Prototype Free Flow CPU Pump Block

fabio

Shrink Ray Wielder
Apr 6, 2016
1,885
4,325
Update 2020-03-03 : Performance tested and it sucks ?, OPTIMIZATION NEXT!




From left to right:
1. DDC, ≈ 6.5 L/M or 390 L/H
2. DDC + IceManCooler Top, ≈ 9 L/M or 540 L/H
3. P5 (same as EK SPC), ≈ 4.5 L/M or 270 L/H
4. P5 + IceManCooler Top, ≈ 4.5 L/M or 270 L/H
5. IceManCooler M1 DDC Pump Res, ≈ 5 L/M or 300 L/H
6. * this project, ≈ 4.5 L/M or 270 L/H
7. * Alpahcool DC-LT should be ≈ 2 L/M or 120 L/H
8. * Asetek/NZXT/Corsair/Cooler Master/DeepCool should be ≈ 1.5 L/M or 90 L/H

For SFF builds running a *single radiator (120, 240, 280, maybe even 360), pump makes insignificant performance / cooling difference, noise is subject (see Optimum Tech video). DDC and D5 makes a difference on performance / cooling when using multiple radiator / block configurations.

I will optimize the water channels to try and get above 8 L/M or 480 L/H.
Great job man!! Seriously! Waiting for the next step! Keep going!
 
  • Like
Reactions: PBJ

Wahaha360

a.k.a W360
Original poster
SFFLAB
NCASE
SSUPD
Feb 23, 2015
2,131
10,697
What for? 4.5l/m is more than enough. I would be happy to buy it in the current state.

Pump Noise!

If you can turn the DDC to 25% speed, it will push out the same amount of water as a DC-LT / Asetek. The reduction noise will be pretty sweet while the temperature will be unaffected.

Also my OCD, the current flow rate is almost half as regular DDC + Pump Top, I'm aiming for 85% of the flow rate.
 

PBJ

Airflow Optimizer
Jan 6, 2019
358
546
Update 2020-03-03 : Performance tested and it sucks ?, OPTIMIZATION NEXT!




From left to right:
1. DDC, ≈ 6.5 L/M or 390 L/H
2. DDC + IceManCooler Top, ≈ 9 L/M or 540 L/H
3. P5 (same as EK SPC), ≈ 4.5 L/M or 270 L/H
4. P5 + IceManCooler Top, ≈ 4.5 L/M or 270 L/H
5. IceManCooler M1 DDC Pump Res, ≈ 5 L/M or 300 L/H
6. * this project, ≈ 4.5 L/M or 270 L/H
7. * Alpahcool DC-LT should be ≈ 1.6 L/M or 100 L/H
8. * Asetek/NZXT/Corsair/Cooler Master/DeepCool should be ≈ 1.2 L/M or 70 L/H

For SFF builds running a *single radiator (120, 240, 280, maybe even 360), pump makes insignificant performance / cooling difference, noise is subject (see Optimum Tech video). DDC and D5 makes a difference on performance / cooling when using multiple radiator / block configurations.

I will optimize the water channels to try and get above 8 L/M or 480 L/H.
Will you be using a DDC pump made specially for this product or using one already on the market?

If so, will it use a 3-4pin fan header?
 
Last edited:

Luke NK

SFF Lingo Aficionado
Nov 17, 2019
126
116
Pump Noise!

If you can turn the DDC to 25% speed, it will push out the same amount of water as a DC-LT / Asetek. The reduction noise will be pretty sweet while the temperature will be unaffected.

Also my OCD, the current flow rate is almost half as regular DDC + Pump Top, I'm aiming for 85% of the flow rate.

This is what I was thinking about ddc improving noise levels if runing at lowest speed. Thanks for going with this pump.

Will you test the efficiency of your block design against, say a Heatkiller IV block?
 
  • Like
Reactions: Sh0ckwaveFlash

flar

Average Stuffer
Jan 19, 2020
78
46
My options for my Ghost S1 seem to be the LT solo or the Apogee Drive II right now. I'm leaning towards the Apogee and thinking of swapping the pump on the Apogee for a lower powered DDC-310. I'd be more interested in how your block compares - flow-wise - to the block in the Apogee as I'd rather go with your block if it is in the same ballpark as the Apogee base.

