What happend to the Nexxos XP review?

[Cathar]

Just to add some data points to this particular graph (mostly for my own interest):

http://img34.exs.cx/img34/5439/nexxus-mcw6000-temps.gif

White Water
Eheim 1046 = 14.21
Eheim 1048 = 13.49
Eheim 1250 = 13.14
MCP600 = 13.02
MCP650 = 13.16

Cascade
Eheim 1046 = 13.74
Eheim 1048 = 13.17
Eheim 1250 = 13.02
MCP600 = 12.86
MCP650 = 13.04

Storm/G4
Eheim 1046 = 13.71
Eheim 1048 = 12.98
Eheim 1250 = 12.74
MCP600 = 12.43
MCP650 = 12.77

Storm/G5
Eheim 1046 = 13.25
Eheim 1048 = 12.62
Eheim 1250 = 12.18
MCP600 = 11.78
MCP650 = 12.17

[Cathar]

Originally Posted by lolito_fr

pH, I think what the average Joe wants to see is something like this:



needs a disclaimer though (in this case at least )

edit: have replaced graph with new one factoring in pump heat as per Cathars data

May be a slight miscalc there for the XP/MCP600 bar. I worked it out as 12.76C. Everything else looks fine.

[pHaestus]

Cathar:

Took your numbers and Lolito_fr's and put the attached graph together. I tried a straight bar graph but it got too complicated to really make sense of. Hopefully this one is a little better?

[Cathar]

Originally Posted by pHaestus

Cathar:

Took your numbers and Lolito_fr's and put the attached graph together. I tried a straight bar graph but it got too complicated to really make sense of. Hopefully this one is a little better?

Looks good. I still think that something's up with lolito_fr's XP/MCP600 data point. A data point of 12.76 is what I worked out. Everything else on his graph I agree with. Suspect he may have mis-eyeballed the flow rate (have done so plenty of times myself). He must have, as the MCP600 and MCP650 give near identical flow rates (within 0.03gpm) for either the XP or the MCW6000, so the relative difference temp differences for the MCP600 and MCP650 for either block should be about equal, and they're not.

What you've done there is a nice way of presenting it.

[DDogg]

Sincere thanks for the Hydraulic Power idea, explanation, and all the work all you guys contribute to the cause.

You mentioned the average Joe (like me I think). It has been very difficult to follow some of you because the data is spread out in many posts as you continue to work on the theories and debunk the myths.

Now that I understand the simplicity of the chart, I am shocked at how this method so elegantly clarifies the complex set of data points that gave me such a major migraine.

I understand the idea has to be progressed through the critical peer review filter, but I very much hope it proves out as a practical way to condense the data. It would be heaven sent for many of us, and if I understand correctly, near impossible to refute by the 'agenda biased'. wtg!

/Add: If possible, could you guys include tabular data with the graphs later on. These line graphs are a bitch to get a decimal from :-) [The bar graph is so much more intuitive <imo>]

[Les]

Originally Posted by Cathar

What you've done there is a nice way of presenting it.

Also like pH's presentation of data..
Should be $40 in pH's PayPal - to buy 2m of 12mm tubing , contribution to a HE120.2, beer or icecream

[pHaestus]

You're too kind Les.

[Les]

Originally Posted by pHaestus

You're too kind Les.

Had a good day at Punchestown

Edit1: "As the market principals got downto business off the home-turn, Paul Carberry was sitting still on Harchibald, who made up the ground smoothly on the inside on the run to the last and was quickly in command.
Noel Meade has always had a high opinion of the winner and may send him to Kempton over Christmas, before targeting the Champion Hurdle." Racing Post

Edit 2 Not sure whether Harchibald is good enough to win The Champion(hurdling).
However,if remaining sound, STRONG FLOW may well win The Gold Cup(chasing). I think a little of the 12-1, readily available, would not be ill-advised

[Cathar]

A "me too" version of Phaestus's graph, with some corrected data points:

[lolito_fr]

Thanks for the thumbs-up and the new graphs, guys.

Cathar, some accuracy was lost in GPM-LPM conversions, eyeballing, and polynomial/power function best fit approximations. Therefore, I'm almost surprised that you only found one point in disagreement
(Your graph will inherently me more accurate than mine.)

