Stefan, I love how your channel shows the applications of practical printing. You stand apart from the rest of the 3d printing channels because you do something unique. Keep it up
@@StopChangingUsernamesRUclips I wouldn't use a metal bracket without a good insulator. With that many important thin traces running underneath the bracket, just a little barb could make a bad short.
@@350606 Stock ones are made of metal. Nothing special in their design; if you can mostly copy what they use, and use common sense, there's no real downside to CNCing one yourself. You could even 3d print a cheap model to put in place to make sure there's clearance all around before making the final product.
@Nugget That's interesting! But, since I'm in IT, it is confusing me. I know the UK used (or maybe still uses) aluminum for the wires at their homes. Of course, that would have a mechanical connection which would likely inhibit the formation of an oxide layer, right? Also, I've tested with a multimeter aluminum PC cases (or they claim they are). Sure, the anodized/galvanized parts are really hard to get something on, but the cut edges, the screw threads... are conductive to the low voltage of my tester. There's some measurable resistance, but it would still cause a short to my power supply. I guess, what I mean to say is, is it that the oxide layer breaks down at 5000V but aluminum (clean) is conductive? Or what is going on? I'm not kidding, you've got me confused here :)
It is great to see videos where the mechanical properties of 3d printed parts are explained. Most reviews focus on surface aspect and visual properties of the print. Thank you for the insights in structural and material engineering :)
My brother gave me a water cooling loop several years ago, but the cooling block was for an intel board and I had an AMD processor. I printed some brackets to make it work just out of PETG lmao. My brackets looked much more dainty than yours. The computer ran for years. I ended up giving it to a buddy. He still plays games on it.
LOL, just went to the trouble of creating my own AM4 to LG1200 heatsink adapter and was feeling to good about myself and then find this video, back to feeling like the noob I am! Thanks a lot for this very useful information. Ill see how to improve my adapter or just buy an LGA 1200 cpu cooler, was fixed on the idea to use AMD Wraith Prism Cooler on my I5 10400 mostly for the nice looks and higher TDP capacity but after watching these, buying a new boring cooler seems the better solution for now.
Yes, i have an Adapter for my water cooling Fan+Radiator. So i can mount the 140mm Fan+Radiator on a 200 mm Case Fan in the Front. And a gpu holder for my Case so it does not sack. Both in PLA. No Problem since over a year!
1:05 lol, here in Hungary, our energy cost is nearly half of it (0.07 - 0.15 eur kwh) however we have lot of other costs in the final electricity bill somehow
Ive been running my graphics card (GTX 970 with a bit of overclock) with a 3d printed refrigeration system for almost 2 years (with 2 different designs that i made myself during this time, and a third one coming). The fun fact is that i used PLA+ and it stills stands besides the deformation haha
Yeah at 100C it won't hurt the CPU at all but it turns off before it gets any hotter to prevent damage. Although most CPUs will stay at 100C for long but throttle performance
Hi Stefan! Ich schaue mir alle deine Videos an und muss sagen, dass sie mir viel Freude bereiten. Ich hatte dir schon mal geschrieben, dass wir einen ähnlichen Werdegang haben und der Anteil an Nerdness dich sehr sympathisch macht, zumal die Interessen sich ziemlich überschneiden. Mach weiter so! Viele Grüße aus dem wunderschönen Allgäu
I personally have printed brackets for my watercooling reservoir! I bought many second-hand components for my loop and the reservoir pump combo didn't come with a bracket!! So I 3D printed a bracket using the Laing D5 bracket design and it appears to be working well! Although the reservoir does seem a little heavy for the bracket, my computer doesn't move so it has enough strength in simple PLA.
I'd have designed a pair of 'truss plates' to bolt on top of the sides of the bracket. Print them laying on their side so one edge has a slight curve. When mounted, that curved edge fits down on top of the mounting bracket so it applies downward pressure to counteract the creep.
When Ryzen hit th market I printed AM4 bracket for Asetek water coolers from PLA, used it for 2 weeks until metal one arrived from manufacturer. It worked ok, but after two weeks it was deformed, but cooler didn't came loose.
yeah pla has some serious creep problems. You need to design everything super overengineered to conpensate for that. I usually add lots of fins as reinforcement on the areas that will be streesed and try to make the part as solid and robust as possible in those parts too even if it already looks good enough.
I designed and printed a support for the far end of my GTX 1070 Ti to avoid stressing the motherboard's PCIe slot due to the weight . I designed it to screw into some handy holes in the motherboard tray of my case.
Hey man, what is the temps they are exposed to? I need to make an AM4 bracket for my corsair h115i watercooler and i'm scared it may melt, will ABS be strong enough?
close to what I was thinking. dropping in small pieces of steel bar/plate into pre made pockets in the print to reinforce it. your idea is better, just print a hollow tube long wise and put a bolt through it.
It works quite well in my experience. I did this for a test bench last year in the motherboard tray and PSU mount, and it's so far been unreasonably sturdy. I added channels to those parts and used a mallet to gently force some 1mm steel rods in. Admittedly, I could've also just made those parts thicker, but it seemed like it'd be more fun at the time.
This leads me to an idea that actually this print can be done with a metal matrice(s) inside. Just need some efforts to make them stick on a layer (on pause) and continue a print. I think that thin metal wire-net can be used here... Maybe some slicing algorythms (to be possible to include thin metal layers inside) could be developed. Although IDK how is it possible without additional touch-sensors... usually a 3d-printer doesn't "see" what's under a nozzle.
@@KiR_3d It's almost definitely in the "more trouble than it's worth" realm, but tin melts around the same temperature as a lot of hotter PLAs at 232C and has a tensile strength of 220MPa compared to PLA's ~35. Provided you have a separate extruder that only handles tin, and print with a shielding gas to keep the tin tinny, I think you might be able to skip any slicer fanciness. Tin's also a pretty soft metal, so whatever hobbed gear's driving it probably won't lose any teeth if everything else in that pile of weird works out.
Another easy solution would be to add small metal strips that are screwed along the edges to give stiffness in the region that failed. You can get metal strips like that at any hardware store.
Great video Stefan, nice to see problem description, design, print, test, post mortem, failure simulation in software, redesign, reprint, and success! Also I've run into something quite similar in trying to find a retaining cooler bracket for slightly out of fashion boards. Good to see that if I had to now, 3D printing can solve this problem.