My ideal flow rate comparison would be:

- Apogee Drive II (with its MCP35X)
- Apogee Drive II base with DDC-310
- Your block with the MCP35X
- Your block with the DDC-310

Looks-wise, I'd much rather have your block than the bottom half of an Apogee.
 

lehman

Trash Compacter
Jan 9, 2019
39
69
My options for my Ghost S1 seem to be the LT solo or the Apogee Drive II right now. I'm leaning towards the Apogee and thinking of swapping the pump on the Apogee for a lower powered DDC-310. I'd be more interested in how your block compares - flow-wise - to the block in the Apogee as I'd rather go with your block if it is in the same ballpark as the Apogee base.

My ideal flow rate comparison would be:

- Apogee Drive II (with its MCP35X)
- Apogee Drive II base with DDC-310
- Your block with the MCP35X
- Your block with the DDC-310

Looks-wise, I'd much rather have your block than the bottom half of an Apogee.

You should look at the EK SPC-60. it flows more than the DDC310 and is already set up to run off a fan header. Uses only 6 watts.

Lehman
 

flar

Average Stuffer
Jan 19, 2020
78
46
You should look at the EK SPC-60. it flows more than the DDC310 and is already set up to run off a fan header. Uses only 6 watts.

Lehman
That does look good. I think I saw it before but didn't really look closely at it because it is marketed in a configuration for a standalone placement with a reservoir and I've been looking at a pump/block combo. Taking another look at the plexi version I think I see how it could be separated and mated to a block. I'm pretty new to searching parts for a loop, so I miss things like this.

As it turns out, it cost only $5 to have PPCs rewire the DDC-310 to a 3-pin fan header when they added sleeving so that part was pretty painless. Total cost including sleeving and rewiring was less than the price on the SPC-60, but it's only 3-pin rather than PWM.
 

Wahaha360

a.k.a W360
Original poster
SFFLAB
NCASE
SSUPD
Feb 23, 2015
2,131
10,697
Update 2020-03-04 : Pump Flowmeter Test



Just tested Gen 6 Corisair AIO pump, it's the Asetek Gen 6 pump. Gen 6 and Gen 7 flow is the same. Less than 2 L/M on the flowmeter or at most 60 L/H.



So I made a chart for reference.
 

emadman

Cable Smoosher
Jan 31, 2020
11
9
@Wahaha360 - this is an exciting project. I'm looking for options to watercool my steck build and this will finally bring a DCC alternative to the apogee drive.

Is there a timeline that you're targeting for getting this out?
 
  • Like
Reactions: Allhopeforhumanity

gwertheim

King of Cable Management
Nov 27, 2017
938
1,555
Update 2020-03-03 : Performance tested and it sucks ?, OPTIMIZATION NEXT!




From left to right:
1. DDC, ≈ 6.5 L/M or 390 L/H
2. DDC + IceManCooler Top, ≈ 9 L/M or 540 L/H
3. P5 (same as EK SPC), ≈ 4.5 L/M or 270 L/H
4. P5 + IceManCooler Top, ≈ 5 L/M or 300 L/H
5. IceManCooler M1 DDC Pump Res, ≈ 9 L/M or 540 L/H
6. * this project, ≈ 4.5 L/M or 270 L/H
7. * Alpahcool DC-LT should be ≈ 1.6 L/M or 96 L/H
8. * Asetek/NZXT/Corsair/Cooler Master/DeepCool should be ≈ 0.5 L/M or 30 L/H

For SFF builds running a *single radiator (120, 240, 280, maybe even 360), pump makes insignificant performance / cooling difference, noise is subject (see Optimum Tech video). DDC and D5 makes a difference on performance / cooling when using multiple radiator / block configurations.

I will optimize the water channels to try and get above 8 L/M or 480 L/H.

 

flar

Average Stuffer
Jan 19, 2020
78
46
Update 2020-03-04 : Pump Flowmeter Test

Just tested Gen 6 Corisair AIO pump, it's the Asetek Gen 6 pump. Gen 6 and Gen 7 flow is the same. Less than 2 L/M on the flowmeter or at most 60 L/H.



So I made a chart for reference.
I'm confused about what we're seeing here. You have various pumps listed, some with tops, but not blocks. From what I've gathered (and forgive me if this is a Noob question), tops aren't really restrictive, but blocks can be. So, if you are running those pumps with tops alone and measuring pressure/flow they will likely perform quite well. Then you have 2 prototypes of the FreeFlow (== Fr33Flow? Has the name changed?) which is a block, but no mention of which pump is being used in those tests.

