[Review] PrimoChill Typhoon III

Quote:

---

Originally Posted by Bojamijams

I finally looked at this thing and I'm having some trouble believing the test results here. 3.47GPM?

I have a hard time believing that it can do double of what the other two can...

---

I think you're probably reading this wrong. The flow rate is not 3.47GPM through any of the components in the loop. No water block sees 3.47GPM, that's the flow rate through the pump only.

It's not a magical feature of the T3. If you did the same with Y's and a D5 or DDC3.2 (or any other pump) you would have a very high flow rate through the pump as well. And if you did this with Y's then the flow rate going from the pump to the first Y would be exactly the sum of the flow rate in the two parallel loops. So adding flow rate in the loop loops is perfectly valid. If you look at a Y like this:
Attachment 97800

then the flow rate in 1 (coming from the pump) must be the sum of the flow rate in 2 plus 3 (going to the parallel loops). It's called the continuity principle and is very well known in physics. If the flow rate of 1 was less than the sum of 2 and 3, it would mean that more water was leaving than entering the Y, which obviously wouldn't work.

I agree about that alachesu but the charts are quite misleading then... it shows it giving double the flow of the other tops.. now the other tops are going to have that flow all throughout the loop.

3.47 will never happen throughout the loop, just at the exit and then immediately split it in half (ie. performance through the loop will be exact same as all other tops)

Except now you're using the warm loop to heat up the cold loop..

so.. what am I missing here? I don't see the point

Yes, I agree with you there. Putting those "through the pump" numbers in the same chart is comparing apples to oranges. The flow rates of the parallel loops are the ones that matter, and they're not much better than the series configurations.

Quote:

---

Originally Posted by Bojamijams

I agree about that alachesu but the charts are quite misleading then... it shows it giving double the flow of the other tops.. now the other tops are going to have that flow all throughout the loop.

3.47 will never happen throughout the loop, just at the exit and then immediately split it in half (ie. performance through the loop will be exact same as all other tops)

Except now you're using the warm loop to heat up the cold loop..

so.. what am I missing here? I don't see the point

---

Check the flow for each flow branch. You actually get a flow increase for the CPU/GPU branch compared to the serial setup. The T3 is basically a convenient way to run parallel loops with a slight tweak to improve high flow performance of the pump.

And I agree, the total flow bar is some what misleading, but if you actually think about what you are looking at I'm sure most people can figure it out :)

Quote:

---

Originally Posted by alacheesu

Yes, I agree with you there. Putting those "through the pump" numbers in the same chart is comparing apples to oranges. The flow rates of the parallel loops are the ones that matter, and they're not much better than the series configurations.

---

That is true. The T3 is not a miracle, it simply allows you to run parallel loops and obtain similar results to a dual loop without the need for a second pump and res. Also as most say here, the water temp doesn't change much more than a degree anyways.

Quote:

---

Originally Posted by millertime359

That is true. The T3 is not a miracle, it simply allows you to run parallel loops and obtain similar results to a dual loop without the need for a second pump and res. Also as most say here, the water temp doesn't change much more than a degree anyways.

---

I agree except for the dual loop part. Heat separation between the two loops is the number of one reason for running dual loop, and parallel loops does not offer that.

Quote:

---

Originally Posted by alacheesu

I agree except for the dual loop part. Heat separation between the two loops is the number of one reason for running dual loop, and parallel loops does not offer that.

---

That is true, but if you are looking for lowest temps possible on CPU, I would think a Chiller would be a bit better.

I keep referring to Norris as he is the only I have seen post results after doing it. He had a dual loop with DDCs and XSPC tops IIRC. He added in a T3 and say a degree improvement over his GPUs and same temp on the CPU.

IIRC, it was I7 920, full board block, and tri-slied 280s. Now he was also running a 320 rad in both loops. It was a XSPC or TC or something, forget the exact brand.

Now it was real world test, but it shows that you can still get similar temps.

As it has been said though, the T3 is not for everyone, but it is a nice option to have if you are a D5 owner. :yepp:

Quote:

---

Originally Posted by Bojamijams

Why would you want to see the numbers? Obviously they would be slightly lower then a stock D5 would be... everyone here can agree to that without me even doing the test. Y fitting do not increase flow (clearly). So what would it prove?

I'm theorycrafting here obviously, but it looks like two single loops were measured and their GPMs added.. which would be wrong.

If thats not what happened then can someone explain the 2x increase over a standard/custom top? The numbers seem improbable.

---

I DID NOT TEST SINGLE LOOPS AND JUST ADD THEM!!! :shakes: I called this out in the initial review...

Quote:

---

Originally Posted by skinnee

Note: Chart and graphs contain measurements for each split loop on the Typhoon III, the Lower and Top Inlet/Outlet flow rates are not a single serial loop. The Lower and Top Inlet/Outlets were measured with two flow meters, one for each Lower and Top Inlet/Outlets.

---

Both legs were running and measured, similar to how the PQ testing was performed.

Whats the best way to show flow rate for parallel legs, subtract them, average them... bake them in a pie? :shrug: Taking the flow rate of each leg and adding them together is the best way to show "System Flow". I've seen a lot of pissing and moaning regarding this, but no one has offered up a better way.

And in hindsight, the charts showing the System Flow rates were not ideal... I fully take the blame on that, live and learn I guess. Additionally, there was some additional testing on the Y'd loops.


http://skinneelabs.com/Pumps/SysLoopFlow-table.jpg

http://skinneelabs.com/Pumps/SysLoopFlow-Loop1.jpg

http://skinneelabs.com/Pumps/SysLoopFlow-Loop2.jpg

http://skinneelabs.com/Pumps/SysLoopFlow-Loop3.jpg


As I stated in the original T3 review, the Single Loop (TITO used in the charts) performance on the T3 isn't earth shattering and does not match the DDC w/ XSPC V3 top or PQ boasting D5 tops. Parallel in general shows how seperating restriction from a serial loop to two legs can boost flow rate for each individual leg, but there is quite a bit of testing that needs to be done to either support or rule out the level of balancing. From what I am seeing on a purely flow rate perspective it doesn't appear that the balancing or planning is as crucial as we once thought.

There is another user who was experimenting with the T3 and some flow meters calling my data fabricated based on his findings, what he doesn't realize is that the flow rates he is seeing actually validates the PQ data and curves. Furthermore his experiments validates that poorly balanced legs can and will effect loop performance potentially decreasing flow to one of the legs and raising temperatures. Again, my tests and reviews are never a be-all end-all nor perfect... each test I run something new is learned and thats why I got into testing in the first place.

The biggest thing to take away here is actually quite simple... with the right components and loop setup, the T3 offers higher flow rates through the components than any other single pump+top can offer, aside from the Iwaki's of course.

And Bojamijams, no worries I'm not pointing a finger at you, just wanted to reiterate the method and findings.

i am sorry but you have to be a ghey penix before you review a product and flame it before you even lay hands on it to have any credibility (I call it Andy-style).

All pictures are gone.. :(