Cooling Which tube diameter ??

Hi, I've got this dilemma over whether to opt for 3/8" tubing (more flexible & easier to set up) or 1/2" (better cooling). The components I'm looking at are DD CPU and GPU blocks, an Eheim 1048 pump and a Thermochill 120.2 or Airplex Evo 240 radiator. I'd also like to use an Aquatube or Tank-o-matic reservoir.

If I opt for 3/8" tubing, I'm not sure if I can replace the barbs on the DD blocks with 3/8" ones, and if I opt for 1/2" the I'd be worried about flow restriction through either of the reservoirs.

I'm also not to sure about my choice of radiator. Would one Black Ice Extreme be better than the Thermochill 120.2 ?

Any other suggestions on good alternatives would be great.
N.B. I'm in this for noise reduction as oppose to overclocking

Cheers,

Olly

 

Hi Olly,

If you use tygon or silicon tubing you can get 3/8" over the 1/2" barbs.... The airplex rad's are supposed to be very efficient, so I hear.

TBH I think the higher pressure with 3/8" is a good thing (awaits flaming) this is a pretty hot topic. There's a lot about it over at the pro cooling fourms.

What I can say is 3/8" (well slightly smaller atm) works great for me and I have a higher powered pump than the one you're intending to use... it works fine with the aquatube as well

 

Flaming at your request.

Only kidding.

Essentially, the higher the pressure drop in the entire cooling loop, the lower the overall flow rate for a given pump. It has been agreed, that for reasonable ranges of flow rates, higher flow is better, up to a point which most people can not expect reach.

It is quite true, that some of the best waterblocks have a high pressure drop. But this pressure drop is a result of dense fins or other features which lead to enhanced cooling.

Let me put it this way.

The pump can only put so much energy into the water.

To get the same flow rate in narrower tubes, the flow velocity will be greater. Hence there will be greater friction against the tube walls, which combined with the overall lower diameter, will sap energy from the flow. This a conversion of kinetic energy into thermal energy at the wall, effectively heating the water.

In an ideal world, we would want zero flow resistance between the components, so that the energy input by the pump is only used to force water through the rad and blocks.

As for the airplex rads. They use very narrow tubing, adding a huge pressure drop to the system.

If you want to use american blocks, which are built the american way, ie bigger is better, where they are designed for as much flow as they can get, then DON'T pair them up with a german rad designed for loops with optimum cooling at very low flow rates.

The trick is to try and get matched blocks and rad, which have a max efficiency at roughly the same flow.

The airplex and the danger den blocks are at the opposite end of the scale as far as flow requirements.

It all comes down to design principles. Many of the german companies have opted for low flow, highly restrictive blocks/tubing rads, and these systems do work, though only at low flow rates. They will not however be able to compete with a white water and a couple of heatercores in parallel with a huge IWAKI pump flowing 180gph.

8-ball

 

The 8-Ball has spoken I stand corrected

Does that mean mu setup is pants?

Blackice Micro
Maze3
Aquatube
Eheim 1250

It seems to cool pretty well (only an o/c'ed 1600) and I'm happy with it at the mo but how would the 8-Ball improve it assuming reasonable financial and size restraints....

/Sorry for hi-jacking Olly

 

You sound very authorative on this subject 8-ball.
Whilst I restrain and compose myself before addressing some of your points, would you care to quantify your statements and make it absolutely clear to us what experience you actually have with the AirPlex Rads?



Would you be referring to the AirPlex or the AirPlex Evo in the statements you made?

You may wish to rethink some of your statements. People look to you for advice when they believe you have a greater knowledge of the subject than they do, so please be careful what you say.

Thanks.

 

The 1250 is quite a powerful pump and will be throttled back by the black ice micro.

The best way to improve that, in my eyes would be add another BIM, in parallel. so the pressure drop for the two rads in parallel would be lower.

If anything, the overall flow rate would actually go up, increasing the efficiency of the block, while dropping the flow in the rads, probably back to a point where they operate most efficiently.

Though there are many ways of doing it, and this might not be the best way to do it. It just seems more sensible to add to what you alredy have rather than replacing parts.

