Danger Den MAG II LE (review & pics inside)

I don't think that Radiators scale as well with flowrates as jet impigment blocks, at least not in a normal setup with a heatload of 100-200watts. Can you give me a link or so with the numbers you mentioned?
of course, optimizing the loop can double or even triple the flowrate. 1/2" tubing instead of standard aquacomputer tubing is a helluva difference! I tink thats most important in a loop, more important than a high priced pump.
Always try to get the best out of the components for little money. Compared to Eheim Pumps, the Mag 2 LE ist a powerful pump. Just because we don't have anymore hydor L30s or Eheim 1250s for the highend but nearly double as powerful Laing pumps, doesn't mean that 2,1m head is not enough anymore. 40$ for a silent pump with 2,1m head is pretty good for entry level and mainstream watercooling.
Its just the range i would choose if i had to build a mainstream complete kit.
Its just that realiability thing that keeps me off from importing this pump.

Quote:

---

Originally Posted by Radical_53

Well, one thing you didn't take into consideration, when talking about the differences, is radiator performance. If you take a look at the tests that were done there, radiators do scale just like modern impingment blocks with higher flow rates, which then would double the performance difference between the very weak and the very strong pump. So then, if you continue to optimize every little bit of your cooling setup, you'd have more than a considerable performance difference :)

---

According to the most recent data from Bill Adams the opposite is true. For instance, using panaflo (medium) fans at 5 volts and a load of 150watts, the output temp of the CR-22T at .5gpm is 2.2C cooler than at 2gpm (10c vs 12.2c). Even at high airflow rates (panaflo@12volts) the output temp at .5gpm is still slightly lower than at 2 gpm(3.9c vs 4.1c).

@flb: Well, I took the numbers from Thermochill's HE series into consideration here.

http://thermal-management-testing.com/Thermo10.gif

Found here: http://www.over-clock.com/ivb/index.php?showtopic=14805

I took that as a baseline. If BillA now comes out with just the opposite, well... makes me curious to be honest! I would like to try that out myself somehow. Do you have a link of where you got that data from?

EDIT: Doesn't really make sense. Both HWLabs rads we had were much worse in the x-flow version. So if low flow was better for rads with low flow fans, that should have been seen just there... *normally*

@davidzo: Of course, you're right. No one needs the most powerful pump unless the rest of the system is as far optimized as possible. I'm sure there are quite many people who think that with a DDC type pump, they can do any cruelty to their loop and still end up with good performance... as that's definitely not true :)

Quote:

---

Originally Posted by Radical_53

Do you have a link of where you got that data from?

---

Here's the link.
http://translate.google.com/translat...language_tools

I don't think higher turbulence, due to increased flow rate, effects rads nearly as much as water blocks because the heat load is being spread over such a large area. In fact if we look at waterblocks like the mcw5000, we already see a major reduction in improvement from increased flowrate.

I think temperature differential, evenness of waterflow distribution between the parallel channels, airflow speed, and evenness of airflow are some of the variables that are effecting the performance of the rads in Bill's test.

It may be that increasing waterflow causes an imbalance between parallel channels. Perhaps, when this is paired with uneven airflow, we see a reduction in performance.

The bottom line is I think there are a few possiblities that no one has really tested for (that I know of) that may explain Bill's results.

Or, it may be that his data is all crap. Until any data is verified by independent sources, I always take it with a grain of salt.

Well, I'm sure that his data ain't crap. I think we can take that as granted, regarding his previous work.
I'll take a look at his setup, maybe there is a way that I can replicate some of it with my test stuff. I'm really curious how that came together, and if so: Is there a way to prevent it.
Would be quite stupid cranking up block performance with higher flow just to loose rad performance with this accordingly, as then the overall performance woulnd't improve at all, rendering it meaningless :(

there is no way that flowrate can have a negative impact on rad performance. Only increased pump heat dissipation could have something to do with these results, so in a loop one should considerate that as well.
I think that the HE numbers you showed here radical, aren't the whole truth too... From what i can estimate, there will be no turbulence in rad chambers due to high cross section and slowly water movement. Due to laminar boundary layer effect, there will be no measurable heat transfer improvement at all in a rad at those flowrates we have in a watercooling loop. Thats what i guess and since other predicates differ a small amount to both sides this must be near the truth.

Whoa, its noisy! gonna pack it in something, but maybe its a bit less noisy when it gets some more restriction than a bit of tubing....

Quote:

---

Originally Posted by davidzo

there is no way that flowrate can have a negative impact on rad performance. Only increased pump heat dissipation could have something to do with these results, so in a loop one should considerate that as well.
...

