Photodiode noise cancelling, how to?

Hello, I am building a simple laser or led communicator with LM386 and photodiode.
I have tested the transmitter and receiver with success.

The problem is that the photodiode receives fluorescent lights fluctuations as well.
I was thinking of using two phododiodes in a noise canceling scheme like this **broken link removed** to avoid using lightproof cabinets (for cost reasons and easiness of construction)

I am thinking of facing one photodiode to the laser (which will also receive ambient light) and the other to the ambient light only (not facing the laser).

Do you think this scheme will work?
I worry about the saturation of the photodiodes from the ambient light, so that they cannot distinguish the fluctuation of the wanted signal (AM laser diode modulation)

neazoi said:

Hello, I am building a simple laser or led communicator with LM386 and photodiode.
I have tested the transmitter and receiver with success.

The problem is that the photodiode receives fluorescent lights fluctuations as well.
I was thinking of using two phododiodes in a noise canceling scheme like this **broken link removed** to avoid using lightproof cabinets (for cost reasons and easiness of construction)

I am thinking of facing one photodiode to the laser (which will also receive ambient light) and the other to the ambient light only (not facing the laser).

Do you think this scheme will work?
I worry about the saturation of the photodiodes from the ambient light, so that they cannot distinguish the fluctuation of the wanted signal (AM laser diode modulation)

Click to expand...

What you need is to make only YOUR light reach the photodiode. Use a telescope-like tube to get a protected line of sight, and use a filter to allow only the transmitter spectrum to get there.

Can you use a light filter in front of the sensor that only passes the LED or laser light wavelength (a deep red filter is typical for IR)?

Also you could use a high-pass analog filter to minimize the signal from the light fluctuations.

Best solution uses daylight blocking filter ( dark plastic ) in PD or IRDA Rx.


YOu can use LM386 for output stage but you should check out SHarp (OSram etc) IR Rx chips with AGC.

SunnySkyguy said:

Best solution uses daylight blocking filter ( dark plastic ) in PD or IRDA Rx.

YOu can use LM386 for output stage but you should check out SHarp (OSram etc) IR Rx chips with AGC.

Click to expand...

As I said the point is to minimize cost and make the construction easier, so tubes to shield from ambient light and filters should be avoided (I know these are solutions btw).
I do not worry about the LM386 not having agc, the 200 max gain of it is very adequate and if a laser and a BPW34 diode is used, the gain is so much that the amplifier blocks. In permanent installations agc is not that much needed, agc is needed in IR data transmission because the user usually moves with respect to the stationary receiver.
The schematic above limits the audio range and this can be done using a discrete BPF, but there is no reason, except if the gear is to be used in DX.

As I said, the thing that worries me is not the amplifier but the photodiodes, because I suspect that if so much ambient light exist, then they could not cope with the lower light intensity signal of the laser beam (saturated). The noise cancellation occurs after the diodes, but do you think diode saturation will occur in such a case? (assuming no extra tube enclosures)

I think it would work although the wanted and unwanted light paths would have to be fixed and the 'unwanted' signal level carefully adjusted to cancel it's pick up on the wanted signal.

In a microphone evironment the two mics would be mounted so that their positions relative to each other was constant, you would have to do the same thing with the photodiodes if they are mobile. Be careful, you may find the signal picked up from one end of a flourescent lamp is different to the other end or middle so the unfocussed light may produce a complex signal which is not exactly the same as the background light on tthe 'wanted' diode. I doubt you would get perfect cancellation but you should see a considerable improvement.

Brian.

betwixt said:

I think it would work although the wanted and unwanted light paths would have to be fixed and the 'unwanted' signal level carefully adjusted to cancel it's pick up on the wanted signal.

In a microphone evironment the two mics would be mounted so that their positions relative to each other was constant, you would have to do the same thing with the photodiodes if they are mobile. Be careful, you may find the signal picked up from one end of a flourescent lamp is different to the other end or middle so the unfocussed light may produce a complex signal which is not exactly the same as the background light on tthe 'wanted' diode. I doubt you would get perfect cancellation but you should see a considerable improvement.

Brian.

Click to expand...

