How do I design a fan powered cooling system for my growbox?

Introduction:
Every serious growing box needs cooling. Most of us use air cooling because it is cheap and very effective. The following steps are used to design a simple fan-cooled box.
This method does not cover active cooling with air conditioning systems or 'CoolTube' designs. It is for grow chambers where the walls are approximately equal to the light pattern, totally enclosed for airflow control, and do not have large radiant heat into or out of the box. Your mileage may vary some for these reasons.

I also picked sane defaults for growing conditions. The formulas diverge if you get too far out of plant growing range. You should be very safe if you are within about 40 to 150 degrees F and 20% to 90% humidity ranges (those are just guesses). Atmospheric pressure was picked as sea level and doesn't really affect anything until about 5,000 or 8,000 feet depending on how accurate you want to get. If extreme conditions apply to you, there may be other FAQ entries with the entire full blown set of pressure/temp/airflow/humidity parameters.

Design
1) Start at the beginning and design this right! Before you ever buy or cut anything for your new project, determine the highest temp (in F) your intake air will ever be when lights run. Get a thermometer and measure it to make sure you have a good value. Call this T(inlet)

2) Use these formulas to determine difference in temp you can tolerate. 81F (27F) is about the optimal for growing, 86F/30C on the higher end.

Tdiff = 81F - T(inlet) (English)
Tdiff = 27C - T(inlet) (Metric)

3) Add up wattage for all power in your box. Lights, pumps, heaters, humidifier, radio, coffee pot, whatever. Add it all up and call it Watts. This will make your number worst-case and therefore a conservative value.

4) Compute the absolute minimum fan power you will need using the following formulas. This is the minimum fan rating you must have to achieve your temperature goals. You will have to increase fan power to compensate for duct constrictions, small inlets, carbon scrubbers, screens, or other items that block airflow.

CFM = 3.16 x Watts / Tdiff (English)
CMH = 2.98 x Watts / Tdiff (Metric)

The formulas are almost identical, due to the counteracting effects of converting airflow from CFM to CMH, and converting temp from Fahrenheit to Centigrade.

formulas can be found on this web page:

(This web site also lists the above formula and uses a constant of 3.16 as shown above)

5) If you have more than one fan, they should be mounted side-by-side rather than inline if you want to add their different CFM ratings.

For inline fans, use the lowest airflow rating of all fans in the path. A fan on the inlet and a fan on the exhaust of the box are considered inline fans. Fans inside the box should not be counted for airflow but must be included in wattage. A standard computer fan is normally right around 30 CFM (50 CMH).

The two lookup charts solve this equation for common lights. Make sure you get the proper one (English or metric). For those of you who are wondering if you did this right, here are a few numbers in English units :


Note: a 30cfm computer fan is trying to cool a 1000w HID bulb, in the 3rd from the last row, as an extreme example

If you are adding any carbon scrubbers or extensive ductwork, this is where you add to the fan size to account for air pressure losses. You have to move this many CFM, or the numbers don't come out right. The deciding factor for these items depends on your exact configuration and is beyond this discussion.

6) When your box is built, buy a thermometer and measure the air blowing out of the box (temp probe or thermometer should be in the air stream just after the fan, outside of the box enclosure) and the temp of the air entering the box (again, from outside the box perimeter). Make sure there is no direct light shining on the thermometers to ruin the measurement. DON'T MEASURE THE TEMP INSIDE THE BOX YET!!!! It's best to do this with 2 thermometers or a single thermometer with a remote probe. Cheap thermometers don't work well because they aren't very accurate. If you only have cheap thermometers, use the same one for all measurements to avoid accuracy issues.

7) Subtract your measured inlet from measured outlet temp. Compare to Tdiff from above. Is your measured difference as good or better than your estimated from step 2? If not, go find out why. Your problems are probably:

A. Heat source you didn't account for (the ballast?)
B. Your fan is overrated
C. You have blocked airflow
D. Your temperature measurement was inaccurate
E. Air leaks into the box (especially around the fan!) that ruin efficiency.

8) Once you get your measured temp difference equal to step 2, measure temps inside the box. Don't let the light shine right on the sensor, it will give faulty readings!! Use a light shield made from a tin can or something. If temps inside the box are higher than your exhaust temp at a reasonable distance from the bulb, you have air circulation problems inside the box. Get some kind of fan to stir up the air in there or look for airflow paths that allow air to travel from inlet to exhaust without spending any time in the box.

9) Always monitor the temperature difference between inlet and outlet temps every time you water the plants. If it varies much more than a degree or two, find out why. I use digital indoor/outdoor thermometer. It tracks high and low for both locations, outdoor probe is on a long wire, $14 at Kmart. No part of the thermometer is inside the box, just in the measuring air blowing in and out from the outside.

Please note that conversion values have been slightly rounded off to make this easy. Using the metric and english formulas will yield slightly different answers if compared. The difference should be less than one percent and can be ignored.

You can use the two load graphs attached if you prefer to do calculations visually rather than using the formulas listed above. Find the line for your light wattage and ignore all others. Each axis is logarithmic, make sure you count along each axis properly. The formulas listed in step 4 were used to make the graphs.

Sidenotes:

You can measure your fan airflow very accurately if you use a standard trouble light with a 60 or 100 watt bulb in it. These are very good test loads for calibrating things.
Just put it in and work through the formulas using a good thermometer to determine airflow. If you doubt the accuracy of your bulb and are really anal about it, you can calibrate the bulb against your electric meter over several minutes. You could also stick in a different brand of bulb at the same wattage and compare results. I haven't tried this, but I would just trust the bulb until proven wrong.

Testing and measuring duct losses:

Ducting losses are hard (to measure) because they rely on knowing your duct material coefficients. You can measure the losses in the duct after it is built and running, if that would help. You could measure a test section to calibrate that material, then extrapolate. Here's how:

Take a known fan (or the fan you will be using) and blow it into a plenum that has a heat source and some of your sample duct mounted to it. To do this, you need a trouble-light or other low wattage known source and a cardboard box to put it in, then mount the fan on the box and stick the duct on the other side.

Calibrate the box by measuring temps without the ducting, then compute CFM. Add the duct, measure the new temp, compute the new CFM. The difference is duct loss. Basically, use temperature and wattage to measure airflow and compute duct loss.

If you have an existing room, just measure inlet and exhaust temps, add up the watts, and then compute effective airflow. I just did this for my box and it's pretty much dead on. I think it varies by about +/- 0.2 DegF for 150 watts and two computer fans.

Once you have the value for your ducts, you can estimate loss by adding up the length. We would have to come up with adjustment for going around corners.

I once saw a Mech Eng. book that had different shapes of pipe listed (Tee, 45 Deg bends, 4-way branches, Y-branches, etc) and then gave an equivalent length of straight ducting they add for flow resistance.