8-ball

 

I admit, I have no experience with the airplex, though I was under the impression that it was a multipass tube + fin design.

This style of rad inherently has a higher pressure drop than a heater core style two pass design.

Lets suppose the rad is 12 pass and is a foot long, that's 12 feet of narrow diameter tubing, compared to a 2 pass heater core of the same dimensions, which has 2 feet of channels of a much greater cross sectional area.

Having argued the point of flow rates affecting performance many times over, as well as studying BillA's experimental data, this is how I think it goes.

In general, most rads have a peak efficiency around 1-1.5 gpm, with the heatercore syle performing at their best at a higher flow rate than the tube and fin ones. However, it is also the case, that the tube and fin rads can be made to perform extremely well, as in the case of the airplex rads. My point is that their efficiency tends to peak at a lower flow rate.

Now comparing blocks, and looking at BillA's data, blocks such as the innovatek perform much better than their amaerican counterparts at low flow rates, though at higher flow rates, the large capacity blocks such as the D-tek's and the Swiftech's come into their own. However, these setups often require big expensive pumps to get the best out of them. People on the two sides of the atlantic have simply chosen to approach the problem from different perspectives.

I am not saying that any of the components mentioned are bad, it is simply a case that pairing off a block which performs best at high flow rates with a rad which peaks at low flow rates is gonna result in a system that will "probably" not perform as well as well matched components.

Your system is a good example of this, as you are always saying that you don't need high flow rates or large diameter tubing. And in your case this is perfectly true. But, to get the best out of a danger den or a swiftech, you need high flow rates, and if that's the case, they are best paired with a rad which also performs "at its peak" at higher flow rates.

Having said all of that. If I am wrong, and the airplex is indeed a heatercore style radiator with a very low pressure drop, then I apologise.

8-ball

 

Are MaxXxpert MXL-MONO rads any good and could it be used with the BM? I guess not as the pressure will be different in the two, would it be better than the BM on its own as I may be able to get one cheap.... Cheers

I thought the Airplexs were really good??

 

I'm not familiar with the maxXxpert rads. If you provide a link, I'll have a look.

As for the airplex.

I never said they weren't good. If I thought that was the case, I would let you know, and I think that's where pug has got me wrong.

I am not saying they are bad radiators, just that you need to pair your equipment carefully.

Again, pug's setup is a perfect example of how the airplex can perform well, provided it is paired with the correct components.

It's like trying to interchange components between typical american engines and modern high efficiency small capacity engines. You won't get the best possible performance, but that's not to say either engine is not good when used correctly with the right components.

Pug,

I have checked the website, and the airplex evo is indeed a tube and fin design, though I can't tell about the airplex. Could you go into some more detail about the construction of the airplex versus the airplex evo.

8-ball

 

It's made by Innovatek I think... most of the stuff about it is in German but here's a link

It' looks like tube and fin design is this a good / bad thing in the context of my setup?

 

Olly, looking at your requirements, I would suggest going for the airplex evo and the eheim 1048, but get some of the matching aqua-computer blocks, as these will offer a greater efficiency than the danger den blocks at the lower flow rates you will inevitibly get with the rad and the pump you specify.

You can get most of this from an aqua-computer stockist, such as www.wizarddesigns.co.uk, cutting down on shipping costs.

Either that or go with the DD blocks and the thermochill rads, and get an eheim 1250 to push both of these into their otimum performance range. Again, you can get all of these from www.pclincs.co.uk.

Note neither shops sell the res you're after. (I don't think anyway.)

8-ball

PS hope everyone (pug) is happy now.

 

8-Ball, there's a better picture here... the kit in the review seems ok but I don't know about the rad in isolation.

Would this be better that one BIM? What about 2?

Cheers

 

Thanks for clarifying 8-Ball.

The thing is, the AirPlex and AirPlex Evo are two completely different designs of radiator.
The Evo is a multipass copper rad whilst the AirPlex is a twin pass Aluminium design (using end tanks).

