Interior light control transistors - how hot is too hot?

[MisfitSELF]

I've been doing a lot of night flying lately so I've been running my lights a lot.  Last night I noticed that the panel was warm to hot around the co-pilot yoke and I traced it up to the two potentiometers (rheostats?)  that control the interior light level.  Turns out the two transistors on the back of that module were scorching hot to the touch.  Is this normal?

 

Bruce

[MikeOH]

Probably!

 It’s a pretty crude circuit where the transistors drop voltage across themselves (controlled by the pots) and dissipate the power. If extra panel lights have been added that would only make it worse.

As a test you can try turning the lights very very dim (reduce current to minimum) and see if they get any cooler. Max bright might be cooler as very little drop but current will be high.

[MisfitSELF]

1 hour ago, MikeOH said:

Probably!

 It’s a pretty crude circuit where the transistors drop voltage across themselves (controlled by the pots) and dissipate the power. If extra panel lights have been added that would only make it worse.

As a test you can try turning the lights very very dim (reduce current to minimum) and see if they get any cooler. Max bright might be cooler as very little drop but current will be high.

If I understand you right that may make sense, though I'm still scratching my head (I'm no EE) as to why the transistors are there in the first place.  I had them set to pretty much as dim as they would go -- so if they (the transistors) are current limiting then they are doing their "thing" and getting hot.  If I get a chance, I'll see if they cool off at max bright.

Bruce

[MikeOH]

2 hours ago, MisfitSELF said:

If I understand you right that may make sense, though I'm still scratching my head (I'm no EE) as to why the transistors are there in the first place.  I had them set to pretty much as dim as they would go -- so if they (the transistors) are current limiting then they are doing their "thing" and getting hot.  If I get a chance, I'll see if they cool off at max bright.

Bruce

@MisfitSELF Basically, they are there to 'drop' the voltage across them to control current to the lights: the lower the voltage on the bulbs, the less the current and, therefore, the lights get dimmer.  The 'problem' is that the transistors get hot doing that!  For example, if the transistor drops 10 Volts so that only 4 Volts gets to the bulbs (I'm assuming 14V buss voltage), and the total current for all the bulbs is 1 Amp, then that transistor would dissipate 10 Watts (10 Volts times 1 Amp)!  It'll get pretty hot at 10 Watts.

[N201MKTurbo]

1 hour ago, MikeOH said:

Basically, they are there to 'drop' the voltage across them to control current to the lights:

Just to be a smart ass, I believe they are wired as a emitter follower, so they actually control the voltage to the lights.

[DCarlton]

I'll just comment that mine get hot too.  Last time I was in the overhead, I noticed the wiring was blackened near the switch dimmers.  I'd like to modernize and upgrade my lighting to LED at some point.  If you're producing that much heat, you've gotta be wasting power.  No idea how much faster you would drain the battery at night if you had an alternator failure.  

[0TreeLemur]

Anyone ever replace those dimmer circuits in a J with a PWM unit?

[N201MKTurbo]

2 minutes ago, 0TreeLemur said:

Anyone ever replace those dimmer circuits in a J with a PWM unit?

I did. It works great. I changed out the instrument lights to LEDs and the analog dimmer didn’t work very well.

[Fly Boomer]

3 hours ago, N201MKTurbo said:

I did. It works great. I changed out the instrument lights to LEDs and the analog dimmer didn’t work very well.

Did you use an off-the-shelf dimmer?  I don't have LED panel lights yet, but ...

[N201MKTurbo]

10 minutes ago, Fly Boomer said:

Did you use an off-the-shelf dimmer?  I don't have LED panel lights yet, but ...

I used this one:

https://www.aircraftspruce.com/catalog/elpages/maxDimUnit.php?clickkey=7737
 

It is FAA approved, so it is 100 times better than the $3 dimmers on Amazon…

[MikeOH]

On 9/8/2023 at 4:55 PM, N201MKTurbo said:

Just to be a smart ass, I believe they are wired as an othemitter follower, so they actually control the voltage to the lights.

