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 even 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

 

[waht]

reduce its gain by changing 2M resistor to something like 100K or 10K.

 

[berkeman]

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 even 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

That is not how you amplify a photocurrent (like from an LED or photodiode). Instead, you should use a current-to-voltage converter configuration for the first stage. Google current to voltage converter photodiode amp, and you should get some representative circuits. Or your textbook or any opamp cookbook will have the circuit.

 

[vk6kro]

I put a red LED across the input of a digital multimeter.

It was able to generate up to a volt when held near a 60 watt lamp bulb.
It had no trouble generating hundreds of mV at greater distances from the lamp.

So, these are not small voltages. Note that the cathode of the LED generated +ve voltage relative to the anode.

You could verify that your LED was generating voltage by using a digital multimeter, yourself.

I would try something like the attached diagram with a lot less gain than you have now. As a start, try making the resistor across the Opamp 100 K and the input resistor 4.7 K. This will give a gain of about 20.

 

[aruna1]

well i use another LED to emitte light so voltage generated on sensor LED is small.

anyway i think i solved it with this circuit

 

[vk6kro]

Looks like that would give a voltage comparator action.
ie your output light would be either on or off.

Is that what you want?

 

[aruna1]

Looks like that would give a voltage comparator action.
ie your output light would be either on or off.

Is that what you want?

yup.just to detect black or white surface

 

[berkeman]

well i use another LED to emitte light so voltage generated on sensor LED is small.

anyway i think i solved it with this circuit

If I were interviewing you for an electronics job, and I asked you to draw me a circuit to detect light, and you drew me that "fixed" circuit, what do you think would happen next?

 

[vk6kro]

I checked the TL084 data sheet and it shouldn't have a voltage less than 7 volts as a supply.

Also it does not say you can operate inputs close to or at either supply voltage rail, so you probably can't.

So, you are probably losing a lot of sensitivity by using this circuit.

However it is your post, so if it seems to be working for you then go for it.

An easy test would be to put a voltmeter across the LED and slowly bring a light close to the LED. Measure what voltage it takes to make the output change state.

 

[aruna1]

i checked the tl084 data sheet and it shouldn't have a voltage less than 7 volts as a supply.

Also it does not say you can operate inputs close to or at either supply voltage rail, so you probably can't.

So, you are probably losing a lot of sensitivity by using this circuit.

However it is your post, so if it seems to be working for you then go for it.

An easy test would be to put a voltmeter across the led and slowly bring a light close to the led. Measure what voltage it takes to make the output change state.

well,i'll increase voltage.anyway do you think this op amp is not suitable to amplify voltage generated in led?.if it can be done with this op amp,can u suggest a circuit?.and also can you suggest a suitable common op amp for this? We can't use 741 here right?.lmc6482 is not available.

 

[aruna1]

ACTUALLY WHAT I WANT DO IS THIS.
I NEED TO MAKE A SENSOR TO DETECT WHITE SURFACE IN BLACK BACKGROUND.SO WHAT I THOUGT WAS BY EMMITING RED LIGHT TO BOTH SURFACES,WHITE WILL REFLECT WHILE BLACK WILL ABSORBE LIGHT.SO IN SENSOR I THOUGT IF LED GENERATES VOLTAGE WHEN RECIEVING LIGHT I CAN USE IT HERE COZ WHITE SURFACE WILL REFLECT LIGHT AND SENSOR LED WILL GENERATE FEW mv WHEN IT RECIEVED THIS REFLECTED LIGHT FROM WHITE SURFACE.SINCE BLACK SURFACE WILL ABSORB LIGHT SENSOR WILL NOT GENERATE ANY VOLTAGE.
SINCE GENERATING VOLTAGE IS LOW AND HAVE ONLY TINY CURRENT I THOUGT FET BASED OP AMP WILL SUIT THIS.BECUSE FET DOES NOT REQUIRE INPUT CURRENT.AND SINCE I WANT DIGITAL OUT(WHITE SURFACE OR NOT) I PUT POTENTIOMETER TO POSITVE INPUT TO SET REFERENCE VOLTAGE.
AND SET SENSOR LED IN REVERSE MODE(positive pin to negative supply and negative pin to inverting input pin).
SO IF VOLTAGE GENERATE ON LED BY REFLECED LIGHT IS LARGER THAN REFERENCE VOLTAGE OP AMP OUTPUT WILL BECOME LOW.OTHERWISE IT WILLBE HIGH.

