Most Stable and Quantifiable Heat Source

[MaxxxRacer]

There are many more questions than asnwers :

Its for reviewing. for XtremeSystems specificly...

not sure of hte purpose of DC verus AC. I know what they are but in this aplication not any real idea why you would prefer one over the other.

[jaydee]

You can get an immersion cartridge heater (goes directly into the water) and make your own little water heater. Insulate the heater unit VERY well and it might be good. At 350watts you will need very good insulation to minimize the secondary heat loss. The water will probably want to boil so you may need some kind of pressure release on it.

[MaxxxRacer]

hmm.. Well i can make a res out of some kind of thermus material..

and yah it would need to be seriously insulated.. if that shorted out... well oh boy would i be in the hospital...

[bigben2k]

Originally Posted by MaxxxRacer

There are many more questions than asnwers :

Its for reviewing. for XtremeSystems specificly...

not sure of hte purpose of DC verus AC. I know what they are but in this aplication not any real idea why you would prefer one over the other.

Ok, that's a start.

So you're not looking to test a water block then?


The reason that DC is preferable, is simply because it's easier to measure, and to replicate from test to test.

If you used AC, you might have some fluctuations in the voltage; you'd be hard pressed to set it to a specific voltage, so that you can test under the same conditions. You'd also be facing the difficult task of measuring AC with similar accuracy as DC, which simply requires a good DMM, and a shunt (all calibrated).

[MaxxxRacer]

well I got a good DMM so that isnt an issue. I dont mind using DC but a psu powerful enough to get me 350watts or more would be insanely expensive. that is my only concern.

and no this isnt for waterblock testing. purely for radiator testing. I have other means (ripped off pH's setup) to review waterblocks.

[bigben2k]

Glad that that's cleared up!

I hear you about the DC supply; I got lucky with my HP supply; max 36 volts, max 5 amps (~180 Watts), for ~$40 plus shipping (+~20).

In your case, as long as you understand that using AC means that you could have some variations on the power output, I guess it'll have to do, but I wouldn't go that way. I'd guess at a variation of at least +/- 5 %, from one day to the next, and I think I'm being optimistic.

[MaxxxRacer]

wheere woudl the variation be? I will measure it with the DMM and clampeter at each setting before i do anything so i doubt it will be an issue.

[BillA]

duh
put the meter on the wall socket
min to hr to day, whats the variation ?
report back
you're heading for a real crap setup

"so i doubt it will be an issue"
your ignorance is speaking here, over and over and over

[MaxxxRacer]

bill, as long as i dotn run at 120volts it wont matter no? If i say get a shunt to where the max i need to run it is 100 volts i should be fine?

If im heading for a real crap setup as you say, please enlighten me as to what would NOT be crap.

bill, lets say i am able to find a DC psu (one of hte lab ones with ampmeter and voltage meter on it) that is poweful enough to do 350watts continuous and I duct it so that its heat is not affecting the testing. Would this work better than the AC source?

One of the reasons i like the AC source is that there is no big psu that generates heat in the room. but if isnt going to be useful then its no matter. BTW, what if I were to build a step down transformer (step ac 120 to like 100 volts) and build in power regulators to keep it at the voltage and do so stabily? Myself, I could not build such a contraption (well i could, i would just need design specs), but I could have it built for me by someone who does this kinda stuff all the time.

[BillA]

Originally Posted by unregistered

put the meter on the wall socket
min to hr to day, whats the variation ?
report back
. . . . .

you want to be a tester ?
one would never know it

what was suggested above ?
so why do you not wish to do it ?
no, you want the endless blah blah blah
learning is by doing, you want something delivered w/o any effort on your part to learn a single thing

I suggest you find a technically competent partner 'till you figure out this stuff
your comments about electricity and water are puerile
best of luck, other's experience and knowledge don't seem to transfer too well to you

[bigben2k]

Originally Posted by MaxxxRacer

wheere woudl the variation be? I will measure it with the DMM and clampeter at each setting before i do anything so i doubt it will be an issue.

Actually, it's a big issue.

The variations will come from many different places. The voltage will most certainly fluctuate, as Bill said, from minute to minute, from hour to hour, as well as from day to day. Everytime one of your neighbour runs a load of laundry or bakes something in the oven, your voltage drops.

