DIY CPU Waterblock




About: I'm a mechanical engineering student and I like to make little side projects in my free time. Of course that's assuming I actually find some free time!

I've been wanting to make a CPU water cooling block for a while, and after watching Linus from LinusTechTips make one in his Scrapyard Wars series I decided that it was about time that I got around to making my own.

My block was inspired by Linus', with a couple of my own tweaks here and there. I decided to use a clear polycarbonate top instead of the original copper plate to display the custom machined block and coolant, as well as a removable mounting system that allowed for a wider range of socket sizes and custom mounting solutions.

I was lucky enough to have access to a fully equipped machine shop for this project, so there are a few machines that I used that may not be very common in a home shop. However, with some creativity and patience the same results can be achieved with some simple hand tools. The only specialized machine that would be required for this project is a CNC mill.

To keep this instructable to a reasonable length, I'm assuming a basic knowledge on the use of machines typically found in a machine shop.

Let's get started!

Step 1: Materials and Tools


  • Aluminum Flat Bar - 2" x 4" x 1/8" thick
  • Aluminum Flat Bar - 2.125" x 2.125" x 1/2" thick
  • Clear Polycarbonate Sheet - 2.125" x 2.125" x 1/4" thick
  • 10-24 UNC x 3/8" Socket Cap Screws Qty. 4
  • 6-32 UNC x 3/8" Countersunk Screws Qty. 4
  • 8-32 UNC x 1 1/2" Pan Head Screws Qty. 4
  • 8-32 UNC Hex Nuts Qty. 4
  • Craft foam sheet
  • Preferred watercooling fittings - I used some compression fittings from Amazon

Note: All stock dimensions are rough cut sizes. Refer to the drawings in the next step for final dimensions.

Also note the material choice for your main block. Be sure to match it to the rest of your water loop to prevent corrosion. (Thanks, ironsmiter)


  • CNC Mill
  • Manual Mill
  • Bandsaw
  • Drill or Drill Press
  • Drill Bits - 0.103", 0.150", 0.2", 0.457"
  • Spotting Drill or Center Drill
  • 2 Flute End Mills - 1/8", 1/2" (Thanks, imakeembetter)
  • Facing Mill
  • Countersink
  • File
  • Utility Knife
  • Ruler
  • Cutting Mat
  • G1/4-19 Pipe Thread Tap
  • 10-24 UNC Tap
  • 6-32 UNC Tap

Step 2: Block Design

I used Autodesk Inventor to create a 3D model of the block to help me determine the final block dimensions and to generate the g-code for the CNC.

The overall design of the block has a clear polycarbonate cover that is mounted to an aluminum base and sealed with a gasket. The aluminum base has a machined pocket containing fins in the top where the water flows through, as well as a contour around the bottom. Eight tapped holes are used to attach the top polycarbonate plate as well as the mounting arms. The watercooling fittings are threaded directly into the top polycarbonate cover.

The mounting arms are removable to allow for the attachment of replacement arms to fit different socket sizes, or a custom mounting system for other uses.

While designing the block I also had to keep in mind the clearance for the motherboard components, as well as the limitations of my tooling. To achieve the proper clearance, I designed the block to have a 3/8" x 1/4" deep contour milled around the bottom perimeter of the block. For the tooling, I decided to use a 1/8" end mill to get as many fins as possible inside the block while maintaining a reasonable depth for the pocket. I'll cover this in more detail later on.

Step 3: Clear Block Cover

I decided to start by making the clear polycarbonate cover for the waterblock. The stock was cut to the rough cut size on the bandsaw, and then clamped into the mill to be squared off and machined to the final size of 2" x 2". Once the block was machined to its final size, I drilled the clearance holes in the corners (0.2") and drilled and tapped the mounting holes for the watercooling fittings (G1/4-19, 0.457" tap drill size) . I used a center loaded into the chuck to align my tap and keep my threads square to the part (last image).

Step 4: Main Block Preparation

With the polycarbonate cover complete, I moved on to the main block. I first took the block down to its final size of 2" x 2" with the mill, then ran a light cleanup pass over the surface of the block to remove any surface imperfections. Take care not to remove too much material during the cleanup pass as to not affect the CNC program later on. If the block is too thin, the cutter will break through the bottom and ruin the part.

Step 5: CNC Milling the Main Block

The zeroes for both of the CNC programs is on the bottom left corner of the part, so using an edge finder I zeroed in the machine. Once the proper tool (1/8" end mill) was securely mounted in the spindle, I loaded up the program to machine the reservoir and let it run.

