optical tach for aircraft propeller

[djacobow]

tl;dr: how might you design a propeller tachometer?

Long:

I'm trying to design an optical propeller tachometer for piston aircraft. Basically, a box you can point at the prop from inside the aircraft and it tells you the RPM. There are several products on the market that do this, but I thought it would be an interesting adventure to make my own.

Most optical tach projects you see on the internet assume that you are illuminating a sensor yourself and have something intermittently blocking the illumination. This case is harder as illumination is coming from the sky. I think an IR photodiode pointed at a propeller will generate a signal because the prop is ambient temperature (warm to quote cold) but the sky behind it is very very cold, but that signal might be quite small.

So, my plan is to use a simple IR diode, amplify/buffer it's output, band pass filter it, and then send the output to the comparator inputs of a microcontroller to time and average the pulses.

The RPM range of interest is from, say 200 rpm to 3500 rpm, which means that the device needs to work from about:

* min 7 Hz   (bottom of RPM range with 2-bladed propeller)
* max 300 Hz (top of RPM range with a 5-bladed prop)

I worked up the attached circuit in LTspice, and it seems to do the business with about the right passband, but my analog-fu is very weak. The LPF is coming from the two RC stages and the HPF is coming from the AC coupling. I have many questions:


First:

 -- is this right level of complexity?

 -- what sort of IR diode is best? I have a few on hand, and though I've only tested with a window fan, they all seem to work, in the sense that I can see a signal on my oscilloscope in the 1-10mV range on AC coupling

 -- can I get away without optics? The diode will be detecting all the light in the general direction in which it is pointed. The variation from the prop will be relatively small. But the background will be changing slowly, so I'm assuming it will easy enough to separate the variation from the static. I can also put the diode in a tunnel to make it more directed. But I'd like to avoid any actual optics.

 -- what sort of op-amp is best for this? The bandwidth required is very low, so I'm assuming any bottom-of-the-barrel device would be adequate. (Also, does LTSpice have any built in models for crappy op-amps, like something akin to LM358?). I don't think noise should be an important factor, but maybe I'm totally wrong about that.

Cockpits are not a great places to fool around with a circuit on a breadboard, so, though I will absolutely make a few trips to the airport (it's not very close) to see if basic assumptions are right, I'll want to build most of this circuit up solidly before I try to fly it. If it weren't for the airport/airplane hassle factor, I'd just go try what I have first rather than come here.

Open to any and all thoughts.

Best,
Dave J

[wasyoungonce]

Might be easier to use an inductive sensor on propeller hub.  I cannot comment on which IR diode, you can test your system maybe on a mock-up, it should work.  Might need a decent reflector on the propeller.  We used to use IR diode setup on Helicopter blades to sense doppler shift in time between blades, aka so we could trim them so they tracked evenly, but yes its used in Aviation.  The tape was that aluminum foil type. Might not last long on a prop though.

You should be able to get away with out optic collimation if the distance is not that large.

[mjs]

You're not measuring thermal radiationwith a photodiode - it's the visible ambient light you're measuring.

You can use a photodiode with lens (like SFH213) and cover it a bit from light arriving from sides. I would have added a cheap laser+red filter for photodiode, but if ambient light works ok, then why complicate things! Just remember that cloudy/dark sky will give much weaker signal.

I would lose R1, R4, C2 and  opamp negative input to form a transimpedance amplifier with 47kOhm gain- you've already got AC coupling with C5. Just make sure the DC light does rail the opamp with maximum light. Adjusting R2/R3 divider can add a bit of headroom.

LM358 is probably fine.

[djacobow]

Might be easier to use an inductive sensor on propeller hub.

Can't modify the airplane. I'm a renter pilot and this is something I'd put into my flight bag.

[djacobow]

I would lose R1, R4, C2 and  opamp negative input to form a transimpedance amplifier with 47kOhm gain- you've already got AC coupling with C5. Just make sure the DC light does rail the opamp with maximum light. Adjusting R2/R3 divider can add a bit of headroom.

LM358 is probably fine.

Thanks! I'll try this. And thanks for setting me straight regarding ambient light versus thermal radiation.

[Benta]

Such things have been available to RC-plane modelers for years and are quite inexpensive. Point it at the prop and you get rpm. Normally they have a switch for 2, 3 or 4-bladed props.

[2N3055]

Why propeller? It is same as engine RPM.

[mikerj]

Why propeller? It is same as engine RPM.

Unless there is a speed reduction unit...

[2N3055]

So what ? Still going to be linear relationship. RPM / constant.

[djacobow]

So what ? Still going to be linear relationship. RPM / constant.

So, the main reason is that I don't own the airplane. But even if I did, attaching anything to an airplane anywhere for any reason is expensive. So this has to be a no-contact sort of thing.

Aircraft actually have tachometers built in but they are notoriously inaccurate and can't be relied on for accurate power settings. Typically, they were installed when the airplane was built in nineteen seventy something something and have not been touched since.

Aviation is strange. Trust me.

[djacobow]

Such things have been available to RC-plane modelers for years and are quite inexpensive. Point it at the prop and you get rpm. Normally they have a switch for 2, 3 or 4-bladed props.

I could try one, but I wonder how well it will work. The main differences I see are that the RPM range itself is pretty different and that the device is much farther from the prop, maybe 6-10 feet rather than 6-10 inches. And you're shooting through some old, crazed plexiglass. OTOH, the prop obviously much larger.

[Benta]

Such things have been available to RC-plane modelers for years and are quite inexpensive. Point it at the prop and you get rpm. Normally they have a switch for 2, 3 or 4-bladed props.

I could try one, but I wonder how well it will work. The main differences I see are that the RPM range itself is pretty different and that the device is much farther from the prop, maybe 6-10 feet rather than 6-10 inches. And you're shooting through some old, crazed plexiglass. OTOH, the prop obviously much larger.

Why not just buy one and try? They're around $30. The rpm range doesn't matter, they show 1 rpm and up. If it doesn't work properly, improving the optics could be an option.

[sokoloff]

I own a genuine PropTach (the $300 aviation specific [read: over-priced] version of this) and a cheap RC tach. Both work. The Prop Tach works in slightly lower light, but neither work at night.

My advice: if you want to measure the RPM, buy the $28 Hangar 9 tach and be done with it. (My RC tach isn't that exact unit, but it looks quite a bit like it and I suspect works the same, which is to say "good enough for purpose".)

Most aircraft engines allow 2% overspeed indefinitely without need for inspection, so a small amount of instrument error is no issue. Gross check the factory tach with the optical tach and then you can use the factory tach operationally. These airplanes flew a lot of years with the old analog tachs. (In many cases, the tach is actually an automobile speedometer, marked in RPM, with the inset tach hour meter being an odometer.)