Every time i operate my upper split air conditioner in my room of my hostel with my remote controls, and one or both my CFLs and/or the HF 21W T5 fixture turned on (Especially if the lamp is warming up shortly after turning on), my air conditioner don't resopnds to the IR signals from the remote control and i have to point the remote control to the air conditioner and even to become closer to the air conditioner (Usually happens to me with the HF 21W T5 ceiling fixture).
This also happens to me with the air conditioner and its remote control in my room at my father home.
The SBMV lamps in my father room don't interference with the remote control of my air conditioner there.
Why fluorescent lamps interferences with IR remote controls (Especially shortly after turning on when the lamps warming up)?

I think its not so much about the IR alone (the SBMV makes a lot more IR), but about working frequency

The remotes usually work at around 40 KHZ

I dont see how the lamp would emit IR that flickers at 40 KHz (the electrodes are hot and emit IR but they cant cool down 40000 times / sec), so either :

The AC sensor is confused by 40 KHZ flicker from the mercury discharge in the green/blue/UV region (isnt it filtered for IR only ? well if it is filtered with woods-glass like filter then it can still get UV, but i dont know whether the diode is sensitive to UV at all)

The interference is entirely electrical (ballast and discharge in the lamp make 40 KHZ EMI, circuit in the AC picks it up) and has nothing to do with the actual light source

Mercury does have two not insignificant spectral lines in the IR region, 794 and 1014 nm, maybe those are causing the trouble. They are not a large distance away from the wavelengths used by IR remotes (around 900 nm). Interference like this is not uncommon with electronic ballasts.

The remotes usually work at around 40 KHZ

Why that the IR LEDs of the remote control would operate at 40 KHZ?
HF ballasts operates at 20 KHZ, the max frequency in which the efficiency of fluorescent lamps increases, and above it, the efficiency of the fluorescent lamp remains the same, while the ballast gets more complex and expensive.

The AC sensor is confused by 40 KHZ flicker from the mercury discharge in the green/blue/UV region (isnt it filtered for IR only ? well if it is filtered with woods-glass like filter then it can still get UV, but i dont know whether the diode is sensitive to UV at all)

The IR sensor of my Electra air conditioner in my room of my hostel, have a dark glass which lets only the near IR and IR wavelength to reach it.
Why that the short wavelengths will affect the sensor?

Mercury does have two not insignificant spectral lines in the IR region, 794 and 1014 nm, maybe those are causing the trouble. They are not a large distance away from the wavelengths used by IR remotes (around 900 nm). Interference like this is not uncommon with electronic ballasts.

SBMV lamps have also these IR spectral lines, as they are mercury discharge lamps. Probably the mercury IR lines don't interference with the sensor, since these IR lines are emitted with double the mains frequency (100hz), and this frequency don't interference with the IR sensor.

I have noted that in many TV and DVB transmitter remote controls, the IR LEDs operates at very low frequency (When pointing them to the digital camera, few tens of Hz), which is visible to the eyes (The frequency when viewing through the camera, not directly with the unaided eyes).
This is probably to prevent fluorescent lamps from being interference with the devices IR sensors.

The circuit in the remote modulates the data (code of the pressed key) on a carrier of about 40 KHZ, and this signal is going to the LED

Since the lamp current is probably not exactly sine, it would seem as emmisions at 20, 40, 60.... KHZ - whatever it takes to sum up to the actual waveform. The sensor would then pick up the 40KHZ part

If the interference is electrical (and not optical), the 20 KHZ of the lamp can directly interfere with the AC electronics even if they are not meant to accept 20 KHZ



Woods glass can pass UV and IR. I dont know what the plastic filter in the AC, as well as the chip case is made of (probably that kind of dark transparent plastic), but if it is similar to woods glass then it would pass UV. If so, it would matter on whether the chip itself is sensitive to UV as well as IR

Also F36t8 said that mercury emits IR, in this case if the chip is sensitive to IR in that bands then it would react to them




The flashing of the LED which you see is repeated transmission of the data stream. For example :

40 KHZ :
10101010101010101010101010101010

Data : 1101, which is at lower frequency, so looks more like :
11111111111111110000000011111111

Modulated (AND) :
10101010101010100000000010101010

And repeating as long as you hold the key down :
10101010101010100000000010101010000000000000000000000000000000001010101010101010000000001010101000000000000000000000000000000000.....

