PDA

View Full Version : Units


gmat
02-27-2003, 03:16 AM
Ok in this thread we'll discuss about units used by the simulator.
As this program is not meant to be used in the U.S.A only, and since the USA are the only place in the world where non metric units are still used, i'm 100% for using metric units only.

Pressure:
mH20 is metric indeed, though not SI. As it's easily understandable, and since one can visualize it easily, i think it's a good choice for *every* pressure. Pressure from the pump, and pressure drops everywhere.
Pa is the SI unit, allright, but we're using water only here, so mH20 is totally relevant IMHO.

Flow:
lph (liter per hour) or lpm (liter per minute) for pumps. lph make sense since the mfgers give that figure, and the 'per hour' is consistent with other units here.
Here in EU fans are advertised with their m^3/h (cube meter per hour) rating. 1 m3/H = 1000 lph.
For examply the Papst 8412 is given at 69 m^3/H.

Temperature: °C (celsius) or K (Kelvin) - they're the same basically. K is the SI unit and only differs by an offset. We can keep °C so ppl with little scientific background can understand...

Power: W (Watts) (i'm amazed that even Americans are using this decimal, metric unit :D )

Thermal resistance: though °C/W is a commonly accepted unit (and metric), i don't think it's the real unit that applies there. I think its Ohms (omega symbol).
But the tool is aimed towards PC watercooling enthusiasts. °C/W is easy to apply to real cases, and work just like ohms (they add up in series etc.) so i vote for keeping it.

---

For you Americans i propose a small (separate) unit conversion program, so you can convert those metric units in gallons, feet and other farenheits.
There are many sites who do that BTW.

Alchemy
02-27-2003, 04:17 AM
I'm against using meters of water as the pressure unit only because it will lead to awkwardly small numbers. Millimeters water would be better, I think.

All the others look fine to me.

The C/W thing bugs me because the actual name of that value is the thermal impedence. Its units are usually (as they will be here) in degrees C per watt. Why people choose a very specific set of units to be the symbol for the number makes absolutely no sense to me.

The units are not ohms because those refer specifically to voltage potential over rate of current, so its units are almost always volts per amp. They have relationships that are similar to those in heat transfer but their units are quite different.

Power the US is almost always in watts (our power bills are in kW*hour) or BTUs per hour. Watts are clearly the more convenient unit.

As for a conversion program for lazy Americans, I wouldn't worry about it. All one needs is to convert ambient temperature and CPU temperature, really, and even here most people dealing with PC cooling tend to use Celsius.

In this thread you might also want to make sure we all use the same name and symbols for pressure, flow rate, thermal impedence, heat transfer coefficients, etc etc.

Alchemy

gmat
02-27-2003, 05:02 AM
Yes consistency is important. Also, the units (and names) whe choose here will apply to all 'plugins' that ppl design for the sim. So for instance if you want to add a pump plugin, you'll have to write data in lph vs mH2O.

Ohms is an universal impedance unit IIRC. During my thermics courses i remember being told that thermal circuits work just like electrical circuit, as per the 'similarity' physics law. Thus thermal impedance has the same 'value' as electrical impedance, Ohms, as they both represent a potential divided by a current.
Now °C/W is maybe not *exactly* a thermal impedance, but it looks like it's very close. My memories are too distant to remember the units involved but a quick google would help for that. (i'm at work right now)

As for mH20 it's quite a standard actually. Numbers are not so small, my pump produces 1.95mH2O of head for example. If ppl really want to we could use cmH2O (centimeters) so 1.95mH20 becomes 195cmH20. I presume most fittings will have backpressures in the orders of a few cmH20.
Up to you guys, what are your thoughts.

(edit) impedance ? resistance? i'm not sure about the english term here.
electrically: U=RI -> R is the 'resistance' (in French), does that translate as an impedance ?
In French 'impedance' is the complex value of 'resistance', called Z and is Z=a+ib (complex number)

gmat
02-27-2003, 05:53 AM
Ah BTW i have not stressed enough the importance of coherent units.
The program will correlate data (using interpolation) based on raw numbers. So for example it will match the pressure / flow of a pump with the flow / pressure of a rad, assuming that figures are all in lph / mH2O.
If units are different, results will be totally wrong, of course.

