looks good though! Something like this awaits my bucket list. Thanks. I am looking forward to overclocking tomorrow if everything goes according to plan. I will say, I am usually a pretty big fan of MSI and perhaps I got a lemon, but I am really not impressed with the quality of this Z390 ACE motherboard. It's only 6 months old and already I've lost the primary PCIe slot for my GPU and the LCD display for post codes literally fell apart into two pieces, so now I cant even read post codes to troubleshoot. Also gets stuck in a reboot loop. Meaning, when I power it on, it wont post until the unit itself does another restart, on its own, automatically. Even after CMOs reset it's happening. Sure, it's still under warranty, I just don't have enough time to RMA it at the moment because I'm leaving to hike the 220 mile long trail that that cuts across Vermont in a north to south direction later next week. So I better not have any OC complications tomorrow... because if I do, this build will have to be mothballed and put on the back burner for a couple months while I go on my hiking expedition in the green mountains of Vermont. RE: PUMP Of course, just my luck, I already went ahead and cut all the extra wiring looms from the PSU thinking (I left enough room to splice it back in, but you know..) what the hell would I need a molex cable for in 2020? Answer: Apparently, at least one freezemod pump lol Unfortunately the pump showed up AFTER I gave my PSU a hair cut. For this particular build I really could have gone two different ways with it. I'm glad I took the road less traveled. In the sense that I was contemplating building one of these giant, alien like, 300lb gaming monstrosities with about 16 fans and 2000mm of radiators towering overhead. (and building the required two ton crane overhead to move it, of course). The idea was catching traction until I realized I didn't really want my computer to resemble the Vegas strip (or a Christmas tree for that matter). I also did not want to be crushed to death in the unlikely event that the machine tips over while I am working on/around it. As far as I have determined, these types of builds do not stay true to form to their lineage and roots, and they scream synthetic and artificial, all money and no vision, because you've seen one, you've seen them all. You might want to look away right now. You can never un-see something like this.  Sleeper won out, and for good reason. I can actually pick it up off the ground without rupturing a disk. At least to me, I get more eye candy out of something like this:  Contrary to what you might think, the case I selected, the Coolermaster Sileo 500, was the last computer I put together with my father many years ago before he passed away. So it does have a good bit of sentimental value to it - and I am well aware of it's flaws and rather "normal" and elemental brushed aluminum finish... and that's the point (nothing to write home about). Anyway, I always had a vision for another build using the same case, except this time with much more powerful internals and upgraded cooling. And that's why I am doing what I am doing. The best place for these overgrown alien-like machines is really your local arcade. Not to use. But just to fill in and brighten the back of the arcade area where the lighting is poor. |
|
|
|
Post by stormchaser on Aug 8, 2020 19:56:14 GMT -5
I seriously cannot believe this BS. I'm literally 5 minutes away from boot and I run into this snag OF COURSE. Arrrg! I seem to have misplaced the molex end connector for the two wires that power the pump (basically I've lost situational awareness of which wire is + and which is - coming off the pump. Keep in mind this is power ONLY, I already have the senor and PWM wire connected properly to the right header on the motherboard. Everything else, including PWM, is ready to rip.
I need to make the distinction ASAP because I'd like to get it up and running tonight. I was thinking trial and error but I have a fear that reversing + and - wires may damage the pump or the PSU? So I'm a little leery to go down this road, and I don't want to burn anything out. Are my fears unfounded? Another option, perhaps... I could wire up an old power supply (and power on via the button on the case) and see if we can make a determination on which wire is which because it's a throw away unit and I don't mind blowing it up if the wires are reversed. Even if something pops, at least I will be able to tell straight away which wire is + and which is - <---- I think this is what I am going to do...
I will get you guys pictures of both wires so you can see what I'm talking about. Can anyone make a distinction on these patterns on the wires?
Freezemod has basically zero documentation online and nothing came with the pump itself. This is not a good thing for a greenhorn such as myself, as this is my first custom loop...
|
|
|
|
Post by stormchaser on Aug 8, 2020 20:30:57 GMT -5
Take a look - this is the last hangup I hope to encounter. These patterns runs the length of each wire. Anyone know the scheme here? Which one is + and which one is - ?? Again this is wiring going directly to the pump motor....  The other one:  |
|
|
|
Post by stormchaser on Aug 8, 2020 23:11:13 GMT -5
We are back in business ladies and gentlemen! Pump is working great and runs very quiet - so I can tidy up and shift gears into overclock mode. FWIW the pump is balanced in a position where there no metal to metal contact with the chassis, which further reduces NVH. Overclocking marathon starts NOW! Im going to set 5.1GHz as the initial goal here. Im a happy camper! |
|