---

The Cooling-Masters review is excellent as usual. Looking at the negative impact of increased flowrate at lor airflow I am beginning to wonder how flowrate was regulated. I see a Iwaki next to the testsetup, did they undervolt it? If they used a adjustable valve to regulate, I wonder how much the pump power draw changes between low flow-max flow.

Oh no, got som rattling in it. sounds a bit like air bibles but not quite. it sounds more like impeller slightly touching something. If i point the inlet down it goes away...

I took one look at it, saw how cheaply built it is (Designed more for looks) and decided not to even waste time thinking about it. Now seing this thread I trust my instincts even more now!

I opened it up and smoothened the outer edges of the impeller slightly, not deforming them just removing som bits that stood out. and i think that did it!

Edit: hmm, still a little rattling, seemes like it comes back after running for a while. If i squeeze it a bit the rattle seems to dissapear, dont know if its from the preassure or that it vibrates less...

The mag2 is useless. Aint going to use it, its really loud and it rattles

guess ill order a ddc-2 in january then...

Quote:

---

Originally Posted by FLB

According to the most recent data from Bill Adams the opposite is true. For instance, using panaflo (medium) fans at 5 volts and a load of 150watts, the output temp of the CR-22T at .5gpm is 2.2C cooler than at 2gpm (10c vs 12.2c). Even at high airflow rates (panaflo@12volts) the output temp at .5gpm is still slightly lower than at 2 gpm(3.9c vs 4.1c).

---

I dont know where you are seeing that at higher airflow rates the lower flow outperforms the higher flow. I see at low airflow that this is the case and if you read that entire review it explains it quite well. With a set amount of airflow you can pump the water through the RAD fast enough that it does not have enough time to cool properly. Increasing the airflow aleviates this problem as can been seen in the graph you showed.

So yes that is something to consider if you are using low CFM fans for a silent computer. You can pump the water too fast...

Quote:

---

Originally Posted by eXa

The mag2 is useless. Aint going to use it, its really loud and it rattles

guess ill order a ddc-2 in january then...

---

same problem that my mag1 has recently developed, and I'm with you the ddc-2 is the way to go

Quote:

---

Originally Posted by Bun-Bun

I dont know where you are seeing that at higher airflow rates the lower flow outperforms the higher flow.

---

Bill's data is based on input temp while I'm using output temp. At .5gpm a 150watt load will drop 1.14c from the input temp, while at 2gpm it will only drop .28C

Quote:

---

Originally Posted by FLB

Bill's data is based on input temp while I'm using output temp. At .5gpm a 150watt load will drop 1.14c from the input temp, while at 2gpm it will only drop .28C

---

And how do you derive that from the graph you attached?

Quote:

---

Originally Posted by Bun-Bun

And how do you derive that from the graph you attached?

---

The temp drop is based on a proven scientific formula

Quote:

---

Originally Posted by FLB

The temp drop is based on a proven scientific formula

---

Which is?

EDIT: I went back and re-read the review again. Bill fixed the input temperature of the water at the inlet thus it makes perfect sense that the DeltaT is smaller the more flow you put through. The faster you move the fluid the less time it has to cool off, however it cools more water at the same time thus disapating more heat. Therefore in an actual loop the overall temperature of the water would be lower becasue more heat would be disapated and there is not a fixed temp going into the RAD.

EDIT: Flow is gph not gpm.

Quote:

---

Originally Posted by FLB

Flow is gpm.

---

lol you posted just as I edited my last post. Yeah I know that equation. Been a few months since I graduated I couldn't remember it off the top of my head...

Quote:

---

Originally Posted by davidzo

there is no way that flowrate can have a negative impact on rad performance.

From what i can estimate, there will be no turbulence in rad chambers due to high cross section and slowly water movement. Due to laminar boundary layer effect, there will be no measurable heat transfer improvement at all in a rad at those flowrates we have in a watercooling loop.

---

Well, I don’t agree with your first statement, but I believe your second statement is dead on. I’m coming up with Reynolds numbers around 930-935 thereabouts for 2GPM. It seems like the new generation of radiators are operating well within the laminar flow range.

There’s one thing that people never seem to pick up on when discussing the type of charts that were recently posted. When flow rate is increased, but the delta between inlet water temp and air temp remains the same, you have effectively increased the amount of energy being delivered to the radiator. It is exactly the same as taking your CPU from idle to full load. The radiator dissipates more heat at higher flows simply because it’s getting more heat!

As far as I am aware, no one tests radiators using a test rig that simulates a PC water-cooling loop. That is, a constant amount of energy transferred to the water regardless of flow rate and water temperature (within reason, of course.) THAT is what happens in our loops.

For years I’ve said that when flow increases, the radiator suffers in performance due to the lower deltaT. But this isn’t necessarily bad if the rad has excess cooling capacity to cover for the loss.