@N201MKTurbo

Both of our statements are correct.  Take a look at the circuit you provided and ask yourself, if the circuit is controlling the voltage across the load (light) then isn't it just as true that it is controlling the voltage across the transistor itself (fixed collector voltage)?  I focused on the voltage 'drop' across the transistor because the topic of discussion was how hot the transistors were getting, not the light bulb voltage.

It's a bit of chicken/egg whether we are 'controlling' current or voltage...what's important is we are controlling the power (lamp intensity) in the load (current times voltage).  I was speaking to load current as that same current flows through the transistors which multiplied by the voltage 'drop' across the transistors speaks to why they get hot.

In practice I believe the actual Mooney control box employs two high-side NPN transistors for some lights, and a PNP transistor to ground as a low-side variable current 'sink' (or voltage)for others; I think the panel lights are on one, and the caution/warning lights on the others....but I'm not certain of that.

[Ragsf15e]

And that’s why I got a mechanical engineering degree….

[N201MKTurbo]

3 hours ago, MikeOH said:

@N201MKTurbo

Both of our statements are correct.  Take a look at the circuit you provided and ask yourself, if the circuit is controlling the voltage across the load (light) then isn't it just as true that it is controlling the voltage across the transistor itself (fixed collector voltage)?  I focused on the voltage 'drop' across the transistor because the topic of discussion was how hot the transistors were getting, not the light bulb voltage.

It's a bit of chicken/egg whether we are 'controlling' current or voltage...what's important is we are controlling the power (lamp intensity) in the load (current times voltage).  I was speaking to load current as that same current flows through the transistors which multiplied by the voltage 'drop' across the transistors speaks to why they get hot.

In practice I believe the actual Mooney control box employs two high-side NPN transistors for some lights, and a PNP transistor to ground as a low-side variable current 'sink' (or voltage)for others; I think the panel lights are on one, and the caution/warning lights on the others....but I'm not certain of that.

An emitter follower will put the emitter voltage VBE below the base, no matter what current it takes. The circuit will put out the same voltage no matter what the load is. If there is one bulb or ten in parallel the brightness won’t change. It is the proper way to drive a parallel string of bulbs. A current controller would be appropriate for a series string of bulbs.

[MikeOH]

52 minutes ago, N201MKTurbo said:

An emitter follower will put the emitter voltage VBE below the base, no matter what current it takes. The circuit will put out the same voltage no matter what the load is. If there is one bulb or ten in parallel the brightness won’t change. It is the proper way to drive a parallel string of bulbs. A current controller would be appropriate for a series string of bulbs.

I guess I should have expected a dive into emitter followers vs. true current sources vis a vis the 'proper' way to drive light bulbs when I just tried to give a simple explanation of why the transistors get hot in a Mooney control box!

You did admit to being a smart ass, so there's that

[0TreeLemur]

What about instruments with internal lighting?  I'm pretty sure the light trays use incandescent bulbs.  Would a PWM dimmer work there?

[A64Pilot]

I am no EE, but a pwm dimmer on regular bulbs shouldn’t get hot and will work.

However not all LED’s are dimmable if you try some flicker and oddly some cheap ones dim just fine and some high quality expensive ones won’t.

I went around and around with this on the boat.

Instrument lights power is very minimal, but if you night fly you should have a back up light. I have both a handheld myself that clips nicely in the overhead light tray and my go to is an inexpensive headlight, inexpensive because you don’t want a bright one, a head light leaves your hands free of course and illuminates anywhere you look, even if you need a chart or something from the back seat.

[A64Pilot]

Oh, beware of LED’s,  many of them are noisy as hell with RF and really can interfere with radios and nav equipment etc. I think it may be their power supplies that are a part of the light, but again I’m no EE, by beware I’m noy saying they are all bad, but some are so be sure you pick “quiet ones”

This is a boat warning but aircraft use radios too. I’m surprised the FAA hasn’t issued something but I guess they believe everyone only uses FAA approved LED’s?

