Light bulbs

[rbp]

16 minutes ago, FlyingDude said:

cameras see the flicker and in the end don't capture the whole frame

aha, now we see what @FlyingDude's issue is.  

my suggestion is that instead of using constant pulse widths, encode the meaning of the sign as morse code, and have your cameras "read" the flash and decode it 

[EricJ]

27 minutes ago, FlyingDude said:

I've worked on autonomous driving, incabin and driver monitoring and machine vision with cameras trying to read dimmed LED road signs. It's a mess because cameras see the flicker and in the end don't capture the whole frame. So I know what I'm talking about 

You'll have trouble at night under normal AC lighting, too, as most traditional lighting technology (incandescent, fluorescent, etc.) powered by 60Hz AC turns off 120 times per second.   That flicker has always been an issue for sensors at night.   This is usually mitigated by having a longer sensor integration time in the camera.    It's a very solvable problem with good engineering.    Most auto headlamps these days, and definitely the LED ones, are pulsed for efficiency, so keeping your image sampling rate (usually the inverse of integration time) lower than the typical PWM rates is how it is usually handled.

That's entirely different than dimming LEDs for cabin lighting.  

23 minutes ago, FlyingDude said:

Which curve? Current vs brightness? Please share the curve.

You'll need to get those from the manufacturers of the LEDs that you're interested in, or make one yourself for a particular LED if you're concerned.   It's not too hard to do with a power supply and a detector.

 

[Ragsf15e]

You guys are awesome!

[N201MKTurbo]

But you are talking about imaging other peoples LED signs. You have no control of other people's stuff. I just built a PWM dimmer for a fluorescence microscope, so the camera will always see the PWM. I found LED dimmer chips that run at 1 MHz. With the long exposures we are using, you will never see it.

There are also current controllers for LED brightness control. They use inductive current chopping to control the current and account for the non-linearity of the LEDs. They don't go as dim as PWM controllers. There are LED controller chips that do both, current control and PWM. Current control for the brightest range and PWM for the dim end of the range.

[0TreeLemur]

16 minutes ago, N201MKTurbo said:

But you are talking about imaging other peoples LED signs. You have no control of other people's stuff. I just built a PWM dimmer for a fluorescence microscope, so the camera will always see the PWM. I found LED dimmer chips that run at 1 MHz. With the long exposures we are using, you will never see it.

There are also current controllers for LED brightness control. They use inductive current chopping to control the current and account for the non-linearity of the LEDs. They don't go as dim as PWM controllers. There are LED controller chips that do both, current control and PWM. Current control for the brightest range and PWM for the dim end of the range.

Cool.  It seems that using a PWM with a frequency well above top end hearing freq. (100 kHz) would be a good way to avoid noise potential for those folks without a high quality audio panel?

[EricJ]

1 minute ago, 0TreeLemur said:

Cool.  It seems that using a PWM with a frequency well above top end hearing freq. (100 kHz) would be a good way to avoid noise potential for those folks without a high quality audio panel?

And low enough to stay out of rf range for any radio equipment.    The really tiny ones, like what would be built into a lamp base, will use smaller coils and capacitors at higher frequencies, so they're motivated to operate the power controller and the PWM at frequencies as high as they can get away with, but many seem to be in the low MHz range these days, which is a happy place to stay out of most audio and rf.   USB power sockets these days seem to do the same.

[FlyingDude]

3 hours ago, EricJ said:

That's entirely different than dimming LEDs for cabin lighting.

I never mentioned in cabin lighting! In cabin monitoring and driver monitoring require shining certain amount of infrared light on people and watching them with cameras. The LED flicker that messes up with cameras is the adaptive road signs, mostly speed limits. Nothing to do with the soft light that helps you find you way into your vehicle! Nothing to do with the 60 or 50 Hz line frequency. 

3 hours ago, EricJ said:

You'll need to get those from the manufacturers of the LEDs that you're interested

I thought you were referring to some nonlinear curve as proof why I'm wrong and you're right. Yeah, I'm telling you. Photon emission is pretty linear to current draw which is controlled by the surrounding circuit once the Vf is reached.