I printed a fan shroud for a 140 mm fan to be mounted on a 120mm radiator using PETG. I hope that it is still functioning XD It has been over 2 month now.
Well, I have done the same exact thing you did. I printed cooler bracked for my Ryzen 7 1700, but did it with just regular cheap PLA and it work perfectly. Also I believe you put too much attention to heat resistant. The air around CPU will never be hot enough to heat such bracket to 60C. My bracked is in direct contact with heatpipes to push the cooler to the CPU (I have added some thermal isolation) and no problem there.
What about cutting a couple of plain metal strips, and making a recess on the top surface of the bracket. When you bolt it down, the strips then spread the load.
With parts that see constant heat and the more air gap inside of the part the more the part tends to warp so the best option would be to do a solid piece if you can afford the time and material.
I think its more a case of moisture causing the nylon to soften rather than creep, could be a combination as well though, but moisture causes nylon to go soft as im sure you are well aware of as well.
Revisiting this video just today because I needed to make custom brackets to attach a regular cooler to an HP motherboard with a hole spacing of their own. Printing it with PETG with a 110% flow made a messy print which needed cleaning but quite strong and resistant to creep. Now just drawn and printed two LGA 2011 80x80 brackets in carbon PETG, 110% flow and 40% infill with 1,6mm walls. They are rock solid. I expect no problems since the regular PETG worked fine. I took this route because i watched a video days ago about printed parts resistance to creep and the two best were ASA and PETG. PLA worked also good until tests rised ambient temperature. Nylon, with no carbon, was the worst so I thought that if the base plastic is already better and worked for me the carbon reinforced one should be perfect. One thing to consider is that I achieved to get Tungsten nozzles that were for sale some time ago and they are eternal and have much better heat transfer than steel. It's weird I've not seen reviews of this nozzles anywere because they are perfect for high temp and abrasive materials.
If you have applications which are going to be long term usage you want to look into 'creep strength' when it comes to plastics. That's when a constant loading failure occurs like books on a shelf that make the shelf begin to sag under its own weight. Plastics are very susceptible to this kind of failure.
The 3 different types of polysulfone filaments which are extremely creep-resistant, they will all hold up (even over time) at temperatures well over 200c. Big drawback is that you need an upgraded hotend (350c), bed (150c) and a heated (100-200c) enclosure.
I highly appreciate the amount of responsibility you spent in this video (i.e. testing it for two months). Nice job, I hope it lasts for many more months to come. 👍
I did not know about creep! Thank you for that. Also, this strikes me as EXACTLY the sort of application that 3d print is made for: custom attachment solution. I too have a custom mount for my PC cooler, but I made it before I had my own 3d printer. I instead made it out of sheet metal. It's not much to look at, but I'm quite proud of it, and it almost looks good enough to have come out of a professional shop ;) But yeah, if I were facing the same problem today, I would try to solve it using 3d printing fo sho.
Subscribed..thanks, I've been watching your channel for a while now and I didn't realize I hadn't already subscribed. Keep it up, I appreciate your hard work.
PETG seems to have the least creep compared to PLA and hybrid nylons. A flex-fit iPhone holder that I designed for use in my vehicle failed with both Black Hatchbox PLA and Black Taulman Bridge Nylon. Only once printing in Black Atomic PETG did it cease to fail (4+ years now in an always-outside E36 BMW M3 with black interior in California :)
You could have designed the bracket to accept metallic washers around the screw holes, and even some flat braces (the sort sold at hardware stores) at selected places, to reinforce it.
You may have been able to concave the design a little (or set the mounting nuts even lower than in this design), and use screw torque to tension against the elasticity. Over time, you might need to further tighten the screws, with care that they do not reach the circuit board out the other side of the nuts.
Printing the part full solid should make for a better thermal conductivity, thus avoiding hotspots. Also you could try printing the part pre-warped, so it will be harder to bend in the opposing direction.
Had the same issue just recently, but solved it in a more traditional craft fashion. I also had an old huge ass thermalright cooler i wanted to use for my ryzen build. And then as I assembled my computer I discovered that the am3-am4 mount holes was different.. But I actually solved it with some simple iron sealing wire. What I did was to cut two sets of 1.2 mm iron sealing wires that i twisted together in pairs - to make them stronger(especially in respect to tightening later in the process). Then I drilled some slightly larger holes in the bundled back bracket of the ryzen cpu, to fit the wires through. After that i applied thermal paste onto the cpu and placed the cooler on top, and fitted the wires from the backplate bracket, then through the motherboard and then into the am3 thermalright front bracket. To tighten it all I used a concrete plier. It worked wonders to me. In stress test of the cpu with prime95 I reached 60c. Then i stress tested the gpu (rx580) and cpu simultanious and got a peak at 74c out of the cpu due to hot gpu air. But I dont get me wrong. I really appreciate your "high tech" cnc attempts tho. Your channel just helped me a moment ago figuring out why my banggood laser cutter would not move due to my decimal format in regional settings. Thanks! Edit: To twist the wires i ran a power drill I had tightened to the wires in one end and used a flat plier in the other end. Edit 2: Did not have a top drill available, as you showed in the end that would have been the most convenient. And again: I get that was not your point of this experiment.
Decent work mate keep it up! People like yourself have saved youtube for me and I'm sure quite a few other techies! To all the thumbs down boohoo to you!
Changing orientation of Case to side solve issue of mechanical stress or change Case to Cooler Master HAF XB or similar additional will also eliminate possible GPU sag issue.
You could've just used a tungsten carbide drill bit and put 4 additional holes in the original mount with the am4 spacing, that's what I did for my axp-100r
I made a bracket for my GPU, so i could use different fans on it. The standard fans for GPU's are usually cheap and the bearings are pretty loud even out of the box. So this was the second best upgrade for me since moving from 3.5" drives to 2.5" drives and SSD.
100% infill and annealing would have probably also prevented it from failure. Especially with CF nylon, I get significantly more strength from annealing.
Heh, this reminds me of the time in my high school "electronics" class, which was really more like a computer repair trade class. Someone way over-torqued screws on a motherboard and I guess they cracked something in the PCB and caused a short because it ran for a few minutes and then CAUGHT FIRE.