As you have already said, it would be useful to have the raw data available somewhere on the site to avoid some of these issues.

Edit: Choice of pump seems quite determinant here
(My only quibble is with the 15w of heat produced by the D4. Is this at 12v? )

[Cathar]

Originally Posted by lolito_fr

Therefore, I'm almost surprised that you only found one point in disagreement

There were others that differed by 0.05C or so, but I didn't feel that was as significant as ~0.25C. The graph above is my take on it all. The E1250 data point was also somewhat out (I felt). Again - small diffs though.

Good work lolito_fr. Am glad when we're all working together to generate more user-friendly data.

[BalefireX]

Cathar: is your graph taking into account the pump's heat output?

[Cathar]

Originally Posted by BalefireX

Cathar: is your graph taking into account the pump's heat output?

Indeed it is.

[DDogg]

Is there a way to extrapolate for a higher load temp? Say double the heat in Cathar's graph above - 142 watts - linear increase? I guess I am wondering if different block designs would have different response curves with double the heat input. Many push the XP-mobiles to 2.0-2.2 VCore range. I've yet to see a way to accurately calculate the resulting heat generated at these VCores, but 142 watts may not be too far off for discussion purposes.

[psychofunk]

If I wanted to stick the Iwaki MD20RZ on that graph, how would I go about doing it?

[lolito_fr]

One of the problems if you want to do this scientifically, is that pH can only test the blocks up to a certain flowrate with his pump. So you'd have to extrapolate somewhat for an MD20RZ (which is not as easy as just reading values off a graph).

Otherwise you could always guess...
My bet is it would look something like the Eheim 1048. Bear in mind it puts out about 31w of heat, which is A LOT compared to the 71w from this cpu. But then if you have a better rad, and more airflow, things may not look so grim

Edit:



Also added Laing DDC based on this PQ graph by Roscal
http://img103.exs.cx/img103/4282/courbespompes.png
Heat output guesstimated at 75% of input power

[Brians256]

Nice job, guys. Makes me wish I had spare time, too. This is exactly what Joe Blow watercooler needs.

[lolito_fr]

I have updated my graph to include the Iwaki MD20rz at 60Hz, and also the new Laing DDC (Swiftech MCP350?).
Have estimated heat ouput at 75% of input power, based on figure of 90% for Laing D4 (MCP650) which uses metal between the motor and pump, rather than plastic for the DDC.
Performance looks promising for such a small pump

[Cathar]

Laing DDC is more like an 8W heat-dump in typical use.

Check out Roscal's graphs. He actually measured it.

[psychofunk]

Wow, so the MD20 would work better for the XP and worse for the 6002. I guess that is another one for "you can't just throw a bigger pump at it and expect it to work better".

[redleader]

Cather:

You're using perl to do those graphics, right? GD::Graph or something else?

[Cathar]

Originally Posted by psychofunk

Wow, so the MD20 would work better for the XP and worse for the 6002. I guess that is another one for "you can't just throw a bigger pump at it and expect it to work better".

It depends on three things:

1) The amount of gain that the waterblock can be expected to receive from the higher flow rates
2) The amount of CPU heat. The hotter the CPU, the greater the CPU temp deltas, thereby the diminished role that pump heat plays in the overall picture.
3) Radiatior performance - a small radiator will hate a hot pump, yet larger radiators (3 or 4-fan) can "shrug off" an extra 20-30W of pump heat with minimal impact, thereby allowing the increased flow rates to the waterblock to have a positive overall impact.

The MCP600 is such a sweet balanced pump for a wide range of blocks.

[Cathar]

Originally Posted by redleader

Cather:

You're using perl to do those graphics, right? GD::Graph or something else?

I'm using gnuplot

[pHaestus]

Should we revise the discussion to reflect the current commercial names of these pumps? The MCP600 isn't currently sold by Swiftech is it? Is Cooltechnica the only worldwide provider at this time (as the AquaXtreme 50Z-DC12), or are there other providers in Europe?

As for the MCP650 is it also marketed as the D4 by Danger Den, or do they also call it the MCP650?

Just striving for clarity here...

[BillA]

it was perfectly clear to me . . . .
lol