While most 1/8" end mills only have a 3/8" (0.375") length of cut, I was able to squeeze an extra 0.025" out of mine and mill the full 0.4" deep pocket in the program. If you don't feel like pushing your cutter then just move the z-axis up 0.025" above the surface of the work and re-zero the machine. That way the program will only cut 0.375" into the material.

Once the program for the reservoir was finished, I flipped the part over, corrected my zeroes and ran the program for the clearance cut on the back of the block.

Note: These g-code files worked on my CNC (Tormach PCNC 1100), but I cannot guarantee that it will work on others. Make sure to check the code before running the program and ensure that it won't crash the machine. I take no responsibility for any mishaps caused by this code.

Step 6: Manual Machining the Main Block

After running the CNC programs, I brought the main block back to the mill to finish machining it.

I first took a light pass with a facing mill to clean up the top of the block and achieve a smooth finish for the gasket. I then spotted all of the holes and drilled them with their proper tap drill sizes (0.103" for 6-32 UNC and 0.150" for 10-24 UNC). With that completed I put the block into a vice and tapped all the holes to their proper size.

Step 7: Machining the Mounting Arms

The mounting arms are machined out of 1/8" thick aluminum, preferably flat stock. However, I had a bit of scrap extrusion and so I machined mine out of that instead. Both methods would produce the same result.

The zero for the mounting arms is also at the bottom left corner, just like the main block. Once the arms are machined I broke them out of their retaining tabs and filed them smooth. The mounting holes to attach the arms to the main block were then countersunk to accept the 6-32 screws.

Step 8: Cutting the Gasket

This step is optional, as a gasket is not really necessary. Some silicone sealant would be more than adequate for sealing the block, but having a gasket allows the block to be disassembled later on and it looks a lot better than a bunch of silicone.

I decided to use ordinary dollar store craft foam to make the gasket for a number of reasons. It is a rather soft material, and just thick enough to allow it to compress and contour to match the shape of the block and the cover plate, achieving a tight seal. It is also readily available, easy to work with, and rather inexpensive.

Pressing the top of the block into the craft foam creates an indent in the exact shape of the block, and I used this outline to cut out the gasket. It's a lot easier than trying to make a template from the block and transfer the shape, and by using the block itself to mark out the cuts there is a much lower chance for error.

Only the reservoir and four corner holes need to be cut, because the smaller 6-32 screws don't pass through the gasket, so it is not necessary to cut out holes for those. Once the gasket is cut out, I placed it on top of the block to double check that everything lines up.

Step 9: Block Assembly

Now that all of the parts are made, it's time to assemble the block!

I started by cleaning off all of the parts to make sure that there won't be any contaminants in my block. Once I was satisfied that everything was clean, I attached the mounting arms with the countersunk 6-32 screws. After those were mounted I aligned the gasket and the clear cover on top. The 10-24 screws were then used to secure the cover, and the fittings were the last to be threaded in. Refer to the exploded diagram in step 2 for the full assembly configuration.

Step 10: Leak Testing

Hook up the block to a standalone water loop, away from any electronics and preferably in a bucket to catch any potential leaks. I put mine in a large salad bowl on a piece of paper towel so I could tell if it leaked at any point in time.

Let the loop run for at least 24 hours (the longer the better) to ensure that there are no leaks in the block.

Step 11: Mounting the Block

First off, before the PCMR community starts pulling their hair out and posting a bunch of comments, I know that this is a stock Intel motherboard and there's no point watercooling it, but I'm just using it as a model and I'm not actually installing a loop on this board. With that out of the way, let's mount the block!

Fit the 8-32 screws through the mounting holes on the motherboard. Apply your preferred thermal compound and then slide the block over the screws. The screws should fit nicely through the slots in the mounting arms. Thread on the hex nuts until they just barely touch the top of the mounting arms, then snug them up finger tight at opposite corners. Ensure that there is even pressure on the CPU socket and that the block is sitting flat on the surface of the CPU. The block should be tight enough that it doesn't move, but not so tight that it flexes the motherboard and/or the mounting arms.

Eventually I might make a proper backplate for the CPU block, but it's good enough for now. If I ever do make one I'll update this instructable with the necessary steps.

Congratulations, you have just completed your very own custom waterblock!

Please feel free to post any questions or comments that you might have down below.