ACCORDING TO THIS I DESIGN ABOVE CIRCUIT.
SO TELL ME WHATS WRONG WITH IT.
THANKS

 

[vk6kro]

well,i'll increase voltage.anyway do you think this op amp is not suitable to amplify voltage generated in led?.if it can be done with this op amp,can u suggest a circuit?.and also can you suggest a suitable common op amp for this? We can't use 741 here right?.lmc6482 is not available.

I attached a circuit above. Post #4.

That opamp is fine but it can't operate with the inputs close to the supply rails. Most opamps can't do this.

The LED generates a voltage but this will vanish quite quickly if you try to draw current from it. It may require a large resistor to work properly but check that after construction.

If you want on/off output, make the ratio of the feedback resistor and the input resistor to the -ve input as high as possible.

Can you use room lighting instead of the other LED? It would make the job easier.

 

[aruna1]

I attached a circuit above. Post #4.

That opamp is fine but it can't operate with the inputs close to the supply rails. Most opamps can't do this.

The LED generates a voltage but this will vanish quite quickly if you try to draw current from it. It may require a large resistor to work properly but check that after construction.

If you want on/off output, make the ratio of the feedback resistor and the input resistor to the -ve input as high as possible.

Can you use room lighting instead of the other LED? It would make the job easier.

no room light.its for line follwer robot.no background light is accepted

and can you explain how your circuit works

 

[vk6kro]

Have a look at this link:
http://upload.wikimedia.org/wikiped.../300px-Op-Amp_Non-Inverting_Amplifier.svg.png

It is a standard circuit.
You can read the full Wikipedia article here:
http://en.wikipedia.org/wiki/Operational_amplifier_applications

With a single power supply, you use a resistive divider to form an artificial ground. So this is the "earth" symbol in the diagram. This is normally bypassed with a capacitor to one of the supply rails if you were using it for audio. In this case, it probably doesn't matter.
This is the wiper contact on the pot in the diagram above.

The LED is a voltage generator with a high internal resistance, so it looks odd in this circuit, but as a voltage generator it is normal to put it there.

Just a typo from above:
"It may require a large resistor to work properly"
This would be a large value resistor across the LED. Maybe 2.2 M.

 

[aruna1]

ok thanks,I'll check

 

[aruna1]

and what is the minimum input voltage for this opamp?

 

[uart]

A couple of quick comments on your circuit.

1. As a passive detector you're usually better off using the LED in reverse biased "photo-current" mode as berkeman suggested.

2. If you do want to use it in the active "photovoltaic mode" as per your circuit then you have the LED polarity wrong. As a photovoltaic device it will generate a positive voltage at the anode in response to incident light. That is, forward biased voltage direction but reverse biased current direction (quadrant IV) in that mode.

3. You should use an op-amp termed "single power supply" which means that it will allow input voltages referenced (typically) to the negative supply. Try LM324 or similar.

4. You should use a very small amount of positive feedback to give the circuit a little bit of hysteresis (not too much though or it will latch up). You'll need to know what you're doing to get this right so you're likely to need further help or research to do it.

 

[vk6kro]

A couple of quick comments on your circuit.

1. As a passive detector you're usually better off using the LED in reverse biased "photo-current" mode as berkeman suggested.

Why is that? He wants to use it in photovoltaic mode.

2. If you do want to use it in the active "photovoltaic mode" as per your circuit then you have the LED polarity wrong. As a photovoltaic device it will generate a positive voltage at the anode in response to incident light. That is, forward biased voltage direction but reverse biased current direction (quadrant IV) in that mode.

No, he wants the output to drop as light level increases.

3. You should use an op-amp termed "single power supply" which means that it will allow input voltages referenced (typically) to the negative supply. Try LM324 or similar.

Good idea. Forgot about the 324.

4. You should use a very small amount of positive feedback to give the circuit a little bit of hysteresis (not too much though or it will latch up). You'll need to know what you're doing to get this right so you're likely to need further help or research to do it.

He hasn't got much input signal. I was reluctant to mention hysteresis.

 

[aruna1]

thanks uart and vk6kro.

 

[uart]

A couple of quick comments on your circuit.

1. As a passive detector you're usually better off using the LED in reverse biased "photo-current" mode as berkeman suggested.

Why is that? He wants to use it in photovoltaic mode.

I didn't say there is anything wrong with using photovoltaic mode (if done correctly) just pointing out that for a simple detector that reverse biased photocurrent mode typically gives a more robust solution with more output swing and better noise immunity. Were these factors not a significant part of auna’s original problems here?

2. If you do want to use it in the active "photovoltaic mode" as per your circuit then you have the LED polarity wrong. As a photovoltaic device it will generate a positive voltage at the anode in response to incident light. That is, forward biased voltage direction but reverse biased current direction (quadrant IV) in that mode.