Then you've got line noise, aka voltage spikes which may not seem relevant, but may have an impact on the power output (more below).

After that, you have frequency shifts; it's usually 60 Hz (in North America), but the electric company doesn't guarantee that. Also, the AC is supposed to be a perfect sine wave, but that's no guarantee either.

All of the above become an issue for two reasons:
1-you really don't know the actual power that you're putting in, and it could be different each time
2-it's really difficult to measure the electrical power, under those circumstances.

So...

For the sake of repeatability, which is important in testing because you really need to make sure that you're being fair, AC is a very poor choice.

Now...

Using DC is infinitely easier to measure; slap a good calibrated DMM on it, and use a shunt as Bill described, to measure the current. Ideally, the DC power supply has very little fluctuation in its output, and I recomend that you at least borrow an oscilloscope to take a look at it. What you'll find, more than likely, is a very small AC signal (aka voltage ripple) at a frequency that's pretty high, which will have almost no impact on the power output of the heater (the voltage sine wave has a low and a high, and they essentially cancel themselves out, usually so fast that the heater doesn't have time to react).

I know that it's not easy nor cheap to get a powerfull DC power supply, so I'd concentrate on finding a source for a used one.


There are other issues that impact repeatability, that fall outside of the whole heater issue. Bill's already hinted that you're going to have an issue with heat within the room. Specifically, you're going to have an issue maintaining a steady air inlet temperature. Your first heat source is the heater (primary and secondary losses) and the second is that most power supplies operate with an efficiency as low as 67%: generating 350 Watts in DC may require over 500 Watts of AC power, and the difference is going to heat up your room.


How are we doing so far?

[BillA]

nice post Ben
linear ~50%, look at Kepco
plus the heat from the rad eh ?

[bigben2k]

Originally Posted by unregistered

nice post Ben
linear ~50%, look at Kepco
plus the heat from the rad eh ?

Thanks Bill!
Yeah, primary from heater.


The only Kepco I found on eBay (that would do the job) is ~$800:
http://cgi.ebay.com/ws/eBayISAPI.dll...tem=7507726455

Be patient, set up your searches and email notifications.

I recently started using a free Sniping tool, and I'll share the link with everyone else here:
http://hammertap.auctionstealer.com/

[BillA]

http://cgi.ebay.com/ws/eBayISAPI.dll...e=STRK:MEWN:IT
NOT a good choice, I use 2 of them for 24V TECs - too much current for a die heater (leads)
go to a higher voltage

I use Cricket Jr

[MaxxxRacer]

10 dollars... with 115 shipping LOL... irony there...

Thanks for the help ben. so DC it is. Ill talk to my friend to see if he cant get a surplus DC power supply. It shouldnt be too hard to come by.

Heat sources
Secondary Heat from shunt - handled with sealed insulating res that shunt is placed in
DC PSU - Hot air vented out window through ducting system
Radiator exhaust and seondary heat loss- have rad exhaust also vented out the window.

Is there any more heat sources? I the pump could be considered, but i think I will use a 50Z for the radiator tests as I will get plenty of flow from it. the iwaki would be excessive as non one can get the flow rates that an iwaki would provide given they had an actual loop. So with that said the 50Z's heat dump is a small 8watts into the water with about 1-2 watts into the air. This will be a constant with all tests done so i think this heat source can be ignored. If I am wrong in my thinking here please lmk.

[bigben2k]

No problem. This exercise has helped me put together the first WBTA article (everything has a purpose, right? ).

There's so much more to do here though.

The next part I'd like to get into, is how you're going to maintain a steady air inlet temperature.

Personally, I favor using a heater with a smart temperature controller, to set the inlet air temp slightly above room temperature, i.e. 25 deg C. Not sure how anyone else feels about that, but I don't believe that it would impact the performance test in any significant way.

The advantage is that it'll be a lot easier to maintain a steady inlet temp, asuming that you can setup the controller correctly. It would have to have a pretty quick response time though.

What you have left, is to measure the radiator inlet and outlet coolant temp, with some kind of accuracy, and that's going to be really tricky, because the temp difference is going to be awfully small.

That aside, you might consider measuring the pressure drop, of both the coolant and the air, if you're looking to fully characterize the radiator. Maybe you can share your thoughts on that too, then we can break it down?


edit: ps: it's not uncommon to find high shipping charges on eBay; it's an easy way to get around taxes...