I wonder if you could have made a hybrid part? Like printing the bracket with a hole that ran through high stress areas. You could insert a short section of CF tube or aluminum tube or steel rod.
Nice video ! Perhaps adding a thin but tall shell section around the part perimeter would make it stiffer, and more creep resistant. Since the top of the shell wouldn't be so close to the CPU, and creep increases non-linearly with temperature. Also, beam stiffness increases with b.h^3/12, so increasing height h by 3 would increase stiffness by 27. Might be worth giving it a try.
Probably already said. But when a new CPU socket comes out, and the cooler manufacturers adds a new mount to their cooler to support it. Sometime, not all the time, if you email the manufacturer, they will send you the new mount you dont have for free, or a small charge.
Hi Stefan, as I commented on an other great video from you: I have printed a retention kit for a water cooling block for may Radeon RX Vega 56 consists out of 2 parts. One holds the water block in place, the other one is a spacer and retention bracket at the back of the video card. To cool the VRMs of the video card, I have placed a 92mm Noctua fan directly over the VRMs and a large part of the water block system. This hybrid construction provides cooling for the retention kit out of Extrudrs Greentec Pro as well. I watched the parts for a couple of days (now weeks) and it seems to work fine. I also printed a housing for my RX Vega 56 out of PETG what seems to work also fine. Finally I managed to print water tied self designed water cooling adapters and fittings out of thick PETG. I used around 3-4mm think walls with very hot printed PETG and a extrution multiplayer of 1.1 to get it tied. Pretty much the same tricks to get prints transparent are also great to get somewhat water tied. Only my PETG cloned Alphacool Eiszapfen 13/10mm to G1/4 Soft tubing fittings fail sometimes when I tied them to much so the G1/4 threads failing by breaking off ;-) I can provide you some fotos for discussion for a future video regarding water tied prints? PS: Thanks for the tipp with the diamond reinforced PLA! I will try it for my next retention kit project! ;-) Mit freundlichen Grüßen Der Marcus :-)
Maybe best of both worlds and use a simple metal sheet thick enough to reinforce each side? Imagining something like a rectangular washer that fits into a recess.
can't hurt to sneak in some steel reinforcing rods. just model in a cylindrical tunnel with a flat ceiling for the rod to sit in, and add a pause print right before that ceiling prints. i use mild steel for this sort of thing, where i live it's pretty easy to get 1/8" or 1/4" round bar, the 1/8" can be cut with bolt cutters and the 1/4" can be cut with a hacksaw. for this application you probably want a harder alloy, you could use a piece of hardened shaft if you have a cut-off wheel to cut it to length, or if not maybe just use some cheap drill bits that are close to the right length and adjust the model accordingly, or torque the heads off some machine screws and use the shafts. to ensure maximum load transfer into the rod, you could put a bit of epoxy in the bottom of the trench before inserting the rods (making sure not to get any near where the printer nozzle's gonna go!)
Hi, at 6:03 it looks like you have loads of depth available (at least on this side). The key to good beam strength is beam depth - much more strength to be gained by changing the depth than by adding width. Bring the part up in the centre and taper to the corners. However this may cause delamination as the tensile loading will be across layers, so perhaps add a thin web rising up topped with a round bar and a thicker column at the ends to transfer the tensile loading down to the mounting bolts ?
I'd be concerned that the print will just fail slower than the previous one. Over time, the gap between the cooler and CPU will slowly widen, decreasing cooling efficiency just like with the first failure. However, the failure will be so gradual that you will be running at slowly elevating temperatures for long periods of time which might reduce the lifespan of your CPU rather than it getting throttled after a couple of days with an decisive failure.
BTW, regarding infill direction, 2 x (sqrt(2) / 2) > (1 + 0), so +/- 45 degree should theoretically be stiffer than 0/90 degree. Or at least, I think maybe... Would be interesting to test!
Stefan, you need an Intamsys Funmat HT. Use PEEK and you'll never worry about too high of a temp ever again, that is unless you try to clock your CPU at 450C
You could if you really want to use PLA use cooling fans or even better have cooling fins on the part. It kinda depends on where the heat originated. If the heat came from surrounding air (convection) cooling fins wouldn't do much. If however, it came from conduction it would. You could also try placing more space between the heat source and the stress points so there is more surface area and longer travel path that then causes more dissipation of heat. If it came from radiation blocking the source will do. Source: mechanical engineer student
ASUS has 3d print schematics for their motherboards on the website you get drivers/bios updates on. It's pretty cool. They have RAM cooling mods etc. The motherboards even come with holes for the 3d prints to fit into.
Great video :) I was aware of the creep issue with PLA (which can even be used intentionally to make springs that auto tune to the correct tension when interfacing with parts that have substantial dimensional tolerances, like AA batteries), but I haven't encountered it yet with CF nylon. Very useful info
Having the CPU in a horizontal orientation (like in a test bench case) would of alieviated some of the stress from mean old gravity. You could if also added an extra support to the end of the cooler so it wasn't pulling on the bracket.
Actually i have a piece of string tying my Scythe Kama Cross to a brace inside the enclosure - thinking of replacing that with a 3D printed clip or something like that. A very top-heavy cooler and about 700g in total, and i had issues where i'd have to re-tighten it once in a while until i did this, or would notice it wouldn't run properly in vertical orientation while it went fine as soon as i laid it down. The string is just taking the weight off the mounts and board and ensures even clamping.
Great video as always! I have also used 3D printed parts for my computers cooling system. I have "custom loop" wich consist of scraped-apart dehumidifier chilling water on my balcony, and tubes going to my room, to radiator (so I can cool my room also) and to my computer. But cuz I'm cheap af, I ordered the waterpump from ebay, it was supposed to be 1500 l/h pump, but I tested it and it came out around 300l l/h, so I designed copy of the "head" of the pump, with bigger chamber, and new impeller, with more and wider blades, now I'm getting much more pressure out of the pump, and more liters per hour!, It's actually pretty amazing what you can do with 3D printing
@@SianaGearz tell me about it... I've got a new gaming pc and while speaking it up, the place I ordered from said no new cases have the drive bays... its a joke they ate so useful.