Indeed, the two photodiodes must receive as much equal unwanted light as possible, in order for the canceling effect to be effective. Since this will be used for a fixed application with a small enclosure, I believe that mounting one diode in face with the laser and the other perpendicular to it and a few cm away, will do the trick. The diodes should then receive almost equal amounts of unwanted light and one of them will receive the laser as well, whereas the other will receive much less laser amount (depended on the diodes distance). This less laser amount will be lost on cancellation, but this is much less than the main beam intensity. Careful adjustment of the amplitude of each section in the cancellation amplifier should be done, to ensure good unwanted light intensity.

If this works, it will dramatically reduce the cost of the overall system, as no filters or shadow enclosures are required.

Another low cost and very simple trick you can try is to cut polarized filters and align them at each end to reduce off-plane pick up. It might work with the noise cancelling circuit to further improve the SNR.

Cheap source of polarized filters:
1. 'Polaroid' sunglasses, cut the lens to size.
2. salavage from small LCD displays such as broken wristwatches and calculators. Peel the plastic front away from the glass and remove any adhesive from it, the plastic is polarizing.

Brian.

I won't expect to achieve acceptable audio signal-to-noise ratio without modulation. Daylight isn't that much of a problem, the DC current increases diode shot noise, but only moderately. DC bias of photodiodes can be handled by suitable circuits. But interferences of fluorescent lamps can't be effectively supressed. For laser light, smallband interference filters (unfortunately expensive) can reduce the crosstalk, but it may still inacceptable.

I would think that there is two major sorts of noise, one a straight forward 50 HZ buzz as the lamps are modulated by the mains voltage and a high frequency hash as the tubes restrike after mains cycle. So a very good LF reject filter to try and get rid of the 50 HZ (and 100 and 150. . .). To deal with the hash a noise blanker might work, take a sample of the signal and pass it through a high pass filter, then using the rate of change to fire a mono stable to "punch" holes in the audio, like a disc scratch reducer. The ear is less upset by bits of silence rather then peaks of noise.
Frank

FvM said:

I won't expect to achieve acceptable audio signal-to-noise ratio without modulation. Daylight isn't that much of a problem, the DC current increases diode shot noise, but only moderately. DC bias of photodiodes can be handled by suitable circuits. But interferences of fluorescent lamps can't be effectively supressed. For laser light, smallband interference filters (unfortunately expensive) can reduce the crosstalk, but it may still inacceptable.

Click to expand...

Why daylight is not much of a problem FvM? Because it is not modulated (in comparison to fluorescent lamps which are mains modulated)?
Is it the modulation only that causes the problem or the intensity of an unwanted source (daylight) onto the diode, possibly limiting it's dynamic range?
Or to put it in another way, if I point a photodiode to the sun and then hit it with a modulated laser, will this still be able to receive the modulated laser signal?
I thought that the red portion of the sunlight will cover up the modulated signal of a red laser pointer if both hit on the diode at the same time.

All asumptions are correct and valid in your scenario. The other major source of interference is from CFL which can produce large signals anywhere between about 20KHz and 60KHz.

You can try sending your signal (which you have not specified but we are assuming is analog in nature) as FM using a carrier of > twice your highest modulation frequency. It isn't complicated to do and has several advantages:

1. The laser is used at peak power all the time so the average received light level is higher.
2. It buys you some immunity from amplitude variation of the wanted signal and some degree of rejection of unwanted signals.
3. You can more accurately design a bandpass filter for only the FM carrier and bandwidth you need.
4. The carrier frequency can be chosen to avoid other sources of signals.

FM is used extensively in my location which is a bi-lingual area and popular with tourists. The local language is Welsh but almost all visitors speak English. Many of the local attractions and community meetings use a translator who speaks into a microphone which frequency modulates arrays of IR LEDs mounted in the corners of the ceilings. The light is received and demodulated into headphones worn by the people needing the translation. The system works well and can even use different channels in other languages simultaneously by switching carrier frequencies on the headphones.

Brian

neazoi said:

As I said the point is to minimize cost and make the construction easier, so tubes to shield from ambient light and filters should be avoided (I know these are solutions btw).
I do not worry about the LM386 not having agc, the 200 max gain of it is very adequate and if a laser and a BPW34 diode is used, the gain is so much that the amplifier blocks. In permanent installations agc is not that much needed, agc is needed in IR data transmission because the user usually moves with respect to the stationary receiver.
The schematic above limits the audio range and this can be done using a discrete BPF, but there is no reason, except if the gear is to be used in DX.

