interfierence with infinite gain op amp

[aruna1]

guys I'm making this color sensor.i have used 2 RED LEDs as both sensor and emitter. I have used TL084 op amp in infinite gain mode because Sensor LED generates very small voltage.
my problem is op amp is too sensitive so it reacts when i tough circuit or negative wire (some times evan my fingers reach the circuit [without even touching])

this made circuit kind a unstable.other than that circuit works fine.
any solution for this?

here is the schematic

thanks

[Admin]

reduce the gain?

attach a grounding wire to your hand?

why use an LED as a sensor? (other than the cool-factor)

[aruna1]

wanted to do something new

[Soeren]

Hi,

I have used TL084 op amp in infinite gain mode [...]

With the component values shown, you have a gain, Av, of between 65x and a very theoretical infinite gain. In the real world, there are several limiting factors (like eg. f_T and slew-rate), so I'd be surprised to get a gain near or above 1000x (it would be utterly useless at any useable frequencies as well).
With the potentiometer in mid-position, the gain is ~100x
Btw. R5/C1 makes up a low pass filter of around 3.6kHz

because Sensor LED generates very small voltage.

And LED1 gets less than 1mA


It is not a very pro design (mildly put), where did you find it?

my problem is op amp is too sensitive so it reacts when i tough circuit or negative wire (some times evan my fingers reach the circuit [without even touching])

this made circuit kind a unstable.other than that circuit works fine.

Could you rephrase the last statement? Sounds impossible to me, either it works fine OR it's unstable I'd say.

[aruna1]

did few changes to circuit.

unstable mean when I set pot to detect white-black.it works bt sensor senstivity varies time to time so i have to adjust it time to tme

[Admin]

Could you rephrase the last statement? Sounds impossible to me, either it works fine OR it's unstable I'd say.

I'm assuming its either/both static electricity or its picking up a slight charge/field about his hand.

Also, moving your hand next to it will change the environmental light contacting the LED sensor, thereby changing the signal . . .

unstable mean when I set pot to detect white-black.it works bt sensor senstivity varies time to time so i have to adjust it time to tme

You need to put a filter on it. For example, the light coming from your window will change throughout the day. Basically I think you want to detect a sudden *change* in the signal, not the amplitude, right?
For example, a differentiator or integrator op-amp circuit, combined with a comparator IC.

[aruna1]

You need to put a filter on it. For example, the light coming from your window will change throughout the day. Basically I think you want to detect a sudden *change* in the signal, not the amplitude, right?
For example, a differentiator or integrator op-amp circuit, combined with a comparator IC.

explain more please

[Admin]

google these terms:

differentiator op-amp circuit
integrator op-amp
comparator IC


The signal average will change with environmental light. You want to remove the slow changing average, and only concentrate on signal spikes. So, you need to filter it wrt time.

[aruna1]

thanks

[Soeren]

Hi,

I'm assuming its either/both static electricity or its picking up a slight charge/field about his hand.

Or it could be as simple as mains hum - I assume a lot of "antennas".
But... In my book, it's impossible for a device to be classified as "working fine" if it's unstable and I was trying to understand the reason for such a statement.

[Soeren]

Hi,

To get a stable output you need to do several things.
Sensor must be shielded from ambient light as well as possible, since it will be swamped otherwise and shielding is very easy to do.
I'd recommend a modulated signal, as that help reduce ambient influence a lot.
To go a step further, a light controlled bias voltage to the sensor will make sure it's operating in its most effective region most of the time.

There is no such thing as an IR light detector that won't be affected by the sun, as the sun is rich in IR.
Using an IR filter will reduce _some_ artificial light, but since it pass IR and the sun emits IR...

[aruna1]

Hi,

To get a stable output you need to do several things.
Sensor must be shielded from ambient light as well as possible, since it will be swamped otherwise and shielding is very easy to do.
I'd recommend a modulated signal, as that help reduce ambient influence a lot.
To go a step further, a light controlled bias voltage to the sensor will make sure it's operating in its most effective region most of the time.

