Be aware that the material used in 3D printing may be transparent at IR wavelengths, or at least far from opaque.

I'll bet you would have more rapid success using an Exacto on some black cardboard and glue.

 

Be aware that the material used in 3D printing may be transparent at IR wavelengths, or at least far from opaque.

I'll bet you would have more rapid success using an Exacto on some black cardboard and glue.

Be sure and paint the inside with flat black paint.

 

Smoke detectors use the same principles as turbidity meters used in water quality measurements - i.e. 90-degree scattered light and often some arrangement with a second detector to determine the falloff in light source performance. Commercial instruments were almost exclusively IR.
To avoid a lot of faffing around with phototransistors and op amps why not try a ready made device like the TSL 257 types that come in visable and IR versions? 5-volts in, GND, analogue voltage out straight out of the box. Cheap, reliable - might get you off the ground, then try the op amp approach. In my experience, optical smoke detectors are best called dust detectors or spider attractants

 

Do you realize that a smoke detector that does this exact function costs less than 10 bucks at Walmart? (or on Amazon if no
Walmart is near you) Why not buy one and cut the guts out of it and be done with it!

ronna

 

Thank you all for information regarding the materials to make one. Ronsoy, i've been searching for weeks to find the chamber and diodes to purchase, so just like tearing a smoke alarm down but without the hassle. I can't find one. I think i may have to resort to your suggestion and then interface the smoke detector IC to the microcontroller or just use the op amp setup i have at the moment.

 

Do you realize that a smoke detector that does this exact function costs less than 10 bucks at Walmart? (or on Amazon if no
Walmart is near you) Why not buy one and cut the guts out of it and be done with it!

Just make sure that it's photodetector type and not an ionization type (which most are).

 

Just make sure that it's photodetector type and not an ionization type (which most are).

Indeed, thank you! I will come back with the results soon as I get something sorted.

 

So i have managed to get my hands on a purpose made enclosure for the optical section of my project.



I have salvaged this out of a photoelectric smoke detector and therefor it has its own phototransistor (I think it is anyway) and the diode already internally. I have attempted to reverse engineer the PCB it came on but i'm struggling to get a decent look on the traces as they're covered with black silkscreen and without removing all the components i can't remove the silk. So i'm reverting back to the circuit that we have been discussing here with the exception of only having a 1M trimpot rather than the suggested 10M. I have set up my prototype breadboard as follows:



I'm seeing a voltage of 2v on the output with no movement when i trim the pot around at all. Would the fact that i only have a 1M trimpot there rather than the 10M trimpot suggested be doing this? There is no movement on the output voltage what so ever, with or without smoke, even while moving the trimpot around. Below is an image of my actual workings in case this could point something out to someone....

 

If the output voltage does not move when the pot is adjusted them something is wrong with the circuit since you should be able to adjust now to zero volts.
Connect the op amp pin 3 to ground. That should give you near zero output.
If not then you have a faulty amp or there's a wiring mistake.

 

What I have seen looking at your circuit:

You have the wrong pinout for the TLC272. In addition, you should not leave unused inputs unconnected. For the op-amp that is not used you should connect the inverting input to its output and ground the non-inverting input. You need to add bypass capacitors on the power supply.

You have not shown all of the pin numbers on your schematic. You should _always_ show every pin number. Having all of the pin numbers shown (including no-connects) is a good way to help catch errors in your design.

If I were wiring this circuit I would only have one power bus and one ground bus. I would put the power on one of the buses along the top and the ground on one of the buses along the bottom. This makes it easier to follow the power and ground wiring.

 

I now notice that the schematic symbol for the OPAMP is indeed incorrect, thank you RichardO for telling me i have the pinout incorrect and making me check. I don't actually have the pinout incorrect, its just the schematic shows it is. I have reqired my opamp in a more clean manor so i can see it properly. I have added a 0.1uF and a 10uF cap to smooth the power to the IC. I'm not sure why the official schematic symbol doesn't show the vcc and gnd connections (Its in a the dual package on the schematic library and that is probably why) but the vcc is 5v and the gnd is gnd and connected on pins 8 and 4 respectively. The below image is the new schematic with the correct pin numbers on the part.



