PrimoChill Typhoon III Laing D5 Reservoir System I will start today by thanking Brian from PrimoChill for providing the lab with our test sample. To say the Typhoon III is one serious example of "outside the box" thinking is almost an understatement. A few lab records were shattered along the way, and one of those was broken by Brian. Normally, I get the first (of many) emails asking when the testing and review will be done within 3 or 4 days of receiving the product (it takes a bit longer to do the type of extensive testing we do after all). Brian's record is never. Not one email. Oddly enough however, I think this made the pressure to get done even worse. I also want to thank Geno (aka: BoxGods) who designed the Typhoon III for making himself available to answer my plethora of questions. I am obviously fascinated by water cooling and having unlimited access to the designer provided a lot of interesting insight into the process, design goals, and technical details of the Typhoon III. Full Review Intro For those not familiar with PrimoChill, they have been actively designing and selling us water heads products since 2001. They have recently begun to push the envelope with several products including their PrimoFlex PRO LRT tubing, unrivaled in terms of bend radius and selection, as well as a full line of fluids and new polycarbonate compression fittings, which are also included with the Typhoon III. I first stumbled upon pictures of early prototypes of the Typhoon III (aka: T3) several months ago while planning Laing D5 tests. In my early discussions with BoxGods, he was very focused on the usability of the T3. Almost adamant, he went so far as to build a small acrylic test stand for me that held the T3 in the same location and orientation it would normally be in when installed in a case. What really stood out about the Typhoon III was the different approach BoxGods took. While most are producing a replacement pump top with maybe some inlet/outlet options, the T3 veers off in an entirely different direction. And from those early discussions it was clear the goal of the T3 was to make a pump and reservoir that was very easy for novice users to install, setup, and use...well noobs need love too right? One of the most touted features of the T3 is the ability to run two independent loops off the same D5 and reservoir. Any hardcore user knows how hopeless that is from a performance perspective, right? One forum poster even claimed he would rather go back to air cooling than run two Y split loops off one pump. In retrospect, I am quite glad I never posted any of the jokes about running parallel loops off the same pump that occurred to me when early images of the T3 began popping up because, while it may have been one of the designer's goals to build an easy to use setup for novices based on the venerable D5, he also stuffed the T3 with enough high flow magic to distill the term "hardcore water cooler's wet dream" down to "priapism". Back to the point. As you're going to see when you get to all the data tables and charts (especially in the dual loop tests), the T3 is going to make a lot of noise in performance circles. So much so that all those usability features (read noob goodness) are going to get lost in the commotion. With that in mind, I have split this review into two parts. The page you're reading now is all about performance. A second page discussing the usability features built into the T3, like the dual front fill ports, overflow basins, etcetera etcetera--is a must read for the novices. Now shoo, the adults need to talk about important stuff. Full Review Dual Loop Capability As I said earlier, every water head knows that if you take your favorite D5 pump top and add a "Y" fitting a few inches out from the inlet and another from the outlet port you are going to split the flow and pressure which means a big hit to your potential cooling. So how does the T3 not only get away with it, but excel at it? I will start with the unique pump expansion chamber, cleverly abbreviated PEC, which is directly adjacent to the pump head’s 360 degree deep-draw helix--close enough that it actually straddles it really. The PEC is proportioned and placed to take the water fed through the primary pump head outlet and let it "expand" before being fed out the back into either one or both secondary outlets. Although I am no mechanical or hydrodynamics engineer, I am told that in addition to splitting the flow evenly between the two ports (no easy feat, +1 T3) it works as a sort of pressure regulator. In order to increase flow into this expansion chamber the primary outlet is almost double what you find on other pump heads. So now we have increased the amount of water the D5 can push out with deeper draw and a larger outlet, and even given it a place to go that helps regulate (and maintain some of) the pressure. Again, you don't have to be an engineer to know that if more water is being pushed out then you need to draw more water in through the inlet or you starve the pump and a nosy neighbor calls pump protective services on you. Feeding the inlet is a large "tank" that is, in turn, fed by the inlets on the back of the T3 that you connect your loop or loops to. The problem is that the inlet diameter is precisely matched to the impeller in the D5 and any change in diameter kills performance. An area that has long been considered the biggest limiting factor on gaining any more performance out of the D5. At this point the discussion turns to laminar boundary layers and feed angles and stuff that Vapor and the engineering types wax on about at