I'm guessing from the cable management that looks doesn't matter. So why not just zip-tie the top of the cooler to the top fan grill of the case to help carry most of the weight of the cooler. I've done this before a couple of times when I've had to install heavy tower coolers with questionable mounting hardware
Stefan this vid reminds me of the markedforge. Fiber reinforced printing. Hmmmm also reminds me of the printing on fabric mesh.... I can’t recall a study video on the topic of a reinforcement mesh inclusion in prints? Perhaps multiple layers of mesh?
If the new bracket creeps, you could print another in cheap PLA, make a 2-piece plaster mold off of the print (or make 2 prints, each one being half the bracket to simplify this step), and then use a high temp / high stiffness 2-part epoxy to make the bracket in your mold. My first thought was to get fancy with dual cylinder / mixing tube type epoxy and use it like an injection mold, to prevent the possibility of air pockets, but that might be overcomplicating things if you can get a good pour in a static mold without going that far.
If you do this there is no reason not to add proper fiber reinforcement as well. E-glass is cheap (ish) and no plain polymer (filled or not) will match a fiber composite part.
Doubling material thickness increases yield strength in that direction 28 fold. (!!!!!) So where the reinforcement was needed, you could have made it come higher for less area coverage. However caps don't tend to run hot, so this is probably fine too. In any case, from the first version it was very obvious it would fail on that spot.
Correct me if I'm wrong, but double the thickness should improve the stiffness 8 fold and the bending stress 4 fold. sigma_bending=M/W, W=b*h²/6, I=g*h³/12
@@CNCKitchen oops you are correct! I was thinking something else :) In any case, 8 or 4 fold increase for 2 times the material is still very good return :)
Annealing Cheap PLA without warping and deformation. 1.) Have an Enclosure 2.) Set the Bed temp to greater 80 C (depending on height The tallest part I printed that way was 62 mm) 3.) Put enough stuff on the build plate so that the print will take longer that 2 hours 4.) do not open the enclosure until the print plate cools down to room temperature. You will have an annealed print you can stick it in the oven afterwards right at 140 C and nothing will happen. It was tested with Hobby King PLA I purchased at $12/kg and with Matterhackers Build Series which I got at $ 19/kg. With an 85 C bed the temperature in my enclusure measured 300 mm above the bed will hold somewhere between 37 and 40 C. The printer is a matterhackers Pulse so A prusa clone. I intend to make a video on it as Soon as robotics season gives me a break. There are already some videos on my channel just not PLA ones. It will be for training my robotics kids as we intend to eventually 3d print the whole 30x30x48 in (about 1Mx1Mx1.5M) robot (in pieces of course)
@@propcutfpv6340 Yes Bed >=80 depending on how tall the part is enclosure temp measuered 250mm above build plate between 36 and 40C depending on build plate temp 40C at 100 bed in my case. Then print and leave bed at that temp until 2 hours are over on prints that are less than 2 hours. Most my prints are way over 2 hours so I don't have to leave it on. Then the bed cools slowly to room temp and so does the print and you printed an annealed part. Safely handles 140C or so. The melt point is still the same so if you stick it in an oven and slowly raise the temp in the case of hobby king PLA at 160 its solid as a rock at 165 it turns into a piece of bubble gum with MH Build PLA it seems to happen about 5 deg C earlier so that why I say 140 should be safe for most PLAs One small drawback on printing PLA with an 80-100C bed is that it seems to me the "elephant foot" on the first layer is a little bit more pronounced than usual. But I can deal and live with that
"The bracket was a piece of sheet metal" for a moment I thought you said something else
I read that right as he said it haha. It did sound like something else
That bracket was a PoS metal.
Same here. I laughed much more than I should have!
@@TheTinkerDad Yup. Y'all are goin to hell for that one. Right along with me lol
lol Exacty!
Stefan, I love how your channel shows the applications of practical printing. You stand apart from the rest of the 3d printing channels because you do something unique. Keep it up
I planned on doing this, but opted for the quicker option of using zip ties
Same here. I was afraid they'd significantly creep over time as well but after 4yrs of running they still hold up nicely 😇👍
Yep. 3 years later it still works
Wouldn't this be a case where your CNC machine could be used to make a part out of aluminum?
To be fair, 3D printers are CNC hammers. When you have one, everything looks like a CNC nail.
@@StopChangingUsernamesRUclips I wouldn't use a metal bracket without a good insulator. With that many important thin traces running underneath the bracket, just a little barb could make a bad short.
@@350606
Stock ones are made of metal. Nothing special in their design; if you can mostly copy what they use, and use common sense, there's no real downside to CNCing one yourself. You could even 3d print a cheap model to put in place to make sure there's clearance all around before making the final product.
If you want to send me the file, I’ll machine it for you out of aluminum. :)
@Nugget That's interesting! But, since I'm in IT, it is confusing me. I know the UK used (or maybe still uses) aluminum for the wires at their homes. Of course, that would have a mechanical connection which would likely inhibit the formation of an oxide layer, right?
Also, I've tested with a multimeter aluminum PC cases (or they claim they are). Sure, the anodized/galvanized parts are really hard to get something on, but the cut edges, the screw threads... are conductive to the low voltage of my tester. There's some measurable resistance, but it would still cause a short to my power supply.
I guess, what I mean to say is, is it that the oxide layer breaks down at 5000V but aluminum (clean) is conductive? Or what is going on? I'm not kidding, you've got me confused here :)
Subbed because you make good videos and asked nice.
> asked nice
there's truly something to that
It is great to see videos where the mechanical properties of 3d printed parts are explained. Most reviews focus on surface aspect and visual properties of the print.
Thank you for the insights in structural and material engineering :)
My brother gave me a water cooling loop several years ago, but the cooling block was for an intel board and I had an AMD processor. I printed some brackets to make it work just out of PETG lmao. My brackets looked much more dainty than yours. The computer ran for years. I ended up giving it to a buddy. He still plays games on it.
PETG has a glass temp of about 2-3x of the material that CNC Kitchen thought was viable (but clearly wasnt)
the water cooling block probably never gets above 60 degrees. PLA would almost work. the air cooler will be getting up around 90 to 105
@@MrOsmodeus No, it's all fine either way. It's just mounting hardware, it doesn't get particularly warm at all! Barely warm to the touch.