2 hours ago, N201MKTurbo said:

They use inductive current chopping to control the current and account for the non-linearity of the LEDs. They don't go as dim as PWM controllers. There are LED controller chips that do both, current control and PWM. Current control for the brightest range and PWM for the dim end of the range.

Yes. You were the first one to say it, so I'm elaborating on your comment. Buck converters with constant current-variable voltage output that can be fine tuned with pwm control. You can go pretty dim with the buck converter but you risk having intermittent conduction, so it's not a great idea.

Guys, this was fun. Thanks. I am leaving this thread. Talk to you at the next thread 

 

[N201MKTurbo]

1 minute ago, FlyingDude said:

I never mentioned in cabin lighting! In cabin monitoring and driver monitoring require shining certain amount of infrared light on people and watching them with cameras. The LED flicker that messes up with cameras is the adaptive road signs, mostly speed limits. Nothing to do with the soft light that helps you find you way into your vehicle! Nothing to do with the 60 or 50 Hz line frequency. 

I thought you were referring to some nonlinear curve as proof why I'm wrong and you're right. Yeah, I'm telling you. Photon emission is pretty linear to current draw which is controlled by the surrounding circuit once the Vf is reached.

Yes. You were the first one to say it, so I'm elaborating on your comment. Buck converters with constant current-variable voltage output that can be fine tuned with pwm control. You can go pretty dim with the buck converter but you risk having intermittent conduction, so it's not a great idea.

Guys, this was fun. Thanks. I am leaving this thread. Talk to you at the next thread 

 

There are many dedicated LED driver chips out there. No use rolling your own with a buck converter.

[FlyingDude]

Before I truly leave the thread, let me share this. 

https://www.aero-lites.com/product-page/ge-89-led-replacement-12vdc-dimmable

 

[Hank]

Ya'll have completely lost me! Can you speak English and tell me which LED bulb fits in the overhead lamp. Dimmable would be nice, since every light in the cabin is on the same dimmer and I can't install a second dimmer just for the overheads.

I'm just a simple mechanical engineer, and can figure out most things, but electrons are too small to see and trying to hold them hurts . . . . .

[rbp]

1 hour ago, FlyingDude said:

Thanks. I am leaving this thread.

As they say, this isn’t an airport: you don’t need to announce your departure

[Andy95W]

1 hour ago, FlyingDude said:

Before I truly leave the thread, let me share this. 

https://www.aero-lites.com/product-page/ge-89-led-replacement-12vdc-dimmable

 

@FlyingDude- good looking possible solution, plus inexpensive. Please let us know how well they work!

[AdamJD]

These worked great for the torpedo lights in my 67F.  They dim fine for my needs.  Not a large range between dim and bright, but again, they work for my needs. -> https://www.superbrightleds.com/led-light-bulbs-universal-finder/miniature-and-subminiature-bulbs/67-led-light-bulb-12-smd-led-tower-ba15s-base+color-Cool~6500K+packamt-2-Pack

I put this in my dome light.  Also works well.  -> https://www.superbrightleds.com/led-light-bulbs-universal-finder/miniature-and-subminiature-bulbs/67-led-light-bulb-12-smd-led-tower-ba15s-base+color-Natural~4000K+packamt-Single

 

[EricJ]

2 hours ago, FlyingDude said:

Before I truly leave the thread, let me share this. 

https://www.aero-lites.com/product-page/ge-89-led-replacement-12vdc-dimmable

 

Yes, that's an example of an LED with the PWM controller built in.   The description makes clear the difference in ordering an LED with or without the dimmable internal PWM controller.

[EricJ]

3 hours ago, FlyingDude said:

I never mentioned in cabin lighting! In cabin monitoring and driver monitoring require shining certain amount of infrared light on people and watching them with cameras. The LED flicker that messes up with cameras is the adaptive road signs, mostly speed limits. Nothing to do with the soft light that helps you find you way into your vehicle! Nothing to do with the 60 or 50 Hz line frequency. 

You started this thread with a question about LED bulbs for in-cabin lighting.

