Go Back   Pro/Forums > ProCooling Technical Discussions > Testing and Benchmarking
Password
Register FAQ Members List Calendar JavaChat Mark Forums Read

Testing and Benchmarking Discuss, design, and debate ways to evaluate the performace of he goods out there.

Reply
Thread Tools
Unread 10-08-2005, 12:31 PM   #26
Incoherent
Cooling Savant
 
Join Date: Sep 2003
Location: Vallentuna, Sweden
Posts: 410
Default

Quote:
Originally Posted by Les
0
What do you predict for 100mm^2?
Incoherent is offline   Reply With Quote
Unread 10-08-2005, 12:34 PM   #27
Les
Cooling Savant
 
Join Date: Oct 2001
Location: Wigan UK
Posts: 929
Default

It on the graph and in the Excel.
Get errors iwhen try to open Excel but downloading with Flashget works fine.
Les is offline   Reply With Quote
Unread 10-08-2005, 12:42 PM   #28
Incoherent
Cooling Savant
 
Join Date: Sep 2003
Location: Vallentuna, Sweden
Posts: 410
Default

Quote:
Originally Posted by Les
It on the graph and in the Excel.
Get errors iwhen try to open Excel but downloading with Flashget works fine.
Can't open it, I'll get it later.
Wait a half hour, I'll post something...
Incoherent is offline   Reply With Quote
Unread 10-08-2005, 01:16 PM   #29
Incoherent
Cooling Savant
 
Join Date: Sep 2003
Location: Vallentuna, Sweden
Posts: 410
Default

First measurement.

I have NOT compensated for heat shadowing, something that may be quite significant on the 10x10mm setup. The TIM is SX-25, value is a little higher than I expected, although I haven't checked back my data.
I have spent all day calibrating the sensors, from the triple point of water (0.01°C)to boiling point (100°C) and generating Steinhart coeficients for the thermistors.
My variation across that range is <0.1° between sensors, two resolution steps.

Dimensions for those who want to calculate the heat shadowing effect.:

10x10mm

Waterblock surface >> fluxblock sensor, 2.23mm
Fluxblock sensor >> FB bottom surface, 2.00mm
Die top surface >> Die top sensor, 2.00mm
Die top sensor >> die lower sensor, 5.00mm
Dimensions to hole centres.
All sensors are in 1.5mm holes drilled 6.00mm.

Les, you're good.

I'm of to kick my heels up. I will be hurting tomorrow.
Attached Images
File Type: jpg Les_compared.jpg (153.2 KB, 18 views)
Attached Files
File Type: zip mcw6000_on_100sqmm_2.zip (234.0 KB, 3 views)

Last edited by Incoherent; 10-08-2005 at 01:49 PM.
Incoherent is offline   Reply With Quote
Unread 10-08-2005, 01:29 PM   #30
Nugit
Cooling Neophyte
 
Join Date: Jun 2003
Location: Aalborg, Denmark
Posts: 85
Default

Getting errors from that zip file Incoherent.

! C:\...\mcw6000_on_100sqmm_2.zip: Unexpected end of archive
! C:\...\mcw6000_on_100sqmm_2.zip: The archive is either in unknown format or damaged
Nugit is offline   Reply With Quote
Unread 10-08-2005, 01:35 PM   #31
Les
Cooling Savant
 
Join Date: Oct 2001
Location: Wigan UK
Posts: 929
Default

Your zip also will not open.
Again OK with Flashget
Think agreement is just as likely to be cancellation of errors as correct parameters.
However it is reassuring that effect of die size seems to be accurately predicted.
Have sent pm and mailed with bit of Excel.

Edit On closer look "effect of die size seems to be predicted(ish)"
Attached Images
File Type: jpg Fantasy14.jpg (46.9 KB, 13 views)

Last edited by Les; 10-08-2005 at 02:23 PM.
Les is offline   Reply With Quote
Unread 10-08-2005, 01:39 PM   #32
Incoherent
Cooling Savant
 
Join Date: Sep 2003
Location: Vallentuna, Sweden
Posts: 410
Default

Quote:
Originally Posted by Nugit
Getting errors from that zip file Incoherent.

