anemometer

just as a tip, nobody cares about the airflow numbers. It's all about temperatures. Over 100 temp readings are possible in modern servers. None of the instrumention is about airflow itself, because it just doesn't matter once the minimums are met. For example, different sizes and shapes of heatsinks in the same flow of air behave differently.

You may need to watch out if their air speeds are high and you want to control air flow with a piece of swiss cheese. Noise and turbulence will be your new friends.

If you want to go real extreme, use a blower and actual tubes/pipes to distribute the air where you need it. Think irrigation system, or old avionics.

Too late now. But getting fairly uniform air flow into 8 tubes could get interesting too.

The baffle is a first approximation to a positive-pressure plenum with controlled-resistance exits.

Yes. For the required accuracy, it was a clever design. Ingenious.

Hard to predict without an exhaust vent profile. if centered in the rear, it can equalise flow through equalised back-pressure.

Looks like the boards are too close to the rear wall, but that can be used to increase influence of the exhaust pattern.

Use of back pressure eventually means less net airflow.

RL

The back panels will have various cutout sizes that depend on how much air flow each type of module needs. That of course complicates things more.

Would a rear plenum help? Our current rear plenum is the world.

Sure. We want some guaranteed minimum flow in the worst slot, 200 f/min as a target.

An interesting number. You mean linear FPM ?

With local scrubbers (which is what you've been touting in the past), your only interest would be delta T of the air, by the time it leaves the box. No numbers on the drawing, but assuming an 18" rack mount, that would mean >> 100 cubic feet / minute of air volume delivered by the fans.

Check the fan spec sheets. What's the air volume capacity of the fans at full speed?

So you're heating about 4Kg ( 3 x 1.276) of air every 60 seconds

- that's .067kg per second - .067kj or 67W for every degree C rise.

If the air volume halves, delta T doubles. If energy absorption doubles, delta T doubles.

With an energy load of x watts, that gives you box air volume requirements in your environment.

Air volume (feet per minute - linear or cubic) will depend on back-pressure presented to the fans. I expect it will be considerably less than 100 cubic feet / minute.

RL