length. Fortunately for most of you, I barely understand it, so you do not have to read my poor attempt at explaining it. BoxGods tried to explain it in layman's terms for me by saying: "Although water is not compressible in the very limited pressure ranges we are working in, it CAN be redirected and moved in a more efficient manner similar to a venturi. I actually got the idea to try it from an article by Martin (from Martin's Liquid Labs =)) where he suggested (amazingly based just on observation) that although the impeller on the D5 was obviously optimized for maximum efficiency, the inlet feed angle didn't seem to be. In simplified terms this means that although the inlet diameter and the impeller are matched for efficiency, the perpendicular inlet angle dictated by the tubing type fitting is actually pretty bad." In the SolidWorks images below you can see that although the outlet is still the same size, it has been "humped" into an efficient hydrofoil cross section and the trailing edge is very sharp. In effect, a low pressure venturi that accelerates the water entering the pump so that a larger volume can pass through. It is also fed into the pump's impeller at a much more efficient angle. All of these tweaks on the inlet, outlet and PEC are balanced to provide the best possible increase in potential flow rate without sacrificing head pressure, and are the true genius behind the Typhoon III. The "Y" fitting method mentioned earlier happens before and after the pump. In the T3, the PEC and reservoir are essentially a part of the pump head, so much so that you can think of the loop split happening inside the pump, which is why you are going to start seeing the term "through the pump" a lot more often from now on in water cooling. Hopefully, with this explanation, you will no longer look at the T3 as just a bay reservoir with a pump attached to it. If you do, you are overlooking a major evolution to our previous LC loops. In the first frame of this SolidWorks image, the PEC is shown with its cover in white, so you get a better view of the relationship between the PEC and the pump head. In the second image, the white cover is removed for a peek inside, and in the third image a .30" diameter circle, shown in blue, demonstrates the size difference between the primary outlet on the T3 and a standard D5 pump head. You also can get an idea of the "humped" shape of the inlet and its razor sharp leading edge. The same area from the opposite side shows the positions of the in and out ports relative to each other and the primary inlet and outlet. Test Methodology Test methods and specifications have been removed from the product review pages to ease the torture my reviews inflict on your mouse wheel finger. You can find the test process, methods and explanations of how to read the charts and tables using the link below. - skinnee labs Pump Test Method / PQ Curves Test Results Okay, I have finished holding back the eye candy you crave from our reviews… Here are the compiled scatter plots for Typhoon III single loop--tested at Speed Setting 4 (D5 Basic) and Speed Setting 5 (D5 Vario). For the single loop configurations, I have included the stock impeller housing or top along with the comparable Koolance top with Front Inlet/Outlet configuration. A large portion of the testing was focused on the dual loop configuration of the Typhoon III and, to be candid, I don't think my scatter plots truly tell the story of how incredible the Typhoon III is with the dual loop configuration--but more on that after the single loop scatter plots. Below each scatter plot is the test data for each inlet/outlet configuration at the specified speed setting. The Performance Results Intro explains how to read the charts and the explanation of the data if you find yourself scratching your head. Setting 5 Top Inlet/Outlet - Setting 5 Lower Inlet/Outlet Setting 4 Top Inlet/Outlet - Setting 4 Lower Inlet/Outlet The single loop charts are nothing new if you have read pump reviews before, as I said, it is easy to get caught up in the performance of the dual loops configuration, but as you can see above, the T3 is competant in the single loop department too. Nothing to write home about, but no slouch either, comparable to the Koolance top or sticking with stock. Though I look at a single loop as someone just waiting to upgrade and go for dual loops. Running two loops off a single pump was taboo on the hard core, more serious cooling forums. Well, the Typhoon III changes the game up bringing the capability of running dual loops without kissing your loop (both of them) performance good-bye. As you are reading this next chart, be sure to keep in mind that these are only the pressure measurements on one of the running loops, the other loop is adjusted to a specific flow rate and maintaining its own head pressure. I could not come up with a good way to display this concept visually, so I have to rely on my non-engineer speak. Separating out one long serial loop into two loops, you reduce the amount of restriction that serial loop has to overcome. Reducing restriction means a higher flow rate which actually boosts the efficiency of the pump back up the performance curve. Vapor and I were trying to get a remixed version of Martin's Flow Estimator ready for this review so you could play with the numbers between single loop and dual loops on the Typhoon III yourself, but we're still working the kinks out even though Vapor has worked the equations every way from Sunday. What he came up with and the explanation are posted at XtremeSystems. We hope to have the Flow Estimator (currently named Loopilator) ready for release soon, and will post on the main page when it is available. As I mentioned previously, thanks to the PEC and its pressure regulating function, the top and bottom ports are so closely matched for output that you can choose whichever one fits your loop layout needs the best--chalk another one up to the PEC. Okay, okay, enough of my blabbering...on to the dual loop PQ Curves. Top Inlet/Outlet, 2nd Loop @ 0.5GPM - Lower Inlet/Outlet, 2nd Loop @ 0.5GPM - Top Inlet/Outlet, 2nd Loop @ 1.0GPM - Lower Inlet/Outlet, 2nd Loop @ 1.0GPM Top Inlet/Outlet, 2nd Loop @ 1.5GPM - Lower Inlet/Outlet, 2nd Loop @ 1.5GPM - Top Inlet/Outlet, 2nd Loop @ 2.0GPM - Lower Inlet/Outlet, 2nd Loop @ 2.0GPM