! C:\...\mcw6000_on_100sqmm_2.zip: Unexpected end of archive
! C:\...\mcw6000_on_100sqmm_2.zip: The archive is either in unknown format or damaged
Same problem as me for Les's. Some pro cooling firewall thing?
Incoherent is offline   Reply With Quote
Unread 10-08-2005, 09:22 PM   #33
Les
Cooling Savant
 
Join Date: Oct 2001
Location: Wigan UK
Posts: 929
Default

Quote:
Originally Posted by Incoherent
.

Dimensions for those who want to calculate the heat shadowing effect.:

10x10mm

Waterblock surface >> fluxblock sensor, 2.23mm
Fluxblock sensor >> FB bottom surface, 2.00mm
Die top surface >> Die top sensor, 2.00mm
Die top sensor >> die lower sensor, 5.00mm
Dimensions to hole centres.
All sensors are in 1.5mm holes drilled 6.00mm.

Les, you're good..
You're joking.
I conveniently ignored last tine round.and did not see others rushing in
However would, if forced with a gun to my head, multiply C/W(Tin) by 1.023 .
Attached tentative effort
Will think again tomorrow.
Attached Images
File Type: jpg Fantasy15.jpg (50.4 KB, 9 views)
Les is offline   Reply With Quote
Unread 10-09-2005, 03:42 AM   #34
Les
Cooling Savant
 
Join Date: Oct 2001
Location: Wigan UK
Posts: 929
Default

Quote:
Originally Posted by Les
,,,,,,,, multiply C/W(Tin) by 1.023 .
......
Crap.
Got it "arse 'bout tit"
Or have I ?
Having logic problems : befuddled(not booze).
Need more time to get my head round
Les is offline   Reply With Quote
Unread 10-09-2005, 05:27 AM   #35
Les
Cooling Savant
 
Join Date: Oct 2001
Location: Wigan UK
Posts: 929
Default

Coming to the opinion: That "the shadowing correction" is only required to estimate conditions(Q and fluxboc/TIM interface's T) if a solid-unsensored-block was used.
A sensored-block is used in the test rig and conditions are described by uncorrected calculations.
Not happy
Help !
Les is offline   Reply With Quote
Unread 10-09-2005, 11:10 AM   #36
Incoherent
Cooling Savant
 
Join Date: Sep 2003
Location: Vallentuna, Sweden
Posts: 410
Default

Quote:
Originally Posted by Les
Coming to the opinion: That "the shadowing correction" is only required to estimate conditions(Q and fluxboc/TIM interface's T) if a solid-unsensored-block was used.
A sensored-block is used in the test rig and conditions are described by uncorrected calculations.
Not happy
Help !
It's a bit tricky.
I've made a little Femlab model showing an isotemperature plot of the Flux die.
The effect of the holes is pretty obvious. (I wonder why I didn't drill all the way through for symmetrys' sake).
Looked at like this there is no obvious problem. There is a compression of the temperature isosurfaces but they are rather flat at the height if the hole centre...


more follows...
Attached Images
File Type: jpg FEM_dieholes.jpg (57.3 KB, 18 views)
Incoherent is offline   Reply With Quote
Unread 10-09-2005, 11:23 AM   #37
Incoherent
Cooling Savant
 
Join Date: Sep 2003
Location: Vallentuna, Sweden
Posts: 410
Default

...following on

A more comprehensive model of the die. With and without holes:



More follows...
Attached Images
File Type: jpg fem-die_big.jpg (183.6 KB, 22 views)
File Type: jpg fem-die_noholes-big.jpg (173.8 KB, 22 views)
Incoherent is offline   Reply With Quote
Unread 10-09-2005, 11:49 AM   #38
Les
Cooling Savant
 
Join Date: Oct 2001
Location: Wigan UK
Posts: 929
Default

So far so good.
Have started my own much cruder play, hoping to get my brain to unfuddle
Will not interrupt further
Attached Images
File Type: jpg Fantasy16.jpg (20.9 KB, 16 views)
Les is offline   Reply With Quote
Unread 10-09-2005, 12:07 PM   #39
Incoherent
Cooling Savant
 
Join Date: Sep 2003
Location: Vallentuna, Sweden
Posts: 410
Default

Quote:
Originally Posted by Les
So far so good.
Have started my own much cruder play, hoping to get my brain to unfuddle
Will not interrupt further
I think you are on to it Les.

...continuing.