No, he wants the output to drop as light level increases.

You've completely missed the point, I'm not referring to the op-amp output. A photo-diode develops a positive voltage at the anode, but auna1 has the anode connected to the negative supply and cathode connected to the op-amp input terminal. When illuminated the cathode will be driven to more negative than that negative supply rail. Do you really still think that is what he wants!

Presumably aruna actually has this correct in his circuit (or it couldn't work) and has just drawn it incorrectly in the schematic (I expect he just tried it either way until he got it to work). I corrected him because I think it's important to get things like that right. Many people (apparently yourself included) think that photovoltaics work the opposite polarity to what they actually do. I hate repeating myself but you're making me do it. In photovoltaic mode the diode develops a "forward biased diode polarity" voltage (that is, anode positive and cathode negative) and any current that it supplies flows in the "reverse bias diode" direction (that is, out of the anode). Can you not see that aruna has this the wrong way around in every schematic he has posted?

4. You should use a very small amount of positive feedback to give the circuit a little bit of hysteresis (not too much though or it will latch up). You'll need to know what you're doing to get this right so you're likely to need further help or research to do it.

He hasn't got much input signal. I was reluctant to mention hysteresis.

That's why I said "a very small amount" :) And note that this wouldn't even be an issue if he was set up properly in reverse bias photocurrent mode.

 

[vk6kro]

I thought you were referring to this circuit:
http://upload.wikimedia.org/wikipedi...lifier.svg.png
In that it doesn't matter which way the LED is connected. You just get opposite output.

You are quite right about the voltage from the LED. I still have one connected to my multimeter and it gives very good output and the anode is positive relative to the cathode..
Yes, Aruna's polarity was wrong, but that was the least of his problems.

I have previously tried this with different LEDs and never got any output, but the one I chose this time works well.
Could be that this multimeter has a 10 Megohm divider on the input where previous ones were possibly 1 Meg or even moving coil meters.

 

[aruna1]

I thought you were referring to this circuit:
http://upload.wikimedia.org/wikipedi...lifier.svg.png
In that it doesn't matter which way the LED is connected. You just get opposite output.

You are quite right about the voltage from the LED. I still have one connected to my multimeter and it gives very good output and the anode is positive relative to the cathode..
Yes, Aruna's polarity was wrong, but that was the least of his problems.

I have previously tried this with different LEDs and never got any output, but the one I chose this time works well.
Could be that this multimeter has a 10 Megohm divider on the input where previous ones were possibly 1 Meg or even moving coil meters.

guys take a look at this
http://www.robotroom.com/ReversedLED.html"

anyway i just brought LM324 and replaced one of TL084 IC in my circuit with it (I use 2 TL084 I'M making a 8 LED sensor array for line follower robot). with running 5V supply for both ICs circuit works same as previous (i had to adjust pots used with LM324) except that TL084 has higher amplification.so it works in bright sunllight compare to LM324.

II used circuit published above (one with wrong polarity-you say) I'm not sure how circuit works if polarities are wrong
any ideas?

 

[berkeman]

Finally! That's the current-to-voltage converter circuit that I was referring to! Maybe you have a chance in the job interview after all. Good.

 

[aruna1]

Finally! That's the current-to-voltage converter circuit that I was referring to! Maybe you have a chance in the job interview after all. Good.

you mean this circuit?

how in this circuit "current-to-voltage converter" thing happens? can you explain me?

 

[vk6kro]

I connected a red LED in series with a 100 K resistor across a 7 V DC supply with a meter connected in series with both, and on the 200.0 uA range.

In the forward direction it drew a current of 54uA. In the reverse direction it drew less than 0.1uA. The reading was ZERO. This was regardless of the lighting on the LED.
Even holding a 60 watt lamp less than an inch from the LED produced no measurable current.
This was the same without the 100 K resistor.

This is the same LED that generates 1 volt when placed near the same lamp.

This LED has a resistance of at least 70 Mohms regardless of light levels, and probably more.

I won't hammer the point, but I know which way I would use this device.

 

[aruna1]

ok guys i did some experiments with follwing two circuits.

both has LM324 op amp

1st circuit

https://www.physicsforums.com/attachment.php?attachmentid=20706&stc=1&d=1253884269"

here i have connected negative pin of LED to negative supply and positive LED pin to inverting input.(forward biased)

result
on black surface-output low
on white surface-output high


2nd circuit

https://www.physicsforums.com/attachment.php?attachmentid=20707&stc=1&d=1253884269"


here i have connected negative pin of LED to negative supply and positive LED pin to inverting input.(forward biased)

result
on black surface-output high
on white surface-output low


so according to you gys 2nd circuit shouldn't work?

so I'm little confused.

any comments?

