Light Sensor schematic check

[gvi70000]

Hello all,

I need your help in checking the attached schematic before I will send the PCB to manufacturing.
The idea behind it is to have a variable sensitivity sensor that will feed the ASC of a microcontroller.
As sensitive element I am using a BPW34 PIN photodiode. The current generated is fed into a TIA that has a variable gain set by the digital pot.
For it to work when no current is flowing I have used a small voltage bias.
In the second stage I am removing the bias from the TIA output.
The third stage is again an amplifier to be able to amplify very weak signals.
The last stage is a buffer and I am using R12+C6 to smooth the output signal.

Please let me know your opinions, will the circuit work as described? Do you have any improvements to propose?

Thank you for your support,
Ion

 

[danadakk]

There are references a lot cheaper than those :

https://www.digikey.com/en/products/filter/voltage-reference/693?s=N4IgTCBcDaIG4HsA2AXAhgcwKYAIBOWAZlgQHYDGWIAugDQgCsUoADlAIz0tuRgMC%2B9AGzMQASwAmUEAFp2ABnYguKacpAoAniyqQQaAM7kQ-fkA

Regards, Dana.

 

[KlausST]

Just to compress/repeat/explain what I´ve already written before.

I tried to focus on the information you gave,
And when I read "detecting meteors"... then for the circuit it means:
--> detecting extremely low light intensity (DC) and detecting extremely low CHANGE in light intensity (AC).
Electrically this means:
--> detecting low current (DC) and detecting very low CHANGE in current (AC).

Where DC means it changes within minutes, years = very low frequency.

Now every circuit causes errors, like DC fluctuation and noise. So your circuit needs to be able to differentiate between these unwanted errors and true light on the sensor.
So the usual way is to keep the errors small and to keep thie light signal high.
In technical terminology: You need a high signal-to-noise-ratio = SNR.

Errors (noise) caused in your circuit:
* using VCC as bias: it is amplified in the TIA, thus it can easily overrule the "meteor signal" by a factor of 100. This makes it unable to detect the meteor. --> you don´t need accuracy(using 0.1% resistors), you need precision, a stable signal with low noise, low drift. Any cheap one will be much better than using VCC.
* noise of resistors, OPAMPs, ADC. Here you need to focus on the "wanted frequency" of the light signal, and use filters to get rid of all unwanted frequencies. Otherwise your ADC values may jump around, you can not rely on them, you can not detect meteors.
To get a high SNR the first idea could be to just amplify the signal from the light sensor. So far so good. But you need to ensure not to amplify the "unwanted" signals the same way --- since then you gain nothing.
Regarding SNR: If you amplify your signal the same way you amplify the noise, then the ratio does not change. No improvement.
--> use filters, like written above.

And another thing to consider: You need to be able to detect the signal.
Imagine a bowl of water. Knock on the bowl and you will see the resulting waves on the water surface.
Now imagine there is just 1mm of water in the bowl. Or even less, let´s say 0.1mm ... this makes it more difficult to see the waves. You surely have a problem when the amplitide of the waves is higher than the water level, because amplitude also goes negative. You can´t get negative water level in the bottom of the waves.
The water level needs a useful distance from ground.

The same is true for electrical signals.To clearly detect them you need a useful distance from ground. A DC offset. With your circuit that completely removes all the VBias to zero, it can´t work for small light signals.
You need at least lift the bottom level high enough to cover errors like noise (peak), OPAMP offsets, offset caused by leakage current, OPAMP regulation undershot, offet caused by the ADC. I did no calculation, but 20--50mVshould be enough, but not removing the VBias to get 100mV offset won´t hurt. (Forget about the loss in ADC input range)

Klaus

 

[gvi70000]

Hello again,

I have experimented with the design and considering Klaus feedback I have the following:
1. Use ADR130BUJZ-REEL7 as 0.5V voltage reference
2. The feedback capacitor value 22nF for a bandwidth of 40Hz or 47nF for 19Hz (100nF for 9Hz)
3. The capacitor in parallel with photodiode (not shown in the screenshots)- I will keep the foot print on the PCB for experiments
4. One solution without bias reduction - with lowest noise and easy to build
5. One solution with bias reduction - has higher noise but reduces the offset from 0.5V to 50mV

Which solution you believe is the best?