So what you see is only :
|<----------- light ----------->|<---------- no light ---------->|<----------- light ----------->|<---------- no light ---------->|....

And what the sensor in the AC (after the demodulation) sees is :
11111111111111110000000011111111000000000000000000000000000000001111111111111111000000001111111100000000000000000000000000000000......

But when the lamp is interfering, all it sees (or the circuit thinks it sees) is :
11111111111111111111111111111111.....

HF ballasts operates at 20 KHZ, the max frequency in which the efficiency of fluorescent lamps increases, and above it, the efficiency of the fluorescent lamp remains the same, while the ballast gets more complex and expensive.

Higher frequency mean all passive power components could be made smaller (lower value capacitors, smaller inductors,...), so actually cheaper. The tendency change at about 100kHz, where the problems with high frequency switching devices control start to bring complications into the ballast design.
Ballasts are usually made to start at about 100..60kHz for preheat and then sweep down to about 25..40kHz for the main operation, while the frequency slightly changes, as the system warm up (either result of thermal drift of timing components, or as the lamp parameter change as it heat up, so the ballast compensate it by adopting the frequency). As remotes are made on 30..55kHz carrier frequency (integrated IR receivers for 32, 36, 42 and 56kHz carrier are usually offered), it is well within the range of ballast operation (the central section of lamp flashes mainly at the double of the operating frequency, but the electrode only at the base frequency). And the phosphor is usually fast enough to propagate this flashing, so as it emit the IR as by-product (some electron transitions fall there), it may be the disturbing IR source as well.
When the discharge operate at low frequency (mains,...), such signal does not pass trough the band-pass filter in the receiver, so does nor disturb the communication.

@Ash: The notation in the RC is light (or carrier) present = "0", no light (or no carrier) = "1", but that does not change the principal workings.

Is the phosphor really fast enough to follow the mercury discharge flicker at 40KHZ ? What is the "half life" of it ?

And can it be result of EMI from the CFL + ballast interfering with the circuit behind the sensor, and not the light interferng with the sensor ?

The 40kHz does not spread electrically so easily, mainly behind the 5V regulator in the receiver.

Most materials respond in big extend in few ns, even when they have afterglow components lasting for minutes.
So they could pass quite significant 40kHz component.
And the direct mercury IR radiation does add to the problem as well.

What may help is to shield the IR receiver from direct light from these fluorescents, the indirect would be attenuated, so the transmitter signal would be significantly stronger.

A little history here - IR remote systems were developed from the original ultrasonic remote control systems, as IR LED technology was improved over time to become cheap and robust. The original system used a audio transducer that was resonant at a frequency between 36 to 40 kHz, depending on size and power rating of the transducer. They were originally developed for burglar alarm use, to detect motion in a room, and were made in large numbers. The transmitter and receiver are essentially identical, being a piezoelectric crystal with metallised faces, attached to a small rigid diaphragm to couple to the air. They were sized to be a mechanical resonant system at the frequency somewhere in the ultrasonic range, so that you could use the system at high power without an audible ( to humans at least) sound from them. Used in matched pairs they were capable of quite long range operation, more than enough to cover remote control operation.

The first IR systems used the existing ultrasonic transmitter designs, as the early integrated circuits were available and cheap to use, and did not pose much trouble to change the method of transmission from sound at 40kHz to light modulated at 40kHz. The main driver was cost, as the ultrasonic transducers were hand made, hand tested, and hand sorted into gain groups and turned into non interchangeable matched sets, and were a lot more expensive than a diode that could be mass produced with a reproducible output. They also were fragile, and prone to failure in humid environments and did not survive rough handling. This was a problem in a remote control that could be dropped or exposed to wet hands, as well it needed a high voltage to drive the transmitter, some needed up to 90VAC to generate significant power. As well they were large devices, easily an inch in diameter and as long. Not easy to fit into a slim case, and you needed a 9V battery to drive them. The IR diode was small, and only needed a 3V supply from 2 cells to operate. A high peak current was used, but with a very low duty cycle, leading to a large reduction in power consumption, to the point where battery life is at least equal to shelf life.