 

[vk6kro]

Just looking at the internal circuit of the 324, both inputs are PNP grounded collector transistors with open bases.

Seems like that would work with negative input voltages?
The current is certainly limited by using a LED as the voltage source. Negative voltages and zero volts are OK for a transistor in this configuration provided you limit the current.

I suggest you go with what works.

You might like to make that 47 K pot smaller if you want better control of the voltage.

 

[aruna1]

Just looking at the internal circuit of the 324, both inputs are PNP grounded collector transistors with open bases.

Seems like that would work with negative input voltages?
The current is certainly limited by using a LED as the voltage source. Negative voltages and zero volts are OK for a transistor in this configuration provided you limit the current.

I suggest you go with what works.

You might like to make that 47 K pot smaller if you want better control of the voltage.

well actually I'm making this as a sensor for line follower robot.my objective is to make a sensor which has less effect on ambient lighting and have high stability.since LDR and photodiodes are too ensitive I though this would work.

do you know or can you suggest any better circuit for line follower with above features? (white line in lack surface with at bright sun light?
thanks

 

[uart]

well actually I'm making this as a sensor for line follower robot.my objective is to make a sensor which has less effect on ambient lighting and have high stability.since LDR and photodiodes are too ensitive I though this would work.

do you know or can you suggest any better circuit for line follower with above features? (white line in lack surface with at bright sun light?
thanks

Yes both of those results make sense (assuming that you are talking about the final output after the 7414N).

In the first circuit the sensor LED anode goes high when illuminated so the opamp output goes low. When the LED is not illuminated the opamp bias current may still give a small forward bias anode voltage, but if it's less than the threshold set at the op-amp non-inverting input then the opamp output goes high.

In the second circuit when the sensor LED is illuminated the cathode goes slightly negative so the opamp output goes high (but only because the LM324 has a common mode range that extends down to VEE - and even slightly below with some small degradation of parameters. Many opamps won't allow this). When the LED is not illuminated the opamp inverting input is essentially floating - and because the LM324 has a pnp input stage the bias currents will make the cathode float high so the output will go low. This really is a very dubious design that only works by good luck really. It is very sensitive to op-amp parameters that may change with temperature or if you use a different opamp type.

Why don't you try the circuit in my attachment. You may have to adjust the value of the 47M resistor to match your LED (lower resistance if too sensitive, higher resistor if not sensitive enough) but it should be a lot less sensitive op-amp parameters and op-amp type.

 

[aruna1]

Yes both of those results make sense (assuming that you are talking about the final output after the 7414N).

In the first circuit the sensor LED anode goes high when illuminated so the opamp output goes low. When the LED is not illuminated the bias current may still give a small forward bias anode voltage, but if it's less than the threshold set at the op-amp non-inverting input then the opamp output goes high.


In the second circuit when the sensor LED is illuminated the cathode goes slightly negative so the opamp output goes high (but only because the LM324 has a common mode range that extends down to VEE - and even slightly below with some small degradation of parameters. Most opamp won't allow this). When the LED is not illuminated the opamp inverting input is essentually floating - and becuase the LM324 has a pnp input stage the bias currents will make the cathode float high so the output will go low. This really is a very dubious design that only works by good luck really. It is very sensitve to op-amp parameters that may change with temperature or if you use a different opamp type.

Why don't you try the circuit in my attachment. You may have to ajust the value of the 47M resistor to match your LED (lower resistance if too sensitive, higher resitor if not sensitive enough) but it should be a lot less sensitive op-amp parameters and op-amp type.

can you explain this circuit please?

is that LED used n photo current mode? and is resistors big like 47M exist? (never heard before)
and is there anyway to add a pot here so i can chage sensitivity-threshhold value?


and can yu read my previous post about my objective of this circuit and give a suggetion?
thanks

 

[uart]

can you explain this circuit please?

is that LED used n photo current mode? and is resistors big like 47M exist? (never heard before)
and is there anyway to add a pot here so i can chage sensitivity-threshhold value?and can yu read my previous post about my objective of this circuit and give a suggetion?
thanks

Yes it's in photocurrent mode. You can get resistors above 10 Meg but some places may only stock up to about 10M. Try it with about 10Meg first up if you like, it might drive it ok. This is the part that's hard to design unless you know the characteristics of your particular LED - just how much photo-current can it supply when reverse biased and illuminated?