I have connect pin 3 to GND like you suggested and i got 0.15-0.9v on the output pin with everything still connected as it is above. Is this close enough to ZERO that you would expect to see @crutschow?

@RichardO - I have done as you requested on the open pins too, thank you.

 

I have added another 1M resistor next to R22 to make 2M resistance and then reduced R25 to 1K rather than 10k. Without any smoke its biased at 0.28-0.3v and the ADC reading of the output is reading anything from 0 up to 50. When smoke is introduced the ADC is reaching around 300, so i think i need to increase the gain even more, to make it more accurate and use the whole range of 1023 of the ADC (0v - 5v), what is your opinion? I could then have the arduino only trigger smoke alarm if it reaches over say 4v or around 900 on the ADC.

Would you expect to see a ADC fluctuate so much with a constant environment (nothing is changing around the enclosure)?

Do you guys have any arduino/adc experience? My multimeter is showing around 0.4v give or take 0.01 fluctuating on the output pin with no smoke, but my arduino ADC is showing 0.00v and spikes up at around 0.4v and back down every 5/6 seconds.

 

.............
I have connect pin 3 to GND like you suggested and i got 0.15-0.9v on the output pin with everything still connected as it is above. Is this close enough to ZERO that you would expect to see @crutschow?

The maximum input offset of the TLC272 is 10mV.
Multiplying that by the gain of 100 you have, gives an output offset of 1V maximum.
But how does it vary from 0.15-0.9V?
It should be a steady DC voltage.
It could be noise spikes or oscillations causing the fluctuations.
Do you have access to an oscilloscope?
What output do you read on the AC scale of the multimeter.

 

The maximum input offset of the TLC272 is 10mV.
Multiplying that by the gain of 100 you have, gives an output offset of 1V maximum.
But how does it vary from 0.15-0.9V?
It should be a steady DC voltage.
It could be noise spikes or oscillations causing the fluctuations.
Do you have access to an oscilloscope?
What output do you read on the AC scale of the multimeter.

Unfortunately i don't have the chance to get an oscilloscope at the moment. I'm guessing its noise of some fashion on the breadboard and jumper leads. Are you asking me to set my multimeter to AC and put it on the output pin?

 

Unfortunately i don't have the chance to get an oscilloscope at the moment. I'm guessing its noise of some fashion on the breadboard and jumper leads. Are you asking me to set my multimeter to AC and put it on the output pin?

That is correct.
If the circuit is oscillating you should see some AC voltage.

 

That is correct.
If the circuit is oscillating you should see some AC voltage.

Okay, soon as i get home i will be checking that and reporting back. Thank you for your continued support dude!

Just to clarify here, i'm putting one probe onto the output and the other onto GND, with the multimeter in AC mode, right?

 

Okay, so with one probe on the output and the other on GND, its reading 00.7VAC. Is this enough of a AC voltage to indicate that its oscillating? The sensor is working fine, i need more gain on it though, i'm using 2M and 10K on the feedback line but its still not enough gain, i have to have a lot of smoke there to send it anywhere above 1V. Its biased to 0.4v without smoke. Would you change the 2M or the 10K resistor to get more gain in this situation? I thought to increase the 2M but its not altering it much, how high would i realistically need to go? :O

The ADC is still fluctuating, maybe this is has some sort of relevance to the 00.7VAC being shown....

When smoke is applied the output is reading what i would expect it to on DC settings, its rising to around 2.5v AC with smoke

 

Hello, I think you might find this article interesting:

http://www.digikey.ca/en/articles/t...sensors-for-monitoring-air-quality-and-safety

 

Hello, I think you might find this article interesting:

http://www.digikey.ca/en/articles/t...sensors-for-monitoring-air-quality-and-safety

Thank you for the linked article. I will read that through in the morning over breakfast.

May i ask what your opinion of my last post is, if you have one that is.

 

Hello, I think you might find this article interesting:

http://www.digikey.ca/en/articles/t...sensors-for-monitoring-air-quality-and-safety

So this is an interesting read this morning. I may refer back and forth to this for some investigations, thank you.

Would you recommend me increasing the 2M resistance or the changing the 10k resistance to increase the gain of my opamp and also, do you think that 00.8v on the AC measuring shows an oscillating output?