I believe Wahaha's version is a "forward" flow design, and he is well into the development, whereas I have just started. Nothing is confirmed on my side, don't mind share my ideas and it's ok for anybody to pick it up from here.
-
Save 15% on ALL SFF Network merch, until Dec 31st! Use code SFF2024 at checkout. Click here!
Production Aquanaut Basic/Extreme - Ultra Low Profile CPU Block & Pump Mount Combo
- Thread starter Nouvolo
- Start date
I just mean you both are going for the open loop CPU pump with sff considerations in mind.
If collaboration isn't your thing, that's ok. I love having more options. I suspect that these products are going to open a new level (open loop water cooling) for more people.
Actually I am open to any possibilities, but I will carry on at my own pace at the same time. Will certainly take a look if opportunity comes along ?
EK-SPC (left) vs Syscooling P67a (right)
Impellers are different : P67a has ridges, whereas EK-SPC is smooth (but ridges are in the pump top instead...)
Pump cavity : diameter 44mm vs 42mm of ddc
Conclusion :
Pump top piece : ddc and spc specific (i.e. 2 ver)
Cold plate piece : ddc/spc share common piece
Update: damn, just going to get a P60a also to finalize dimensions for SPC pump top.
Impellers are different : P67a has ridges, whereas EK-SPC is smooth (but ridges are in the pump top instead...)
Pump cavity : diameter 44mm vs 42mm of ddc
Conclusion :
Pump top piece : ddc and spc specific (i.e. 2 ver)
Cold plate piece : ddc/spc share common piece
Update: damn, just going to get a P60a also to finalize dimensions for SPC pump top.
I'm curious, are you going to design your own cold plate too or are you planning on designing your top so users can screw in coldplates from existing CPU blocks on the market?
I have sourced some cold plates and mounting brackets that are similar to EK Supremacy (please check earlier posts for photos). But I have not closely compared mine with EK's.
Although some have suggested that I can just provide the acrylic top that is compatible with other brands, I would prefer provide a "complete" product. Although with less items to worry about, there are disadvantages with providing the top only, e.g. after sales support for different cold plates, no control of others products (discontinuation).
It could be confusing for you all to decide whether this product is good or not, if there are different performance/results from different reviews, as each uses different cold plates, microchannels, fins sizes, (so many variables), etc. And to the point even I, as creator, would not know where to start to make changes to improve.
In short, I want to limit the "variables" in a product, just my design approach.
Additionally, don't most cold plates have different mounting holes patterns? They are not interchangeable. Some have fins protruding off a flat plate (skieved fin I believe it's called) and some have fins milled into a thicker plate. I don't think it's possible to have compatibility with more than one at a time?
Yes, if I were to go "just the top" path, I can only support 1 single water block out there, and it will be the most common one only. Then as said, it will be at the mercy of that water block brand, that totally not ideal situation for me, if I have to keep stock of this. (maybe not such big problem for users point of view ?, but hey don't leave me out in the cold)
Development Update
I can confirm the following :
2 types of tops : DDC/VTX & compatible, (EK) SPC [only, not compatible with Syscooling P60a/P67a]
2 types of mount brackets : Intel 115x / AMD AM4
Prototyping : 80% complete
Outstanding :
- CNC demo for final measurement and fitting test. Held up due to lockdown
- Actual flow rate tests
- Potential design adjustment? (or scrapping the project)
Sourcing :
[Done] O-rings, cold plates (+jet plates), packaging, screws(M4)
[Outstanding] CNC tops
Bulk production : dependent on completion of prototyping
I can confirm the following :
2 types of tops : DDC/VTX & compatible, (EK) SPC [only, not compatible with Syscooling P60a/P67a]
2 types of mount brackets : Intel 115x / AMD AM4
Prototyping : 80% complete
Outstanding :
- CNC demo for final measurement and fitting test. Held up due to lockdown
- Actual flow rate tests
- Potential design adjustment? (or scrapping the project)
Sourcing :
[Done] O-rings, cold plates (+jet plates), packaging, screws(M4)
[Outstanding] CNC tops
Bulk production : dependent on completion of prototyping
Because they are so few options regarding pump mount combo this is hugely appreciated! Looking forward to the final prototype. I hope that they will be more colors than just orange ?