[MaxxxRacer]

Ahh, very nice article ben. btw, how do those shunts work. I dont see any viable way of using them in a setup i would use without electrocuting myself??

As to measuring inlet outlet temperatures, from pH's tests those do seem rather futile so i gave up on trying to do those a long time ago, but that did seem the first route that i wanted to go with.

My thinking now is that I will measure the water temp on the radiator outlet and also measure the air temp. I am eyeing a RTD probe setup that has .001C res and .01C accuracy which would really help out in this endeavour as it comes with two RTD probes all ready to go.

The heater idea is interesting, but the only thing is that as more heat gets dumped into the room the rad inelt temp would get hotter. the only way to keep a perfect rad inelt temp is as bill mentioned is a chamber.. but none of us testers have anywhere near the fluid assets to buy one of those things. I guess if you had a thermostat on the heater that was connected to a thermocouple or RTD probe you could get the temperature relatively stable though.

BTW, my room is around 26C without any heaters. its this god forsaken computer that acts like a space heater...

I am going to measure the pressure drop of the fluid flow of the radiator, but I am not going to test the air pressure drop. What I will do is test with different sets of commonly used fans and see the performance levels with each fan at 5,7, and 12 volts. Doing this will give less of a quantitative number as to the pressure drop for the airflow, but will be a more realworld test i believe. It will help the end user choose their fans more wisely as most users WILL NOT go through the trouble of caclutlating the cfm a fan will give them by looking at the static presure curve of a fan related to its voltage and flow and looking at the pressure drop numbers of each radiator.

again, thanks for your input ben.

[bigben2k]

Here's a link to the spec/install sheet of that Datel shunt:
http://wbta.us/bigben2k/datel_shunts.pdf

It covers a lot. Yeah it's open, so you don't want to touch it unless you're isolated from ground, especially in your case, because you're looking at a (relatively) high DC voltage, which can be very dangerous. I'd try to cover it up with some sturdy but vented box, plastic or wood (or whatever electrical isolation of your choice).

So you'd measure the dT between air inlet, and water outlet? That seems reasonable. As long as everything is calibrated, and repeatable, you ought to be able to tell which radiator is better, with confidence.

The only issue I see, and maybe you can share your thinking here too, is how (and if) you would address the difference in size of the radiators.

Note Bill's article, here: http://overclockers.com/articles481/

Bill created a test setup where the pressure drop is the same for all radiators (which of course results in various airflow rates and speed). Of course it makes no reference to actual fans typically used, and as it turns out, some of these pressure drops cannot be achieved with a typical single axial fan (hence why blowers are starting to crop up).

The difference here though is that Bill set the system to maintain a dT of 10 deg C, then simply measured the heat level put into the coolant loop.

The 10 deg C was an arbitrarily set point, as it was (still?) representative of an average water cooling setup. I'm not sure that it would apply in yours though.

[MaxxxRacer]

Yah i would measure the dT between air inlet and water outlet. When i test it and it isnt repeatable and donesnt work im gonna break down in tears and need to be commited though.

Im not sure I understand ur comment on the sizes of radiators. From what i can see it wont make a difference. I would just use the appropriate amount of fans that are needed and comence with testing.

Bills article you mentioned is quite an amazing and thorough aritlce but the one caveat i have with it is that he does make the pressure drop the same for all radiators which is great for analytical data but for something more tangibe to the average reader it isnt as useful. If everyone in the WCing community had test setups like bills and we were all PHD's in thermodynmaics then sure its the best way. but most watercoolers arnt anywhere close to that.

10C sounds about right. well that is if you have a well balacned system. alot of kits these days dont achieve 10C delta between ambient and water temp as they dont have the rad capacity, but 10C sounds about right.

Looking at that PDF manual i dont see any sensible way of heating up the water with one of those shuts as any exposure to water and i am gonna blow up alot of gear. So my question is how do i get the heat from that shunt into the water?

[jaydee]

Originally Posted by MaxxxRacer

hmm.. Well i can make a res out of some kind of thermus material..

and yah it would need to be seriously insulated.. if that shorted out... well oh boy would i be in the hospital...