System Flow Rate Tests I was having a hard time figuring out the best way to properly display the System Flow (through the pump) improvements the Typhoon III unleashes when you are using dual loop configuration in a format similar to the way it has always been done. The reality is that the T3 is going to rewrite a lot of those "old school methods". After much debate with Vapor and a lot of input from various forum members I decided to just plumb up some loops using different radiators, CPU blocks and GPU blocks. This is a very simple test, one that we all have done by building our own loops. Just add the loop components, fill/bleed and use the "system". These tests are the flow rate for the loop. For the regular D5 and DDC tops there is one serial loop. Thanks to the dual loop configuration on the Typhoon III, we can split the loops. I decided to split the loops how I would inside a case. Lower Inlet/Outlet for GPU blocks with radiator(s) and Top Inlet/Outlet for CPU and radiator. The loop order does not matter, the individual components will not change their restriction properties by being in a different spot in the loop, or on a different loop altogether. While this is a simple test, it is a full disclosure of pumping performance. 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. Here is the data table that feeds the three System Flow charts. Test Loop 1 The first loop is a simple CPU and single GPU loop with a radiator after each block. I tossed in a Swiftech GTZ, EK-9600GT and 2 XSPC RX360 radiators. On the Typhoon III, the Top Inlet/Outlet had the GTZ and 1 RX360 while the Lower Inlet/Outlet had the EK-9600GT and a RX360. The single pump serial loops are still well in the target range for a good flowing loop, the Typhoon III boosts System Flow up to 3.47GPM. Yes, I said 3.47GPM!!! Test Loop 2 After running the first loop, I wanted to see an extremely restrictive loop...let's stomp on the T3 a bit and see what happens. I went a little crazy and decided to run a Koolance CPU-350 and EK Supreme for blocks with a Swiftech MC320 and HWLabs GTX360 for radiators. Flow rate on the serial loop was actually higher than I expected, I was predicting lower than 0.5 GPM in serial. For the Typhoon III, the CPU-350 and MCR320 were on the Top Inlet/Outlet and the Supreme with GTX360 took the Lower Inlet/Outlet. Typhoon III comes closer than 0.25GPM of tripling the System Flow Rate (flow through the pump) achieved by the other pumps with aftermarket tops--Yes--nearly tripling the system flow rate. If your heart’s not beating noticeably faster by this point...you're either not a hard core water cooler, or you are so hard core you passed out after the first graph. Test Loop 3 Now that I have the crazy lab technician torture the T3 loop out of my system, let's load up an SLI loop...only with a bit extra. This time we loop a HeatKiller LT, RX360, two DangerDen GTX295's in serial, a Swiftech MCR320, and let's add a HWLabs GTX360 in for kicks. This is a lengthy serial loop and although I personally would not run all of that off one pump, we see people do it. Anyone shocked at this point that the Typhoon III is able handle the loops with no problem, cranking out 3.16 GPM or more than doubling the serial flow rates of the other 3 pumps? As I was packing up all the test gear and entering the data into Excel, the thought occured to me that it would have been interesting to crank the power setting down on the D5 vario until the T3's performance matched the other pump top's flow rates. Not that anyone reading this review would ever run it at less then full power, I am curious. Ah well, that is a test for another day. Conclusion From the initial forum photos and debate to the end of testing, the PrimoChill Typhoon III has been one surprise after another. As I mentioned in the intro, I was one of the many (pretty much everyone) that were skeptical of running two loops from one Laing D5. Skeptical--ok, let’s be honest...we called it insane and other colorful terms that amounted to pointing and laughing. After the time I've spent on the bench with the T3, I will gladly admit I was wrong--the dual loop capability and performance are a truly revolutionary leap forward with the D5. The proof is in the data, and the last three charts show exactly what the Typhoon III can bring to your loop. You can easily run two full traditional loops off ONE Typhoon III with D5. That takes a whole bunch of the loop cost and sticks it right back into your wallet (for more water cooling parts of course). You can snag a major performance improvement that costs less than providing pump power and reservoirs for that second loop. I do need to take a step away from performance a bit because if this review was only about the numbers I would be doing all of you a disservice since the Typhoon III is really a collection of many well planned ideas which were