  • Sew Tough Challenge

    Sew Tough Challenge
  • Barbeque Challenge

    Barbeque Challenge
  • Stone Concrete and Cement Contest

    Stone Concrete and Cement Contest

32 Discussions


Question 1 year ago on Introduction

Sorry for resurrecting a 2 year old thread, but if possible could you please make a schematic for the mounting arms for the LGA 771 Socket J? Im trying to watercool my workstation and its proving rather difficult.


1 year ago

Dang! You nailed it! ;) I logged in just to tell you that. Absolutely awesome. I was looking for mounting bracket for my Corsair Hydro Series H100i v2 Extreme 240mm aio, started even looking for some CNCd ones and I found this. Well done, sir! :D

1 reply
tinker botfawkulce

Reply 1 year ago

Hey Fawkulce!

Thank you very much! I'm glad you like it


2 years ago

Hey, can you make a DIY GPU Waterblock too ? (that covers the whole videocard, not just the chip)

1 reply
tinker botniki292002

Reply 2 years ago

Hey niki292002!

Thanks for your comment! Unfortunately I'm fairly busy in the foreseeable future so complex and time consuming projects such as this and GPU waterblocks will most likely be few and far between. Also, unlike CPU sockets, GPU board layouts vary widely between manufacturers, card models, graphics processor series, etc. making it rather difficult to make a universal block that would be compatible with most or all GPUs. However, if I ever get around to making one I'll be sure to make an instructable on it and post it here!


2 years ago

Hi tinker bot!

I'm starting my graduation project here in the Swedish equivalent of high school, which will include the making of a DIY waterblock. Is it okay if I use your blueprints as a baseline? Unfortunately I can't use all of it due to the difference in standards and units, we're stuck with meters here haha...

Anyways, it was a good read and definately one of the most professional looking blocks I've seen on the net thus far. I hope I'm able to make something like this, since school provides me with all CNC-machines and tools necessary.

Would I be good off with a copper block instead of aluminium? I'm concerned about the corrosion. Any tips on radiator/pump combos - this won't be the fanciest of loops.

If I come across something that doesn't make sense to me, would it be possible to ask you about it?

1 reply
tinker botFredrikN4

Reply 2 years ago

Hey FredrikN4!

Thanks for your comment! I would be honoured to be a source of inspiration for your graduation project. Feel free to use any part of my instructable in your project!

If you have the materials, I would definitely recommend building a copper block rather than an aluminum one, as copper blocks and radiators are more common in the watercooling community, giving you a larger selection of accessories once your block is built and in use. Also, since you are designing a brand new block from scratch I would suggest that you use the smallest cutter you can get your hands on and mill very thin channels and fins to maximize surface area in your block. That would increase the transfer rate of heat between your block and coolant, making your block more effective. Just make sure you use a high spindle speed and keep the block well flooded with coolant when machining it!

As for radiators and pump combos, any pump with a decent amount of flow rate and a reasonably sized radiator should be able to cool a system. There's a guy on YouTube called JayzTwoCents who does a lot of watercooling videos. Take a look if you want any useful tips and pointers!

If you come across anything that doesn't make sense, would like clarification on something, or just want to talk CNC machining, feel free to PM me!

Also, please comment a picture of your block once you're finished. I would love to see how it turned out!


3 years ago

Hey man this is awesome. I'm planning to make a CPU and two GPU blocks as soon as I finish enclosing my CNC so that coolant doesn't spray all over my work room. Any tips on milling speed, cut depths and spindle RPM? I tried to mill some wood and ended up setting it on fire so I have rather low confidence in my choice of speeds and feeds.

2 replies
tinker botSuperfish1000

Reply 3 years ago

Thanks, Superfish1000!

Feeds and speeds depend greatly on the machine that you are using, cutter specs and material. The more rigid your machine and the better your cutter the faster you can go. What I usually do is just put everything into an online feeds and speeds calculator and take the values that it gives me. Sometimes I turn them down a little bit just to be on the safe side, but the original values are usually pretty good. Once you have the program made and the machine running, you can also watch how the machine's behaving and make any fine adjustments based on that.

For your wood project, you probably had too low of a feed rate. When I cut wood I usually keep a high spindle speed and a high feed rate, and then run shallow passes. For reference, I had a part that I made with a 1/4" HSS end mill running at 4500 rpm (max rpm for my machine), 55 ipm feed rate and a 0.1" cut depth.

