jmke

My russian ain't too good, sorry to the crappy translation, but the photo's do tell the story

To undertake the given series модов me has induced the competition organized by a site overclockers.ru together with Glacialtech Inc.. Not being the supporter of compromises in cooling, I before practically was not interested in production GT, and in competition to participate firstly did not think. However моддерский the reflex has prevailed: having started to read a condition, in half an hour I have caught myself that instead of it, вперившись in the monitor a unseeing sight, actively I consider variants of updating Igloo . Coolers of the bottom price range, for the clear reasons, have been considered by the first - and rejected: to make something explanatory with a simple piece of aluminium possible it was not represented. But research of a fresh ruler 5700 \ 7700 has revealed so interesting prospects, that next day, not looking at a Saturday hang-over and the general безденежье, has been bought experimental - Igloo 7700 MC. Interesting feature of this cooler is that edges on tubes sit внатяг, nothing fixed. So they can be removed and... Well, нап

__________________

jmke

More translation fun:


Three six-millimetric tubes (useful length of 10.4 sm), 45 edges and 92мм the fan with fashionable prosaws in a framework (all effect from which with interest is blocked by a crash of the motor - good children work in GT, here if only on suppliers saved less...).

Has first of all turned off from it all superfluous and in such kind has put on primary testing.


As термоинтерфейса it was used термопаста Thermaltake №1 (analogue КПТ8). Прожиг the processor it was traditionally carried out by program S*M 0.3.2b with exposed manually реалтаймовым a priority, the temperature of a kernel was not less traditionally traced МВМ. The temperature in a room during testing remained more or less a constant, floating between 22 and 23 degrees.

Has included, has started... In half an hour экзекуцию it was necessary to stop: the temperature of the processor has reached 95 degrees and to stop did not gather. Well, it is not necessary to be the greater expert to understand, that the radiator with such intercostal distances (1.5мм) for cooling natural конвекцией does not approach absolutely. It to me also should be corrected. Has first of all armed with the tool made quickly...

... Also has removed edges. In case someone will want to repeat - it is done approximately so:

First we tease the grinded aluminium strip, we expand a crack that it was possible to thrust a plug and, operating with last as the lever, we shift an edge upwards. In some receptions: first on the one hand it is a little, then with another, then again with the first - yet слезет it is final. A glove - an absolute must, without it fingers will quickly исколоты the sharp ends трубок. That business went more quickly, tubes I advise to grease something slippery (and to grease in process of removal of edges). Procedure, by the way, it is necessary to repeat 45 times so nervous these to be engaged I do not recommend.

Having removed all edges, has counted half and, having armed with scissors, has cut half-and-half.

Then has spread all back here thus:

As it is possible to notice, at such way of assembly the backlash between edges essentially increases, at the same time in height the radiator does not grow. Непонятки arise only with fans: 92мм - it is not enough, 120мм - it is a lot of. As has shown experience, is better (though and it is not ideal) the variant 2х80мм approaches. One more opportunity for completion was in parallel found out: engineers GT have definitely gone too far in maintenance keep-out zones on a vertical from what the idea follows to shift all package of edges downwards on tubes (on how many to shift "will prompt" rods of memory), and on the released place to spread still. It also has undertaken - in half an hour of an edge have been shifted. Places it was released not so that much, but nevertheless...

Here, I think, it is meaningful to interrupt and bring an intermediate result for it is a maximum that it is possible to make with a cooler without additional financial expenses. So, what I have made? I пересобрал a radiator in such a manner that, at preservation of the same area of heat exchange, aerodynamic resistance have decreased almost twice. Now we shall look, that it has given. We collect the stand, запускем S*M and with satisfaction наблюдем stabilization of temperature on a mark of 70 degrees. Круть! Perhaps, it is possible to start to be proud and bend fingers: who else can brag, what its completion has helped to throw off 25С?)) All right, it is time to go further: in spite of the fact that such design has quite the right to independent existence (naturally to shift edges downwards in this case it is not necessary), the soul asks something heaped more up. For reduction of creation final harmony it was necessary to buy the second Igloo 7700 and 120мм the fan then work has proceeded.

Свежекупленный the cooler раздербанил on spare parts also has finished, as well as gathered, edges up to the top. Has put on the stand..

Also has received... The same of 70 degrees. On is indignant a question asked to " what [censored]? " The distinct answer it is found and was not. There can be on "hotter" processor a difference would be shown? However, we any more do not learn it: to assort all on-new, again to urge on the left unfinished variant, I hardly shall find forces in myself. It is necessary to learn only that costs "dispersed" thus 7700 in comparison with the recognized grandees of cooling - for this purpose on the stand has been established Scythe Ninja.

62 degrees. Мда - how many the wolf do not feed, and at the elephant all the same more thickly. However, about feeding I, perhaps, paltered: reserves still are. If to think, edges it is possible to spread even more (two thirds on the second radiator have remained not used): for example, to not cut external in half, and насаживать entirely.