Generaly when I make 3D printed parts to heavy loaded applications I epoxy some metal rods in them.
interesting technique
LOL, just went to the trouble of creating my own AM4 to LG1200 heatsink adapter and was feeling to good about myself and then find this video, back to feeling like the noob I am!
Thanks a lot for this very useful information. Ill see how to improve my adapter or just buy an LGA 1200 cpu cooler, was fixed on the idea to use AMD Wraith Prism Cooler on my I5 10400 mostly for the nice looks and higher TDP capacity but after watching these, buying a new boring cooler seems the better solution for now.
Some of the best content on RUclips. Thanks and keep it up!
I had 3D printed a CPU mount for my PC, 1 year ago. I used ABS and had no issues with creep. :)
Yes, i have an Adapter for my water cooling Fan+Radiator.
So i can mount the 140mm Fan+Radiator on a 200 mm Case Fan in the Front.
And a gpu holder for my Case so it does not sack.
Both in PLA. No Problem since over a year!
1:05 lol, here in Hungary, our energy cost is nearly half of it (0.07 - 0.15 eur kwh) however we have lot of other costs in the final electricity bill somehow
Ive been running my graphics card (GTX 970 with a bit of overclock) with a 3d printed refrigeration system for almost 2 years (with 2 different designs that i made myself during this time, and a third one coming). The fun fact is that i used PLA+ and it stills stands besides the deformation haha
Stefan, I love what you are cooking in your kitchen. From SoCal, keep it cooking!
"and energy cost is something to consider here in Germany"... that's what happen when you prematurely close nuclear power-plants.
You can't really kill a CPU by not cooling it. Most turn off at around 100°C to prevent damage, though some tolerate higher temps.
Yeah at 100C it won't hurt the CPU at all but it turns off before it gets any hotter to prevent damage. Although most CPUs will stay at 100C for long but throttle performance
This exact project has crossed my mind several times... Never been brave enough to try it.
Nice design, I am impressed. I created a PCI bracket fan mount to cool my GPU.
Hi Stefan! Ich schaue mir alle deine Videos an und muss sagen, dass sie mir viel Freude bereiten. Ich hatte dir schon mal geschrieben, dass wir einen ähnlichen Werdegang haben und der Anteil an Nerdness dich sehr sympathisch macht, zumal die Interessen sich ziemlich überschneiden. Mach weiter so! Viele Grüße aus dem wunderschönen Allgäu
I personally have printed brackets for my watercooling reservoir! I bought many second-hand components for my loop and the reservoir pump combo didn't come with a bracket!! So I 3D printed a bracket using the Laing D5 bracket design and it appears to be working well! Although the reservoir does seem a little heavy for the bracket, my computer doesn't move so it has enough strength in simple PLA.
I'd have designed a pair of 'truss plates' to bolt on top of the sides of the bracket. Print them laying on their side so one edge has a slight curve. When mounted, that curved edge fits down on top of the mounting bracket so it applies downward pressure to counteract the creep.
When Ryzen hit th market I printed AM4 bracket for Asetek water coolers from PLA, used it for 2 weeks until metal one arrived from manufacturer. It worked ok, but after two weeks it was deformed, but cooler didn't came loose.
yeah pla has some serious creep problems. You need to design everything super overengineered to conpensate for that. I usually add lots of fins as reinforcement on the areas that will be streesed and try to make the part as solid and robust as possible in those parts too even if it already looks good enough.
I want to print a AM4 bracket for corsair h115i, will ABS be a good material??
I designed and printed a support for the far end of my GTX 1070 Ti to avoid stressing the motherboard's PCIe slot due to the weight . I designed it to screw into some handy holes in the motherboard tray of my case.
Can you just CNC an aluminum block with the design you have?
He should do that with the first model, but sometimes one just wants to experiment and have some fun.
@@soundspark the base brackets are all made of a sheet steel so well no
i actually have 3d printed parts in the water loop of my desktop, regular pla, have been working perfectly for almost 6 months now.
Hey man, what is the temps they are exposed to? I need to make an AM4 bracket for my corsair h115i watercooler and i'm scared it may melt, will ABS be strong enough?
You could put some very fine metal tubes in your 3D printed part as reforcement.
close to what I was thinking. dropping in small pieces of steel bar/plate into pre made pockets in the print to reinforce it. your idea is better, just print a hollow tube long wise and put a bolt through it.
It works quite well in my experience. I did this for a test bench last year in the motherboard tray and PSU mount, and it's so far been unreasonably sturdy. I added channels to those parts and used a mallet to gently force some 1mm steel rods in. Admittedly, I could've also just made those parts thicker, but it seemed like it'd be more fun at the time.
or mill it from alu....
This leads me to an idea that actually this print can be done with a metal matrice(s) inside. Just need some efforts to make them stick on a layer (on pause) and continue a print. I think that thin metal wire-net can be used here... Maybe some slicing algorythms (to be possible to include thin metal layers inside) could be developed. Although IDK how is it possible without additional touch-sensors... usually a 3d-printer doesn't "see" what's under a nozzle.
@@KiR_3d It's almost definitely in the "more trouble than it's worth" realm, but tin melts around the same temperature as a lot of hotter PLAs at 232C and has a tensile strength of 220MPa compared to PLA's ~35. Provided you have a separate extruder that only handles tin, and print with a shielding gas to keep the tin tinny, I think you might be able to skip any slicer fanciness.
Tin's also a pretty soft metal, so whatever hobbed gear's driving it probably won't lose any teeth if everything else in that pile of weird works out.
When I want to have mechanical stiffness I add a simple steel hard wire 1 to 2mm diam and I'm sure it's good. It acts juste like steel for concrete.
Another easy solution would be to add small metal strips that are screwed along the edges to give stiffness in the region that failed. You can get metal strips like that at any hardware store.
Great video Stefan, nice to see problem description, design, print, test, post mortem, failure simulation in software, redesign, reprint, and success! Also I've run into something quite similar in trying to find a retaining cooler bracket for slightly out of fashion boards. Good to see that if I had to now, 3D printing can solve this problem.
I printed a fan shroud for a 140 mm fan to be mounted on a 120mm radiator using PETG. I hope that it is still functioning XD It has been over 2 month now.
Well, I have done the same exact thing you did. I printed cooler bracked for my Ryzen 7 1700, but did it with just regular cheap PLA and it work perfectly.
