That's marketing for you.
G
That's marketing for you.
G
for who?Originally Posted by Master_G
Will just be interesting to see what pops out at the end of the day. This "Next Gen" range has been on the cards for how long now? It's been somewhat amusing watching the game of cat'n'mouse between the various radiator manufacturers as they all jostle for market mind-share.
One thing which I find somewhat of a misnomer "Power Users Demand - We Deliver!". Really? That was mentioned 3 months ago now and has anyone been delivered anything yet?
Just being "tongue-in-cheek" about it. No offense intended. At the end of the day the end-users will benefit and that's the important thing.
Water-cooling improvements are coming to the end of the road though. Managed to eke out as much performance as possible within the single 120 form factor without going to some radical changes (forwards looking), but it got to the point eventually where we were just picking balance points with the traditional radiator manufacturing principles. PA120's heavily focused on <=70cfm fan performance although they were not bested by anything else in a 120mm form factor up until around 120-130cfm of fannage. Got to a stage where no matter which way we twisted we simply couldn't get more out of a 120mm 10-30dBA fan performance range (10dBA being a Yate-Loon equivalent at 5v, 30dBA being a Panaflo L1A), and in the end had to make a decision on whether it was better to tune purely for that range (<=30dBA), or to tune more broadly up to 45dBA say. We decided it served the end-user community better to provide radiators that best worked in the manner in which water-cooling should be used, which is quiet & excellent cooling performance. The new PA rads still kick butt at higher airflows (as stated above), but rather than make them even better at insanely loud fan volumes, making them better at more tolerable volumes seemed to be more productive and useful.
As for single-pass vs dual-pass. At the sorts of flow rates that people see in systems it's also a matter of swings and roundabouts. In a 120mm form factor, for single-row cores single-pass can work better for flow-rates above 4-6LPM, but dual-pass is arguably better across the board. The 160.1 was single pass because of the low-air-flow restriction nature of the design and it benefitted it well. For smaller radiators it seems to close up somewhat. For dual or triple row cores, dual-pass is always better and flow restriction doesn't really come into it. The vast bulk of the flow restriction in these types of cores all comes from the fittings so being single or multi-pass internally makes very little pressure drop difference, and certainly next to completely insignificant when waterblocks are involved, so long as the radiators have correctly designed end-tanks.
There are other advances which we're looking into further down the road. Don't want to play all our cards at once and certainly don't want to reveal the whole hand to the wider market on a platter, suffice to say radiator performance improvements are not totally done-for yet, but certainly close. Probably another year or two away for those advances, again due to the complexity of eking out the last drops of performance, as well as the need to further investigate other avenues.
Still, this whole radiator development thing has been a real eye-opener for me, if only to open my eyes at the evident lack of adequate research that various companies have been putting into proper radiator design up until the last few months, as well as lack of a properly focused market goal. Simply saying that a radiator fits into a PC case is not enough, the radiators also need to perform well in the fannage scenarios that best match the desired goals of water-cooling's noise-performance benefit.
Worry not, Cathar, much has been accomplished in the 5 years we've been working on the matter. Our many tests and tweaks, however, simply aren't yielding the heat rejection targets that we've set, hence the monumental engineering efforts we are currently undertaking.
The Black Ice® XFlow radiators are an update stemming from requests by many customers seeking a more flow-optimized radiator for their increasingly constricted systems. We're quite astonished with the response though.
Most heat exchanger designs have been...well pushed to its technological limits. And no application has tested this realm even more than what we're all doing. It is almost like the formula one of heat exchangers. So much effort for so little yield.
That is why targets and practical compromises on noise and heat dissipation are being made by radiator designers right now.
We're working on a different paradigm, however.
Typical radiator manufacturers have little incentive to push their limits as the markets they serve do not have the constraints that computer radiators work in, not to mention the margins to support high level R&D.
I sincerely wish you chaps the best with the PA120.
Viva la difference!
Last edited by hwlabs; 09-24-2005 at 08:52 AM.
hwlabs (mind if we actually know your real name
I belive the market is long overdue for a quiet and powerful radiator.. One that you can attach fans like the yate-loons (nexus) and get good enugh airflow to where your performance is good.
Im just throwing out an idea here, but a lower FPI, double pass radiator might be the thing to do.
If its not too much trouble, make a BIX3 or BIX2 with 12FPI and slap either the yate-loon on there or one of thermaltakes thunderblade fans. run the same size BIX (regular) with those fans and compare.. If you get higher heat disipation from the 12FPI version.. congrats.. you have a product that will sell.