executed very nearly to perfection. After countless installs, fills, air bleeds, draining, rinse and repeats, I can confirm that the T3 has really been designed with ease of use as a primary goal. No black magic involved, just good old fashioned attention to the little details by a designer that has obviously built a lot of machines over the years and I believe genuinely cares about the guys new to water cooling. To back up that claim, I will point you to www.boxgods.com and tell you to check out the article index for how-to articles. It is a nice change of pace to see a product hit the bench that is not just about pure performance, but one that also recognizes usability as an integral component of real world performance. If it goes like crazy but makes you crazy using it...well, as long as it goes like crazy we're all good. Seriously though, not cursing the whole install is a nice change. Be sure to read Full Review for all the details that go into making the Typhoon III user friendly. Pros Dual Loops from 1 Pump Blow your mind flow performance from a D5 Easy installation, fill/bleed Breakthrough design of performance and usability Amazing system flow rates (yes, its worth mentioning twice) The Typhoon III that hit my bench was a great test subject, but there are some minor things I have to point out. There is a scratch-like mark on the right front port as seen from the front that is repeated on the inside and outside of the front that looks very much like a crack. I did not notice it until It was brought to my attention (by PrimoChill) because, although you can see it if you look closely when the T3 is empty, it is invisible when filled. According to PrimoChill, "Although, it looks like a fine crack, it is actually a flaw in the mold that needs to be polished out. Unfortunately, the first 1,000 parts were delivered before anyone noticed it so it's there till that batch is gone." Also, on the edge where the top meets the face, toward the right side, there is a very slight imperfection called a molding spru (an air passage in the mold that lets entrapped air escape as the molten plastic is injected). You can feel it with your finger more than you can see it, and it's not visible when installed. My test unit did not have this, but seeing user photos posted in the forums there is a bonding strip that wraps around the two pieces of the T3, this change was made to quell the outbursts by some community members who want to use non-PrimoChill UV reactive coolant that contains Ethylene Glycol. I understand the need for it, but the amber colored bonding strip IS visible. It will not show in 95% of the installs because you normally only see the front and back, but anyone running a bench type system or a clear case needs to be aware. If I did use a points system, I would subtract a point for the bonding band. At some point in the future PrimoChill may find a clear method that works as well. If they do, I will gladly give the point back. One last little thing, to do with fittings...I use Bitspower High Flow barbs, and had no problems threading in the barbs, but some people are experiencing pains with the o-rings on some brands of fittings. If you want to use fittings other than BP High Flows or the Ghost Compressions that come with the Typhoon III, you may need thicker o-rings. They can't design for every fitting. Cons Slight molding imperfections Alternate fittings may need thicker o-rings Amber bonding strip The tests and review you just completed reading was a challenge, a great learning experience and a fun three weeks. Yes, from start to finish the whole thing took a little over three weeks. A bit longer than anyone would have preferred I know. From the initial threads that popped up once photos started to appear and discussions of this taboo parallel flow beast to the present, the Typhoon III is finally here and ready to be unleashed on your loops. If you have not been able to tell from all of the above, I think the Typhoon III is a great product, one that has the users' needs in mind as well as a major shift in loop design with a kick in the performance pants that will leave a mark. If you are just planning a build and debating on whether or not to run two loops, no need...pick up the Typhoon III with D5 Vario and be on your way. I am still in awe over how the Typhoon III brought forward so many new design concepts to reality...we might have to initiate a yearly award system just to crown the PrimoChill Typhoon III the Most Innovative Water Cooling product of the year. I know the year is only half over, but I cannot imagine something else this flat out innovative stealing that crown...if this were to happen, it would be another design by BoxGods I suspect as I have seen the new Radiator line (code named RainMaker) he is working on and I will be stunned if these radiators don't blow your socks off even more than the Typhoon III. I seriously doubt that when RainMaker pictures start popping up in forums anyone is going to be "pointing and laughing". Congratulations on Typhoon III and a Well Done to PrimoChill! Thanks for reading! If you missed the several links already... Be sure to read the Full Review.
Hang on, dual loop off one pump without sacrificing pressure or cooling capabilities? This I would love to test in person.
I don't think it's physically possible to double the flow rate with one pump like this thing supposedly does. Adding up the flow rate from each "loop" doesn't give the total flow rate. That's just adding it up. I'm not so sure about this. I shall wait for more reviews before I believe the flow rate of this thing.
