Stren's Titan Water Block Roundup

[stren]

Heatkiller block finally arrived yesterday though the box is pretty beaten up. Hopefully the block isn't damaged. 2.5 Weeks in the hands of DHL will do that I guess. Results coming this week.

[LeoAndrade]

Stren, sorry if I missed that at some point, or if that it is kind of an off-topic, but... have you ever tested the difference in performance (flow restriction and deltaT) between changing the flow direction inside of the waterblock?

[stren]

Stren, sorry if I missed that at some point, or if that it is kind of an off-topic, but... have you ever tested the difference in performance (flow restriction and deltaT) between changing the flow direction inside of the waterblock?

Hey Leo - I'm not planning on it because I don't expect any noticeable change. Much of the restriction should be from the core itself which is symmetrical. The only slight difference is the order. Either the cooler water hits the core first or the VRMs first. Either way the VRMs will make little difference to coolant temperature I would have thought. For testing purposes I always have the cooler water hitting the core first.

Again this is all speculation and not backed up by any testing

[stren]

Well I got the Swiftech block in:



Internals are the same as the hydrocopper, but externals are classier:



Accessories



The insert plate has a brushed aluminum style finish and looks way better than the tacky hydrocopper insert:





I'd still prefer a one piece top, but this is definitely an improvement over the hydrocopper. The base is chrome plated (not nickel):



The bridge no longer says evga and the swiftech logo glows green instead of red



The edge also has stripes. Overall quite an improvement and $20 less for the swiftech version compared to the hydrocopper. You also get a backplate, but don't get included compression fittings:



I do think the logo is too big though, and would have looked better if it was etched out of the Aluminum



It is pretty meaty though.

I also got a package from watercool:



DHL took nearly 3 weeks to deliver and it arrived in perfect condition lulz



Luckily it was packaged well so everything survived:



The block was in a protective bag, but I removed it for the unboxing video:



Pre cut thermal pads and socket cap screws:



Hole edition:



Plastic standoffs, actual thread is in the copper though:



bridge:

[musicfan]

Thanks Stren. Great story-telling with excellent photos. Glad Watercool packages their contents well - that DHL photo made me cringe.

[stren]

Thanks musicfan Got the first heatkiller run done - performs solidly in the middle. I think the vram temps are good because the vram pads were so thin - the VRM pads on the other hand are thicker.





I added Koolance 0.7mm thick pads for the inductors/sense resistor type things as these weren't included. Now that I've completed all the blocks that have been shipped, I think I'll start on the 2nd/3rd runs. For the 2nd run I will keep the thermal pads on the inductors, but add on my new temp sensors to the back of the board for more accurate vrm/vram measurements. For the 3rd run I will remove the pads if they weren't instructed to be there. This way we get three good core runs to average, one good run of vrm/vram data with extra thermal pads and one good run of vrm/vram data with thermal pads as instructed

[LeoAndrade]

Congrats for the hard work Stren! Much appreciated.

I've just bought myself a aquacomputer block, and it looks fantastic! Will be installing it next week, I'll share the results here.

[stren]

Got some time to do the restriction plots:



I'm surprised by two things:

1. Why is the swiftech block so different to the EVGA? I wonder if this is a bad block - I haven't done thermals on it yet. Internals should be the same as the EVGA.

2. Why is the AC block pressure seemingly tailoring off at high flow? I've never seen that happen in any restriction plot?

For reference here are curves from other blocks and rads plotted on the same scale axis:





So in other words for restriction rads are generally the least restrictive, and some of the GPU blocks are similar in restriction to some of the cpu blocks

[LeoAndrade]

The problem with the AC graphic should be only measurement/equipament error. The curve varys with the flow squared (constant*Q^2), and simply there's no way for that to happen... probably if you get more data the graphic will show as it is expected to be.

[stren]

The problem with the AC graphic should be only measurement/equipament error. The curve varys with the flow squared (constant*Q^2), and simply there's no way for that to happen... probably if you get more data the graphic will show as it is expected to be.

Yeah agreed, shouldn't be the manometer going weird because it was fine afterwards, maybe a hose was working loose and was starting to leak. I'll retest when I get a chance. I'll take the komodo apart to see whats going on, Gabe confirmed it should be the same as the hydrocopper.

[Martinm210]

Got some time to do the restriction plots:



I'm surprised by two things:

1. Why is the swiftech block so different to the EVGA? I wonder if this is a bad block - I haven't done thermals on it yet. Internals should be the same as the EVGA.

2. Why is the AC block pressure seemingly tailoring off at high flow? I've never seen that happen in any restriction plot?

For reference here are curves from other blocks and rads plotted on the same scale axis:





So in other words for restriction rads are generally the least restrictive, and some of the GPU blocks are similar in restriction to some of the cpu blocks

#2 Likely due to "flex" in the block. I have seen this in rads like the HWlabs GTX series and a few others. As pressure increases the chambers/tubes whatever expands causing the normal curve to flatten. GTX rads use very thin and flat and restrictive internal tubes. Som block tops probably just don't have as much bolt restraint or the thin copper base could be flexing. It happens..

Great work as usual. Hope be done moving in the next month so I can do more playing as well. Now the proud owner of three mortgages..