I for one am tired of having to use 38mm thick fans.. they are just too flipping big. That combined with the fact that You MUST run them at 7V or below or go deaf makes all of this very annoying.
oh one other thing..
This is a request to both HWlabs and to Thermochill.
How about standardizing testing of the radiators so the end user can get some USEFUL data, and not the rather pointless dribble that is speewed out now.
Standardize the heat load, fans (maybe multiple fans), DeltaT of water to air, and flowrate.
~200watt hea load
5C Delta T (the current 10C i have been seing is rediculous unless you have a room that is like an ice cube, or are running 4x prescott based xeons)
1.2 to 1.5gpm (this seems to be where the average users flowrates are hitting with current hardware)
Not unsurprising to me at all. Radiators are already pretty decent items and are a mature technology. There are certainly no HUGE gains left to be had, but this would've been fairly obvious simply through calculation of the C/W of the air-flow vs the C/W being achieved with the radiators presently in use. These claims of "a heatercore will always perform better than a PC optimised radiator", while true some years ago, is definitely not true any longer.
Yes, single-pass is now purely a market-driven buzz-word where people are hungry for what they perceive will perform better, but not that which actually will and is proven to perform better, and this is heavily evident within the mind-share of the water-cooling community. People believe that single-pass is better and will buy it in droves, regardless of the suitability of single-pass for higher performance. A shrewd marketing person will indeed take advantage of this and provide what people want to buy. It's a choice between making money, and making something better.The Black Ice® XFlow radiators are an update stemming from requests by many customers seeking a more flow-optimized radiator for their increasingly constricted systems. We're quite astonished with the response though.
It's long overdue that radiators would be optimised for PC use. Even with 100cfm fannage, the moment you start sticking things in a case and getting choked by case orifices, the actual air-flow is much lower. For far too long radiators have been optimised around 100-150cfm fan rated air-flows and it simply is the wrong way to be doing things.Most heat exchanger designs have been...well pushed to its technological limits. And no application has tested this realm even more than what we're all doing. It is almost like the formula one of heat exchangers. So much effort for so little yield.
That is why targets and practical compromises on noise and heat dissipation are being made by radiator designers right now.
We're working on a different paradigm, however.
Typical radiator manufacturers have little incentive to push their limits as the markets they serve do not have the constraints that computer radiators work in, not to mention the margins to support high level R&D.
Doesn't concern me either way. I make no money from it. I just want end-users to have a better water-cooling experience. Having that as a number one goal means that I am unconstrained by monetary pressures to design what the market feels that they want (buzz-word marketing), as opposed to honestly pushing progress along in the direction that the market actually needs. As I often say to many companies who approach me, I value my independence - it allows me to stay focused on the end-goal. Annoys the crap out of most people though as they don't understand someone who does something for no monetary gain.I sincerely wish you chaps the best with the PA120.
Competition is always good. Enough words though. Let's see the goods Willie.Viva la difference!
His name is Willie (Wilmont Yugue), and he's the CEO of HWLabs.
Last edited by Cathar; 09-24-2005 at 12:53 PM.
I thought that Thermochill was doing a pretty good job of the above.
Cathar before he deleted his post.
G
Cathar, you simply have no idea what it takes to bring things to fruition even after recognizing the limitations that have been reached.
And we can easily say that no one else in this "industry" has spent more time and resources to bring real development into this field.
But good things come to those who wait.
Mate, that's Wilbert and not Wilmont.
It's not just about fin density. You can ram as much fins into a 120mm space to increase surface area but you get so much pressure drop you'd end up killing airflow.
Too little and it becomes laminar and you don't end up with anything.
Last edited by hwlabs; 09-24-2005 at 11:24 PM.
Well there you go, offhandedly dismissing the work of others, while espousing your own. For years and years people have directly asked you for the C/W curves of the HWLab radiators, and at what air-water delta and air-flows that the mythical heat dissipation ratings at HWLabs sites were achieved with, and for years this requests have been ignored. Once via PM at OCAU you admitted that you did not even have that data. I still have PM response from you BTW.
Excuse myself and others for being a little hesitant to believe claims of "intensive research". Such claims hold about as much weight as the heat dissipation ratings in the HWLabs advertising material.
Forgive me. My honest apologies. Don't know where I got that from. Heck I even checked it right before I typed it in, and still typed it in wrong.Mate, that's Wilbert and not Wilmont.
This is common knowledge as has been discussed publically for a long time now.It's not just about fin density. You can ram as much fins into a 120mm space to increase surface area but you get so much pressure drop you'd end up killing airflow.
Huh? Air-flow is always laminar in short of at least 200CFM through a 12x12cm radiator. We'll have to agree to disagree on that point then.Too little and it becomes laminar and you don't end up with anything.