The Femlab model gives a gradient through the die which deviates a little from nominal even without holes I think because of the uneven flux density as it enters the riser from the baseplate. This is probably valid.

more follows...
Attached Images
File Type: jpg temp_grad.jpg (56.8 KB, 8 views)
File Type: jpg Dev_from_nom_grad.jpg (66.5 KB, 11 views)
Incoherent is offline   Reply With Quote
Unread 10-09-2005, 12:30 PM   #40
Incoherent
Cooling Savant
 
Join Date: Sep 2003
Location: Vallentuna, Sweden
Posts: 410
Default

...so.


Edit, added calculated Watts error to chart
Attached Images
File Type: jpg grad-error.jpg (97.6 KB, 15 views)

Last edited by Incoherent; 10-09-2005 at 01:01 PM.
Incoherent is offline   Reply With Quote
Unread 10-09-2005, 12:39 PM   #41
Les
Cooling Savant
 
Join Date: Oct 2001
Location: Wigan UK
Posts: 929
Default

Fine
Then I lose it when looking at the big picture ,shown very roughly.
Will have to think my own play through.
Attached Images
File Type: jpg Fantasy17.jpg (35.8 KB, 11 views)

Last edited by Les; 10-15-2005 at 05:38 AM.
Les is offline   Reply With Quote
Unread 10-09-2005, 01:44 PM   #42
Les
Cooling Savant
 
Join Date: Oct 2001
Location: Wigan UK
Posts: 929
Default

OK at last.
Still to dot some "i"s.
That was hard work. dunno why.
Ta Jonathan.
Where were we?
Les is offline   Reply With Quote
Unread 10-09-2005, 01:51 PM   #43
Incoherent
Cooling Savant
 
Join Date: Sep 2003
Location: Vallentuna, Sweden
Posts: 410
Default

Quote:
Originally Posted by Les
Fine
Then I lose it when looking at the big picture ,shown very roughly.
Will have to think my own play through.
Staying with just the die for now...

Obviously you can't correct this error with a temperature offset, that is dependent on heat flux.
If you add corrections to the hole positions in the calculations (for watts q=k.a.dt/l, for die surface temp t2- (t1-t2)/(l1-l2)*l1), you can trick the equations into giving you the correct numbers. This is what I mean by "effective length"

Edit: added comments to chart
Attached Images
File Type: jpg New holes.jpg (94.4 KB, 15 views)

Last edited by Incoherent; 10-09-2005 at 02:10 PM.
Incoherent is offline   Reply With Quote
Unread 10-09-2005, 02:01 PM   #44
Incoherent
Cooling Savant
 
Join Date: Sep 2003
Location: Vallentuna, Sweden
Posts: 410
Default

Quote:
Originally Posted by Les
OK at last.
Still to dot some "i"s.
That was hard work. dunno why.
Ta Jonathan.
Where were we?
Hmm, good question.
Maybe I need to add these corrections to the DAQ scripts, I haven't calculated them independently yet. I am not sure how much I trust the FEMLab model for generating these numbers, the above charts are based on it. For sure it is better than anything I can hack up but ... you know.
Also my access to FEMlab is limited (by totalitarian licencing) so I try to avoid using it to much for home projects. Plus it crashes frequently on more complex models.
Would really appreciate some independent calculations for these corrections Les.
I am not sure if the temperature at the die center is the whole story.
Attaching Excel for perusal.
Attached Files
File Type: zip fluxdie_temp_grad_3d.zip (38.3 KB, 9 views)
Incoherent is offline   Reply With Quote
Unread 10-09-2005, 02:09 PM   #45
Les
Cooling Savant
 
Join Date: Oct 2001
Location: Wigan UK
Posts: 929
Default

Oh god its magic.
Will peruse
Quote:
Originally Posted by Incoherent
I am not sure if the temperature at the die center is the whole story..
Ah! The dark grey bits. Could not fathom what those temps were(not readable through dark grey without glasses) only discovered when played with..

Last edited by Les; 10-09-2005 at 04:30 PM.
Les is offline   Reply With Quote
Unread 10-09-2005, 05:36 PM   #46
Les
Cooling Savant
 
Join Date: Oct 2001
Location: Wigan UK
Posts: 929
Default

Still feeling my way round .
Note the gold bit is 400w/mc Cu ( I use 400 as default)
With what are the holes filled ?
k appears to be ~214.5, Al?