@ talladega: An original Laing DDC-1T has max. 500l/hour, but this is never achieved in any loop, as the flow in a loop is reduced by the blocks and radiators etc. So if you have a loop with 250l/hour, then you can fire two loops of this kind with a Laing DDC-1T and have basically doubled the flow. EDIT: Take an Eheim 1046 for example... these have 300l/hour in comparison and are usually limited allready by the loop.
I understand that doing two loops you lower the pressure on the system which will increase the flow rate, but it does not double it. That's like putting a Y on my loop and saying i've doubled my flow rate. I do believe it can increase flow rate because it does reduce pressure but I do not believe the results in the review.
You have a single loop, with a 240 radiator + block, which limits the flow to 250l/hour. It doesn't matter if your pump can deliver 500l/hour, as it can't pump it that fast through the loop. By adding a second loop to your pump consisting of a 240 radiator + block, you now have 2x 250l/hour = 500l/hour, which is double the flow basically. However, you should make sure, that the loops have not more capacity then the pump, or one of your loops will not recieve enough pressure. It's also preferable, that both loops have the same pressure.
The System flow Rate comparisons I ran will be added to my Pump Comparison pitching both D5's and DDC's in one set of scatter plots, we'll finally get to see PQ's together. Its going to be an eye chart, but it will show the full sprectrum of PQ. I've already been IM'd numerous times by someone who wants RD-30 testing included and really wants this whole Y/parallel loop discussion to be over. But there is still more testing and information that needs to be done before this parallel discussion can be boxed up and set on the shelf... For each loop the configuration and components can be found in the OP. Y loop were split up the same for the Y tests as the were for the T3. On to the system flow rate comparisons... 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...but we'll find out for sure soon enough as PQ testing on the Y configs has to be done now.
This may be the performance king, but in the quest for performance there have been a number of major oversights! 1. No instructions Not too tricky but come on! 2. Plugs for inlet/outlets when not running dual loop which cannot be tightened using the provided allen key You need to find your own! 3. No accounting for cases with drive rails...you need to break these off Would it really be difficult to recess the are where the drive rails are the same on every damn case? 4. You cannot mount the reservoir flush, you have a huge gaping piece of plexeglass haning out of your drive bays It can be easily moded by filing off the excess plexi on 3 sides, but it would nice if this was not necessary especially when it looks 10 times better flush mounted! 5. No indication of the inlet/outlets Easy enough to figure out, but its all about the details! 6. When mounted the case needs to be tipped on its rear/front to fill/empty it! Whilst this may be a good way to bleed your loop...this is laughable when put into practice. They now have a version with a fill port which allows for filling the loop with the reservoir installed in the drive bay. Who the hell tested the functionality of this reservoir before asking people to pay for it! 7. An worst of all by far is the sheer volume of manufacturing defects. There are cracks, nicks, excess plexi all over the place. On mine there are two nicks around the mounting point for the pump. This means I cannot create a seal when using the plate to secure the pump. Water drips out of the bottom regardless of how well I tighten the screws. Every feasible method has been tested to see if it's because the Oring is not in the correct position. Putting all of this aside there is an additional factor seldom accounted for when testing flow rates. We see all of these graphs showing the advances in flow rate, however what happened to testing the effect on temperatures? I have a D5 obviously, yet regardless of howrestrictive my loop is the best temperatures are achieved by having it on setting 1...yep the lowest! Of course it's an advantage to have good flow through your system, but if it passes too fast then its not going to cool effectively. You may think that my loop would then contain litte restriction buuuut: 1 x Swiftech GTZ CPU block 2 x EK 4000 series Full Cover GPU blocks 1 x HWLabs BI 360 GTX 1 x HWLabs BI 240 GTS So its not quite as simple as increasing the flow rate for incredible gains, nope because you may just end up with higher loop temperatures. If I increase the flow rate of my D5 to setting 5 then within 2 minutes I've gained 8c
Very nice review. Do you think it will be possible to mount the T3 fillports facing up in the top of a case? That seems to be more practical. However... I've been researching watercooling for a while now to put something together and you my friend have hit the nail on the head! I've not yet come across anything looking into this while it seems quite obvious. A thorough 'ultimate guide to' all aspects of watercooling would be quite a good series of indepth articles for bit-tech to do (and probably get massive views if done well). More so now they are recommending WC for the top tier monthly builds.
agreed. i've seen other tests of this 'thing' that show it isn't the most amazing thing like he claims. infact it is pretty junky.
