# Complimentary Photodiode Amps

Status
Not open for further replies.

#### dknguyen

##### Well-Known Member
Hey all I have some questions about photodiode amps. I'm thinking about an amp that uses two photodiodes where one is used to measure an incoming signal and the second is used as a reference to remove the influence of ambient light.

1. What makes more sense?
(a) Giving both photodiodes their own amplifier and subtracting the amplified outputs from each other (more components but less unknowns)
(b) Connect both photodiodes anti-parallel to each other to a single amplifier (this feels like the superior way but I feel I might be missing a problem with this setup)

2. Seeing as how I want to have two diodes to eliminate common mode signals, is there any way electrical (no optical filters) to stop them from saturating when exposed to intense light that would otherwise saturate a single lone photodiode ? The only thing I can think of that even has any chance of doing this is anti-parallel photodiodes but I don't think it does.

Last edited:

#### ronsimpson

##### Well-Known Member
What signal are trying to detect?
What light are you trying to reject? Sun light?

Welcome back

#### KeepItSimpleStupid

##### Well-Known Member
Can you chop the known signal. i.e. know when it's there? If so, then this is a job for a lock-in amplifier. See the DSP Lock-in amps from SRS.

That's what I did.

#### dknguyen

##### Well-Known Member
Welcome back
Thanks. It's been a while. Life happened.

Can you chop the known signal. i.e. know when it's there? If so, then this is a job for a lock-in amplifier. See the DSP Lock-in amps from SRS.

That's what I did.
I am chopping the signal (see below) but would still like to implement an optical common-mode rejection just to experiment around with and give me more options.

What signal are trying to detect?
What light are you trying to reject? Sun light?
This is intended for a onboard propeller tach so I mostly want to reject sunlight. The plan is to illuminate the propeller blade with an LED and detect the reflection with a photodiode. I'm going to be modulated things for immunity, but just for kicks and experimentation I'd like a second LED for common mode rejection too. The second LED would also let me, with minimal modification, experiment with using ambient light for illumination rather than the LED.

These are visible-opaque IR photodiodes.

Last edited:

#### jpanhalt

##### Well-Known Member
Model aircraft tach's use a single photodiode and rely on ambient light. They work in front of or behind the engine. Now, if you want to use one at night, a single, always-on LED might suffice.

Why complicate it with the other photodiode?

John

#### dknguyen

##### Well-Known Member
Model aircraft tach's use a single photodiode and rely on ambient light. They work in front of or behind the engine. Now, if you want to use one at night, a single, always-on LED might suffice.

Why complicate it with the other photodiode?

John
The handheld ones do, however I have read from some people who have said onboard ones that work off ambient light don't work so well. Part of the reason I want to add the photodidoe in is that if it doesn't require significant circuit changes, it's easier to remove than it is to add after the PCB is made. Especially if it's something like an anti-parallel LEDs. If I just don't mount the second LED it's identical to a single LED system.

#### tomizett

##### Active Member
I think what John has in mind above (#6) is that if you differentiate the signal from a single diode - that is to say, you only take notice of sudden shaddows caused by the moving blades, not slower changes caused by the change in attitude to the sun - then you have effectively compensated for ambient light. You can just use a single diode to measure ambient light and then look for sudden dropps in intensity with respect to that baseline.
Of course, your remarks about avoiding saturation of the diode or any associated circuitry are still well made. As a human, it's very easy to overlook the massive variation in light intensity that we experience daily - our eyes are very good at compensating for it.

I should add that I've no experience of building this kind of system, but that's my 2 pence worth anyway.

#### jpanhalt

##### Well-Known Member
dknguyen

I think compensating for the range of brightness of the sun versus your reflected intensity will be quite difficult.

You might also consider a typical IR detector approach with carrier frequency of 38K or higher. Even with an RPM of 10K (2-bladed prop), you should be able to differentiate modulated IR from background IR.

The easiest approach for an on-board sensor given your concerns might be a Hall detector with a magnet on the crankshaft. If it is an electric ignition model, that probably already exists. If fully electric, then go with the back emf. If compression ignition, then add a magnet.

John

#### MikeMl

I took a low $model-airplane tach flying in my Cessna. It works remarkably well.... You just aim it at the back of the prop arc out the windscreen. I think the secret is that it uses a high-gain, AC-coupled, DC-blocked amplifier behind the single photo detector. #### dknguyen ##### Well-Known Member Most Helpful Member I took a low$ model-airplane tach flying in my Cessna. It works remarkably well.... You just aim it at the back of the prop arc out the windscreen. I think the secret is that it uses a high-gain, AC-coupled, DC-blocked amplifier behind the single photo detector.
Hmmm. Maybe I'll just take a closer look at AC coupling. This is definitely one of the few applications for the stuff I do where DC coupling isn't necessary. I think making it AC and edge-triggered should allow it to inherently work in in the dark and in the sunlight without having to switch modes, as long as the reflection of the LED is sufficiently different from the ambient lighting.

dknguyen

I think compensating for the range of brightness of the sun versus your reflected intensity will be quite difficult.

You might also consider a typical IR detector approach with carrier frequency of 38K or higher. Even with an RPM of 10K (2-bladed prop), you should be able to differentiate modulated IR from background IR.

The easiest approach for an on-board sensor given your concerns might be a Hall detector with a magnet on the crankshaft. If it is an electric ignition model, that probably already exists. If fully electric, then go with the back emf. If compression ignition, then add a magnet.

John
The initial plan was to send a pulse train to drive the LED and only accept received pulses that were synced up when the LED was supposed to be on. That seems like the absolute lowest form...so low that I don't think that really counts as modulation though?

It's full electric but I'd like to avoid back EMF or phase-tracking just because there's a project in the future that requires knowing the absolute position of the propeller blades. An encoder would work for that too, but it'd be nice to be able to blindly mount the propeller and not have to calibrate to align it to the encoder, and the only solution I can think of that does that is measuring the time interval between blades at a known point and assuming the blade velocity is the same between measurements.

Last edited:

#### dknguyen

##### Well-Known Member
I did find this:
http://www.ee.ryerson.ca/~phiscock/papers/photo-detector.pdf

With page 8 and 9 being the parts of interest. I am not familiar with the how that modulator and LP filter in the last stage works. Does anyone know what it's called? The description on page 9 seems to imply that the modulator switching at 2000kHz and the LP filter with a cutoff at 5Hz results in a pass band of everything outside of 2kHz +/- 5Hz, though I'm not sure why that is.

The stacked "current steering" setup of the photodetectors is also not really what I expected. I was literally going to put them in parallel with each other in opposite directions but maybe I overlooked some issue with that configuration. It doesn't allow for bipolar operation but I don't need that. Perhaps there is some issue with common-mode photodiode currents circulating within the loop formed by the photodiodes.

#### jpanhalt

##### Well-Known Member
I apologize if I am mentioning the obvious. Various manufacturers, of which Vishay is one, make IR receivers that suppress any modulation out of the center frequency. Here is a link to a Vishay 56 kHz IR receiver: http://www.vishay.com/docs/82667/tsdp341.pdf 38kHz is very common. The reception band is quite broad.

Although, I don't think your application needs the complication of an emitter LED (see Post #10) you seem committed to doing that. I used a 38 kHz modulation TX/RX pair to detect movement a 3/16" diameter wire in bright sunlight to dusk conditions. They work. Detecting a prop should be easy.

John