As for me not knowing how long it takes and the effort required, excuse me, I must have amnesia with respect to my involvement in assisting ThermoChill for the last six months where we managed to develop 3 separate radiators that fairly conclusively blow the alternatives away across the range of fan powers that PC users use. You'll see them hit the market soon enough.
There you go again, being as dismissive of my work and efforts as you were three years ago over at OCAU. Be careful Willie, you're dealing with someone who's a lot more aware of what's going on than you'd like to believe.
I do so enjoy the verbal sparring that you and I engage in Willie, but the difference for me though is that I have nothing to lose because I don't earn an income from this, so I'm not compelled to inflect my words with marketing spiel.
Last edited by Cathar; 09-25-2005 at 12:50 AM.
Good of you to remember our encounters downunder, Cathar, but I do recall it was our work that was being dismissed in favor of a heatercore you saw somewhere. Big arse impressions do remain as stains on an ethical cloth. But I've considered that water under the bridge.
Cathar, just wait, mate.
People like someone you know in particular have driven us to enact measures to protect our product's designs. Releasing our stuff sooner rather than later is something that would be falling to the same level of naive trust.
It may be easier for you to deride and criticize efforts because of your independence, likewise dismissing efforts in bringing something beneficial to end users by veiled taunting remarks is just as easy. It goes both ways, mate.
Nonetheless, I do admire your work, and I hope this didn't come across as patronizing.
No disrespect to you, Cathar, I'm sure you've spent a lot of time in determining the optimum fan capacity in relation to noise and derived your preferred core configuration. And by my impressions you've done so without any form of compensation.
Certainly 6 months is a lot of time, but we've been at it far longer and the real bills we've paid are substantial. From integrating engineering disciplines around the world, hundreds of hours of wind tunnel testing, to getting support from key industry persons, and not to mention obtaining financial backing.
At this point its no longer about just tweaks anymore for us, it's re-engineering, hence next-gen. Otherwise it'd be somewhere in between.
If true improvement is the direction that you're looking at for end users, then we have nothing to disagree about.
The engineering changes to our new units are radical. The equipment required to produce parts for these are incredibly more complex (expensive), and toolings alone are so custom and precise that a simple regrind would be an easy $30,000 if a chip of copper gets stuck in between.
If these things are likewise "dismissed" as marketing banter then there's little merit in anything that we do.
I'd be more than glad to send you some of these for your own assessment when the time comes.
But do understand, Cathar, we're guarded simply because not everyone has the same kind of altruism as you do.
Potato tomato..., the reason why fins are outfitted with louvers is to create turbulence on the surface of the fins. If you have an unusually wide gap such as 12fpi in between you get a lot of airflow but you certainly don't get to use all the air for heat transfer where it counts.
Last edited by hwlabs; 09-25-2005 at 01:48 AM.
It's all good Willie. I always look forwards to seeing better products for end-users. Right up until I concluded work with ThermoChill I was still commonly recommending the HWLabs BIP3 as the high performance radiator of choice, even while assisting Thermochill, just as an example of my altruism on this.
I'm quite aware of who you're talking about with respect to your hand being forced. Heck, he even takes pride in it. Would love to see you answer in style and make him eat his pride. That would give me joy.
I've never had anything against HWLabs, as I stated 3-4 years ago, and as is true today, my goal has always been to see reliable information on performance and that's what sparked off that discourse back then. I would be one of the first in line to openly applaud HWLabs if you finally proved me wrong and stepped up to the plate with actual performance data, which is all everyone has been requesting for the last 4 years.
Good to see that HWLab's historically high profit margins have been put to good use though. I engaged a local heater-core/radiator manufacturer here in Australia and even to contract them to make something like a BIP1-BIX3 in quantities of 1000 would cost around $30-45AUD per unit for a raw core, so it is good to see prices finally coming down to more sane levels. I definitely applaud HWLabs for finally doing that, although it must be said that it did take some serious competition to get that to finally happen eh? Don't worry, I have no aspirations to be a radiator manufacturer, but I did want to be aware of what the costs were, even in a fairly "high-cost-of-manufacture" country like Australia. On top of that I am also very much aware of the ~$20K AUD tooling up charges for a final complete design, but if it's possible to use established end-tanks then the prices are much lower. Still, amortised over a production run of 10000 we're still talking manageable costs.
At the end of the day, top-performance and low costs are what people want. Sadly Thermochill doesn't fit the latter too well, so they have to focus on top-performance.
