A challenge!

[Incoherent]

Originally Posted by pHaestus

BTW do you normally mount that wb with the angled outlet going up or down? I had some disastrous mounts (pulled PSU plug at 75C) with barb mounted going up. Just remounted with outlet facing down and temperatures seem a good bit better. Maybe some temp compression... I need to check a couple things out on that test system now that I have configured it a little differently and make sure all is well this evening

For me, down too. Damn hadn't thought about the hose stress aspect, that angled solution was always a bit temporary.

[pHaestus]

all is well no worries etc etc. Give me about an hour and I'll post a "C/W" vs. flow graph.

[BillA]

schedule on C/W vs. Head Loss ??

we are ALL waiting

[Cathar]

Originally Posted by unregistered

schedule on C/W vs. Head Loss ??

we are ALL waiting

Hmmm, given that pump head pressure varies dependent on flow rate, I'd just be happy with:

1) C/W vs Flow
2) Head Loss vs Flow

[BillA]

ok, how is that different ?

C/W vs. Flow + Flow vs. Head Loss = C/W vs Head Loss
no ?
'cuz thats how I've been doing it

[Cathar]

Originally Posted by unregistered

ok, how is that different ?

C/W vs. Flow + Flow vs. Head Loss = C/W vs Head Loss
no ?
'cuz thats how I've been doing it

Sorry Bill, my intentions here are to keep achievable flow always on the graph where people can see it.

The issue I have with a C/W vs PD graph is that when people start linking to it without flow information, then it can present very low P.D. blocks in an overly favorable light. If a low P.D. block matches the performance of a high P.D. block at 1mH2O, but the high P.D. block was pushing 2LPM, while the low P.D. was pushing 15LPM then that puts a remarkably different perception on the pump that is needed than merely looking at P.D. alone. One cannot easily map such information back to the commonly available pump PQ curves.

With a Flow vs Head loss graph, one can very easily grab a pump PQ curve, overlay it, look at the intersection of the two curves, find the flow rate, and then match that up to the block's C/W on the C/W vs flow graph. Yes, I know that there will be other restrictions, but it'll help if stops the users from looking at the outmost right end of C/W vs flow curves, and start focusing on the best that they could achieve given a pump in hand.

i.e. A C/W vs flow and a P.D. vs flow graph is more immediately useful, and inherently easier to use and understand, for the lay-person, with less chance of them making false conclusions due to mis-mapping a C/W derived from a P.D. vs C/W graph back onto the C/W vs flow graph.

Pump's PQ curves are already head vs flow, and it makes most sense to me to keep to that data-style presentation since people are already used to it, and even more so, can finally get to see a real-world use for it.

[Cathar]

As a follow up, it would be then nice to see on a mixed graph, much like the present flow vs performance graphs at Procooling something like:

[graph]

Left side column of radio buttons:
Selection of pumps

Right-side column of radio buttons:
Selection of block

Pick two (or more) buttons, and it overlays the pump PQ curves over the block PQ curves, and the end users can easily see the intersection, to then apply to a block C/W vs flow curve.

I see this as a means to an end to help users, and the first step towards producing a pseudo-simulator that allows people to "build" a virtual kit, and then analyse the predicted performance.

[pHaestus]

Just settle for this:



As far as head loss #s go I am fiddling I am fiddling. Hell I'd be happy to jump right on that if I could get Labview to just output a txt file from the current I can currently see in the 3478a vi.

[Les]

Think Incoherent's unfinished-die-simulator is giving sensible results.
Roughly the block gives similar results to the WWs ?
Incoherent's "C/W" = 0.182c/w @ 2.73 LPM
Adjusting thermistor offset to 2mm and die area(109mm^2) to 100mm^2 get ~ 0.22c/w
Bill gets "C/W" ~ 0.2c/w @ 2.5-3 LPM for the WW

[pHaestus]

I really should have done more points on that LRWW test, but I figured since it was no longer commercially available I'd cut a corner or two...

Ask and ye shall receive; Bob sent me a vi that lets me log current from the 3478A. Now perhaps I should fiddle with the counter on that 3497A and see about full on automation of dP/dQ curves*

*or log mA in Labview and flow with the CF633 and merge the logs manually

[Les]

Near enough the same at ~0.75gpm(2.84LPM)

[Incoherent]

Originally Posted by pHaestus

I really should have done more points on that LRWW test, but I figured since it was no longer commercially available I'd cut a corner or two...

Ask and ye shall receive; Bob sent me a vi that lets me log current from the 3478A. Now perhaps I should fiddle with the counter on that 3497A and see about full on automation of dP/dQ curves*

*or log mA in Labview and flow with the CF633 and merge the logs manually

Lovely. Thank you pHaestus.

A little better than anticipated but very restrictive. (head loss, head loss... )I take it that the maximum flowrate point is what was achievable.
Rue-ing the mistake that forced me to block a quarter of the tunnels. Probably wouldn't make too much difference on the small die though.
Very very happy to have a baseline. I will return to my numbers in a few days after the holidays, so some better comparisons can be done, but at this point the 0.182C/W@2.73lpm number stands.
Hard to quantify the temperature compression. We wait for your diesim setup.
Have you decided on a die size, material, probe placement?

Thanks heaps

[Groth]

Originally Posted by Cathar

Pick two (or more) buttons, and it overlays the pump PQ curves over the block PQ curves, and the end users can easily see the intersection, to then apply to a block C/W vs flow curve.

Working on it. Given sets of polynomial coefficients, I can overlay a block[+tubing[+rad]] PQ curve on the pump curves.

Raw data for block and pump PQ gladly accepted.

[BillA]

just the wb P-Q data ?

following the lead of H, we have an equation that relates pressure to temperature
do you want that too ?

[Groth]

Wb, rad, and pump PQ data, all donation accepted. I can, uhh, handle the tubing on my own.

[H]eck, yeah! If you [h]ave that equation, scribble it on a [h]undred dollar bill and mail it to me.

[Blackeagle]

Groth would you please include the Iwaki MD15 & MD20RLZT pumps?

Thanks

[WAJ_UK]

I don't know of this will help you at all Groth but I have got some raw data together. It's hardly comprehensive but it'll give you something to play around with at least

I just thought I had better add that I can't claim credit for the data in the excel file its from:

1m of tubing, using http://www.tasonline.co.za/toolbox/pipe/velfirc.htm to calculate pressure drop

a ww, using figures from http://thermal-management-testing.com/white%20water.htm

a FEDCO 2-274 using figures from http://forums.procooling.com/vbb/sho...&postcount=108

and pump info from http://forums.procooling.com/vbb/showthread.php?t=5068

[Groth]

It'll be a couple days before I have a chance to play with that, but it looks like good stuff.

Thanks WAJ!

Edit: Odd, it didn't increase the number of views to 1 after I downloaded it.

[kronchev]

As I see it...

At low flow rates, European systems win. Thats why everyone touts them as "quieter", because you can use smaller pumps. Not sure how fans got ignored in that theory, but whatever.


At high flow rates, American blocks will win, while European designs don't gain anything.

That is the big difference and I can see how people can compare that to American vs German cars, as in, the American design is more about displacement than efficiency. I would tend to believe this if the European blocks gained from high flow rates, but they DONT. They sure do look nicer, they use smaller tubing, smaller pumps, and smaller rads, but I have NEVER seen any perform better.

I really can't wait for this.