Orange is just the 3D printing material for the prototype, final product should be in common transparent acrylic or POM.
Great work!
I'm interested in which pump gives us the lowest clearance. Will any combo get a person close to the 50mm mark in height?
I'm interested in which pump gives us the lowest clearance. Will any combo get a person close to the 50mm mark in height?
I just responded on my thread, tldr, can't do it at the current stage, I'm working with two other company in water cooling already, so it's too far down the road.
Never say never right, so if things on that path doesn't work out, and I'm unlocked form previous commitments, then I will reach out.
There is a concern though, neither of us has the resources for leak QA and mfg, I consider myself a newbie in water-cooling parts manufacturing. So any sort of partnership really need some third party with much experience, or else this product is going to ruin people's motherboard and CPU, this is not some Distro Plate or Reservoir far from your key components, a leak is going to be really expensive.
That is true, also reason why I don't want to raise expectation too high, though I wouldn't take this on if I don't feel I could overcome those challenges myself.
So far progress is fine and has now got to the stage of being stalled by the lockdown. Need to personally visit some suppliers and then we will see. There is also flow rate test which I don't want to make any presumption at this stage, best to carry out real life tests with actual prototypes (CNCed not 3D printed).
I was thinking about how you might certify the units against leaks, and in many industrial applications it usually boils down to a pressure test on a unit by unit basis. Understandably, this becomes tricky for a product like this given the multiple pump blocks and vast number of potential fittings, not to mention the labor involved in having to test every unit.
I wonder if a large amount of hassle could be avoided if specific fittings were included such that a standard test stand with quick-disconnects could be used to rapidly certify units in a parallel fashion. This wouldn't preclude the user from switching fittings at their own risk, and would certainly increase the unit material cost, but might end up breaking even from a labor/time-cost perspective during the QA process.
It will be impossible for me at this stage to arrange 100% testing. I wonder if others really carry out 100% testing. It is water leakage test, and not something that can be tested with some electronic probes & software e.g. smartphones. The added cost for manual labour testing each and every product, I have no idea how they (EK and others) can achieve it economically, unless there is some secret sauce to it.
Then again, if all the parts are CNC'ed the risk of defects is small. I wonder if it is about taking enough samples from the bulk to get to say 95% confidence level (statistics). Even testing 100% does not mean zero risk of leakage for the batch, testing is manual process, some lousy workers can make matters worse.
My understanding in the industrial cooling sector, is that the testing is done with air pressure and not water. Basically, they seal the outlet, and then pump air into the component though a pressure gauge until reaching 3-4x the standard water pressure. Then they seal the inlet of the pressure gauge and make sure that the component can hold pressure for a couple of minutes.
In the CPU block space, I can't say whether they test 100% of the units. My intuition says that it's unlikely from a "scale of production" and "cost minimization" standpoint, however the trade off would seem to be how much risk you're willing to incur from an RMA perspective. Maybe it's worth giving one of the larger commercial manufacturers who do "custom designs for small batch runs" a call and see if they're willing to share their QA/QC process. It's basically what I had to do to get any information on these topics in the power-electronics market.
Just as a though on testing... Could you seal up the input/output ports with stop plugs and instead of the pump just bolt on a blanking plate with an o ring, to fully seal everything. Then dump them into a tub of water overnight? It's like a reverse leak test. I don't know anything about how it's done in the industry but my thought process is if water does not leak in... It should not leak out?
Definitely an inexpensive partial solution. However, the issue with this is that during operation the fluid will be pressurized, which could result in leaks that your test wouldn't catch at ambient pressure.
Similar threads
