I'm assuming he means the water only does a single pass instead of two passes. Therefore it's in and out of the rad quicker... but since the water is entering at a constant the point is mute assuming you have a more capable 120MM fan (louder).

Max: research, NOW! *Sound of a whip cracking*

(Where's the slap my forehead smilie?!?)

The single pass will be less of a flow restriction.

So...

Let's assume that you're able to double your flow rate, with a single pass (for the sake of the simplicity of this explanation).

While it's true that the water will be in and out of the rad faster, if you consider a specific amount of time, the water would pass twice, where the double pass core would allow the water to pass only once.

So the water spends the exact same amount of time in the core.

Any Q's?

Woot here we go.....could be fun!

*SMACK!*

sorry guys.. only spent 10 seconds thinking about it in my head.

I forgot to consider the fact that the water gets split up to twice as many flowtubes and hence the velocity of the water is lower so it therefore spends the same amount of time in the rad.

hmm.. thinking about it, would there be any downisde to single pass. im talking about at the exact same flow rate would the performance be greater or less all other things being equal.?

As with most things, there are pros-cons to doing things either way.

In a single-pass implementation the per-tube water velocity is half that of a double-pass implementation. This reduces the convectional transfer efficiency between the liquid and the tube surface. That is single-pass's con. The upshot of single pass is that all of the core surface area gets engaged with the highest temperature delta between the air-flow and the water temperature.

In a dual-pass implementation the per-tube velocity is twice as high as a single-pass. This improves the thermal transfer efficiency, but now only half the core surface is being engaged in operating on the maximum temperature delta between air and water. The second half of the core receives water pre-cooled by the first half.

A way to explain it simplistically in a mathematical sense is to think of it in terms of ratios.

Let's say that at some given set conditions that a single-pass radiator removes 20% of the heat from the liquid, bringing the liquid to a temperature delta that is 80% of what it was when it entered. Now if the water goes up and down over half the core each time, then it loses 10% of its heat on the first pass, and 10% of its heat left over after the first pass on the second pass back. 0.9 x 0.9 = 0.81. The single pass radiator removes 20% of the heat, but the double pass equivalent removes 19%. The single-pass implementation is therefore 5% more efficient (in this example scenario) if we exlude the effects of water velocity based convectional differences.

So it then comes down to which effect is the greater of the two.

However, I believe that much of the above is somewhat irrelevant and not the main cause of differences seen between single pass and double pass. I believe that it is more complicated than that and has to do with how the water flows internally and whether or not that flow is balanced among the tubes availble. I suspect that much of the difference seen is more due to flow imbalance issues for dual pass radiators. Let me explain.

For a dual pass radiators the flow all comes down one side, enters the bottom end-tank and has to flow across to the other side and back up and out again. Where I think the MAIN problem for double-pass radiators lies has to do with that it's harder for the water that's moving down the edge of the radiator to make its way across the length of the end-tank to get to the outlet pass, while the water flowing down the more centrally located tubes is subject to a much lower pressure-drop resistance to make its way to the outlet pass.

I believe that this sets up a cascading set of conflicting pressure drops across the width of the outlets of the first pass, thereby imbalancing the flow. At lower flow rates this effect is relatively minor and at lower flow rates the performance drop-off that we see is primarily due to convectional liquid-metal efficiency. At higher flow rates the pressure-drop imbalance becomes more severe, and this is why we then see a peak in the performance curve for various radiators above which higher flow rates result in decreasing performance.

A single pass radiator with opposing inlet/outlets and evenly configured tanks is not subject to this style of flow imbalance occuring (or is substantially less impacted by it), and this is why I believe that a single-pass configuration is quite often superior - not due to it explicitly facilitating better heat dissipation, but rather due to the flow-design deficiencies of various dual-pass radiators.

All this doesn't quite apply to the PA160 though, because it also loses the typical "second row", so the per-tube velocity of the PA160 is the same as if it were a dual-pass dual-row core, so it retains the dual-pass per-tube velocity, but adds the single-pass benefit of a more even flow regime across the tubes.

Very interesting theory cathar. I would love to be able to see some actual testing were we could see how the water flowed through the rad. That would be very enlightenting.

Not that i am against single pass, but do you think that there is any way to optimize the flow inside of a dual pass radiator to eliminate this performance asymptote. Obviously this would need to be done inexpensively so that radiator companies could actualy build this design and still turn a proffit.

I think i know of a way to do it and i dont think it would be too much more expensive than the current designs of thermochill and hwlabs radiators. my idea is to have a bunch of flattened tubes going from the down flow to the upflow. conneccting inner to inner, going out. it would look a bit like a rainbow.

but now that i think about it it would still have the problem that the tubes closer to the center of the rad would be shorter in total distance and thus disturb the flow of the radiator as a whole..

either way though i think that my redesign would reduce the problems with dual pass rads, but not solve them..

