Radiators and Fans

[BillA]

yea, ball valves are worthless

all of the flow meters are of the DC pulsed mag type
have Sparling, Foxboro, Yokogawa, and Danfoss

a typical accuracy description is below



at 0.3gpm with a 1/4" flowtube I'm at ~1.5%, maxes at just under 4gpm

note the oscillations described, several hundred 1~ readings should be averaged

[Since87]

If control needs to be via GPIB, (due to software) the most practical way appears to be to use the 8 bit digital I/O port of the Keithley 706. It would probably be a fairly slow interface, but there's not much need for speed here.

If RS-232 or USB could be used there would be a lot more options. Interfacing to LabVIEW or some other canned software might be more problematic though.

Originally posted by unregistered

the coolant temp is stable (±0.02°C fluctuation), but subject to drift as the outside temp changes form night to day
- the pump, chiller, and cross-flow brazed plate heat exchanger in the chiller bath are all outside
(too much heat and noise)
I suspect that with a DAQ board having a 16 bit analog output it would be simple enough to gin up an interactive chiller control
- the chiller is now remotely controlled with a precision microvolt (calibration) source, but such devices with GPIB control are VERY expensive (0.0001 VDC = 0.01°C)

Hardware wise this isn't difficult. Getting the control right could be a real problem. I expect the chiller bath has a long thermal 'time constant'. It would be a long tedious process ensuring that an automated control loop wasn't doing more harm than good. (Causing very low frequency oscillations.)

Originally posted by unregistered

coolant flow is a can of worms as the system length is ~11' and I have 3 oscillations of different periods
the flow rate is quite stable but understanding the actual range (to 'set' the mean can be vexing)
- I know of no single throttling (needle) valve whose effective range spans my flow rates (0.3 to 4 or 5 gpm),
nor with the sensitivity needed in an 'automated' controller
??
BTW, it does need to be set accurately for the 10 mountings as these are single point readings,
otherwise no, not at all critical - so long as accurately known

How about four electrically controlled on/off valves. Each one followed by a manual 'restrictor' valve. All four sets of valves would be connected in parallel. For maximum flow, all four electrically controlled valves would be opened. For minimum flow, only the electrical valve connected to the most restrictive 'restrictor' valve would be open. Intermediate flows could be achieved by opening combinations of the four electrical valves. It would require some initial setup of the manual 'restrictor' valves to get the right range of flowrates for any given DUT.

Perhaps something like these dishwasher valves could be used. (These look more restrictive than desireable, but the price is right. $9.50 per five at this moment.)

Originally posted by unregistered

air flow can be difficult to stabilize, at specific flow rates the fans may sync, the flow start pulsating, really bizarre
(the compressibility of air is quite apparent)
I started with mechanical dampners, went to huge pots from welding machines (10A of fan current is no joke), to 115VAC fans with Variacs
- don't think its worth the effort (not actually running enough hours, at least for now)

Have you seen fluctuating flowrates with a single fan or only with combinations of fans?

I haven't found a single fan that is spec'd to cover the full range of the rads in your article. It seem like what is needed, is a fan run by a stepper motor, to allow it to cover a wider range of RPM's than is achievable with brushless DC fans.

It wouldn't be difficult to create a computer controlled variable output switching supply to drive 10A worth of fan current, but it still leaves the issue of a fan that covers a wide enough operating range. Can you cover the full range with an AC fan on a variac?

Originally posted by unregistered

I have a different question for you:
in running 4 - 20mA transmitters, I have a huge settling time (as in hours) to read at the 100pA level (Setra units, Foxboro only to 1µA)
which I believe is caused by the transmitters and not the meter, a Keithley 195A
-> can I switch in and out a resistive dummy load in the transmitter current loop when its not being read ?

I'm guessing you meant nA? (nanoamps) I don't believe the Keithley 195A can give you very meaningful data at 100 picoamps.

Are these transducers powered through the 4-20 loop? (I did some looking around on the web, but wasn't sure if I was looking at relevant stuff.)

I'd say a dummy load is worth a try. (Probably even a dead short would be ok.) What would be most likely to give the current output fits, is an open circuit. I'd suggest trying a 100 Ohm resistor and set up a 'make before break' switching arrangement. (i.e. connect both the dummy load and the Keithley in parallel, and then open the connection to the dummy load as a second step.)

If the only electrical connection is the current loop, having the dummy load, (and power constantly applied) will probably make a big difference in settling time.

[BillA]

one post on topic, then I'll go back to testing stuff

the rads from Since87 and Blackeagle arrived today
and deeppow had loaned me a '87 Chevette core (AND his new MD-20RLZT - obviously a high roller here !)

some rads



and some pumps



who knows when . . . .

EDIT: Hydor pumps and black rad from Bruce at CoolTechnica
Sicce and unk rad from Joe Citarella at overclockers
and the 'dual' rad from Brian at Be Cooling

[Blackeagle]

Lots of neat toys.

Is the dual rad a single, dual or quad pass type?? What's it out of & is it alum. or SS?