3 hours ago, FlyingDude said:

I thought you were referring to some nonlinear curve as proof why I'm wrong and you're right. Yeah, I'm telling you. Photon emission is pretty linear to current draw which is controlled by the surrounding circuit once the Vf is reached.

There's been a lot of experience, some mentioned in this thread, of people plugging regular LED bulbs into dimmed circuits and finding that the dimming curve of an LED is highly unsuitable for use in that application.   This is why PWM controllers are so widely used with LEDs, because they pretty much suck otherwise due to their sadly not-linear control curve in a circuit made for dimming incandescent lights.   Try it and you'll see.    It is possible to linearize the brightness control, but then you wind up with a very electrically inefficient solution, which is the other aspect that the PWM controller solves.

3 hours ago, FlyingDude said:

Yes. You were the first one to say it, so I'm elaborating on your comment. Buck converters with constant current-variable voltage output that can be fine tuned with pwm control. You can go pretty dim with the buck converter but you risk having intermittent conduction, so it's not a great idea.

I don't think anybody suggested that as a solution.   That's not what a PWM controller does in this application.   

3 hours ago, FlyingDude said:

Guys, this was fun. Thanks. I am leaving this thread. Talk to you at the next thread 

 

 

[PT20J]

All this reminds me of one day when I was in the light bulb aisle of a hardware store and a bewildered elderly woman exclaimed with a note of resignation, “All I want is a sixty watt light bulb.”

[FlyingDude]

8 hours ago, AdamJD said:

These worked great for the torpedo lights in my 67F.  They dim fine for my needs.  Not a large range between dim and bright, but again, they work for my needs. -> https://www.superbrightleds.com/led-light-bulbs-universal-finder/miniature-and-subminiature-bulbs/67-led-light-bulb-12-smd-led-tower-ba15s-base+color-Cool~6500K+packamt-2-Pack

I put this in my dome light.  Also works well.  -> https://www.superbrightleds.com/led-light-bulbs-universal-finder/miniature-and-subminiature-bulbs/67-led-light-bulb-12-smd-led-tower-ba15s-base+color-Natural~4000K+packamt-Single

 

I had used superbrightleds products on other non-work related projects and they work pretty well. Here I went with Aero-lites merely out of marketing reasons: they mention aircraft on their webpage whereas the other guys don't. Aero-lites also mentions that they are not PMA, so it's truly a marketing gimmick but tilted the arrow as I was on the fence between the two. Installation is up to AP discretion and he said OK.  I wouldn't be surprised if they sourced from the same place. I'll give pireps on this supplier down the road. 

[Pinecone]

14 hours ago, EricJ said:

Yes, because you want to drive them at just the right minimum current for full brightness with every PWM cycle.

The only thing you need to avoid flicker is to have the PWM period higher than human flicker sensitivity, so better than 30 Hz or so, which is very easy to do.  Varying the duty cycle from 0% to 100% above that frequency gives you full dimming range with no flicker. 

Many people can perceive 60 Hz flicker. 

In the days of CRT monitors, I "fixed" a lot of issues with users and eye strain by bumping the refresh rate to 72Hz or higher.

[Pinecone]

14 hours ago, EricJ said:

You'll have trouble at night under normal AC lighting, too, as most traditional lighting technology (incandescent, fluorescent, etc.) powered by 60Hz AC turns off 120 times per second.   

In candescent don't normally have the issue as much as the filament can't cool fast enough for the light output to decrease a lot.

[A64Pilot]

9 hours ago, Hank said:

Ya'll have completely lost me! Can you speak English and tell me which LED bulb fits in the overhead lamp. Dimmable would be nice, since every light in the cabin is on the same dimmer and I can't install a second dimmer just for the overheads.

I'm just a simple mechanical engineer, and can figure out most things, but electrons are too small to see and trying to hold them hurts . . . . .

Methinks they certainly know what they are talking about but is building a nuclear powered wrist watch worth it to get rid of a few small bulbs that last decades due to lack of use and draw an insignificant amount of power?

