Lamp Brightness with Varying AC: Understanding the Consistent Output

[AbuBekr]

AC is varying.So why does the brightness not vary?

 

[PeterDonis]

How do you know the lamp's brightness does not vary?

 

[AbuBekr]

It looks constant to me

 

[PeterDonis]

It looks constant to me

How does that prove that it actually is constant? What if it were fluctuating faster than you could see?

 

[AbuBekr]

that is possible too

 

[PeterDonis]

that is possible too

Does that make you want to reconsider the question you originally asked?

 

[mathman]

AC is usually 50 or 60 cycles per second. This is much too fast for ordinary vision.

 

[Nugatory]

Many years ago, I was hanging out with an EE friend of mine, and this topic came up.

Being total hopeless nerds, nothing would do except that we pull a photovoltaic cell out of the clutter on the bench and hook it up to a scope, then expose it to a light powered by 120V 60 Hz (that US line voltage) AC, and another powered by 48v DC. No question, the AC light intensity was fluctuating... fine clear 60 HZ signal. The amplitude never fell to zero and the shape of the signal was not sinusoidal - instead there was a fair amount of hysteresis which we attributed to the temperature of the filament not exactly tracking the voltage at the filament.

 

[William White]

nice image of city lights
https://simple.wikipedia.org/wiki/Alternating_current

50 or 60 Hz is slow enough to see if the brightness varies considerably (try flicking between a bright white and black screen at 50Hz) but incandescent lamp filaments cannot cool quickly enough for our eyes to notice any dip in brightness (the voltage supply is gradually decreasing then gradually increasing rather than blinking on and off - its only 0V for an instant)However, I cannot stand the stroboscopic effect of fluorescent lamps (migraine) and have had to resort to buying expensive daylight incandescent bulbs off the internet (almost impossible to buy in shops nowadays because of energy reduction laws)

 

[PeterDonis]

there was a fair amount of hysteresis which we attributed to the temperature of the filament not exactly tracking the voltage at the filament.

It's actually the current that's the key; the current is what drives the temperature and hence the light output. In an ideal purely resistive AC circuit, the current would be in phase with the voltage; but a real light bulb filament is evidently not very close to an ideal pure resistor.

 

[William White]

I think the filament of a light bulb IS very close to being a pure resistor

It's a coil, but the inductance at 50Hz is neglibigle.
http://www.66pacific.com/calculators/coil_calc.aspx

The impedance of the circuit obviously includes the circuit all the way back to the transformer (i.e the PF you are getting at your cutout is going to be 0.9ish)I've never connected a single lamp to a pure 50Hz AC supply to measure the impedance of the circuit. I'd guess the PF is as close to 1 as you can get.

 

[sophiecentaur]

The flicker from a mains voltage 50Hz light bulb can be sufficient to give a strobing effect with rotating machinery and to make it appear stationary. So much so that it is common practice to use a low voltage (say 24V AC) powered lamp on a drill or lathe. The reason that this solves the problem is that low voltage filaments are more massive and their temperature doesn't vary much over the cycle - and the flicker is much less.

The change of resistance of a tungsten filament when cold can be a factor of ten, which can produce strange V/I characteristics. The thermal lag and changes in resistance at lowish frequencies can produce a change in phase between V and I waveforms. I once did a very instructive experiment with an AF signal generator and a small light bulb. Putting measured V and I values on the X and Y inputs to an oscilloscope would give an S shaped trace at low frequency, which would develop a hysteresis 'hole' as the frequency was raised, due to this thermal effect. Easy to do with simple equipment and worth trying.

 

[Nugatory]

It's actually the current that's the key; the current is what drives the temperature and hence the light output. In an ideal purely resistive AC circuit, the current would be in phase with the voltage; but a real light bulb filament is evidently not very close to an ideal pure resistor.

The current drives the power output and hence the change in temperature; but the light intensity at any moment will depend on the temperature at that moment, while the power output drives the rate of change of that temperature. A further complication is that the resistance of the filament varies by more than a factor of ten across the operating temperature range, so even though an incandescent lamp is fairly close to a purely resistive AC circuit (impedance effects are negligible and the power factor is near as no never mind 1.0) a change in voltage causes an out-of-phase change in current and power output as the temperature climbs. Together, these effects make for a very complicated relationship between the applied voltage and the light output.

 

[PeterDonis]

the light intensity at any moment will depend on the temperature at that moment, while the power output drives the rate of change of that temperature.

Ah, right.

the resistance of the filament varies by more than a factor of ten across the operating temperature range

I hadn't realized it varied by that much. Oof.

 

[sophiecentaur]

I think the filament of a light bulb IS very close to being a pure resistor

Apart from the negligible reactance, due to the coil, it is 'resistive'. But, despite being a metal (tungsten), it is not operated under 'ohmic' conditions (i.e. not constant temperature) and its VI characteristic is not linear. It is clearly not 'reactive' in the conventional sense but, at around a critical range of supply frequencies, the I will measurably lead the V when increasing and decreasing. This has a 'Power Factor' look about it and it certainly could have repercussions in the losses in supply resistance - which is where PF is relevant to Power Engineers.
This is hardly a relevant consideration for real cases because the bulb would need to be operating at frequencies well below normal 'mains' supplies but it is of interest (to me at least).