Also I believe you put too much attention to heat resistant. The air around CPU will never be hot enough to heat such bracket to 60C. My bracked is in direct contact with heatpipes to push the cooler to the CPU (I have added some thermal isolation) and no problem there.
Hey man, i want to do that for my corsair h115i watercooler, soo, the bracket never gets that hot? Maybe i should use ABS just to be sure???
What about cutting a couple of plain metal strips, and making a recess on the top surface of the bracket. When you bolt it down, the strips then spread the load.
With parts that see constant heat and the more air gap inside of the part the more the part tends to warp so the best option would be to do a solid piece if you can afford the time and material.
I think its more a case of moisture causing the nylon to soften rather than creep, could be a combination as well though, but moisture causes nylon to go soft as im sure you are well aware of as well.
Revisiting this video just today because I needed to make custom brackets to attach a regular cooler to an HP motherboard with a hole spacing of their own. Printing it with PETG with a 110% flow made a messy print which needed cleaning but quite strong and resistant to creep. Now just drawn and printed two LGA 2011 80x80 brackets in carbon PETG, 110% flow and 40% infill with 1,6mm walls. They are rock solid. I expect no problems since the regular PETG worked fine. I took this route because i watched a video days ago about printed parts resistance to creep and the two best were ASA and PETG. PLA worked also good until tests rised ambient temperature. Nylon, with no carbon, was the worst so I thought that if the base plastic is already better and worked for me the carbon reinforced one should be perfect. One thing to consider is that I achieved to get Tungsten nozzles that were for sale some time ago and they are eternal and have much better heat transfer than steel. It's weird I've not seen reviews of this nozzles anywere because they are perfect for high temp and abrasive materials.
If you have applications which are going to be long term usage you want to look into 'creep strength' when it comes to plastics. That's when a constant loading failure occurs like books on a shelf that make the shelf begin to sag under its own weight. Plastics are very susceptible to this kind of failure.
The 3 different types of polysulfone filaments which are extremely creep-resistant, they will all hold up (even over time) at temperatures well over 200c.
Big drawback is that you need an upgraded hotend (350c), bed (150c) and a heated (100-200c) enclosure.
I highly appreciate the amount of responsibility you spent in this video (i.e. testing it for two months). Nice job, I hope it lasts for many more months to come. 👍
I did not know about creep! Thank you for that.
Also, this strikes me as EXACTLY the sort of application that 3d print is made for: custom attachment solution.
I too have a custom mount for my PC cooler, but I made it before I had my own 3d printer. I instead made it out of sheet metal. It's not much to look at, but I'm quite proud of it, and it almost looks good enough to have come out of a professional shop ;) But yeah, if I were facing the same problem today, I would try to solve it using 3d printing fo sho.
Subscribed..thanks, I've been watching your channel for a while now and I didn't realize I hadn't already subscribed. Keep it up, I appreciate your hard work.
PETG seems to have the least creep compared to PLA and hybrid nylons. A flex-fit iPhone holder that I designed for use in my vehicle failed with both Black Hatchbox PLA and Black Taulman Bridge Nylon. Only once printing in Black Atomic PETG did it cease to fail (4+ years now in an always-outside E36 BMW M3 with black interior in California :)
Would be a nice experiment to load some printed rods/hooks for several weeks and measure the displacement through creep for various materials! ;)
You could have designed the bracket to accept metallic washers around the screw holes, and even some flat braces (the sort sold at hardware stores) at selected places, to reinforce it.
Interesting method to design an object with reference images. But is there any App that let´s you take photos in parallel to another surface ??
You may have been able to concave the design a little (or set the mounting nuts even lower than in this design), and use screw torque to tension against the elasticity. Over time, you might need to further tighten the screws, with care that they do not reach the circuit board out the other side of the nuts.
Printing the part full solid should make for a better thermal conductivity, thus avoiding hotspots. Also you could try printing the part pre-warped, so it will be harder to bend in the opposing direction.
Had the same issue just recently, but solved it in a more traditional craft fashion. I also had an old huge ass thermalright cooler i wanted to use for my ryzen build. And then as I assembled my computer I discovered that the am3-am4 mount holes was different.. But I actually solved it with some simple iron sealing wire.
What I did was to cut two sets of 1.2 mm iron sealing wires that i twisted together in pairs - to make them stronger(especially in respect to tightening later in the process).
Then I drilled some slightly larger holes in the bundled back bracket of the ryzen cpu, to fit the wires through.
After that i applied thermal paste onto the cpu and placed the cooler on top, and fitted the wires from the backplate bracket, then through the motherboard and then into the am3 thermalright front bracket.
To tighten it all I used a concrete plier. It worked wonders to me.
In stress test of the cpu with prime95 I reached 60c. Then i stress tested the gpu (rx580) and cpu simultanious and got a peak at 74c out of the cpu due to hot gpu air.
But I dont get me wrong. I really appreciate your "high tech" cnc attempts tho. Your channel just helped me a moment ago figuring out why my banggood laser cutter would not move due to my decimal format in regional settings. Thanks!
Edit: To twist the wires i ran a power drill I had tightened to the wires in one end and used a flat plier in the other end.
Edit 2: Did not have a top drill available, as you showed in the end that would have been the most convenient.
And again: I get that was not your point of this experiment.
You could use the PLA piece as a mold to cast that same piece of aluminum.
ruclips.net/video/8HlXLEVZGBk/видео.html&ab_channel=ArtByAdrock
Decent work mate keep it up! People like yourself have saved youtube for me and I'm sure quite a few other techies! To all the thumbs down boohoo to you!
Changing orientation of Case to side solve issue of mechanical stress or change Case to Cooler Master HAF XB or similar additional will also eliminate possible GPU sag issue.
You could've just used a tungsten carbide drill bit and put 4 additional holes in the original mount with the am4 spacing, that's what I did for my axp-100r
I made a bracket for my GPU, so i could use different fans on it. The standard fans for GPU's are usually cheap and the bearings are pretty loud even out of the box. So this was the second best upgrade for me since moving from 3.5" drives to 2.5" drives and SSD.
Nice CPU choice!
I 3D printed a pump/res mount for me build a few months ago. Saved me like $30 :D
Learning how to incorporate piece of metal into your print will get rid of this kind of failures. Tip: PLA melts easily, small piece of metal don't.
