How do you measure headspeed

[MicroMan]

the cheap way without a Tach, is there such a method.

[RSL_Mongoose]

Theres a headspeed calculator here that you can Download, I'm not sure how acurate the thing is but its cheaper than a tach, let me know how it works out for you, it is for electrics.

http://www.helitown.com/Downloadpagepipoinheadspeed.html

[MicroMan]

Thanks RSL.

[kgfly]

Check out this thread for some interesting ideas. There is a way to get an estimate using a digital camera and little maths:

http://www.helifreak.com/viewtopic.php?t=18538

OR for a shorter/friendlier version:

http://www.archeli.com.au/phpBB2/viewtopic.php?t=8642

[Jermo]

kgfly,

can you upload your audio file ?
Seems to me a few fundamental assumptions make this resolvable with audio but will require some filtering of the audio using logic.
Let me spatter some thoughts and see if I can set this up correctly and come to a conclusion. If this works on paper we could easily write an application to run on a laptop to tell us the speeds.

The approach I'm taking is based on filtering out everything but the blade sound based on the relationships between each sound source. What I expect is a wideband audio proportional to the RPM of the blades that varies in frequency from the inner most tip of the blade(s) to the outermost tip of the blades. Using software to do the filtering and calculations we can use the estimated minimum frequency of the blades, the maximum frequency of the blades and the lenght of the blades to closely estimate speed of the blades at any given point which would directly give us the headspeed.
Needed items are:
Directional Microphone (take a mike, put it in a paper tube..viola! directional mic,
recording device OR laptop,

software to do the analysis.
t = time
f = frequency
Mf = Motor frequency
Mr = Motor RPM
Bmr = Main Blade RPM
Bm = number of main blades (typical heli has 2)
Bt = number of tail blades (typical heli has 2)
Bmf = main blade frequency/Bm
Btf = tail blade frequency / Bt
Pf = Pinion frequency
Hr = Head RPM
RPM= Rev per min

Physical relationships and characteristics:
Motor f should be semi narrow band, noisy, and higher than Pf, Bmf, and Btf
Pinion f should be narrow band, quiet, and lower than Mf
Main Blade f should be very wideband, louder than pinion f but quieter than motor f
Tail Blade f?? no clue I would guess it's wide band, quiet, lower than everything

How does that look so far?

At the end of the day it should be simple for the modern soundcard to filter out down to the primaries and calculate the values we seek. It's all relational.
Jermo

[Pinecone]

Much easier way:

http://www2.towerhobbies.com/cgi-bin/wti0001p?&I=LXPT31&P=ML

:)

[ClayK]

I wouldn't use that particular gauge though. You have to stand over the helicopter in order to "shoot down" onto the blades to pick up the headspeed. Horribly unsafe for a variety of reasons. :WOW

[kgfly]

Jermo,

Interesting ideas. I don't have a recording, I did my experiment by playing a youtube video of a TRex hovering while running the spectrum analyser. Not a good experimental design due to the many distorting factors between the original heli, recording, compression and playback. However I did not have a safe way to hover my TRex anywhere near my PC. I might have to try again using a laptop to record the audio. Right now it will have to wait until I can get the spares I need to fix my TRex :oops:

[Jermo]

kgfly,
I found some code for audio analysis. It's a good start, Uses some basic fourier transforms. Once I can get an ideal audio sample to start with I'll know more about identifying secondary audio sources..etc. Heck technically we could fingerprint each heli from it's audio signature but that's a bit extreme :shock: .
The main concept sounds valid. IF it I can make it a viable solution it could probably be coded for a pocket pc. I don't know enough to build a stand alone version otherwise but it could likely be done on a microcontroller device. The algorithms and filtering are the keys.
Jermo

[eeeyal]

:WOW Cheep way to get head speed: borrow your buddy’s tach

[kgfly]

So there you go MicroMan,

The practical suggestions for "cheap way without a tacho" are limited to:

a) Borrow a tacho (should give fairly accurate results)
b) Digital camera and a little maths (will give a useful approximation)

Other less imaginative suggestions that don't quite meet the constraints:

c) Buy a tacho

And finally the difficult and possibly useless but intellectually challenging

d) Use audio spectrum analysis and signal processing

Some folk seem to lose sight of the fact that this is a hobby and doing stuff for fun is OK even if it's harder or more complicated. This sound analyser thing is simply an interesting intellectual challenge. Might lead nowhere. Might prove unworkable. Might end up being helpful. In this case the destination doesn't matter, it's the journey that counts.

[eeeyal]

it's the journey that counts.
:YeaBaby:

[spork]

b) Digital camera and a little maths (will give a useful approximation)

I just posted this digital pic to the other thread you mentioned. Due to the way cameras shutter the blades look curved rather than blurred. With a carefully taken pic you can still do some math and get a good approx. of head speed.

Some folk seem to lose sight of the fact that this is a hobby and doing stuff for fun is OK even if it's harder or more complicated.

I know exactly where you're coming from. Most of the really fun and interesting stuff (for me anyway) comes from the most ludicrous of pursuits. For me it's more about learning than doing (but you've seen me fly, so you can probably understand that :mrgreen: )

[kgfly]

Thanks Rick. I didn't see this affect with my digital still camera, maybe the scan rate is different between still and video modes ?

[spork]

Oh no. Now I've got us jumping between two threads.

Nice pic. If you used a flash you wouldn't see that effect (although it doesn't look like you did).

Also, the effect will be minimized if the shutter "traverse rate" is high relative to the shutter duration. The CVS cam is peculiar in that it's traverse time is always 1/60th. However, I have seen this exagerated effect in many pictures.

By the way, if you aren't finding a significant peak in the spectrum analyzer you need to stick a baseball card in the blades! :mrgreen:
(Weird Science!)


EDIT:

Looks like you shot that at 1/250th?

[Pinecone]

YOu can also set the tach up in a roll of duct tape under the blades pointing up. Just do it outside looking at clouds or a clear blue sky.

For smaller helis you can strap the thing down to a platform and even check speed at various pitch settings.

[AV8TOR]

I wouldn't use that particular gauge though. You have to stand over the helicopter in order to "shoot down" onto the blades to pick up the headspeed. Horribly unsafe for a variety of reasons. :WOW

Or you could strap it on the tail boom looking up, bring it into a tail in hover and read it.

[kgfly]

EDIT:
Looks like you shot that at 1/250th?
Yup. No flash since that would give (a) too short an arc increasing the error margin and (b) an unknown actual exposure time so we couldn't do the calculation.

[Jermo]

If you mount the Tach to the Heli don't you change the effective hover head speed due to higher load on the Heli? If I recall the throttle and pitch are both on a curve. A heavier heli requires more throttle to hover which would move the curves up.

Does that sound right?

Jermo