As I said, the thing that worries me is not the amplifier but the photodiodes, because I suspect that if so much ambient light exist, then they could not cope with the lower light intensity signal of the laser beam (saturated). The noise cancellation occurs after the diodes, but do you think diode saturation will occur in such a case? (assuming no extra tube enclosures)

Click to expand...

There are a few incorrect assumptions in your reply.

1. Daylight blocking filters are free and available in almost every IR optical detector, IRDA Rx etc.
2. Light shrouds for 5mm parts cost about a penny using 1cm of heat shrink
3. These are all low cost solutions, if you know where to look.
4. Photodiodes are stable and can easily provide 0.4A/W.

Saturation from Sunlight or other stray light sources must be avoided for reliable reception using the 2 almost free methods that I indicated before and repeated here.

FM modulation of a laser diode, that's new to me. I wonder if there are simple means of achieving this and if this can be done with LEDs as well

I've never opened one to look inside but my guess is they use a simple timer, possibly a 555 in the transmitter and feed audio to the frequency control pin. At the receiver they probably use a CD4046 or similar as a simple PLL FM demodulator. For audio links you can probably get away with a carrier frequency as low as 20KHz although it needs to be at least twice the highest modulating frequency to work properly. Bear in mind that most IR remote controls work with carriers of 38KHz or 56KHz so driving an IR emitter is quite easy at those frequencies.

Brian.

betwixt said:

I've never opened one to look inside but my guess is they use a simple timer, possibly a 555 in the transmitter and feed audio to the frequency control pin. At the receiver they probably use a CD4046 or similar as a simple PLL FM demodulator. For audio links you can probably get away with a carrier frequency as low as 20KHz although it needs to be at least twice the highest modulating frequency to work properly. Bear in mind that most IR remote controls work with carriers of 38KHz or 56KHz so driving an IR emitter is quite easy at those frequencies.

Brian.

Click to expand...

Oh so you mean pulse width modulation at a frequency higher than the highest audio frequency.
Thanks Brian

neazoi said:

Indeed, the two photodiodes must receive as much equal unwanted light as possible, in order for the canceling effect to be effective. Since this will be used for a fixed application with a small enclosure, I believe that mounting one diode in face with the laser and the other perpendicular to it and a few cm away, will do the trick. The diodes should then receive almost equal amounts of unwanted light and one of them will receive the laser as well, whereas the other will receive much less laser amount (depended on the diodes distance)...

Click to expand...

I think your assumption that two diodes mounted in this way will receive an equal amount of ambient light is unjustified. The amount of ambient light received by the two diodes depends on the reflectivity of the walls or objects in view. Imagine a white-painted wall straight ahead and a brown couch to the perpendicular side. Diodes pointed in these two directions will receive very different amounts of ambient light. The only diode that is guaranteed to receive the same amount of ambient light as the main signal diode is the main signal diode itself. That is why using an IR carrier frequency works. The process of demodulating the carrier effectively compares the light when the laser is on to the light when the laser is off. The ambient light is cancelled out because it appears in both instances. But you will never get a separate diode to reliably receive the same ambient light as the first one, especially if it has to point in a different direction.

Tunelabguy said:

That is why using an IR carrier frequency works. The process of demodulating the carrier effectively compares the light when the laser is on to the light when the laser is off. The ambient light is cancelled out because it appears in both instances. But you will never get a separate diode to reliably receive the same ambient light as the first one, especially if it has to point in a different direction.

Click to expand...

This is an interesting point thank you!
However total cancellation is not absolutely needed, if most of the light is canceled this should be fine.
Indeed the technique you mention is optimum. I would be very interested to see an example of it.

FM not PWM. You can use PWM but that might introduce the interference problem again.

Using a PLL to recover the frequency modulation gives a high degree of amplitude change immunity.

Brian.

neazoi said:

This is an interesting point thank you!
However total cancellation is not absolutely needed, if most of the light is canceled this should be fine.

Click to expand...

You don't understand. Your attempt at canceling the noise may actually make the noise worse, if the ambient light in the side-looking diode is more twice the light the forward-looking one. I seriously doubt that you will be able to use this method to make a difference between working and not working.