I can't believe it!!!!!
 

It's there

Pirate!!!!!!!!!!!!!!!
 

what a radiator

I believe that the problem/issue is directly linked to the amount of space available in the end-tanks.

Thermochill radiators have a fairly generous amount of space in the end-tanks. All it takes to alleviate the problem near totally would be 10-15mm of end-tank depth from the outlet of the first-pass tubes. If we look to Bill's tests of the two-pass thermochill radiators we can clearly see that they did not exhibit performance peaks then performance drop-away as one moves upwards through the flow-rate range that is typical in water-cooling loops.

Where I think that the issue is most prevalent is for certain radiator manufacturers who for the sake of making their radiators as compact as possible have perhaps sacrificed too much of the necessary end-tank cavity, and have thereby created the issue. I believe that adequate testing by the radiator manufacturers for which the issue is evident, and which should be fairly easy to do, would very quickly determine if my theory is anything more than fancy.

I believe that it would be wise for any testers who attempt to compare the difference between single-pass and dual-pass performance would do well to first assess that the dual-pass configuration is not crippled in this fashion, and so a better apples to apples comparison can be made.

Comparing a crippled dual-pass radiator to an otherwise near identical properly implemented single-pass version of the same radiator is most likely always going to lead to the conclusion that single-pass is always better, when that may not be the case if the dual-pass radiator that it is being compared to is not implemented correctly.

thanks for the insight cathar. I will have to put this into my radiator testing articles. I will be comparing HWlabs and thermochill rads so i will be able to comment on the difference in tanks and the plateus in performance vs. flow.

It's just a theory of mine, not fact, but it's the only explanation that I can possibly think of after looking very closely at the physical characteristics of these different radiators that exhibit the "middling-flow-peak" effect.

Sorry.... all of that is making my head itch this early in the day... anyways... here goes my stoopid question. Based on the above theories, which would be preferable...

1 - Take an existing HE120.3, knock out it's endtank baffles, block up one hole. Duplicate on the other end - leaving you with single pass thru a two row core...

OR

2 - Refab the HE120.3 to single pass single row?

Remember, once we get towards the point of having to discontinue the current HE series, it effectively (in my mind) makes BillA's original review of them null & void in a way if the rads are changed, or are no longer available in the specification that pairs up with that review.

If you were in a manufacturer's position, would you replace the HE series with singlepass-singlerow / singlepass-dualrow (based on your answer to the above question) or would you bring it out as a separate series... and keep the HE as they are in full production?

S'not that I'm gonna immediately walk off and do it, but it is certainly worthy of consideration at this point for later in the year... (we can only put so many new projects underway at once - currently working on the "bling" side of things designing fan grills for all the ThermoChill rads, which are a more pressing immediate requirement at the moment)

Well in my mind the 120.3 has it's end-tanks implemented correctly, or close enough that isn't an issue that anyone has been able to observe/isolate.

Given that the 120.3 is a dual-pass dual-core radiator, I am not wholly convinced that making it single pass will make it better because we're immediately halving the per-tube velocity which will have a significant negative performance impact going by what we can observe within BillA's results. What all of the above was trying to say was that just because a different manufacturer's implementation of a single-pass version of their more typical dual-pass radiator might turn out that the single-pass version is clearly superior, this may not necessarily be true of Thermochill's 120.3.

If you're asking for my advice, I would say keep it in mind, and perhaps get a single one made up purely for evaluation purposes. Shouldn't be too hard to grab a raw core and stick suitable end-tanks on it for single-pass config (in exactly the way that the PA160 is implemented) and give it a go. The issue of course comes with finding a suitable testbench to verify the findings. I don't know what Thermochill has in the way of a test-bench, but assuming that you do, I guess it's something to consider. If and when you do decide to do it, and if it turns out worse then just let people know that, if only to get them off your back. ;) If it turns out the same and/or slightly better then it all comes down to costs vs return.

Nothing that I'd consider suitable for our purposes.... but that's something I'm trying to sort...

Cheers for your thoughts on the matter! :D

I'm aware that by using a shroud on the intake side, the radiator wouldn't be getting the optimum airflow, but I'm not really excited about cutting a case that cost over $300, especially when it's already setup to accept a given line of radiators (namely, the Thermochill 120.1, .2, and .3 units).

If anyone does a test of the PA160 (along with a comparison of other purpose-built PC cooling radiators), I'd like to to suggest that the test be performed in both configurations - as shipped, and with the additional shroud being used as an adapter to an existing 120mm fan mount.

If it can out-perform a HE120.2 and/or a BIX2 when using the 2nd shroud, you can bet I'll jump on one.