Awesome! I modelled and 3D printed an entire case for my PC. It's called the Laier One.
Dont worry brother, I have been subscribed since the beginning! hahaha Keep up the great work!
100% infill and annealing would have probably also prevented it from failure. Especially with CF nylon, I get significantly more strength from annealing.
Heh, this reminds me of the time in my high school "electronics" class, which was really more like a computer repair trade class. Someone way over-torqued screws on a motherboard and I guess they cracked something in the PCB and caused a short because it ran for a few minutes and then CAUGHT FIRE.
That's it a feature. You've put it in 'Campfire' mode. Grab some sausage!
Nice. When a 200 pound gorilla does eletronics. It just doesnt work.
LOL
Nice work man! You are awesome 💪
I wonder if you could have made a hybrid part? Like printing the bracket with a hole that ran through high stress areas. You could insert a short section of CF tube or aluminum tube or steel rod.
Nice video ! Perhaps adding a thin but tall shell section around the part perimeter would make it stiffer, and more creep resistant.
Since the top of the shell wouldn't be so close to the CPU, and creep increases non-linearly with temperature.
Also, beam stiffness increases with b.h^3/12, so increasing height h by 3 would increase stiffness by 27. Might be worth giving it a try.
Probably already said.
But when a new CPU socket comes out, and the cooler manufacturers adds a new mount to their cooler to support it.
Sometime, not all the time, if you email the manufacturer, they will send you the new mount you dont have for free, or a small charge.
Hi Stefan,
as I commented on an other great video from you: I have printed a retention kit for a water cooling block for may Radeon RX Vega 56 consists out of 2 parts. One holds the water block in place, the other one is a spacer and retention bracket at the back of the video card. To cool the VRMs of the video card, I have placed a 92mm Noctua fan directly over the VRMs and a large part of the water block system. This hybrid construction provides cooling for the retention kit out of Extrudrs Greentec Pro as well. I watched the parts for a couple of days (now weeks) and it seems to work fine.
I also printed a housing for my RX Vega 56 out of PETG what seems to work also fine.
Finally I managed to print water tied self designed water cooling adapters and fittings out of thick PETG. I used around 3-4mm think walls with very hot printed PETG and a extrution multiplayer of 1.1 to get it tied. Pretty much the same tricks to get prints transparent are also great to get somewhat water tied. Only my PETG cloned Alphacool Eiszapfen 13/10mm to G1/4 Soft tubing fittings fail sometimes when I tied them to much so the G1/4 threads failing by breaking off ;-)
I can provide you some fotos for discussion for a future video regarding water tied prints?
PS: Thanks for the tipp with the diamond reinforced PLA! I will try it for my next retention kit project! ;-)
Mit freundlichen Grüßen
Der Marcus :-)
You should have tried printing a fan adapter for the stock cooler. That way you can use any other fan.
subbed because your worth it!
Thank you!
Sorry Stephan, subscribed :D
Much appreciated!
Nice work, very interesting seeing the overall results.
Maybe best of both worlds and use a simple metal sheet thick enough to reinforce each side? Imagining something like a rectangular washer that fits into a recess.
can't hurt to sneak in some steel reinforcing rods. just model in a cylindrical tunnel with a flat ceiling for the rod to sit in, and add a pause print right before that ceiling prints. i use mild steel for this sort of thing, where i live it's pretty easy to get 1/8" or 1/4" round bar, the 1/8" can be cut with bolt cutters and the 1/4" can be cut with a hacksaw. for this application you probably want a harder alloy, you could use a piece of hardened shaft if you have a cut-off wheel to cut it to length, or if not maybe just use some cheap drill bits that are close to the right length and adjust the model accordingly, or torque the heads off some machine screws and use the shafts. to ensure maximum load transfer into the rod, you could put a bit of epoxy in the bottom of the trench before inserting the rods (making sure not to get any near where the printer nozzle's gonna go!)
i think its not creep. its plastic deformation when you put something into stress.. it reach the yield point of the material.
A spring on those two bolts would have either taken out all the creep or prolonged the longevity of the first bracket for quite some time.
Same thing happened to me with the stock bracket I got with the fan. It was made of plastic, and had snapped under the heatsink's own weight.
Hi, at 6:03 it looks like you have loads of depth available (at least on this side).
The key to good beam strength is beam depth - much more strength to be gained by changing the depth than by adding width. Bring the part up in the centre and taper to the corners. However this may cause delamination as the tensile loading will be across layers, so perhaps add a thin web rising up topped with a round bar and a thicker column at the ends to transfer the tensile loading down to the mounting bolts ?
Yeah, there have always been those safety margins where use of a 3d printed part comes into question. Good effort and thank you for sharing. :)
I'd be concerned that the print will just fail slower than the previous one. Over time, the gap between the cooler and CPU will slowly widen, decreasing cooling efficiency just like with the first failure. However, the failure will be so gradual that you will be running at slowly elevating temperatures for long periods of time which might reduce the lifespan of your CPU rather than it getting throttled after a couple of days with an decisive failure.
I have used HDD caddy made out of standard PLA for over a year with no deformation and the drives reaching 50-60*C daily.
Use stainless tig rods to reinforce your 3D prints. Works very well
Wouldn‘t give me a good feeling to have a high-performance workstation kept together by a unstable plastic part. I am waiting for a follow up ;-)
perhaps adding an embedded piece of metal rod across the creep area as reinforcement would keep it stiff enough over time.
BTW, regarding infill direction, 2 x (sqrt(2) / 2) > (1 + 0), so +/- 45 degree should theoretically be stiffer than 0/90 degree. Or at least, I think maybe... Would be interesting to test!
Good point! Need to run the math on that.
Stefan, you need an Intamsys Funmat HT. Use PEEK and you'll never worry about too high of a temp ever again, that is unless you try to clock your CPU at 450C
You could make a mold from the printed bracket and make a new bracket with some strong heat resistant resin.
You could if you really want to use PLA use cooling fans or even better have cooling fins on the part. It kinda depends on where the heat originated. If the heat came from surrounding air (convection) cooling fins wouldn't do much. If however, it came from conduction it would. You could also try placing more space between the heat source and the stress points so there is more surface area and longer travel path that then causes more dissipation of heat. If it came from radiation blocking the source will do.