Since87
02-27-2003, 01:18 PM
Originally posted by gmat

(edit) impedance ? resistance? i'm not sure about the english term here.
electrically: U=RI -> R is the 'resistance' (in French), does that translate as an impedance ?
In French 'impedance' is the complex value of 'resistance', called Z and is Z=a+ib (complex number)

The convention here (at least in electrical engineering) is to use thermal resistance with units of C/W.

The extremely curious can see an example in the lower left corner of page 6 of this component datasheet. (http://www.linear-tech.com/pdf/1844i.pdf)

"Impedance" is technically reserved for circuits which are at least partially 'reactive'. (i.e. complex, or purely imaginary)

In actual conversation between EE's "Impedance" is frequently used somewhat sloppily, and is often used to refer to purely resistive circuits.

Alchemy
02-27-2003, 01:41 PM
Originally posted by gmat
Ohms is an universal impedance unit IIRC. During my thermics courses i remember being told that thermal circuits work just like electrical circuit, as per the 'similarity' physics law. Thus thermal impedance has the same 'value' as electrical impedance, Ohms, as they both represent a potential divided by a current.
Now °C/W is maybe not *exactly* a thermal impedance, but it looks like it's very close. My memories are too distant to remember the units involved but a quick google would help for that. (i'm at work right now)

The units for ohms are almost always Joule*seconds/coulomb^2. They do not involve temperature or



(edit) impedance ? resistance? i'm not sure about the english term here.
electrically: U=RI -> R is the 'resistance' (in French), does that translate as an impedance ?
In French 'impedance' is the complex value of 'resistance', called Z and is Z=a+ib (complex number)

Pretty much the same here. Ohm's law is V=IR.

If there are capacitive or inductive properties in the circuit, you use impedence Z=a+bi.

The inverse of resistance is conductance and the inverse of impedence is admittance, if I recall correctly.

The fact that there is a similar relationship between the forces of heat transfer and the forces of electrical current does not mean they share the same units. These same relationships - where a transfer of something is proportional to some sort of gradient - are common in heat transfer (Forier's and Newton's laws) electrical current (Ohm's law) differential mass transfer (Fick's Law) and momentum transfer (Newton's law).

Clearly, this does not mean that heat transfer coefficients, electrical conductance, diffusivity coefficients, and viscosity all share the same units.

Alchemy

murray13
02-27-2003, 02:36 PM
IMHO if this is to be used by 'anyone' not just the technically minded. The program should be able to take input in either format and produce output in either format. As an example look at this page: LMNO Engineering Darcy-Weisbach (http://www.lmnoeng.com/darcy.htm)

If you were just looking for opinions on what units to use here are what I use most of the time:

Water pressure = ft-H2O (feet of water)
Water flow rate = gpm (US gallons per minute)
Temperature = °C (degrees celsius)
Power = W (watts)
Thermal resistance = °C/W (degrees celsius per watt)
Air flow rate = cfm (cubic feet per minute (ft³/m))

Personally I can do whatever conversions are necessary. I agree that consistancy is a good thing. Everyone using the same terminology. Maybe there could be a page with definitions and conversions. Just a thought.

Either way this was a great idea. It will take a lot of hard work but I think the results will be well worth it.

Since87
02-27-2003, 03:08 PM
Eventually user input would just be menu selections like:

Eheim1250
TC-4
BIX
etc.

The end user doesn't need to know the units. Only model makers need to get the units right. And, if model makers can't do the unit conversions, how useful is their model likely to be?

Edit: For waterflow, I'm inclined to stick with the units Bill is using for his waterblock testing, which are lpm and mH2O.

gmat
02-27-2003, 06:00 PM
Since87: Exactly.
I've got a degree of EE and have some experience, and i see the terminology is the same :p 'impedance' used everywhere.

Murray:no one outside the USA knows what the hell is gpm or cfm... heck i don't even know exactly how long a foot is in meters. Every international project i've been on (not many, but a few already) involve only metric units...