[N201MKTurbo]

28 minutes ago, A64Pilot said:

Oh, beware of LED’s,  many of them are noisy as hell with RF and really can interfere with radios and nav equipment etc. I think it may be their power supplies that are a part of the light, but again I’m no EE, by beware I’m noy saying they are all bad, but some are so be sure you pick “quiet ones”

This is a boat warning but aircraft use radios too. I’m surprised the FAA hasn’t issued something but I guess they believe everyone only uses FAA approved LED’s?

I was recently involved in the design of an LED controller for a microscope illumination system. The LED control chips have made a lot of progress lately, mostly due to improving the operation of the LED flash in your cell phone. It is a lot easier to build a noisy power supply and PWM controller than a quiet one. The advantage of optimizing a circuit for efficiency is it usually reduces the noise emissions. It takes power to make noise and that power should go towards powering the LED, not radiating noise.

[EricJ]

32 minutes ago, N201MKTurbo said:

I was recently involved in the design of an LED controller for a microscope illumination system. The LED control chips have made a lot of progress lately, mostly due to improving the operation of the LED flash in your cell phone. It is a lot easier to build a noisy power supply and PWM controller than a quiet one. The advantage of optimizing a circuit for efficiency is it usually reduces the noise emissions. It takes power to make noise and that power should go towards powering the LED, not radiating noise.

Also, as the power requirements diminish, so does the power in the noise, so that helps.

When making circuits to power radio frequency (rf) circuits in radios the power supply needs to be very clean so that it doesn't corrupt the signal, especially in a receiver.    So it used to be normal to have to use linear supplies (like the transistor control for the lighting circuit), even though they are inherently very inefficient (and make heat).   Switching regulators (like PWM) have been around forever since they're far more efficient, but were too noisy to use for applications that couldn't tolerate it.   That has been changing over the years and probably ten or fifteen years ago the little switching DC-DC converters that people put on circuit boards started to get clean enough that the good ones could even be used for sensitive rf circuits.   That trend has been continuing and really good, really efficient, and really quiet regulators, supplies, and DC-DC converters have become common.

Most circuits of all kinds these days run on fairly low voltage for both speed and efficiency, so they're all sensitive to noise sources.   When I was still doing this it had gotten to where the most siginificant noise sources coming from a system, including the supplies, were typically the system clocks that were driving the digital circuits.   It got to where the best way to pass FCC Part 15 (unintentional radiation) requirements was to intentionally dither the frequency of the clocks to spread the energy around and reduce it's intensity.

So getting quiet supplies and control circuits these days is readily doable, and even small LEDs that are dimmable sometimes have built-in PWM control.   If a device like an LED or USB supply or dimmer is making noise, it's only because some cost-cutting or other simplifying corners were cut, not because it's not possible or stretching the state of the art.

The comparatively ancient, crude linear transistor dimming circuit in question is beyond obsolete.    The heat they make is inherent in how they work, and it'll definitely seem foreign and wrong to somebody used to using modern stuff that just doesn't do that.    Worries about noise should not be an issue with modern circuits that aren't excessively cheap-ified, but that does become a challenge sometimes.   

[A64Pilot]

41 minutes ago, EricJ said:

The comparatively ancient, crude linear transistor dimming circuit in question is beyond obsolete.    The heat they make is inherent in how they work, and it'll definitely seem foreign and wrong to somebody used to using modern stuff that just doesn't do that.    Worries about noise should not be an issue with modern circuits that aren't excessively cheap-ified, but that does become a challenge sometimes.   

A whole lot of LED’s bought off of Ebay, Amazon etc are junk, cheap and noisy as hell.

I had a high freq radio on the boat used for over the horizon comms, same as used to be used on intercontinental aircraft back in the day, it was particularly susceptible to noise, I had to turn off the fridge to use it, the Adler Barbor variable frequency drive compressor power supplies made a noise similar to a space video game.

Again boat stuff, but Marine Beam tests their lights etc for rf output and their stuff is quiet, if someone is going to roll their own lights etc they are an excellent source for high quality RF quiet lights etc. Often boats come out with tech before aircraft, 406 ELT’s came out in boats before aircraft and are called EPIRBS, boat AIS is very similar to ADSB and pre-dates it by years etc.

https://store.marinebeam.com