air flow sensor on PCB

Metallic RTDs are quite stable, and the PTFD form (in the datasheet above) can stick up 10 mm, which is about the same as a TO-92. One can compare a SMD version with a nearby PTFD version, canceling local ambient temp variation.

But the RTDs are ~10 times more expensive than 7 cents. Don't know how much difference that makes in this application.

Joe Gwinn

Another possibly interesting approach would be to use a loop of flex circuit with a SMT thermistor in the middle. A bit of copper pour on the ends of the thermistor, with little skinny serpentine traces going to the board, could make the sensor temperature actually track the air quite well, I'd expect.

Cheers

Phil Hobbs

In a given crate, the air flow might go in any direction, which could affect the loop. The classic below-the-boards fan tray can create bizarre flow patterns. I'd expect the flow speed and direction to vary in different places on the board, too, so I guess I'd go for the middle. A TO92 is a reasonable approximation to a cylinder.

The weather has been really weird here lately. Yesterday we had three distinct cloud layers, moving in different directions. Air is evil.

That's a good price. We have an essentially identical Minco part in stock that costs $4. But it would be flimsy to mount on a board, sticking up into the air flow. And it needs to self-heat to measure flow.

I guess I could do my sequenced heat and cool time constant measurement on it.

Not a big deal, but may as well save money if it doesn't take a huge hassle.

That's insanely complex. Not hardly "simple, low cost."

Do you have any idea how the piezo sensors measure air flow?

You still need a magnet to operate the Hall sensor, but yeah you could loose the stator and have an anemometer. BTDT

It's your gizmo, of course, but to me, a sensor that has a 90:10 air vs board temperature seems likely to be less bothered by mild to moderate airflow pattern disturbances than one that's 10:90.

The old National Semi temperature sensing handbook has a lot of good info about using plastic packaged ICs (mostly TO-92) for air temperature sending.

. Cheers

Phil Hobbs

An LM35 could be persuaded to self-heat I guess, load it pretty hard and vary Vcc. But I don't like to run them above 5 volts.

Shift in time-of-flight. The projects I've seen use four in a square, and measure TOF across the diagonals. Needs temperature and humidity compensation to be accurate.

Both of the piezos in the TI diagram seem to be receivers. And the signal path includes rectifiers and "smoothing." That doesn't look like TOF.

Horrible app note.

Incorrect as usual.

The reason the fans start at high speed is that the PWM controlled fans that Dell installed do not startup at anything other than a high duty cycle PWM waveform. Once spinning, they can be throttled down to their usual RPM sufficient to maintain cooling. But to initially get started spinning, they require that high speed jolt.

The PWM fans the we have used will start spinning at low PWM duty cycle. We start them at powerup at some minimum quiet speed, and then slew up gently to max if measured temperature is high. That avoids rude noises.

May be they make them smart enough to do just a few periods of 100%, see them (can they?) started without going to high speed and then just do the pwm you have set? I am wondering because I have not had a PWM fan in my hands (did not know such existed, really). Some 30 years ago I did PWM on a normal

12V fan on the first nukeman (the second, copied that) and I had to go for a period of something like a whole second IIRC, the thing would not move otherwise. And I did have to start it initially at full power so it would start; and yes, the very low PWM frequency was audible... if this can be called pwm, it was more an on/off thing. I'd have to check those PWM fans, lately (on the tld readers) I just did on/off, long time on/off that, is some 1C hysteresis etc.

We have older boxes where we used non-PWM DC fans and controlled speed with the DC voltage into the fans. They also start up slow. I think we decided on a minimum of 20% voltage on those.

Good fans seem to work well at low speed.

Our algorithm is

Once a second,

If the box is too hot, increase fan speed by 2% of full scale else Decrease by 2%

But limit to 20% and 100%.

That works well without drama.

Probably, but air flow could be blocked and fan blades would still flap the sensor. May as well just use the free tach.

I would guess that a TO-92 and a PTFD RTD are equally fragile in practice. Don't know about the Minco. Part number?

Isn't that true in general?

A 100-ohm RTD will be easier to heat from 5 volts than the 1000-ohm RTD, but one tenth as sensitive to temperature change, so there is a tradeoff to be made.

Sort of like a ring-down measurement - fit the log of the offset versus time to a line, which can be done pretty fast - fever thermometers do this and the algorithm is simple. So it could work.

How fast does loss of airflow need to be detected to avert damage?

Yeah.

Joe Gwinn

My inventory program shows

MINCO S262PF MINCO S247PF12 MURATA TRFA102A OMEGA F3141

for the ceramic slab 1K thin-film RTD.

The 100 ohm 1206 is

ENERCORP PCS 1.1503.1 VISHAY PTS1206M1B100RP100

I'd probably have 12 or even 24 volts available in the case I'm considering. That would toast a 1K RTD.

It could be very simple, just a couple of timed limit checks. But with

6 GHz of ARM available, we could get fancy.

I'd think that one test per minute would be enough. We could probably manage two. With two fans, nothing will fry soon.

The VME crate controller card can only raise a flag and light a red LED. It may be in a crate where it can't do much else.

Our general fault plan is to shut down for 1 minute and try again. We'd do that for individual plugin modules too, when something terrible happens. If we can.

Am I missing something: you want to check airflow so that you can prevent the board from overheating, right? If so, why don't you just measure the temperature directly? There must be spots on the board that would be the first to overheat and give the alarm.