Author Topic: Light testing - the results  (Read 72779 times)

simonp

  • Omnomnomnipotent.
Re: Light testing - the results
« Reply #75 on: December 01, 2008, 08:19:03 pm »
I plugged a SON into an oscilloscope, once (I was investigating building a charging/lighting circuit so you could charge by day and have SON + battery boost at night).  I didn't get a sine wave.  The waveform will depend on the nature of the load, so I think using an AC source might be unrealistic.


Re: Light testing - the results
« Reply #76 on: December 01, 2008, 08:19:10 pm »
Pffft!  ::-)

You lot belong in a lab, wearing white coats.



The real cyclist, real world darkness of the yACF comparison was much more relevant than some lab experiment.

Charlotte

  • Dissolute libertine
  • Here's to ol' D.H. Lawrence...
    • charlottebarnes.co.uk
Re: Light testing - the results
« Reply #77 on: December 01, 2008, 08:27:13 pm »
It's all about the "bin" of the LED, innit?  Even same spec LEDs will have different characteristics, so manufacturers rate them according to quality.  The better the "bin", the lower the power consumption and the better the output.  I have learned this on CPF  :)

What it would need to get better, consistent results, is a test rig as TimC says.  For best realism I reckon a SON-equipped front wheel, a DC motor, some turbo trainer spare parts, an old pair of front forks, appropriate mounting hardware, and some control gadgetry to regulate the speed (servo controller maintaining a constant speed rather than position).

Could be done for under £100 if you already have the wheel and the spare forks I reckon.

If the setup was clever enough you could even have some indication of power usage - if you can measure the current required to maintain the constant speed then you should be able to work out the wattage.


Oddly, I was considering rigging a spare SON wheel I have into my sturdy wheel jig and using a belt drive round the rim to an old drill or something.  Whacking a multimeter across the terminals would tell me when it was up to a constant output and we could use that to power the lights.

But on balance, I thought I had less chance of braining myself on the rollers...
Commercial, Editorial and PR Photographer - www.charlottebarnes.co.uk

Re: Light testing - the results
« Reply #78 on: December 01, 2008, 08:36:34 pm »
Really you want some stepper motors to point the light as you sweep it around, a light meter with a suitable interface, likewise a current clamp, and obviously everything plugged into a computer to control it.  You just plug a light into it, go away for some tea and cake, and come back to a screen full of results.  Of course you also need a controller for the motors and servos.

There are a few widgets around which allow this sort of interface, the various Bee things would do the jobs for a Windows box, so long as you don't mind writing the code to talk to the C DLL they supply.  I'm sure there are similar things for Linux.
Actually, it is rocket science.
 

Re: Light testing - the results
« Reply #79 on: December 01, 2008, 08:50:24 pm »
Brilliant stuff. I did a lot of trawling for this sort of thing a few months ago and this is better than anything I found.  :thumbsup: What I did find was a lot of pictures of people's back gardens. ???

One question - was the Fenix L2D or P2D? (label in graph is P2D)

Justin(e)

  • On my way out of here
Re: Light testing - the results
« Reply #80 on: December 01, 2008, 08:52:32 pm »
Echos applause and plaudits for those responsible.

Re: Light testing - the results
« Reply #81 on: December 01, 2008, 09:40:00 pm »
It's all about the "bin" of the LED, innit?  Even same spec LEDs will have different characteristics, so manufacturers rate them according to quality.  The better the "bin", the lower the power consumption and the better the output.  I have learned this on CPF  :)


According to one of my former colleagues who did the electrical assessments, it wasn't unknown for there to be a bit of fiction of the figures when running an electrical assessment of some of our wafers. We were often behind schedule with the orders and customers can get impatient. I never told anyone that OK ;)

Zipperhead

  • The cyclist formerly known as Big Helga
Re: Light testing - the results
« Reply #82 on: December 01, 2008, 09:45:16 pm »
One question - was the Fenix L2D or P2D? (label in graph is P2D)

It was an L2D, I've just fixed the caption.
Our son does know who Boz Scaggs is, we've done ok as parents.