There is no such thing as an IR light detector that won't be affected by the sun, as the sun is rich in IR.
Using an IR filter will reduce _some_ artificial light, but since it pass IR and the sun emits IR...

about modulating; wont excessive light simply saturate sensors so it wont recognize any kind of signal?

and "a light controlled bias voltage to the sensor will make sure it's operating in its most effective region most of the time." please explain this

[Admin]

and "a light controlled bias voltage to the sensor will make sure it's operating in its most effective region most of the time." please explain this

Not sure what Soeren had in mind but . . .

Get another LED sensor and mount it a few feet away. Then connect it to a voltage difference op-amp. You effectively subtract the environmental average. This is commonly done for AFMs (atomic force microscopes) to subtract temperature swing effects on the probe.

[aruna1]

what about sensor saturation and modulation?

[Admin]

what about sensor saturation

Put a filtered lens over it.

[aruna1]

hmm how thats gonna help? lets say IR filter+lenz it will stillsaturate sensor with IR from sun

[Soeren]

Hi,

about modulating; wont excessive light simply saturate sensors so it wont recognize any kind of signal?

The receiver/sensor still have to be screened from light, as if it's totally saturated by another light source, there's no chance of detecting anything of course.
With a partial "light influence" from another source, it can still be made to handle well, see below

and "a light controlled bias voltage to the sensor will make sure it's operating in its most effective region most of the time." please explain this

Consider a "voltage divider" setup with a sensor as the lower leg and a resistor as the upper leg.
With more light, the sensor will open more and the midpoint will go lower, so there's less potential left for detecting the wanted signal.
This can be changed by using a lower value resistor for the upper leg - i.e. you can tailor the resistor to the ambient light at any given time.

To make it work under different lighting conditions, you have to vary the current going through the sensor, so if eg. a transistor is controlled by the output of the (entire) detector circuit, but only responding to slow (relative to the signal) changes, the bias will be self adjusting. This takes a bit more circuitry of course, but might be the only solution in some cases.

As a first try, use a good screen and mount the sensors pointing down and only "seeing" the are hit by the IR you control.

[Soeren]

Hi,


Big oops!

I must have been sleepy - you can't use a TL084 on 5V!
Better go for something like a TLC272 or similar.

[aruna1]

Hi,


Big oops!

I must have been sleepy - you can't use a TL084 on 5V!
Better go for something like a TLC272 or similar.

I know but amazingly it can be,sensor circuit worked with 5v,don't know how

[Soeren]

Hi,

Well it wasn't stable, according to yourself and it will never be, even if it was made correctly - try a more suitable op-amp or raise the voltage - or don't, if you think it's working as you want it 

A different approach, could be using an integrated IR-receiver (of the kind that allows a continuous wave) and adjust the power of the modulated LED to adjust sensitivity. This ought to be much easier for you to get working, but you'll have to go infrared for this to happen, as the integrated receivers are IR.

[aruna1]

Hi,

Well it wasn't stable, according to yourself and it will never be, even if it was made correctly - try a more suitable op-amp or raise the voltage - or don't, if you think it's working as you want it 

A different approach, could be using an integrated IR-receiver (of the kind that allows a continuous wave) and adjust the power of the modulated LED to adjust sensitivity. This ought to be much easier for you to get working, but you'll have to go infrared for this to happen, as the integrated receivers are IR.

what are they,any exxample?

[Soeren]

Hi,

The integrated IR-sensors?
Here are some from Vishay http://www.vishay.com/ir-receiver-modules/

[aruna1]

oh these,they don't allow continues IR signal

[Soeren]

Bottom of the page: "IR Receivers with constant gain for sensor and light-curtain applications
TSOP4038   TSOP5038   TSOP58038".

[aruna1]

well will they responde to IR in sun?

[Soeren]

The more ambient IR, the less sensitivity of course, but as I mentioned, you need screening to control the amount reaching the receiver.

It's similar to the remote controls used on eg. TV and HiFi systems, so you can test how much sunlight it will handle - use a mirror or two to direct it towards the receiver and see how the range of the remote goes down with increasing amounts of sun incident on the receiver.

[aruna1]

ah i see