Thanks

[BillA]

dual rad ? - upper left by Be Cooling
Cu tube with Al turbulators inside, Al fins
very good for low noise (low air flow resistance w/good cooling)

I have your data, I'll try to get it into Excel tonight

[Blackeagle]

Many thanks Bill, look forward to seeing the comparison.

I hope the data will be of some use to you for the sim as well.

[BillA]

absolute values could be off by 10%, relative quite correct



Blackeagle and Since87 need to fill in the blanks

somewhat more meat here





from this we should be able to back out in.³ (or cc) values to characterize any hc with this construction

gmat
you want the excel file to play with ?

[Since87]

Woohoo! I was right.

My HC has about 4 times the flow resistance of Blackeagle's.

No wait. That sucks.

Thanks a lot Bill.

[Blackeagle]

I'd like to add my thanks to Sense87's Bill.

Is the data gained going to be of much help with the sim for you guys working on it? I sure hope it is.

And your suggestion about adding meat to what you've shown us here will give me a good deal to think on for a while. I hope I can determine a good combination of pump, fans and blocks that should go well with this rad.

I have some ideas in each of those areas, but I'm going to take the time to rethink each before buying, I wan't to get as close to a uber system as I can first time.

Once again, thank you very much Bill ! !

[Since87]

To fill in the details:

First off, there is discussion of some of these HC's here.

[I hope I've got the following correct but with heatercores4u down, I'm not certain.]

Mine (Sean's) is a 2-274:
9.875" X 6.125" X 2", 13 flat tubes, two pass
two 5/8" tubes

Blackeagle's (Craig's) is a 2-342
9.5" X 6.125" X 2" 13 flat tubes, single pass
one 5/8" tube and one 3/4" tube

So, aside from a 3/8" difference in length, and the (relatively insignificant?) difference in inlet/outlet tubes, the difference between the two is the number of passes.

It looks like my HC has about 3.5 times the flow resistance of Blackeagle's.

I had expected mine to have more than 4 times the flow resistance of Blackeagle's due to the fact that my HC consisted of:

seven 9.875" 'paralleled' flat tubes in series with six 9.875" 'paralleled' flat tubes

whereas Blackeagle's consisted of:

thirteen 9.5" 'paralled' flat tubes

It appears that there is more 'overhead' flow resistance associated with the inlet/outlet tubes and the tanks than I had guessed. I'm assuming this accounts for the difference between the greater than 4X difference I'd expected, and the 3.5X difference measured.

Bill, if you want to send me the excel file, I'll take a shot at coming up with a breakdown of the relevant factors. (I haven't seen anything from gmat in a while. He'd said something about a vacation coming up.)

Edit: grammar

[BillA]

well, for the OCers out there (as in Obsessive Compulsive, eh ?)
here is my take

Since87's rad, 2-274


2-pass, 1/2" CTS connections (5/8" OD) on same end
effective dimensions: 9 3/4" long, 6" wide, 1 15/16" thick
13 corrugated tubes, with folded and louvered fins @ 10/in.

Blackeagle's rad, 2-342


1-pass, 3/4 and 11/16 OD connections on opposite corners
effective dimensions: 9 1/4" long, 6 3/8" wide, 1 15/16" thick
14 corrugated tubes, with folded and louvered fins @ 10/in.

deeppow's '86 Chevette rad (# not known)


2-pass, 5/8 OD tubing connections (one cut back from 3/4" end)
effective dimensions: 6 1/8" long, 6 3/8" wide, 1 15/16" thick
14 corrugated tubes, with folded and louvered fins @ 10/in.

I'll send the file

I'm working on a new 'consumer friendly' test/reporting procedure for rads
will post as we may want to get a confirmation on dissipation

[Since87]

Originally posted by unregistered
well, for the OCers out there (as in Obsessive Compulsive, eh ?)
here is my take

Thanks. I was afraid I didn't have some of the dimensions right, but I couldn't find the relevant stuff in the old thread.

[pHaestus]

It appears that there is more 'overhead' flow resistance associated with the inlet/outlet tubes and the tanks than I had guessed. I'm assuming this accounts for the difference between the greater than 4X difference I'd expected, and the 3.5X difference measured.

Could that be due to the inlet and outlet coming into the tanks at a right angle? I have this lovely:



which is I think the same size as the radiator Blackeagle sent but with fittings in the ends of the heatercore instead...

[koslov]

Bill, are you still taking HCs for testing? If so, I may have one I can send to you. And it might help us determine another thing: the accuracy of manfacturer spec claims (Lytron).

[Blackeagle]

Well one area that would help explain the differance in Sense87's first figures and the real world result is that the increased velocity in the system lines in the set up with the single pass would partly offset the reduction in headloss in the rad it's self. I don't know how long BillA's lines are in his test rid, but this would have to account for at least a small portion of the differance.

Also the change from 3/8" inlets/outlets to 5/8"+ would also cause some changes. And perhaps due to the higher flow rate through the rad's finned/tube area, this would also cause a increased headloss at the inlet/outlet area in the single pass.