At the receiver end the photodiode is used either in avalanche or in short circuit mode. Most are operated in short circuit, where the current developed by the incident light ( over a large range depending if there is a IR filter in the encapsulation or not) is converted to a voltage and then amplified by a large amount. Then it is passed through a band pass filter to only pass the desired range of frequencies ( 36 or 40kHz) and then this signal is demodulated ( either by a simple diode peak detector or a more complex method) to give the original modulation. The receiver is very insensitive to slow changes in light level, and to low frequency light changed, like sunlight or an incandescent lamp running on the mains and generating a small IR signal at 100/120Hz. The CFL ballast generates a output that can be within the passband, and this will be detected as well, so the remote signal will not be easy to decode, or will generate spurious signals from the remote receiver. A ballast operating at between 18 to 20 kHz can easily give a signal that will overload the IR receiver, especially if the photodiode is operating at a point where the amplifier is running out of headroom due to being exposed to high IR light levels. Most commonly nowadays the entire receive system is in a small black 3 pin assembly, requiring only 5V power and giving the demodulated signal out to the decoding circuitry.

I made a distance meter at school which was based on ultrasonic detector (send pulse and receive it, counting the time)

I used a transmitter which is about 13x11 mm in size and has shape that resembles a capacitor mic. And i just drived it with 12V AC (made by connecting it to 555 timer and inverting transistors to gt 24Vp-p). What was the requirement for high voltage for ?

Also the transmitter/receiver were available to buy in bulk in the local equivalent of Digikey without any matched pairs, so what was the matched pairs thing for ?

The matching was necessary, whenever narrow band operation was desired, as the main selectivity was in the receiver piezo itself.
And narrow band operation is necessary, if you want to avoid the system being disturbed by foreign signals, on TV mainly the ~15.6kHz of horizontal deflection and it's harmonics. Without this selectivity the ~47kHz (3'rd harmonics) would overload the receiver and you would not be able to extract any data.

The receiver selectivity mean, then the transmitter band (so carrier and sidebands) was necessary to match into the quite narrow and of the receiver. As at that time the manufacturing was not accurate, the resonance frequency varied a lot. There is possibility to fine tune the resonator, but such step was quite costly, so it was better to be avoided. So another approach was chosen: Select two units, what happen to have the resonance frequency close enough to each other and use self oscillating power stage, what use the transducer itself as the frequency control element.
Note, then the transmitter side used different design then the receiver (it had to adopt the different power levels, as the system is a bit nonlinear), but the sets were sorted so, the matched pair was able to work together.
The problem with humidity/dirt was, then it shifted the resonance frequency, so the transmitter no longer operated in match with the receiver.

The distance meter is a bit another task: You have to measure the time quite accurately, so you need fast response path. And fast response mean wide frequency band, so for that task you need wideband transducers. If the bandwidth is so wide, the center frequency tolerance is of no problem anymore (it is way lower then the bandwidth), so even units at opposite sides of the tolerance area are able to work together, so you do not have to frequency match the components or so.

The bulglar alarms work in a bit different way: They rely on the Doppler effect to shift the frequency of the reflected signal.
The signal from the receiver is mixed with the transmitter one, producing low frequency component what's frequency is the frequency difference between transmitter and receiver signal. So when nothing moves in the detector range, the received frequency matches the transmitted one, so the result is only the zero frequency (DC) mixer output. Once someone enter the room, his movement cause frequency shifts, so low, but nonzero frequency component appear on the mixer output, what is detected by the following circuitry (LF bandpass filter with a detector) and an alarm signal is triggered.
So this is again a narrow frequency band system, what need the transducer's resonant frequency to match.

I had a similar problems with the remote control of my satellite receiver. After installing a new Panasonic LCD flatscreen (before I had a Panasonic Plasma) suddenly the remote control of my receiver didn't work correct. Sometimes it worked, sometimes not. At first I thought the HDMI disturbs the receiver but also using analog connection didn't solve it.

I was not able to find the problem and quite nerved so I switched on the main light of the room and suddenly the remote worked normal. After long try and error I found out, that the backlight (ccfl) was dynamically dimmed by the TV and on a specific level it had exact the frequency of the remote. As I switched on the light the sensor of the TV increased the back light and the remote worked.

The only solution I could find was that for the sat receiver the backlight is fixed at full power (less good black level) using "True cinema" mode of the TV.