[A64Pilot]

11 minutes ago, Pinecone said:

In candescent don't normally have the issue as much as the filament can't cool fast enough for the light output to decrease a lot.

I put a flasher on the third brake light on my Miata, I only pulse it for about 1 sec then it goes solid, but I had to go LED for exactly the reason you mention. 

I hate those the continuously flash their third brake light, it’s obnoxious and defeats the purpose.

I tried dimming the cockpit lights on my boat with a PWM, once they dimmed much they would flicker, so I ordered a very high frequency PWM, still flickered, so I tried several resistors in line, made an effective little heater, but LED’s still flickered, so they ended up as very bright engine compt lights. Power draw on a sailboat is a big deal so resistors for dimming is not really what your after.

I found some LED strips that were meant as daytime running lights for cars, they would dim down to almost no light for some reason and not flicker, but the salt would get to them and they wouldn’t last more than a few months.

 

[rbp]

6 minutes ago, A64Pilot said:

Methinks they certainly know what they are talking about but is building a nuclear powered wrist watch worth it to get rid of a few small bulbs that last decades due to lack of use and draw an insignificant amount of power?

this is exactly why I didn't change my panel lights. and also because everything is backlit. the only things that need light are mechanical switches

[A64Pilot]

4 minutes ago, rbp said:

this is exactly why I didn't change my panel lights. and also because everything is backlit. the only things that need light are mechanical switches

Still a backup light is not a bad idea, just in case the internal lights go out. Military we flew with “lip lights” little green LED mounted to the mic so that you pushed your lip out and that lit the LED, worked well, but it’s use was primarily not to illuminate yourself so you didn’t get shot, but a good back up in the civilian world.

Now that everything is glass and you can’t turn those off, we had “bat wings” folding accordion blackout curtains that covered the side windows so you weren’t seen from the ground, but of course you couldn’t see out with your naked eye either.

No more lip lights, I guess the newer pilots never thought about flying with all cockpit lights off.

When I flew nights a lot I carried a tiny LED headlight, performed the same function as the lip light in that every where you looked was illuminated. It was just a back-up of course, but better than a flashlight as it left your hands free.

If you fly at night try turning off the cockpit lights, after a few minutes you’ll be surprised how much better you can see

[rbp]

9 minutes ago, A64Pilot said:

 

If you fly at night try turning off the cockpit lights, after a few minutes you’ll be surprised how much better you can see

my primary instructor made me do all 10 night landings/takeoffs without lights

and believe it or not you can get a https://www.aircraftspruce.com/catalog/pspages/bitealite.php

[FlyingDude]

1 hour ago, A64Pilot said:

o I tried several resistors in line, made an effective little heater, but LED’s still flickered, so they ended up as very bright engine compt lights

put an RC filter upstream of your LED strip or a diode/inductor to pretty much turn it into a quasi-buck converter.  You'll still get some wavering intensity, but it'll be much less invasive than the 100-0% flickering...

 

55 minutes ago, rbp said:

my primary instructor made me do all 10 night landings/takeoffs without lights

that's part of the training actually... 

 

55 minutes ago, rbp said:

and believe it or not you can get a https://www.aircraftspruce.com/catalog/pspages/bitealite.php

That must be a joke!  Walking around with metal rod in your mouth, with all the saliva drooling around.  I made a joke like this when my PPL instructor had said "for night flights, carry a flashlight with D-cell batteries"... 

I have something like this on the plane : https://www.grainger.com/product/492T31?gucid=N:N:PS:Paid:GGL:CSM-2295:4P7A1P:20501231&gclid=CjwKCAiAleOeBhBdEiwAfgmXfys8nqXJ0hZ1sb9FXRQv1icu0suGNM_-OWij3UTrULL6uuLR4gCqohoCdDUQAvD_BwE&gclsrc=aw.ds

Very helpful than a flashlight during preflight or fueling.

1 hour ago, rbp said:

because everything is backlit.

I don't have full-glass.  2x G5s and steamgauge.  They don't make a nu-lite for the 6-pack Mooney instrument cluster... I point the overhead torpedo light towards that.