Phototransistor VS Photoresistor
idbruce
Posts: 6,197
Hello Everyone
I want to be able to measure low light levels very quickly, and I was wondering which would be better, a phototransistor or a photoresistor. This thread may include discussions about connectivity of these devices to the Propeller chip.
Bruce
I want to be able to measure low light levels very quickly, and I was wondering which would be better, a phototransistor or a photoresistor. This thread may include discussions about connectivity of these devices to the Propeller chip.
Bruce
Comments
Sorry, you know how the forum works
To the guys at CERN, femto seconds is 'quick', so you will need to clarify that.
At low light levels, sensor area and spectrum start to matter more, so those may decide for you.
Photo transistors are less ideal than photo diodes, but can give a solution, depending on what your 'low' really isq
I've seen solar cells used, as zero voltage current sources, for low light sense.
(this removes junction leakage from the equation, and gives a large collection area)
Yes, I know how the forum works
Okay, first off, I really do not need to measure the light, I only need to detect any light level above a weeny teeny little bit of total darkness to direct sunlight (so I suppose low was not a very good description), and I do suppose that the light source could be something other than natural lighting, such as lighting from an incandescent, halogen, or fluorescent lamp.
Here is a perfect example of similar situation
you gave a picture that I can imagine. But are these the specs?
I try to specify: level of light some milli-Lux (a night light)
responsetime: I'm not a biologist specialised on snake eys. But I would gess it takes something around a 50-100 millisecs for the snake to become aware of the light.
Maybe you want to film the snake. So I guess one second would be still enough to start the camera filming.
So the specs would be 0,01 Lux of "brightness" and responsetime 100 milliseconds.
I guess the bigger problem will be light-level. At such low levels phototransistors and photodiodes will have only a very small response.
as I was a student I made a project that used a laserdiode sending light to a wheel in a usual watermeter counting the turnings of the wheel.
Even for that I used a transimpedance OpAmp circuit to get a signal that I could use.
If you got the chance to measure the lowest light-level with a luxmeter measure it.
What do you estimate in milliseconds how long will the light-pulse be?
Is this a superprivate project you want to hide carefully? If not just tell us what you really want to do.
best regards
Stefan
Thanks for the response.
I apologize for being vague, but to answer your question, yes it is one of those "superprivate projects".
I know the defintion of a millisecond, but trying to imagine the speed breakdown is beyond my human comprehension. However, if I was to guess, I would think that I am shooting for about 20 ms or somewhere in that ballpark, because referring back to my vague example, even though the light in the bathroom will remain fairly constant, the snake may sneak out very quickly and into the shade provided by the toilet seat.
As for the specs of the lums, all I can say at the moment is that it could be pretty dim. Wish I could be more specific.
Bruce
For photo resistors the number of charge carriers is a balance between pairs created and pairs that recombine - the slower the recombination the more sensitive the device is (again low light means slower performance).
Also there may be ambient-light monitoring chips that already handle low light well - worth searching for.
Don't forget that LEDs are nice photovoltaic sensors too.
Here are several circuits:
Beam Circuits
Some of these can see a candle across a room.
Duane J
Here are two LDR circuits that I have been looking at:
http://www.youtube.com/watch?v=LQImmpcFDBw&feature=related
http://www.analog.com/library/analogDialogue/archives/45-07/light_sensor.html
Thanks for the link Duane
Try this:
This is very sensitive.
The 10M resistor is my Digital Voltmeter on the volts scale.
Sensitivity is easily adjustable by varying R1.
This is a photovoltaic cell, (Actually a red LED), running into a relatively
constant voltage of about 0.6V and is quite linear and repeatable.
Unlike the non-linear Cadmium Sulfide photo-resistive cells.
And don't age.
The spectrum is from about red and up if the case is clear.
If interested in a narrower spectrum slice use an LED with a colored case.
I just tested the circuit with cheap 2N3904 NPNs. However, if high temperature
operation is desired a low leakage current transistor should be used in Q2.
"Try it, you'll like it."
Duane J
Thank you very much for taking the time to provide a schematic. I sincerely appreciate it. I just happen to have all the necessary parts, so it will be the first circuit I test.
It's people like you that make this a wonderful online community.
Bruce
LEDs, depending on the package, have what is essentially lensing effects.
Maybe you want a narrow field of view.
If you file the end down the acceptance angle gits wider.
You can also use multiple LEDs in parallel and angled in different directions
to get an omnidirectional acceptance angle.
Duane J
There are a couple of ways to hook up directly to the Prop. The following is a simple light meter that uses the green LED as both a light meter and an indicator. The LED flashes at a rate proportional to the ambient light level. It can work down to low light levels, on the 1 lux level, but is not "fast" as you define it. I'm attaching a couple of demo programs for that.
Higher speed can be had with a sigma-delta connection:
The sensitivity can be set with R2. It is considerably harder to detect a small optical signal riding on a large ambient.