There are no "additional shrouds" for the PA160. Every rad is supplied with one, and a second one is not required, and the shrouds aren't available separately. To run it with 2x shrouds, you'd have to buy 2x PA160s and throw away one but keep it's shroud. As I've already said, mount it with fan pushing instead of pulling... will likely offer better performance than pulling thru 2x shrouds with inlet restriction, and means NO cutting required. You're really making life difficult for yourself pursuing multiple shrouds, and completely missing the point of the radiator. If you're going to restrict the airflow of the PA160, then you may as well use an existing HE series that is already restricted at the core by default, rather than getting an ultra-low restriction rad and then restricting it by adding the 2nd shroud.

As I said, why obsess about having fan pulling air thru the rad to conform to "proper" watercooling fundamentals, if you're then going to go against fundamentals by adding the 2nd shroud and restricting air inlet. Just mount the rad the opposite way round with fan pushing air thru rad instead, or pull air from inside the case to outside the case as opposed to vice versa. Both are suitable solutions that would be cheaper and offer improved performance over PA160 with 2x shrouds, and doesn't involve any case cutting.

You have that much room in that case that there shouldn't be a problem using internal heat vented to outside over outside air vented to inside.

What you're wanting to do is akin to buying a superb pump, then using 1/4" tube on it's inlet with 1/2" on it's outlet. Defeating the object and pointless. :shrug:

EDIT: *sigh* fair enough.... I concede...

I've just spoken to the Factory Production Manager and asked them to do me a separate run of shrouds to be made available separately. They're gonna try and squeeze it in between production runs... (which are planned a month in advance usually)

Initially these'll only be available from O-CuK til International Distributors pick up on it (they've already ordered basically)...

So what's the guesstimate of danger den's retail price on these, taking into account the dismal value of the USD?

Honestly not a clue... sorry!! We've stopped issuing an MSRP to resellers and distributors and leave them free to set whatever pricing they want. S'a shame the watercooling industry isn't big enough to support loss-leaders, then they could knock em out dirt cheap!

I sent them an email and asked, but all I gt back was "watch the website".

Hey Marci, I know it's not as efficient to have the shroud on the intake side, it's just that I don't want to cut a 160mm hole in my case (besides, last time I tried cutting something, I severed all the tendons in my left hand, and I'm a little leary of anything that ends with the word "saw" or "cutter". :)

I currently have a BIX and was considering moving to a Thermochill, but I've never seen a direct comparison between the two (I've seen one with a BIP, but not a BIX), and I hate making a purchase without at least being partially informed.

Ouch! Hope you're OK now...
You could use aviation snips. They usually come in sets of three:
- cut straight
- cut curves to the left
- cut curves to the right

And as one of the folks having said that pull seems to be slightly more cooling-for-noise efficient, I really did mean slightly - and only for the particular heater core I was trying the different configurations with. There may be no difference between push and pull with, say, a Nexus 120mm fan and the PA160. It has much lower air resistance and all bets are off - or at least all old measurements are meaningless...

Cathar I'll keep in mind the comments you made about your theories and state that they are theories. (i cant spell).

Of course this alll depends on whether or not this phenomenon shows up in my testing. If it doesnt and the rad performs better as flow increases in a linear fashion then i dunno what to say.

One thing i will note though is that on the BIX and BIP here they both have very thin chambers at the bottom of the radiator. when compared to my 2-199 heatercore i was quite surprised. i mean it makes it look nice but at what price.

Just want to make it clear that the behavior I'm talking about is a radiator that exhibits a performance peak at say 4-6LPM, then the performance drops away after that.

Radiators that achieve a performance plateau as flow rates increase are perfectly "normal".

There's so such thing as a linear performance curve as flow rate increases.

Maybe it's just the terminology you're using and my understanding of it, but wanted that to be clear.

yah i know it isnt linear.. its a curve that looks like Ke^(-x) for "normal" radiators. Trying to think of an equation for the performance drop at 4-6lpm rads but im too lazy. ahh i know. a forier series with an interval of 3/2pi.. lol.. im bored.

as to any radiator plateuing i wasnt entirely certain on this. I had seen from bills tests that some rads drop in performance and others plateu after a certain flow rate. but when i talked to bill about it he mentioned that it could have been an area of stagnent water where his temp sensor was which might have accounted for the drop in performance.

btw 4-6lpm is relatively average for a decent watercooling system (non storm based system :p as the storm is too restrictive to allow thise flow rates with an "average" pump) so that would mean alot of people out there are over the peak performance of their radiator.. if this is the case there are alot of rads out there that were not very well throught out in design.

in any case I will have to check for these tendencies when i start rad testing.

Hmph - them's fightin' words.

Can achieve 3.5LPM easily with a Storm and an Eheim 1048.

Will achieve > 4LPM with a Laing DDC.

Can achieve > 5LPM with a Storm and a MCP600/D4. Over-volt either a small amount and 6LPM is obtainable.


If you said that 7+LPM was out of the picture for typical pumps, I'd have agreed.

FWIW - am running ~9.5LPM through my Storm/G5 here in regular use with the RD-30.