 

[KlausST]

Hi,

the schematics now come closer to what I did recommend.

But with both schematics I see some issues:

Solution1: this includes offset reduction
It looks like a simulation, so did you check that the offset really becomes reduced?
What I expect: (on 0 uA input signal)
* VOUT1 = 500 mV
* VDIFF = about 45 mV
* but: VOUT = about 227mV (amplified by 5)

*********************

Solution2: (without offset reduction)
What I expect: (on 0 uA input signal)
* VOUT1 = 500 mV
* but: VOUT = about 2.5V (amplified by 5)

I like the more simple solution2. Just connect R6 to VREF instead of GND.
Then the offset is 500mV.
(And add pads for a capacitor in parallel to R5 for additional noise reduction)

Also you may play with much higher values for C5.

I expect no improvement by adding C2, since the node is the input of a TIA and should not move.
So adding C2 increases output noise instead of reducing it.

This circuit will be an enormous improvement in detecting low light signals compared to your first approach.

**********************
If you change the feedback resistance of the TIA you change it´s gain. So far so good. But you also change it´s upper cutoff frequency.
With low gain you get higher frequency response and vice versa.

In my application I kept the TIA gain constant but changed the gain of the second stage. So I had constant cutoff frequency.

Klaus

 

[danadakk]

Get rid of offset, use 3.3V Vref to power OpAmp and supply a Vref :

Phiotocurrent swept from 0 to 20 uA.


Regards, Dana.

 

[gvi70000]

Hi,

the schematics now come closer to what I did recommend.

But with both schematics I see some issues:

Solution1: this includes offset reduction
It looks like a simulation, so did you check that the offset really becomes reduced?
What I expect: (on 0 uA input signal)
* VOUT1 = 500 mV
* VDIFF = about 45 mV
* but: VOUT = about 227mV (amplified by 5)

*********************

Solution2: (without offset reduction)
What I expect: (on 0 uA input signal)
* VOUT1 = 500 mV
* but: VOUT = about 2.5V (amplified by 5)

I like the more simple solution2. Just connect R6 to VREF instead of GND.
Then the offset is 500mV.
(And add pads for a capacitor in parallel to R5 for additional noise reduction)

Also you may play with much higher values for C5.

I expect no improvement by adding C2, since the node is the input of a TIA and should not move.
So adding C2 increases output noise instead of reducing it.

This circuit will be an enormous improvement in detecting low light signals compared to your first approach.

**********************
If you change the feedback resistance of the TIA you change it´s gain. So far so good. But you also change it´s upper cutoff frequency.
With low gain you get higher frequency response and vice versa.

In my application I kept the TIA gain constant but changed the gain of the second stage. So I had constant cutoff frequency.

Klaus

Hello Klaus,

Thank you for you patience and explanations offer to me. You helped me understand the basics of TIA and what to consider when designing sensitive sensors.

 

[KlausST]

Hi,

looks good so far.

It would be nice if you give some feedback after testing.
This circuit now gives good quality signal to the ADC.
For a first test you may use the ADC output as is, but for sure it can be improved by simple software filters.

Klaus

 

[gvi70000]

Hi,

looks good so far.

It would be nice if you give some feedback after testing.
This circuit now gives good quality signal to the ADC.
For a first test you may use the ADC output as is, but for sure it can be improved by simple software filters.

Klaus

Hello Klaus,

For sure I will post an update after I will build and test the sensor. Thank you again for your great support.

PS: I need to build also a sound sensor - see https://www.edaboard.com/threads/sound-sensor-schematic-check.408178/ - can you support me here also?

 

[gvi70000]

Hello Klaus,

The PCBs have arrived and I have assembled the sensor. After first test everything looks great.
Thank you again for your great support.