As there is less than 1 cubic inch of differance between the two rads this couldn't have made much differance.

[Roscal]

BillA > When your article about new rads will be available on overclockers.com ?

[BillA]

koslov
I never have been accepting rads for (free) testing
Since87's and Blackeagle's were flow tested only, as a favor
the Chevette hc is deeppow's and he loaned me his Iwaki 20RLZ as well

recognizing that this simulator activity could not get off the ground without some real rad data,
I ran the Chevette hc as a freebie, the complete curves such as are useful to the product or systems designer
- I am hoping that from this we can generally characterize all hcs of similar construction AND thickness

to tie some of this together I will be testing a 'system' using the heat die simulator and the now calibrated rad pressure drop as a totally non-intrusive flow meter (slick, eh ?)
- and I hope to try 2 wbs, 2 rads, and 3 pumps
all I need is the time, a kinda substantial amount of it for this project

Blackeagle
if you look closely at the connections on the Chevette hc you can see the pressure taps, same size as the rad connections
- the numbers posted are the actual net values for just the component itself

the jackal
several different things going on (as usual)
I am completing a large study of ThermalDesigns rads which will be furnished to them - only
I suspect they will make much of it public

as a separate activity I will be using their data (should I get permission to do so) to write a user oriented article on rad/fan selection
and this data will be somewhat different than that developed for the rads designer (ThermalDesign)
this is at least several weeks off (have a fistful of new wbs to test)

there is of course much that could be written about rad design, and testing, and application
not sure that I will ever do this (even though the old article is quite deficient in many regards)
- I will be posting a more detailed description of rad testing on my site in the next several days
one program for rad (design) performance characterization,
and another for user performance guidelines

[Blackeagle]

BillA,

Thanks for pointing that out to me. That eliminates that theory. Means that the tank differances and the restrictions caused there by turbulence and the inlet/outlets have to be it like Sense87 posted.

I had thought that the larger tanks of the single pass would, if anything, offer a advantage. It seemed to me the larger tanks would allow the water to move a bit slower and thus have less turbulence. Wrong again. :shrug:

[OTMOPO3OK]

Very interesting discovery....thank you for taking it to this point...apparantly i had some wrong ideas also....

[Since87]

Originally posted by unregistered

to tie some of this together I will be testing a 'system' using the heat die simulator and the now calibrated rad pressure drop as a totally non-intrusive flow meter (slick, eh ?)

Yeah.

It had occurred to me that with your test results, and a differential pressure sensor I've got a flowmeter without added dP.

Cheap differential pressure sensor here.

Digikey price for MPX2010DP (Qty 25) = $11.15

Anyone want to go into business selling Chevette heatercores with flowrate indicator installed? Add on flowrate indicator kits?

I'll sell you a circuit design that will definitely give "good enough" results.

J/K (mostly)

[Blackeagle]

Would not a inline psi gauge, of good quality and accuracy, located between the pump and whatever restriction is first in a individual system offer a method of rough flow rates? You'd get a read out of the back pressure caused by all the systems fittings and blocks.

Couldn't you then use that psi reading along with a pumps charted performance to reach a systems flow?

Can someone point out flaws in this reasoning? In a way it seems to easy to me, but don't know why.

Thanks

BE

edit: Gauge in mind is on page 491 of McMaster's, digital grade 1A, only digital on that page, resolution of .01 psi and 1% accuracy. I'd be interested in the lowest scale of 0-30 psi.

[Since87]

Originally posted by Blackeagle

Couldn't you then use that psi reading along with a pumps charted performance to reach a systems flow?

The biggest problem with this is that you don't necessarily know the pressure at the pump's inlet, and you need the inlet/oulet pressure differential to be able to use the pump's PQ curve.

Then there is the issue of how well the pump actually matches the spec'd PQ curve, the difference between the voltage at your wall outlet vs the supply used to generate the pump curve, etc.

[bigben2k]

What you really want is a flowmeter, as well as two pressure gauges, but that can get pricey. Two pressure gauges would probably serve you cheaper, with some lack of accuracy.

[pHaestus]

Originally posted by bigben2k
What you really want is a flowmeter, as well as two pressure gauges....

Flowmeter is going to equate to substantial pressure loss. The advantage of using pressure differential as suggested is that provided one has data for the rad one can backcalculate flow rate reasonably well in their cooling loop without adding more resistance. Sure we could throw thousands of dollars at the problem of measuring flow rate directly in our system's loop, but Sean's suggestion would get reasonably close without much money being expended.

[bigben2k]

Originally posted by Since87
Cheap differential pressure sensor here.

Digikey price for MPX2010DP (Qty 25) = $11.15

Anyone want to go into business selling Chevette heatercores with flowrate indicator installed? Add on flowrate indicator kits?

I'll sell you a circuit design that will definitely give "good enough" results.

J/K (mostly)

Not bad. How 'bout p/n MPX2050GP-ND, MPXM2053D-ND and MPXM2053GS-ND? $15.85, $7.90 and $8.07 respectively.

Circuit diagram?