My idea was just a consideration. I don't know if it is good or not. I didn't mean insulation from electrical shock, I meant insulation from heat loss. What ever res you got with the heater in it will be sheading a lot of heat at 350 watts. You don't want that. If there is a lot of heat radiating off your res or heat die then your electrical measurments turn nearly useless as a lot of that power is not going directly to the radiator. Your numbers will not be accurate. You want as much of that heat to be contained in the loop. If you loose 50watts of the heat at 350watts just in secondary losses you will get bad numbers when you add in your electrical measurments for how much heat the rad is actually dissipating. You would be 50watts off. Your secondary losses will probably not be liner through the wattage range either. At 350watts you may loose 50watts, but at 250watts you may only loose 20 watts ect depending on insulation thickness and type.

If you don't think you will loose much heat then touch a 100watt lightbulb (don't really do it, HOT!) and see how hot it is. That is heat radiating off of it. You have to contain as much of that heat as possible in the loop IMO.

[Chew_Toy]

Originally Posted by MaxxxRacer

1
Radiator exhaust and seondary heat loss- have rad exhaust also vented out the window.

I dont know how you would be able to do that without it affecting the results of your tests. Ducting (depending on how far) could add a fair amount of added air flow restriction and depending on how its setup and wind speed and direction could actually add flow to what the fans could deliver on thier own.

Reason I asked is I have thought of in the future trying some rad tests of my own (for personal use) and the only way I thought of doing it was having the rad and fans only in 1 room and all the other equipment in an adjoining room with the exhaust from the rad going into the 2nd room through a hole in the divider wall (1st room is large enough in volume, its aprox 50'x14' finished basement that temps should be fairly stable as long as the furnace/AC is off). And at the other end of the room they are open through a standard doorway and short hall so the the pressure between them should be relatively equal, also neither has any windows so that outside wind shouldnt have any effect either.

The other way I thought would be the same setup only with everything in one room and just drawing air from the main room.

I believe that you will need a faily large room to draw your air from to be able to have fairly stable temps for the testing. I think getting a stable air temp to the rads for testing could be quite a problem also.

[jaydee]

Originally Posted by MaxxxRacer

Looking at that PDF manual i dont see any sensible way of heating up the water with one of those shuts as any exposure to water and i am gonna blow up alot of gear. So my question is how do i get the heat from that shunt into the water?

Correct me if I am wrong, but the shunt resistor is not a heating device. It is used to measure the current.

[MaxxxRacer]

chew toy, I could find some 3 foot diameter flexible ducting that would work. lol.. would be big but it would work. but i do see your concerns. I will have to wait and see what happens i guess.

[bigben2k]

Originally Posted by jaydee116

Correct me if I am wrong, but the shunt resistor is not a heating device. It is used to measure the current.

Yep.

Not sure how that got mixed up.

The shunt is something that you want to plug in series, between the PSU and the heater cartridge. The shunt, if calibrated properly, will give you a proportional but very small voltage reading, usually in the range of 0 to 50 mV (millivolt). It's that measurement that you use to quantify the current that goes to the heater.

Assuming that you also have the voltage measurement (at the cartridge), you apply:
P = V * I

where I is current, and V is voltage.

Nice and simple, no?


I'm glad that you see the pressure drop issue.

Check out this setup, by Joe Citarella (www.overclockers.com, a WBTA member):
http://overclockers.com/articles750/

That setup is quite expensive, but the expensive parts are the air nozzles, because they're the calibrated parts that allow one to measure actual airflow (i.e. if you know the dP, you can calculate the air flow rate).

The setup has flow straighteners, in and out (easily replicated). The pressure gauges are simple manometers.

I'm sure that you're thinking about using the same fan for each rad that you're testing, but you really ought to make sure that your airflow is going to be the same, and to do that, you ought to consider measuring it.

[bigben2k]

Originally Posted by MaxxxRacer

jayde i think your correct in ur asumption but it gets really hot and is a measurable source of heat???? i think that is why bill and ben reccomend it for heating.

...

I hope you figured out by now that the shunt is not a heater.

Bill recommended using a shunt, for the purpose I described in my previous post. He also stated that it would be best to use it at no more than 30% of its rated current, because if you go above that, the shunt will warm up; when it warms up, it starts resisting more, and that introduces an error margin, in your current measurements (an irrecoverable error, I might add!).