Unless in their this variant it is necessary to cut a little on the one hand, differently rest against a wall the chassis. Most likely, at such length problems with heat conductivity will already begin, but, judging about dynamics of warming up половинок, the prize in any case should be [a voice from the future: - ah as it was mistaken!]. Anyway, loss of eight degrees to the monster with twice lot трубок and on third greater area of a surface of special affliction at me has not caused.

To the second experimental became Coolermaster Hyper 6 - rolled without business in a box why to not test? Having intercostal distance 2.5мм, the radiator hardly, but has kept the processor within the limits of 89 degrees. And its all copper with паянными edges has not helped, that, in general, a surprise for me did not become. At Ninja, to a word, between edges 4мм, at ours свежесобранного - too (not including two enough wide areas where edges go вперехлест: and has not dared to cut them, probably in vain). Probably, also it is necessary to consider this distance minimal for a passive cooler.

Whether will give something its further increase? After short considering I have considered, that this question достоин separate check - the blessing on a table just rolled half peeled the second 7700. Superfluous edges have been removed, and remained are distributed on tubes with intervals in 6мм. As it is simple to count, knowing length трубок, from initial 45 plates remains only 16, in what kind the radiator and has been set up on the stand.

Also what you would think? 67 degrees! On three degrees it is less, than at earlier collected " the senior brother " with quantity of edges equal already 54. We go further. Half an hour of boring work and result: 24 edges with intervals 4мм. I start - that for devilry? Still 67 degrees! The similar case with the first radiator then in a head was born classical though also absolutely useless conclusion was there and then recollected: something here is not pure. What exactly - to guess it was possible long, but one became absolutely clear: experiment was covered with a copper basin. And I, staying in the upset feelings, villages to complete the big radiator.

To leave it in a naked kind the aggravated feeling fine has not allowed, therefore removed in the beginning of work the casing has been altered in view of the increased dimensions and pulled down back.

After it, after small completion of fixture, on a place has risen GT12252BDL-2. Вуаля, meet Igloo 7900!

By the way... And why to not try пободаться with Ninja in such kind, for interest having added in the company "grown thin" 7700? Not a question, it is told - сде... The Pancake! To measure материнку I, certainly, has forgotten and now the severe reality severely interfered with my impulse in the form of one of планок memories. It was necessary to come back in a workshop and to alter at 7900 вентиляторный fixture, as a result стодвадцатка has risen with small displacement. Ask, why I simply have not rearranged a rod in next слот? In half an hour I too have asked to myself this question, but was already late. Well and all right, at old Igloo fans too were on the party moved down. Continuity of generations, you understand. In general, the cooler таки has risen:

As you can see, "brick" has turned out rather big, for which reason to develop a radiator on 90 degrees will not leave any more - rests against a videocard. As well as to hang up the fan on the other hand, however here already all depends from concrete материнки. The lonely condenser which has ruined to me all идилию, at someone in this place can and to not be (the truth, with the same probability it can "grow" there, where располгается the fan in my configuration). Proceeding from all it to recommend such design to a batch production of a reason is not present (though at Ninja too normally to hang up стодвадцатку that is possible only on the one hand from four, and not on all mothers). But private самодельщики to which do not have need to care about всеплатформенности, can quite not hesitate.

All right, to tests. Our Trinity (7700, for a dense forest of presentation of results, participated in a 16-costal variant):

For maintenance of the best warming up it was necessary to change some iron. The new configuration looks so:

- Athlon64 3500+ S939
- MSI K8N Neo2
- 2x512 OCZ
- GeForce 2 MX 200
- 200Gb Samsung SP2004C
- FSP 400-60PN

Actually, it was already other computer, therefore and the set of a software has changed: S*M last version and Speedfan 4.26 as the monitor. The hardware too has added confusions: new материнка obviously conceded K8NS in accuracy of reading of indications термодиода. In addition, at it "feature" repeatedly noticed at series Neo was showed: serious overestimate Vсore, from for what poor Newcastle on a thermal emission has closely come nearer to Прескоттам. To what I speak this all: the figures received in the second series of tests do not correspond in any way with figures from the first, the difference between experimental has value only.

In the first call I have fondly tried to check up, what maximal capacity can be removed, not resorting to installation of fans: has unscrewed on a maximum and so already задранное the pressure and was dispersed with the processor up to 2500МГц. "Heater" which has turned out as a result in a few minutes has tired out for 100С Ниндзю then I have understood, that has got excited and has got into a BIOS, to expose something more humane. However, far from it: материнка, for not clear reason, impudently ignored all attempts to set a pressure below the certain threshold. In general, with tests in a passive mode it was necessary to fasten.

Experiments with active обдувом were spent on regular frequency (some degrees of weather did not do) and with the pressure unscrewed on a maximum (how many there it was real - it is not known, but by heating, it is not less 1.8В). And here, at last, the difference in quantity of edges was showed.