Source: mechanical engineer student
Fun video. Love the practical use.
ASUS has 3d print schematics for their motherboards on the website you get drivers/bios updates on. It's pretty cool. They have RAM cooling mods etc. The motherboards even come with holes for the 3d prints to fit into.
Aren't those prints strictly decorative and have just 1 corner on the MB available to be screwed at?
Great video :) I was aware of the creep issue with PLA (which can even be used intentionally to make springs that auto tune to the correct tension when interfacing with parts that have substantial dimensional tolerances, like AA batteries), but I haven't encountered it yet with CF nylon. Very useful info
Having the CPU in a horizontal orientation (like in a test bench case) would of alieviated some of the stress from mean old gravity. You could if also added an extra support to the end of the cooler so it wasn't pulling on the bracket.
Actually did that in my tower some years ago. Without any bracket (forgot to order it and couldend wait fir it xd
Actually i have a piece of string tying my Scythe Kama Cross to a brace inside the enclosure - thinking of replacing that with a 3D printed clip or something like that. A very top-heavy cooler and about 700g in total, and i had issues where i'd have to re-tighten it once in a while until i did this, or would notice it wouldn't run properly in vertical orientation while it went fine as soon as i laid it down. The string is just taking the weight off the mounts and board and ensures even clamping.
A bold experiment. And you eventually made it work!
Great video as always! I have also used 3D printed parts for my computers cooling system. I have "custom loop" wich consist of scraped-apart dehumidifier chilling water on my balcony, and tubes going to my room, to radiator (so I can cool my room also) and to my computer. But cuz I'm cheap af, I ordered the waterpump from ebay, it was supposed to be 1500 l/h pump, but I tested it and it came out around 300l l/h, so I designed copy of the "head" of the pump, with bigger chamber, and new impeller, with more and wider blades, now I'm getting much more pressure out of the pump, and more liters per hour!, It's actually pretty amazing what you can do with 3D printing
If you've got a spair dvd drive bay, you can print a draw for the front of your pc. I keep all my sd cards in mine
I love optical drive bays! I don't know why manufacturers are so intent on killing them off.
@@SianaGearz tell me about it... I've got a new gaming pc and while speaking it up, the place I ordered from said no new cases have the drive bays... its a joke they ate so useful.
Actually its really hard to burn down a CPU, motherboards have a lot of fail safes to prevent that from happening
Impressive job
Thanks for sharing😀👍
Great video, lots of valuable information and experience
I'm guessing from the cable management that looks doesn't matter. So why not just zip-tie the top of the cooler to the top fan grill of the case to help carry most of the weight of the cooler. I've done this before a couple of times when I've had to install heavy tower coolers with questionable mounting hardware
Stefan this vid reminds me of the markedforge. Fiber reinforced printing. Hmmmm also reminds me of the printing on fabric mesh.... I can’t recall a study video on the topic of a reinforcement mesh inclusion in prints? Perhaps multiple layers of mesh?
what about adding an internal metal structure? Like an aluminum square
If the new bracket creeps, you could print another in cheap PLA, make a 2-piece plaster mold off of the print (or make 2 prints, each one being half the bracket to simplify this step), and then use a high temp / high stiffness 2-part epoxy to make the bracket in your mold. My first thought was to get fancy with dual cylinder / mixing tube type epoxy and use it like an injection mold, to prevent the possibility of air pockets, but that might be overcomplicating things if you can get a good pour in a static mold without going that far.
If you do this there is no reason not to add proper fiber reinforcement as well. E-glass is cheap (ish) and no plain polymer (filled or not) will match a fiber composite part.
Doubling material thickness increases yield strength in that direction 28 fold. (!!!!!)
So where the reinforcement was needed, you could have made it come higher for less area coverage. However caps don't tend to run hot, so this is probably fine too.
In any case, from the first version it was very obvious it would fail on that spot.
Correct me if I'm wrong, but double the thickness should improve the stiffness 8 fold and the bending stress 4 fold. sigma_bending=M/W, W=b*h²/6, I=g*h³/12
@@CNCKitchen oops you are correct! I was thinking something else :)
In any case, 8 or 4 fold increase for 2 times the material is still very good return :)
Annealing Cheap PLA without warping and deformation.
1.) Have an Enclosure
2.) Set the Bed temp to greater 80 C (depending on height The tallest part I printed that way was 62 mm)
3.) Put enough stuff on the build plate so that the print will take longer that 2 hours
4.) do not open the enclosure until the print plate cools down to room temperature.
You will have an annealed print you can stick it in the oven afterwards right at 140 C and nothing will happen. It was tested with Hobby King PLA I purchased at $12/kg and with Matterhackers Build Series which I got at $ 19/kg. With an 85 C bed the temperature in my enclusure measured 300 mm above the bed will hold somewhere between 37 and 40 C. The printer is a matterhackers Pulse so A prusa clone. I intend to make a video on it as Soon as robotics season gives me a break. There are already some videos on my channel just not PLA ones. It will be for training my robotics kids as we intend to eventually 3d print the whole 30x30x48 in (about 1Mx1Mx1.5M) robot (in pieces of course)
nope , it warps thin prints
@@propcutfpv6340 What warps thin prints and how thin is thin. The printing on a high temp bed or putting it in the oven afterwards?
wait , are you printing it at 80c bed temp?now it makes sense , i thought you were annealing with the bed after the print@@martinpirringer8055
@@propcutfpv6340 Yes Bed >=80 depending on how tall the part is enclosure temp measuered 250mm above build plate between 36 and 40C depending on build plate temp 40C at 100 bed in my case. Then print and leave bed at that temp until 2 hours are over on prints that are less than 2 hours. Most my prints are way over 2 hours so I don't have to leave it on. Then the bed cools slowly to room temp and so does the print and you printed an annealed part. Safely handles 140C or so. The melt point is still the same so if you stick it in an oven and slowly raise the temp in the case of hobby king PLA at 160 its solid as a rock at 165 it turns into a piece of bubble gum with MH Build PLA it seems to happen about 5 deg C earlier so that why I say 140 should be safe for most PLAs One small drawback on printing PLA with an 80-100C bed is that it seems to me the "elephant foot" on the first layer is a little bit more pronounced than usual. But I can deal and live with that