Zipperhead

  • The cyclist formerly known as Big Helga
Re: Light testing - the results
« Reply #83 on: December 01, 2008, 09:46:22 pm »
I've just modified the pictures in the original posts to change the layout of the images. Hopefully it will mean that you can read the captions without having to download the pictures.
Our son does know who Boz Scaggs is, we've done ok as parents.

LEE

Re: Light testing - the results
« Reply #84 on: December 01, 2008, 09:49:46 pm »
Really you want some stepper motors to point the light as you sweep it around, a light meter with a suitable interface, likewise a current clamp, and obviously everything plugged into a computer to control it.  You just plug a light into it, go away for some tea and cake, and come back to a screen full of results.  Of course you also need a controller for the motors and servos.

There are a few widgets around which allow this sort of interface, the various Bee things would do the jobs for a Windows box, so long as you don't mind writing the code to talk to the C DLL they supply.  I'm sure there are similar things for Linux.

Are you sure that would work?  It just seems too simple.

Re: Light testing - the results
« Reply #85 on: December 01, 2008, 10:26:54 pm »
It depends on how complex you want things.  You would need to work out some mechanics to drive the lights motion from the stepper motors.  I'd probably just use some microswitches for the end stops at ±90°, and then use that to autocalibrate the angular movement.  You would also need some sort of control to drive the motor, and probably a speed sensor, so you can drive the dynamo up to a constant speed whatever the load it exerts on the dynamo.  Getting a calibrated light level out might be interesting, although you could just use a photodiode or LDR and calibrate them against a decent light meter (although this probably wouldn't deal with complex spectral effects very well).  I think one of the older current clamps we have in the lab just chucks out a voltage level, so you could trivially interface that to a calibrated ADC.

How fast you could do a sweep of the beam pattern would depend on the sensitivity of the light sensor, it's settling time, and how detailed you want the measurements ie ±90° in both axis with 5° of resolution would produce 1296 data points.  If the sensor measurement and/or movement took 10 seconds per measurement, it would take almost 4 hours to produce a beam plot of one lamp!  In practice you probably only want detailed measurements near the central axis, and cruder measurements at the more extreme limits.

It's not rocket science.  That over there <fx: points across lab> is. ;D
Actually, it is rocket science.
 

Hummers

  • It is all about the taste.
Re: Light testing - the results
« Reply #86 on: December 01, 2008, 10:31:00 pm »
Really you want some stepper motors to point the light as you sweep it around, a light meter with a suitable interface, likewise a current clamp, and obviously everything plugged into a computer to control it.  You just plug a light into it, go away for some tea and cake, and come back to a screen full of results.  Of course you also need a controller for the motors and servos.

There are a few widgets around which allow this sort of interface, the various Bee things would do the jobs for a Windows box, so long as you don't mind writing the code to talk to the C DLL they supply.  I'm sure there are similar things for Linux.

Are you sure that would work?  It just seems too simple.

ROLMFAO.

H

Re: Light testing - the results
« Reply #87 on: December 01, 2008, 11:46:49 pm »
Rather than making some elaborate gadget to move the light sensor or light around to get multiple readings, I would have thought you could get fairly accurate results by taking several photographs at different exposure levels to create a very high dynamic range image. You could calibrate it with the light meter, but then you'd ba able to measure the light level at any point on the beam image.

Re: Light testing - the results
« Reply #88 on: December 01, 2008, 11:52:09 pm »
You could do that, but you would need a large flat white surface, with the lamp mounted relatively highly up, to give a sufficiently wide angle, and you would need a fairly wide angle lens to get a wide enough view angle.  I don't have easy access to either of those, whereas I probably could knock up a testing rig such as I outlined.
Actually, it is rocket science.
 

Pingu

  • Put away those fiery biscuits!
  • Mrs Pingu's domestique
    • the Igloo
Re: Light testing - the results
« Reply #89 on: December 02, 2008, 01:15:43 pm »
Did anyone try reversing the polarity?

Charlotte

  • Dissolute libertine
  • Here's to ol' D.H. Lawrence...
    • charlottebarnes.co.uk
Re: Light testing - the results
« Reply #90 on: December 02, 2008, 01:20:33 pm »
Did anyone try reversing the polarity?

