I would like to do some thermal characterization on a heatsink assy and it would be helpful to get a "somewhat" accurate air flow measurement. I've thought of building some ductwork and variable speed fans to control the air across the fins but need a way to measure the air speed, etc. I know I could spend some $$ for a lab flow meter but it occurs to me that cars have a air flow meter in the intake system. Could I get one of those cheaply from a salvage yard and would it work for the sort of low air speeds (0-8mph) I want to test at? I would need to get the hookup connection and what the signal vs flow cal information is if its available..... The design I am working on has a heatsink about 9" wide with 1" fins and typically we would be running
150W or so with it. I need to collect the temp rise vs air speed to verify the design.
You could use a hot-wire anemometer based on a temperature sensitive filament run at constant temperature. I've not personally used one of these, but they're reckoned to work well at low flow speeds. For some reason I can't select a fixed font for this reply, so I'll have to describe the basic circuit:
Op-amp output to base of an NPN transistor. Transistor collector to +ve supply voltage. Transistor emitter to two resistors, R1 and R3. Other end R1 to amplifier inverting input. PTC resistive element, Rf, from inverting input to 0V Other end R3 to amplifier non-inverting input. Adjustable resistor, R2 from non-inverting input to 0V. Appropriate + and - supplies to the op-amp.
The sensor output is the voltage on the transistor emitter and is non-linear with air speed, sensitivity being greatest at the lower speeds. R2 is adjusted for optimum sensitivity at the flow rates of interest. You may also need to add additional component(s) to ensure that the circuit initialises at switch on, depending on the initial status of the op-amp output.
The air flow meter on my car is fairly stiff. You need a lot of CFM to get these going. There must be a cheap meter around. You could make it with a resistive sensor attached to a vane. Make it for $10. Cal it yourself. Air speed is one thing, but CFM in a tunnel is another thing.
I've thought about this too. A hot-wire aenomometer might work nicely, if you could find a way to calibrate it. Could be done in a car if your probe is on a long enough pole to get away from the car's slipstream -- not compatible with traffic! Might be able to calibrate it "close enough" with a shop-made pitot tube, a simple slant manometer made of a bit of glass tubing and some colored water, a leaf blower with variac or lamp dimmer and some duct.
formatting link
It would probably be good to pack the duct upstream with some bits of tube, maybe a foot or so of length, to reduce turbulence and get closer to laminar flow. Bits of 1/2" EMT (conduit) might work nicely for that and EMT is very cheap -- about a buck for 10 feet of it.
A simple cheap way to make a pretty good hotwire aenomometer is to break the bulbs on a couple of automotive marker light bulbs to expose the filaments. (I'm not making this up, I've done it.) Shield one from the flow, excite both with same current, measure the differential voltage. They're a bit fragile, but they work quite well at low flow rates like you mention. Might also be able to do it by maintaining constant resistance with suitable elex (low-level AC current excitation, AC control loop to keep sensed AC voltage at setpoint) and then measure the DC current or voltage necessary to maintain that temp and resistance in the presence of airflow.
Volumetric air flow in ducts as in HVAC systems is often measured with a distributed pitot e.g. a ring of tubing with a bunch of holes in it, or a distributed hot wire as in some mass air flow sensors in throttle bodies. Thing is, the velocity profile across a cross section of duct is far from uniform.
Or, just go with the pitot tube and manometer and fuggedaboud the electronics!
"JH" wrote in news:1158071090.457168.90670 @p79g2000cwp.googlegroups.com:
The problem is calibration. Get a large container of known volume - say a
3 gallon water bottle or a small trash can you can measure the volume of. Cut several large holes in the sides and bottom of container. Fit a collapsed trash bag over the fan. Insert trash bag into mouth of container. Turn on fan and measure time it takes trash bag to fill container. Turn fan off, collapse trash bag, repeat and take average of several measurements. This gives you the aproximate average flow rate. Modify to suit needs, season to taste.
There are lots of ways to make a device to measure the speed of airflow. But if you don't have a calibrated "airflow meter" (which is what you are trying to avoid buying), you will have to also devise some way to calibrate whatever you build.
I have read some of the other posts, which mention some possible calibration schemes that sound like they should work.
But another possible calibration scheme just occurred to me (I've never tried this, and am just speculating): Assuming you can get your hands on a heatsink for which you can also get a good/detailed manufacturer's datasheet, i.e. which has some data for the temperature-rise/watts/airflow-speed relationships, you could then try this: Dissipate a known power level with/into the heatsink, measure the temperature (or temperature change) of the heatsink (per heatsink's datasheet's measurement-location/methods/specs), and adjust the airflow speed (and/or power and temp) until your "knowns" measurements/values match datapoints that are available from the heatsink's datasheet, which should then give you the airflow speed. Then probably repeat that over ranges of power levels and airflow speeds.
I don't know how accurate that method might be. You might also want to talk to the heatsink manufacturer's engineers, first. But, on the plus side, it sounds like it might be pretty easy to try, using hardware and information that you might already have, or could easily get.
Good luck!
- Tom Gootee
formatting link
"He who lives in a glass house" should not invite "he who is without sin".
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.