Edit : Actually the LM324 bias currents might a bit on the high side if you're using too higher value resitors. Try 10 Meg and if that still has enough sensitivity then you might even be able to go a bit lower. With this circuit you'd probably be better off with a FET input opamp like you had before.

well actually I'm making this as a sensor for line follower robot.my objective is to make a sensor which has less effect on ambient lighting and have high stability.since LDR and photodiodes are too ensitive I though this would work.

Not sure about this. I assume that what you need are illumination and sensor LED's that are narrow band to minimize the influence of ambient light. I'm not very familiar with the use of a LED as a photo-detector. I'd guess that as a sensor it would respond most strongly at the wavelength that it emits in normal LED mode, but I'm not 100% clued on this aspect.

 

[aruna1]

thanks

 

[vk6kro]

Looking at your second circuit again, the 47K pot would never be able to give a negative output, so both states of the LED should be more negative or equal to the output of the pot.
So, it shouldn't be switching the opamp used as a comparator.

It is probably just working because of a lucky input offset voltage.
Even using a different opamp on the same chip may not give the same result.
So, this is a risky circuit as it stands.
The first circuit, though would be OK if you can get enough sensitivity.

With your line follower, do you have 3 sensors so that if the robot goes off line, the robot controller can work out which way to turn to get back on track?

You can get better rejection of ambient light if you use a pulsed light source.
You can get 38 KHz receiver chips intended for TV remote controls. These give a low output if a signal of pulsed Infra Red light is received but high output otherwise.
They are good chips and easy to use.
http://au.mouser.com/Search/Refine.aspx?Ne=254016&N=1323038+11326342+4294930559
A 555 can drive an Infra Red LED at 38 KHz.

 

[aruna1]

Looking at your second circuit again, the 47K pot would never be able to give a negative output, so both states of the LED should be more negative or equal to the output of the pot.
So, it shouldn't be switching the opamp used as a comparator.

It is probably just working because of a lucky input offset voltage.
Even using a different opamp on the same chip may not give the same result.
So, this is a risky circuit as it stands.

that explains why it is changing time to time.

The first circuit, though would be OK if you can get enough sensitivity.

how to increase sensitivity?

With your line follower, do you have 3 sensors so that if the robot goes off line, the robot controller can work out which way to turn to get back on track?

i have 8 sensors in array

You can get better rejection of ambient light if you use a pulsed light source.
You can get 38 KHz receiver chips intended for TV remote controls. These give a low output if a signal of pulsed Infra Red light is received but high output otherwise.
They are good chips and easy to use.
http://au.mouser.com/Search/Refine.aspx?Ne=254016&N=1323038+11326342+4294930559
A 555 can drive an Infra Red LED at 38 KHz.

well then i have to send bursts of 38KHz IR signal right? that means 2 555 ICs per sensor (-->2*8=16 555 for complete array)
and with it how can i adjust sensitivity?


cant we use LM567 tone decoder with IR receiver diode?

 

[vk6kro]

The first circuit, though would be OK if you can get enough sensitivity.
how to increase sensitivity?

The obvious way is to get more light on the white line. There are some very bright LEDs available. I would try that before anything else. If you get a brighter subject and more output from your detector, the whole thing will be less noisy and more predictable.



You can get better rejection of ambient light if you use a pulsed light source.
You can get 38 KHz receiver chips intended for TV remote controls. These give a low output if a signal of pulsed Infra Red light is received but high output otherwise.
They are good chips and easy to use.
http://au.mouser.com/Search/Refine.a...342+4294930559
A 555 can drive an Infra Red LED at 38 KHz.

well then i have to send bursts of 38KHz IR signal right? that means 2 555 ICs per sensor (-->2*8=16 555 for complete array)
and with it how can i adjust sensitivity?

You probably only need one 555 and a bunch of switching transistors to turn the IR LEDs on and off at 38 KHz.
You don't need bursts of 38 KHz. Each LED should be just switching on and off at 38 KHz.


cant we use LM567 tone decoder with IR receiver diode?
Sure, but 567s are not as sensitive as these receiver chips. I tried one with a TV remote control and I couldn't get far enough away from it in the room to stop it receiving the signal.
These receivers have nice filters and amplifiers in them so they are fairly immune to ambient light. They also include the diode while a 567 would need you to supply one.

Not sure why you want to reduce sensitivity.
You can put up a shield so that the line does not get sun on it where it is being viewed by the sensors.
You can have each LED and each receiver looking down a black tube so it only covers a small area at the end of the tube.