Results "facilitated" 7700 can be characterized as a full failure: 46 degrees without loading and 88 under S*M. Regular 92мм the fan has been exposed on 1500RPM (on these turns the noise level approximately corresponded blowing competitors стодвадцатке), but not in it business. The forty-degree difference idle-burn speaks for itself: the radiator simply did not have a working surface for effective tap of such quantity of heat. A conclusion: прореживание edges - business useful but to take a great interest it is not necessary.

7900 has acted much better, having proved, that I таки have knowingly ruined all of one and a half hour on переборку a radiator: 43С - 71С.

Ниндзю to butt (or even to catch up) and it was not possible. Despite of received on fault материнки гандикап (the fan did not manage to be established so that the bottom radiator was blown), "Japanese" has confidently won competition, having kept practically former отрыв: 37С - 61С.

By the way, the second series of tests was made with horizontally established материнкой so any influence of type applied трубок on results is excluded.

At last I shall mention one more interesting moment revealed during dancings with бубном in attempt to understand, why the radiator does not react to change of quantity of edges. In fifth time обжегшись at removal of a radiator about heated бэкплейт, I have recollected, that in passive systems, unlike active, the quite good result gives cooling textolite материнки from a underside from a socket. Whether not considered watts there left? To check up a guess, in a quickly way has altered regular бэкплейт Igloo 7700 in a fixing plate...

To not short-circuit a leg of a socket, has put two layers теплопроводной substrates, like that, that is applied in power units to isolation of power elements from radiators (is on sale sheets 20х20см in ù¿»*ä¿»Ñ). For the given specific target, we shall directly tell, not the best variant (I am usual for manufacturing термопрокладок use glue "Radial"), but on безрыбье... Regular screws (under рожковый a key) have appeared коротковаты and have been replaced on extended usual - under a screw-driver. After the radiator was fastened for 20 seconds instead of usual pair minutes, there was an idea to go and be killed by a head ап a wall - how many it would be possible съэкономить for time if I have made it at once? Well with it, it is time to learn a horse-radish, that I this all have achieved. Having fastened K8NS on operatively rivetted стендик for vertical installation, has started S*M. The temperature has slowly crawled to 43 degrees... Also has stopped! Anything to myself, I have told to myself... Still 27С? Where there the Nobel Prize give out? However, the pleasure was short: having put Hand on the basic radiator, I hardly could keep it there. Sensations were obviously not on 40 degrees and I обреченно потопал behind a multimeter. Gauging of temperature on a cover of the processor has confirmed, that miracles do not happen - 62 degrees. Transition a kernel-cover - still degree three, total 65. Мда as this is regrettable, but I only have cooled a thermode. And the happiness was so close... However, strongly to be afflicted in any case of an occasion was not: though experiment also has not justified the hopes assigned to it, 5 degrees too on road do not roll . And, most likely, dimensions of a back radiator can be reduced lost-free in temperature: at least, replacement Igloo 7700 on 5600 (х 4мм) any changes has not brought 17 edges. To experiment with radiators of the smaller size of an opportunity was not from for absence normal термопрокладки ("Radial" in ù¿»*ä¿»Ñ, to сожаленнию, will sweep away more quickly, than it there have time to deliver) and problems with fixture. Probably, in the future I to it still shall return.

---------

Ideas and conclusions:

1. We had a new toy-designer. New "needles" - not unique coolers with such landing edges so opens enough the big open space for creativity. There would be a patience.

2. Edges are desirable for rearranging no more once since during this procedure are unbent sealing пояски. During experiments with quantity of edges I on this raker have come: after the second переборки the part of plates began to dangle freely on tubes. It was necessary to take тонкогубцы and to draw in.

3. Radiators of family 5700\7700, considering the moderate price, can become a quite good source of edges for кастомных projects, and here as to their use as objects of completion... Conclusions from this history with identical temperatures arise sad. The impression is made, that in a passive mode to a cooler edges on it насажено are absolute all the same, how many: if only was соблюдено minimally necessary for конвекции intercostal distance. And it means, that its maximal productivity is limited not with the area оребрения, and any features of an internal structure and to squeeze out of it something over regular parameters it is possible to not try. However, concerning cold processors and regular parameters it is quite enough for passive cooling.

4. From for вышеупомянутой features experimental to find out optimum intercostal distance to me, alas, and it was not possible. If someone will not be too lazy to check up on other radiator backlashes from 3 up to 5мм and will share results - I shall be grateful.

5. As we see, to cool younger Атлоны on naked конвекции it is possible without special problems. The only thing, that in this case will demand serious completion - the case.

6. In case of with passive systems it is not necessary to forget about heat removal from the back party материнки - it is possible to throw off degrees a little. The main thing to be convinced, that you have thrown off them менно from the processor, instead of from a thermode.

__________________