For sure, I'd love to have and test a radiator of your next-gen models. If it works better than what Thermochill and I've designed, you can be sure that you'd see me recommend it when people ask in forums, just as I've done with HWLabs radiators in the past. I'm a fair and honest person and I like to think that I've always shown that and always shown that I've had the end-user's best interests at heart. After all, what I want now is the same thing that I wanted 4 years ago. The best possible water-cooling setup. I'm just prepared to build/design/influence whomever or whatever to make it happen if at all possible.
No worries, mate. We're good.
We're working on the cost issue too. It is often easier to intro a product with little or no research by just standing up on the shoulders of those who laid the foundation.
Users will get good stuff at better prices. But know that retailers also play a role in wanting to stock up an item and sell it at a reasonable profit. The sweet spot for rads of varying quality will be determined by the market.
I guess just because we don't publicize what we do with what we've "earned". I doubt people would be interested.
But one of the major things we've done to increase quality and lower cost is to set up our own tool and die facility. I've got 2 CNC mills doing that along with a toolroom compliment and a crew of 4 people.
Tooling right now isn't as much an issue for us as the real technology and techniques we've developed for the new radiators.
The usual tool and die industry "just doesn't get it" when it comes to our requirements.
Likewise we cater to a lot of custom work, needing low volume items for manufacture.
It isn't as much as competition as preparation and scale that got prices lower from us. Our distros know that every saving we can derive from operations is passed on to them. But we still prefer to give users a choice that conforms to their budgets and performance targets.
Despite the usual flak and the ocassional uninspiring mime, I still love what I do.
We've come a long way from transmission oil coolers and this industry moving forward is a good thing to know.
Stew my comment on perfrmance data was, admittedly aimed more at HWlabs as they have never (not to my knowledge anyway) released legit performance numbers in the entire history of the company.
The part of my comment geared towards thermochill was to work with HWlabs to decide on predefined test conditions as they already do release all of the performance data / perfomrance predictions (that would be your area stew)
The problem there is always going to be on getting agreement.
BillA used to calculate of the average of the inlet/outlet temps as per more general radiator industry standard, but for PC water-cooling use the values this method gives I think isn't as useful as another method (which I'll describe in a bit).
Les at Procooling convinced Bill that he should then be calculating around the inlet temperature, because that can be controlled fairly steadily. Measure how many watts it takes to hold the inlet temperature at either 5C or 10C (or whatever) above ambient, and then the C/W is just an easy division of the wattage by the temp delta. Still though I don't feel that this gives a value that PC users actually want though, and in truth I believe that this method understates the actual C/W.
I prefer the method of calculating the difference between the radiator outlet and the air intake temperature. Why? Both of the above methods require you to know what the thermal capacity of the water is and calculate the discharge temperature of the water from the radiator. I like the wattage divided by (water-outlet - air-intake) version of C/W better because this allows the users to know what the temperature of the water coming out of the radiator will be in a simple and usable fashion, and not have to calculate it dependent upon what the flow rate is (i.e. the thermal mass rate through the radiator).
The way I see a test loop is like this:
radiator out -> temp sensor #1 -> pump -> flow controller (tap) -> flow rate sensor -> temp sensor #2 -> water heater giving fixed wattage -> temp sensor #3 -> radiator in
and for the air:
fan -> intake plenum + temp sensor #4 -> radiator -> outlet plenum + temp sensor #5 -> exhaust
Now the tester using the above setup can determine the total number of watts being dumped into the water from the pump + frictional restriction from the tap, from the heater, and overall by subtracting sensor #3 from sensor #1, and multiplying by the thermal mass flow rate of the liquid. In this way we know exactly how much wattage has been added from the radiator outlet to the radiator inlet. This is pretty much like a normal cooling loop where the radiator cools, and everything else adds heat.
So basically we set the heater such that 200W of heat, or as close to it as possible, is caclulated to be added to the water between the radiator outlet and inlet.
Now by providing what the water-out minus air-in differental is (at various flow rates) then we give the end users an exact way to calculate how warm the water will be coming out of their radiator (and into their waterblocks) depending on whatever heat-load they have. The first two methods require the users to calculate the thermal mass-flow rate of their liquid and calculate the outlet temperature, whereas the method I propose above (and use in my testing) gives a simple way to work out the outlet temperature of the radiator for a given heat-load, flow rate, and fan-type.
The above method also gives substantially lower C/W values that the first two methods, but this is just a side-effect of where the calculations are being taken from. When you calculate from the lowest water temperature (the radiator outlet) the water-air delta is always going to be smaller, but ultimately I believe that this value is more useful, and further, it allows for a way to directly calculate the radiator efficiency.
Just me 2c, and the method that I use to calculate C/W's for a radiator, and it also explains why the C/W's that I calculate are a little bit lower than BillA calculated.