Aqua Computer have actually dropped the older AirPlex from their range in favour of the newer Evo due to its popularity having all copper internals but I have requested that they keep the older AirPlex available for me to stock because it performs so well in a variety of setups for many of the reasons you state (using blocks as diverse as the Maze to the White Water).
With the end tank construction, the flow is split between seven twelve inch long tubes on each pass, each with approximately 8mm diameter, which allows very good cooling using the full surface area of a pair of 120mm fans.

In my experience, it is possible for this rad to make more efficient use of low cfm fans than the alternative copper multipass design.
I am currently trying to source some test equipment to corroborate my findings on this but thought I'd mention it anyway.

I also like to use the AirPlex when configuring lightweight LAN party rigs due to its Alu construction.
The mixed metal horror stories are easy to avoid by using the correct inhibitors and flushing your system (like servicing a car, only lots cheaper) every six months or so.


Regarding the point you made about small hoses, the Rads can be supplied with barbs for hose of 1/2" Inner Diameter (without the use of thread convertors), hardly what I would class as small.
Just 'cos I use 8mm OD, doesn't mean it's the only option.

 

Thanks for clearing that up.

I apologies for any confusion I may have caused by my earlier posts.

I get the impression that the AC blocks are not quite as restrictive as the innovatek blocks and could probably benefit from a little more flow as a result, so a 2-pass design would be more effective for this.

As for the galvanic corrosion, I guess that comes down to whether or not you trust yourself to remember to flush the system every once in a while. If the state of my car is anything to go by, I think I'll stick to all copper just to be on the safe side.

That said, I probably won't be moving my case that much.

As for hose diameters, the lower the flow rate, the less of an issue the tubing size becomes, but if you're going for big flow rates, the kind you'd need to get the best performance out of many of the american blocks, you're gonna need 1/2" ID tubing.

8-ball

 

Jetset.

The maxxxexpert rad is typical of the style of watercooling rads used several years ago. To be honest, you can do better with many of the newer additions to the market, such as the Black Ice range, thermochill range, and many of the german radiators, or a heatercore from a scrap heap if you're trying to save money.

8-ball

 

Forgot to mention -
The AirPlex is available with 1/2" barbs and has no problem whatsoever being paired with any current blocks on the market, including American offerings.
The radiator itself is actually made by Serck and has just been modified by AC for improved fittings and mounting options.

Oh and I do stock (and use) both reservoirs mentioned (Aquatube and Tank-o-matic).

 

8-Ball....

I am thinking of changing all the barbs in my 3/8" setup to 1/2" but still staying with 3/8" tubing as I find it easyer to work with. Will this help increase the flow rate?

Also I have been offered two of the maxxpert rads cheap and have been thinking of setting them up in parallel, but am wondering what sort of performance these would provide or whether I'd just be better off with a single black ice extreme or some thing?

Thanks

 

Rather than the black Ice rads, I would go with the thermochills.

BillA has been testing them as part of his last write up before starting work at Swiftech. He was very impressed with them and is now using them in his rig.

As for his credentials, BillA has done some of the most thorough radiator testing ever, with his initial radiator roundup requiring 6 months of data collection!!!

I think I will be going with one of these.

As for the tubing, running 3/8 tubing with 1/2 fittings will remove some resistance as 3/8 fittings have an ID much less than 3/8, while 1/2 fittings have an ID near that of 3/8 tubing. Make sense?

However, it will still not have as low a resistance as 1/2 tubing throughout, though I get the impression you understand this.

8-ball

 

Though it does depend on the type of tubing you get as to whether it will stretch over the fittings.

Oh and the thermopchill rads are cheaper than the Black ice rads.

Available from pclincs.co.uk

8-ball

 

Cheers.... Yes I understand that 1/2" tubing would give less resistance over all. What sort of absolute performance improvement would be gained by going totally to 1/2"?

I take the Maxpert rads are a total no no... You mentoned before that they were an old design... is there lack of performance to do with the way the tubing snakes arounf the rad rather than using a tank at each end of the rad?

Also which thermochill rad would you recommend or is bigger just better?

Thanks again