Don't cross the streams!

 :o :o :o :o :o :o :o :o :o :o
Commercial, Editorial and PR Photographer - www.charlottebarnes.co.uk

Tim Hall

  • I want to eat the fleeting shade of your lashes
Re: Light testing - the results
« Reply #91 on: December 02, 2008, 01:45:49 pm »
Seeing as this awesome work is in the public domain, do you have any objections if I point readers of urc to the results?
There are two ways you can get exercise out of a bicycle: you can
"overhaul" it, or you can ride it.  (Jerome K Jerome)

Zipperhead

  • The cyclist formerly known as Big Helga
Re: Light testing - the results
« Reply #92 on: December 02, 2008, 01:47:11 pm »
Seeing as this awesome work is in the public domain, do you have any objections if I point readers of urc to the results?

Not as long as you exclude Judith!
Our son does know who Boz Scaggs is, we've done ok as parents.

Re: Light testing - the results
« Reply #93 on: December 02, 2008, 10:24:34 pm »
I've just caught up with this - great work, everyone

Re: Light testing - the results
« Reply #94 on: December 02, 2008, 11:29:43 pm »
Did anyone try reversing the polarity?

Yes, it went very dark ;D


IGMC

Re: Light testing - the results
« Reply #95 on: December 05, 2008, 10:46:04 pm »
No intent to derail, but....  I've a SON hub and a IQ Fly light.  Peter White's website says that two Flys can't be run together (assume parallel or series).  I've written to B & M and asked about wiring two Flys together (can it be done, if so how, what limitations or concerns) and after a week or so, have received no reply.  Anyone here done it or have any advice?  Reply to private message or to hamilgs at hiwaay dot net is fine.  Now back to your regularly scheduled program.--george

11 Dec.:  Received short reply from  Busch & Müller KG

 "it is possible to run two IQ fly with the hub dynamo, connected in series."

I've asked for a bit more info, will share if I get it.--george

16 Dec.:  B & M wrote back and replied "up to 20 km/h the two IQ Fly are as bright as only one headlight, also together with a dynamo rearlight.  It is possible to connect an additional switch (ON/OFF) parallel to one IQ Fly, to switch it off driving at lower speed."

My interpretation of his English (many times better than I could even attempt in German!) is that you can run two in series, but that below 20 kph (12 mph), you might be better off with just one IQ Fly running, and that could be accomplished with a "secondary shorting switch" in parallel with the "secondary" IQ fly (there is not a formal secondary IQ Fly from B & M, like there is for the E-6 halogen). The function of this switch is to bypass the second IQ Fly, and let the "primary" one run at full (low speed) power. This is what the "secondary E-6" halogen light does (the switch on it either lets current go through the E-6, or around it-"off" bypasses the current around the lamp). This also works fine when running a dyno taillight, as I do.

So all I need to do is buy & install another IQ Fly. Anyone have a used one they'd like to sell?--george
Tour Easy Recumbent
To recline is devine

border-rider

Re: Light testing - the results
« Reply #96 on: December 05, 2008, 10:51:37 pm »
Anyone here ... have any advice?

What peter white said

These modern LED lights need to have "exclusive" use of the dynamo.

Cudzoziemiec

  • Solar powered, tea fuelled cycle-wol
Re: Light testing - the results
« Reply #97 on: December 06, 2008, 11:21:05 am »
The graph on the first page is only labelled with the light names. I presume the blue line is the measurement in the centre of the beam and the red one the measurement 1m from that?

Apart from that... Wow!
"Pick a flower on Earth and you move the farthest star." Dirac.

Re: Light testing - the results
« Reply #98 on: December 06, 2008, 11:29:03 am »
Anyone here ... have any advice?

What peter white said

These modern LED lights need to have "exclusive" use of the dynamo.

Except they can run a rear light too. The IQ CYO has wires to connect up a rear.  I snipped them off.... hope that wasn't what caused the thing to fail (can't see why it would)

simonp

  • Omnomnomnipotent.
Re: Light testing - the results
« Reply #99 on: December 06, 2008, 11:42:14 am »
Did you read the bit in the manual that said not to do this?