 

[Redbelly98]

Many years ago, I was hanging out with an EE friend of mine, and this topic came up.

Being total hopeless nerds, nothing would do except that we pull a photovoltaic cell out of the clutter on the bench and hook it up to a scope, then expose it to a light powered by 120V 60 Hz (that US line voltage) AC, and another powered by 48v DC. No question, the AC light intensity was fluctuating... fine clear 60 HZ signal. The amplitude never fell to zero and the shape of the signal was not sinusoidal - instead there was a fair amount of hysteresis which we attributed to the temperature of the filament not exactly tracking the voltage at the filament.

{Bolding added for emphasis}
The light intensity will vary at double the mains frequency, reaching a peak (near) when the voltage is at +/- the full amplitude, or twice per mains cycle ... so, it should have been at 120 Hz. And when the voltage goes through a zero-crossing, the filament (if it is a filament bulb) does not have time to cool down very much, so you get a relatively small-amplitude intensity variation superimposed on a relative large DC intensity.

At least, that is what I have seen when looking at a filament lamp's brightness on an oscilloscope.

 

[Nugatory]

{Bolding added for emphasis}
The light intensity will vary at double the mains frequency, reaching a peak (near) when the voltage is at +/- the full amplitude, or twice per mains cycle ... so, it should have been at 120 Hz.

Dammit! You're right. Thanks.

It was thirty years ago, and the takeaway memory is "We saw the frequency we predicted given a 60Hz input" not "We saw 60Hz"... somehow over the years I let the one slide into the other until you reminded me. We even discussed that factor of two at the time.

You're also right about the small signal on top of the larger DC offset, but of course with a scope it's easy to dial the DC offset away. That's what we did, as we were trying to see something interesting in the AC signal.

 

[tech99]

Dammit! You're right. Thanks.

It was thirty years ago, and the takeaway memory is "We saw the frequency we predicted given a 60Hz input" not "We saw 60Hz"... somehow over the years I let the one slide into the other until you reminded me. We even discussed that factor of two at the time.

You're also right about the small signal on top of the larger DC offset, but of course with a scope it's easy to dial the DC offset away. That's what we did, as we were trying to see something interesting in the AC signal.

A small flash lamp bulb can carry speech and music quite successfully. The audio signal is impressed on a DC current which keeps the lamp at about half brightness. Using a photo transistor or photo diode with focusing lenses, the sound can be heard across the lab.

 

[256bits]

A small flash lamp bulb can carry speech and music quite successfully. The audio signal is impressed on a DC current which keeps the lamp at about half brightness. Using a photo transistor or photo diode with focusing lenses, the sound can be heard across the lab.

Now, that is interesting. I would have thought there would have been too much "noise" ( wave modification, to get the original signal back.

while this is not part of the original discussion, I found this intriguing,
Yesterday, on the radio, the conductor/director Kent Nagano, of the Montreal Symphony Orchestra, interviewed by CJAD 800 AM, let it be aware that while answering a question of reading music, let it be known that he can hear the AC oscillations directly from a light bulb. I do hear the buzz when they are under stress, but not when they are in normal operating conditions, at least if he was talking about a say 60W home light bulb, but he did not elaborate.

 

[CWatters]

The light intensity will vary at double the mains frequency, reaching a peak (near) when the voltage is at +/- the full amplitude, or twice per mains cycle ... so, it should have been at 120 Hz. And when the voltage goes through a zero-crossing, the filament (if it is a filament bulb) does not have time to cool down very much, so you get a relatively small-amplitude intensity variation superimposed on a relative large DC intensity.

I have some 50Hz LED lights that visibly flicker. I'm wondering if the reason is the lack of an equivalent to the thermal mass of the filament or is it just because they do/don't have full wave rectifiers in them? Guess I'll have to take one apart to find out.

If you look directly at the LED the flicker isn't obvious but if you look away so that they are visible in your peripheral vision (which is more sensitive to motion)they flicker so badly that I've had to remove them.

 

[sophiecentaur]

they do/don't have full wave rectifiers in them?

That would / could still produce a ripple in their light output. You would need to have reservoir capacitors and / or voltage regulators to reduce the AC ripple on a DC supply voltage to a suitably low level. A cheaper and more efficient solution could be to use tiny inverters to supply the LEDs with AC with a frequency of several hundred Hz.

 

[Redbelly98]

I have some 50Hz LED lights that visibly flicker. I'm wondering if the reason is the lack of an equivalent to the thermal mass of the filament or is it just because they do/don't have full wave rectifiers in them? Guess I'll have to take one apart to find out.

If you look directly at the LED the flicker isn't obvious but if you look away so that they are visible in your peripheral vision (which is more sensitive to motion)they flicker so badly that I've had to remove them.

It would be good if you had a photodiode to check the light output. That would answer both questions, whether the light output goes to (or close to) zero, and whether it's at mains frequency or 2x mains frequency.

(I've heard that an LED can be used as a photodiode as well, but have never used one this way myself.)