Since you are producing a low profile block to mate to a pump, the comparison that I would want to see is how your block compares to other blocks with an emphasis on blocks designed to mount a pump directly. Perhaps a standalone pump not driving a block included as a reference, but the apples to apples part of the comparison would be pumps mounted to blocks. I don't think that is what I'm seeing here, but perhaps I'm not familiar enough with standard parts and practices to understand it?

As far as I can see, depending on which pump you used on your block in the tests, your block is performing quite well already! If I had a comparison of that same pump on, say, the base/block of an Apogee, I'd have the information I needed to decide if I want to wait for this.

In my case, I'm only looking at pump/block combo solutions since I don't have an obvious place to put a separate pump/reservoir in the case. While I'm curious how the planned setup might compare to a non-space-restricted loop design, I really only care about how various pump/block combo options compare to each other with respect to what I finally install in my Ghost.
 

Wahaha360

a.k.a W360
Original poster
SFFLAB
NCASE
SSUPD
Feb 23, 2015
2,131
10,697
I'm confused about what we're seeing here. You have various pumps listed, some with tops, but not blocks. From what I've gathered (and forgive me if this is a Noob question), tops aren't really restrictive, but blocks can be. So, if you are running those pumps with tops alone and measuring pressure/flow they will likely perform quite well. Then you have 2 prototypes of the FreeFlow (== Fr33Flow? Has the name changed?) which is a block, but no mention of which pump is being used in those tests.

Since you are producing a low profile block to mate to a pump, the comparison that I would want to see is how your block compares to other blocks with an emphasis on blocks designed to mount a pump directly. Perhaps a standalone pump not driving a block included as a reference, but the apples to apples part of the comparison would be pumps mounted to blocks. I don't think that is what I'm seeing here, but perhaps I'm not familiar enough with standard parts and practices to understand it?

As far as I can see, depending on which pump you used on your block in the tests, your block is performing quite well already! If I had a comparison of that same pump on, say, the base/block of an Apogee, I'd have the information I needed to decide if I want to wait for this.

In my case, I'm only looking at pump/block combo solutions since I don't have an obvious place to put a separate pump/reservoir in the case. While I'm curious how the planned setup might compare to a non-space-restricted loop design, I really only care about how various pump/block combo options compare to each other with respect to what I finally install in my Ghost.

I don't have other pump/blocks atm, so the chart is just showing what I am able to test as reference.

With the Apogee Drive being a close alternative, and this BlockPump + DDC being more expensive, if the pump block design is half as powerful as a separate pump reservoir combo, this product won't be popular enough to even hit MOQ. So the product viability in some ways depends on it.
 
Last edited:

flar

Average Stuffer
Jan 19, 2020
78
46
Another comparison issue to account for...

Pretty much any AIO and the DC-LT were designed for silent operation out of the box. They have low flow because they aren't running very fast. But they are very quiet.

Nearly every review I've seen of putting a full sized (18w?) DDC pump in an SFF case mentions that it is very loud when at full tilt and that the reviewer dialed it back to about 30% speed to get it quiet. At that point, it is not going to produce the flow that running it full out in an empty loop will see anyway. If you reduce power on one of the DDC's listed in the chart, how does its flow compare to the DC-LT/AIOs?
 

Wahaha360

a.k.a W360
Original poster
SFFLAB
NCASE
SSUPD
Feb 23, 2015
2,131
10,697
Another comparison issue to account for...

Pretty much any AIO and the DC-LT were designed for silent operation out of the box. They have low flow because they aren't running very fast. But they are very quiet.

Nearly every review I've seen of putting a full sized (18w?) DDC pump in an SFF case mentions that it is very loud when at full tilt and that the reviewer dialed it back to about 30% speed to get it quiet. At that point, it is not going to produce the flow that running it full out in an empty loop will see anyway. If you reduce power on one of the DDC's listed in the chart, how does its flow compare to the DC-LT/AIOs?

DDC at 25% is more quiet imo and has a higher flow rate than DC-LT and Asetek variants.

Where it really matters is multiple radiators.
 

lehman

Trash Compacter
Jan 9, 2019
39
69
Some noise results from testing by VSG.

This first group is through a full loop. Raystorm Pro CPU block, a Swiftech Komodo R9-LE GPU block, a HWLabs Nemesis GTS 360 radiator and 2 sets of Koolance QD3 quick disconnects to go with 6 Bitspower 1/2″ x 3/4″ fittings .


I believe this is an Asetek pump.


DC-LT pump




Hope this helps some.

Lehman
 
Last edited:
  • Like
Reactions: Allhopeforhumanity