Interesting. that is the same way i ended up doing the testing of the BIX and CoolRad.. Obviously minus the 3 extra sensors.
but in any case I do find the way you outlined to be the best way to give useful results to everyday users..
I can think of one very good reason........XXXX number of end tanks to use up! !
LOL!!
IMO, fin count optimisation is going to do more for performance than just a switch to single pass. Single pass with a optimised fin count would be of interest, but not single pass alone.
Edit:
Another thought, a 120.3 with lower fin count might just be interesting to have go head to head with a monster core. A interesting match up......could the 4th fan & greater surface area of the core make the differance still? Air side performance would go to the 120.3 for sure by a large margin.
Hmmmm, this could cost me money......again. LOL! !
Last edited by Craig; 09-27-2005 at 03:36 PM.
Cathar posted that the new Pa120.x type rads will be set up with fin counts to best work with fans of 20-70cfm (20cfm = YL @ 5v?).
And considering that HWLabs says it trys to provide what power users want.......
We can perhaps look forward to the new gen HWLabs product being set up for higher CFM rates for the power users (compared to the Pa120.x rads). If they make the new gen rads to perform best with fans of from 60-110cfm, then between the two companys we'd have something for everyone.
Will be interesting to see the new Thermochill's and HWLabs "new gen" rads & how they compare & where each new products sweet spot is for both noise & performance.
Craig, as I tried to impart, the PA120.x, when faced with a choice between ekeing out difference of ~0.002C/W at high fan noise vs lower fan noise, we went with the lower fan noise focused tweak.
Even when between 60-110CFM I still think you'd be hard pressed to find anything that's going to trounce a PA120.3. There's certainly nothing on that market today that will, and that includes these heater-cores that various people seem to think are "always better than any PC-radiator". Coupled with 110cfm fans the PA120.3 would still handily beat out a HE120.3. It's only as you get much above 110cfm is where the older style radiators start to come good, and that's a heck of a lot of noise.
It will be interesting to see what HWLabs brings to the table with their next-gen rads though.
I've always believed that if a radiator can offer close to as good performance with 3 x 70cfm fans as another radiator with 3 x 100cfm fans, then this still serves "power users" just as well. "Power Users" now have a choice though. They can enjoy the same cooling power for less noise, or they can keep the same noise levels and enjoy better performance.
Splitting hairs here people. I can count the number of people who will can tell the difference on my fingers, do I need my toes as well?
Again I play advocate for the poor guy as I am happy enough with BIP price and performance.
Ah well, this is how advances come along I guess. Will HW's new line be replacing current products or are they in addition to them?
ls7corvete, you have a habit of always dissing advances, although I'm not sure why. If people are after a "good enough" solution they'll stick to the stock heatsink that comes with their CPU, and certainly not bother with water-cooling. The stock heatsink that came with my 5yo daughter's 4400X2 machine does a decent enough job of keeping the CPU cool and is fairly quiet, or at least I can't hear it over the noise of the stock 6800GT cooler fan.
You are quite correct - this is how advances happen. Making bleeding edge advances is neither cheap, and certainly not targetted for the "good enough" crowd. Two years down the road the "good enough" crowd will probably end up using it anyway when the prices come down, but this is the here and now, and we're discussing the leading edge of advancements here and now. Can one thread go by without you moaning constantly about the cost of leading edge technology that benefits what the poor person gets to use sometime down the road?
Shall we not advance anything? Would that make you happier?
I've got a bad case of the flu, am tired, sick and grumpy. Excuse me.
No dissing, just saying what you said in the second paragraph. I dont expect prices to go down from the point they are at and I dont expect performance to increase from where they are at. I just think people should be more aware of the differences.
The comment on performance was more focused on the discussion rather than my veiws on price and performance. You both got good products but when it comes down to it I know I dont have the knowledge to tell/measure the difference in perfomance.
Heh, I seem to hit a nerve with you. We do have similar thoughts on many issues.Can one thread go by without you moaning constantly about the cost of leading edge technology that benefits what the poor person gets to use sometime down the road?
I think the following says more than anything though. The answer gives a big insight into the flow from cutting edge to mainstream for these new rads.
Will HW's new line be replacing current products or are they in addition to them?
you make a lot of sense stew.... but what i dont' get is why does your five year old daughter get such a kick arse rig T_T
"You can do anything, only after you lost everything" - Fight Club
Yea, whats that put you on? Last I heard it was a socket A. You should have enough parts sitting around to get her rig WCed, though I shouldnt say anything about that as Im not even WCed at the moment.
Posting Permissions
Reply With Quote
Bookmarks