Here's the feeds and speeds calculator that I use:

Good luck on your water blocks! Please post a picture when you're done. I would love to see how they turn out!

Superfish1000tinker bot

Reply 3 years ago

My feed was definitely too slow. My max feed is 16"/min at the moment. Going to try and fix that but for now it's way too slow. I need to key the shafts on my steppers and switch them to 1:1 stepping instead of 1:8. I don't need 0.0012mm accuracy. I mechanically don't have anywhere near that. I think my first major mill will be replacement parts for my CNC to fix it's various flaws and not a water block.


3 years ago

This is the most professional looking DIY waterblock I've seen and a giant improvement over Linus's!! lol I plan on using this design for my custom loop instead of buying a $80 EKWBs one. but I guess it's up to me to design a GPU block

1 reply
tinker botschildress4

Reply 3 years ago

Thanks, schildress4!

That sounds great! Please post some pictures when you're done, I would love to see how your blocks turn out!


3 years ago

I hope you did take the Aluminium only for practice. You have such fantastic tools , why don't you try it with copper. Now you have a 1st generation design which was outdated 8 years ago. I bought a remnant of that but in copper for 8€ on Ebay. And you can get that aluminium crap for 10$ on alibaba but better made.

Thin fins are only for air cooling, for water it must not be taller than 1:3 (width to height).

Try to duplicate a watercool heatkiller with a copper baseplate, the most demanding would be the milling of the very thin channels into the baseplate.

PS: never use acryl it looks good but cracks unpredictably, good when you buy the used 1st generation cooler with cracks for change and can make a new cover in 2 hours.

2 replies
tinker botHerbertD1

Reply 3 years ago

Hey HerbertD1!

To be perfectly honest I made this block more for the fun of making it rather than to actually have a functioning water block. The reason I used aluminum was because I had a rather large amount of scrap aluminum that I could use, so that decision was made mainly out of convenience. If I were to use copper, I would have had to order and have it shipped in, which would be significantly more expensive.

Thin channels could be machined without much difficulty. All that is required is a smaller end mill with a fast spindle and a lower feed rate.

I didn't use acrylic for the clear cover on my block. It is made of polycarbonate plastic (makrolon/lexan), which is the same stuff used in safety glasses. I mentioned this in the materials list, as well as the step regarding the clear cover and the applicable drawing. I chose polycarbonate specifically because it has more flex than acrylic, allowing it to take much more stress without cracking or breaking while still being stiff enough to compress the gasket enough to maintain a good seal in the block.

HerbertD1tinker bot

Reply 3 years ago

>>Thin channels could be machined without much difficulty. All that is
required is a smaller end mill with a fast spindle and a lower feed

I thought Copper is a pain in the as for milling. I got copper from the local scrapyard, they only charged the double market price for raw copper.


3 years ago

Thanks for including all the documents! Fantastic work! Nothing like making high end stuff for yourself ;)

2 replies
tinker botevrenucar

Reply 3 years ago

Thanks, evrenucar!

Are you going to make one yourself? If you do, please post a picture here! I would love to see it if you do!

evrenucartinker bot

Reply 3 years ago

I just wanted to examine it and take a look at your design procedure. I`m into industrial design and stuff. So It`s quite interesting to see it all together! Thanks again...


3 years ago

Looks really good man, but I wonder if it would be worth it, they say that making your own block doesn`t really give that low temps. Was tempted myself to do it, when I saw Linus` video, but after he said it isn`t really good, I stopped thinking about it. Do you have temps while under stresstests?

Either way, looks good man

1 reply

Reply 3 years ago

"LOW" is relative.

a well made and designed homemade water block will equal or better a commercial block. Remember, water cooling is just a deferred and relocated form of air cooling.

Instead of big blocks of aluminium or heat pipes, we are using water/coolant to move the heat to a remote location for easier and quieter dissipation. if the rest of the system isn't designed to cope with the total thermal load, it doesn't matter how good your block is. Likewise with big enough tubing, fast enough pump, and large enough radiator/fan, even a mediocre block will keep things fairly well under control.

My last water cooling block was mainly a copper plumbing cap soldered onto a copper plate. No internal baffling, no fancy flow rate calculations. 120mm fan/radiator and a good strong pump kept me at 5 degrees F above ambient pretty much forever, regardless of cpu load or overclocking. A too-small pump, in a friends rig, with $300 worth of commercial water cooling gear and he could barely surf the internet without overheating.