[Mobsters]

Anyone knows what the difference is with EK FC Titan and EK FC Titan SE blocks? Are they same with performance?

[stren]

Anyone knows what the difference is with EK FC Titan and EK FC Titan SE blocks? Are they same with performance?

Should be similar, my SE is arriving soon so first results should be in a few days after it arrives

[Razor Time]

Anyone knows what the difference is with EK FC Titan and EK FC Titan SE blocks? Are they same with performance?

The internals are slightly different around the VRM area and the fins are meant to be finer. I have an EK SE block on my GTX780 and so far I am pleased with the results.

[stren]

The internals are slightly different around the VRM area and the fins are meant to be finer. I have an EK SE block on my GTX780 and so far I am pleased with the results.

This, despite the finer fins restriction looks similar, though I don't have the plot yet. Here are some pics:

SE vs Vanilla:





SE in nickel plexi covers the entire pcb much like the vanilla XXL did

[m_jones_]

Keep up all of the hard work, not the biggest fan of the plexi overhang on the EK block.

[stren]

Keep up all of the hard work, not the biggest fan of the plexi overhang on the EK block.

Thanks It does look better in real life than the photos, and it has LED holes too, so it might look better when lit up. I do miss the fish though

[mojododo]

hi stren, great write-up!

Silly me grabbed a swiftech block before searching the net for comparisons! Someone might have an answer for me, i went from 2 ek gtx580 blocks in paralell @ 4.2 L/min and now have one swiftech titan block @ 2.6 L/min. I didnt think the swiftech block would be that restrictive! Anyone got any thoughts?

[Church]

Hmm, somehow i dig looks of new EK block. AC's still best look wise, but this is unexpectedly close. Damn, how i wish old EK waterbridges back though.

[stren]

hi stren, great write-up!

Silly me grabbed a swiftech block before searching the net for comparisons! Someone might have an answer for me, i went from 2 ek gtx580 blocks in paralell @ 4.2 L/min and now have one swiftech titan block @ 2.6 L/min. I didnt think the swiftech block would be that restrictive! Anyone got any thoughts?

The swiftech should be very low restriction, similar to the hydrocopper restriction data I took earlier in the thread. I'd expect it to be similar to the old GTX580 blocks, but because you had 2 in parallel vs one now therefore you should see restriction increase. However that sounds like a bigger change than I would expect, having said that, it does depend what else you have in your loop. I wonder if your swiftech block is running at much higher restriction like my sample is. I just took mine apart and I don't see anything wrong with it, but I'll send Gabe some photos to see if he can see anything wrong. If not I'll take the hydrocopper apart to compare.

Hmm, somehow i dig looks of new EK block. AC's still best look wise, but this is unexpectedly close. Damn, how i wish old EK waterbridges back though.

I know those old bridges were much sexier, still I feel like they're on a upward trend in terms of style

EK SE (second edition) data:

The finer fins in the cooling engine did not alter the poorer performance at low flow. However I was able to get a better TIM spread and at very high flows the block now performs at the top alongside the XSPC:



This time I didn't apply TIM to the VRMs/VRAMs which may explain the slightly poorer performance. However the VRM section design did change.



Still to do:

- verify Koolance performance when you don't warp the card
- stick my new temperature probes to the back of the card to log VRM/VRAM temps more accurately
- run the EK SE vs flow (expected to be similar to the older EK block based on flow numbers)

Still waiting for
- Bitspower sample (Redesigning I think)
- Alphacool sample (OOS)
- Phobya sample (Not even sure if this is still happening)

[Jhors2]

Stren,
This is outstanding work, solidified my choice in an EK block for some GTX 780s. Any chance you can post pictures of the setup? Of the EK block with some LEDs? I know this is a performance thread but I'm just curious.
Thanks!

[stren]

Stren,
This is outstanding work, solidified my choice in an EK block for some GTX 780s. Any chance you can post pictures of the setup? Of the EK block with some LEDs? I know this is a performance thread but I'm just curious.
Thanks!

Hey - I've actually taken the block off already in order to check the TIM spread was good. I'll take some when it's on again for the next run. I have some of the XSPC blue LEDs as well as some red ones from Monsoon that I can try.

[Razor Time]

Thank you for posting the EK-SE results Stren. Makes me pleased I went for one.

[Church]

stren: BTW, what do you think of testing out something like this? I liked a lot idea of universal block that fixes most of cons keeping pros (still universal, as cooling plates are exchangeable and cuttable as needed, bridge-edition style enables usage of water bridges enhancing looks, and users get also vrm liquid cooling unlike with other universal blocks). I don't like that i was unable to persuade other vendors with potential of such product though

[stren]

Thank you for posting the EK-SE results Stren. Makes me pleased I went for one.

Anytime

stren: BTW, what do you think of testing out something like this? I liked a lot idea of universal block that fixes most of cons keeping pros (still universal, as cooling plates are exchangeable and cuttable as needed, bridge-edition style enables usage of water bridges enhancing looks, and users get also vrm liquid cooling unlike with other universal blocks). I don't like that i was unable to persuade other vendors with potential of such product though

Yeah it's a good idea, do they have a titan/780 version though I didn't see one? I'm willing to test whatever the manufacturers are willing to send I do think those mini tubes are a bit ugly, but it's a nice idea over just a universal block.