Don't bother answering until have added further queries - which I probably will.
Les is offline   Reply With Quote
Unread 10-09-2005, 06:28 PM   #47
Incoherent
Cooling Savant
 
Join Date: Sep 2003
Location: Vallentuna, Sweden
Posts: 410
Default

Quote:
Originally Posted by Les
Still feeling my way round .
Note the gold bit is 400w/mc Cu ( I use 400 as default)
With what are the holes filled ?
k appears to be ~214.5, Al?

Don't bother answering until have added further queries - which I probably will.
The holes in the real world are filled with: the thermistor, some AS Ceramique and a 3mm wooden plug drilled through the centre (for the wires) 0.5mm. In the model values of these are:
k=1, density=1000, heat capacity=1000
Guesses, the results are almost exactly the same with nothing in the holes. The gradient from which I guess you are calculating the k is established more by the temperatures at each end than the material itself.

Here's a more comprehensive model raw data result from WB interface down, 0mm is FB/die interface. I have used k copper =392 in this model. Nom gradient 1.65°C/mm.
I will use this model I think, it'll save a bit of work despite my reservations expressed above. I'll stick with die, FB centerline temps as well unless something else comes up. I'll have something based on this tomorrow.

Edit: Added working for final position corrections
Attached Files
File Type: zip Heat-shadow-corrections.zip (114.4 KB, 7 views)

Last edited by Incoherent; 10-10-2005 at 02:04 PM.
Incoherent is offline   Reply With Quote
Unread 10-10-2005, 02:15 PM   #48
Incoherent
Cooling Savant
 
Join Date: Sep 2003
Location: Vallentuna, Sweden
Posts: 410
Default

I have established sensor position corrections for the whole heat path. (updated attachment above)
As follows with position 0 being the fluxblock-die interface, + towards heatsource:
WB-FB I/F
-4.005 (reality -4.23)
-1.988 (reality -2.00)
0.000
FB-die I/F
0.000
2.081 (reality 2.00)
7.158 (reality 7.00)
v
v
heater.

This seem to have had the result of lifting the curve upwards.
I will attach a logfile in a wee while, this one was too big for the system to accept
Attached Images
File Type: jpg hscor6000.jpg (156.7 KB, 8 views)
Incoherent is offline   Reply With Quote
Unread 10-10-2005, 02:56 PM   #49
Les
Cooling Savant
 
Join Date: Oct 2001
Location: Wigan UK
Posts: 929
Default

I will use
(C"/W")= 1.02670158(C/W)

Not given Fbloc any d"
Attached Images
File Type: jpg Fantasy18.jpg (11.9 KB, 5 views)
File Type: jpg Fantasy19.jpg (38.7 KB, 6 views)

Last edited by Les; 10-10-2005 at 04:16 PM.
Les is offline   Reply With Quote
Unread 10-10-2005, 03:21 PM   #50
Incoherent
Cooling Savant
 
Join Date: Sep 2003
Location: Vallentuna, Sweden
Posts: 410
Default



A more condensed, cleaner run attached.

Uncertainties now are the water temperatures, I have only checked them briefly, not a full on steinhart recalibration. I am seeing odd bumps in the calculated flowrate every now and again, caused by a fluctuation in Water out, not sure what it is, electrical?, don't think it's a blockage, invisible in Water in and PD.

Les, what is the laminar-turbulant flow transition flowrate for 10mm ID pipe? I am seeing something at about 4.2lpm which I have seen before but assumed was my pump sequencing , but it is not I have since discovered. My flowrate errors seem to change sign at this point. Visible in the PQ plot. Not so obvious in the attached logfile, clearer with the higher sampling rate. (chart attached)
Curious.
Attached Images
File Type: jpg pd.jpg (59.2 KB, 14 views)
Attached Files
File Type: zip HS-corrected_2.zip (44.3 KB, 4 views)
Incoherent is offline   Reply With Quote
Reply


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 11:29 PM.


Powered by vBulletin® Version 3.7.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
(C) 2005 ProCooling.com
If we in some way offend you, insult you or your people, screw your mom, beat up your dad, or poop on your porch... we're sorry... we were probably really drunk...
Oh and dont steal our content bitches! Don't give us a reason to pee in your open car window this summer...