As for doing unit conversions, it will be a hassle for me and a nest for future bugs. So currently it's a "NO", we'll see later. The program won't convert anything and will take raw numbers (or interpolated ones) as true values.

As for lpm vs lph: i don't care, as long as everyone is consistent with that. Pump mfgers and aquarium enthusiasts seem to like lph.

Alchemy
02-27-2003, 08:19 PM
Originally posted by gmat
Murray:no one outside the USA knows what the hell is gpm or cfm... heck i don't even know exactly how long a foot is in meters. Every international project i've been on (not many, but a few already) involve only metric units...

Sometimes I think half of the work of an engineer in the US is converting units from English to SI and vice versa.

Damn the English and their blasted units! They aren't worth a slug!

Alchemy

BillA
02-27-2003, 10:29 PM
this is a pretty well tilled field, but jus' one mo' time

gmat
I have been told that the axial fan P-Q curves typically show 'both' units
well, crap
pabst fan curve (http://www.ebm.com/series.asp?SeriesID=FC040)
the gd thing is using Pa !
so much for mH2O representing a consensus

my question/point was that it seems 'all' fan curves have dual units
AND
those curves are going to have to be understood, and used, by the user to define the fan's output (given the rad's resistance)
OR
all those curves (of the relevant fans) are going to have to be tabulated/transcribed so that the 'calculator' can do the work

hmmm . . .

I "measure" air volume by calculation given its velocity (profile) in a duct, can be expressed in anything desired

I measure air pressure in Pa because I have an inordinately sensitive instrument which outputs only that, obviously this is converted to the units of choice

I read liquid pressures in µA, displayed however, converted to whatever

I use lpm because I have several direct reading flowmeters and that gives me the most resolution
- it is also a simple conversion in the head to gpm

if one is to focus on utility, then the units of choice should be selectable by the user
if I can calculate such, certainly the 'calculator' should be able to do so
(lets eschew the punishment of the SI impaired)

Alchemy
read up on how those Frenchies came to 'measure' a metre
lol

murray13
02-28-2003, 12:39 AM
Thermal resistance vs. Thermal impedance:

As I understand the terms they differ in their units.

Thermal impedance units: °C-m²/W

Thermal resistance units: °C/W

I think thermal impedance is used to describe a material not an object. And thermal resistance is used to describe an object.

Maybe I've got it all basackwards.:shrug:

Alchemy
02-28-2003, 01:06 AM
Originally posted by unregistered
Alchemy
read up on how those Frenchies came to 'measure' a metre
lol

Old news. Nowadays the meter is determined by the speed of light, seconds by the frequency of a Cesium atom, and kilograms by the big old block of metal they keep over there in France.

So France has the distinction of being the only country holding an object that is a universal standard for a unit of measurement.

A lot better than having us measure things by the length of King Louis XVI's arm, I think.

murray13: Heat transfer applications are usually run in terms of heat transfer coefficients, which are W/C*m^2. This is usually multiplied by area when the area isn't known, thus eliminating it.

The terminiology of the inverse of the coefficient - impedence, resistance, etc - isn't widely agreed upon in the texts I've read. So long as we set a name for it here we'll be fine.

Alchemy

gmat
02-28-2003, 02:44 AM
Originally posted by unregistered

the gd thing is using Pa !
so much for mH2O representing a consensus


Well it only makes sense, as if one is talking about air pressure, there's no point into considering meters of water, thus using the SI unit (Pa).

We can perfectly use diffent units for water and air. After all air backpressure is only used at one place (rad, air side) and not in relation with any other figure in the sim.
On the other hand, some could consider it's throwing away the consistency, but consider that's air pressure, and it's different from water pressure.

So - what do we choose - use Pa for fans (as per the mfgers specs), and mH20 for water ? Or Pa for both ?
Remember the end user won't see any figure, apart from the output of the solver (°C/W, °C and lpm).

Les
02-28-2003, 03:29 AM
Originally posted by gmat
Remember the end user won't see any figure, apart from the output of the solver (°C/W, °C and lpm).
If user sees a dPwater output then my play suggests:
"bar" is viewable for dual axis graphs with "C/W"
"m(H2O)" is viewable for dual axis graphs with LPM