Thanks for the response and schematics. After looking at numerous circuit examples, I came to the conclusion that I really did not need to use a uC to accomplish my task, however I can certainly understand that there will be times when a light sensor should be monitored by a uC, and I am sure the schematics given can prove useful to someone. I am now leaning towards a circuit similar to the one shown in Section 7, Page 9 of the following document: http://www.produktinfo.conrad.com/datenblaetter/175000-199999/192296-an-01-en-CONRAD_LERNPAKET_ELEKTRONIK_BASIC.pdf
Thank you very much for your input Tracy
Bruce
I have done extensive tests and experiments with LEDs and photo diodes. (Since I use them as sensors in my solar trackers.)
There is little material difference in the current flowing out of the anode of the device whether reverse biased or
self forward biased as long as the voltage stays less than the LED's junction voltage.
Compare my circuit and that of the one presented in your reference.
Notice, in both cases the anode is connected to the base of the first transistor.
In both cases the current eminates from the anode and enters the base in the same direction.
Hmmm, the same direction!!!
As an experiment, I have connected the cathode of the LED to a variable power supply, and a sensitive nanoamp meter in series.
I find the photo current is essentially independent of the voltage bias across the LED whether reverse biased or forward biased.
Of course, the range of this voltage is reverse breakdown and forward PN junction voltage where the junction starts emitting light.
In conclusion, there isn't really any difference between a photovoltaic diode sensor and a photoconductive diode sensor.
There is one disadvantage of reverse biased mode though. The greater the voltage the greater the leakage current and the leakage current is highly temperature dependent.
This leakage current is independent of the photo current but is indistinguishable from it.The effect is to limit the lower sensitivity of the sensor. And the temperature thing means the lower light level calibration drifts with temperature.
Conversely, as the voltage across the sensor approaches 0, the leakage current approaches 0, and calibration errors due to leakage also approach 0.
Ok, my circuit has the sensor voltage at about 0.6V (1 emitter base junction) so the leakage current is quite low.
Anyway, I just wanted to explain the finer details of the operation and characteristics of these interesting devices.
Duane J
Thanks once again for all the responses.
As mentioned in numerous other locations throughout the forum, my knowledge of electronics is highly deficient, and I do intend to remedy this deficiency in the near future, but I currently do not have the time to set aside for an in depth study of electronics. When schematics and part values are given, I have no problem throwing some circuitry together, but if there are any unknowns, then I start to have problems. That being said, I did attempt your circuit, however I must freely admit that I was having problems understanding the full concept, as well as knowing where to monitor for changes and what changes to expect. Additionally, I could not comprehend the necessity for both a battery and a VDD supply.
As for the other circuit that I referenced, the only reason that I lean in that direction is that it provides a visible indication of the presence of light and it does not require a uC for monitoring. However, it is worth noting that the referenced circuit is far from perfect. In an ideal situation, I would prefer to have an ideal trip point with very low lums, in which case, if light is detected, it supplies power to another LED to show that light has been detected, and if darkness is detected, then it disconnects the power to the indicating LED. In order to obtain this kind circuitry, I will have to do a little experimenting and I am sure quite a bit more research.
Thanks again for all your input Duane
Bruce
These posted schematics were intended to be circuit "Snippets" representing concepts more than final circuits. Basically the LED generates a small photo current proportional to the light intensity.
This current is applied to the emitter/base of Q2 which amplifies it and applies it to Q1 which amplifies it further. Probably about 10,000 times total. Plenty of current for input to a high impedance Prop pin.
However, if driving an LED in a dark environment another gain stage is required.
This circuit adds Q3 and an LED indicator.
Note, you need to experiment with the value of R1 for setting your required sensitivity. The higher the value the greater the sensitivity. Their one and the same. I just wanted to label the battery as VDD if you were using the 3.3V from a Prop supply.
It's common to add labels in schematics for clarity.
Actually this circuit can operate from any voltage from about 2V to 12V or or more. Your referenced circuit probably wouldn't work for you as it requires a fairly high intensity to turn on the LED. It would need another gain stage for your dim light application
Sharp trip points are a bit harder. This example is analog in nature so LED2 varies in brightness.
Do you really need a trip point? I could work up a Schmitt trigger circuit with 2 more transistors if you need this.
If you want to try the LED sensor in reverse biased mode in my circuit simply move the cathode of LED1 to Vdd.
(Be careful with higher voltages though as some LEDs have limited breakdown voltage ratings.) That is what forums are about a give and take of information.
Duane J
The big advantage of reverse biasing a photo-diode (or LED) is that it reduces the diode's parasitic capacitance. This is very important if you're designing a circuit for speed.
I've personally played a lot with the standard Op-amp circuit for amplifying photo-diodes. (see the attached schematic) My favorite photo-diode to use is the BPW34 as it's large. Really, any LED, solar-cell, or even a glass-case switching diode will work. Op-amp selection is important for best results with this circuit. I've had good results using a OPA350 FET input rail-rail op-amp. The feedback resistor sets the gain of the circuit. A 90Mohm feedback resistor and BPW34 photo-diode can detect my movement in a room lit only by a small LCD screen. With a cheap DMM that reads to 0.1mV it will pick up 2-3pW or greater changes in incident power. Bandwidth of this circuit is about 10KHz, but that can be cut down with a capacitor in parallel with the feedback resistor.
Another favorite circuit is an amplified PV sensor. I again use a BPW34 and either directly measure the open-circuit voltage with a DMM, or amplify the open-circuit voltage with a non-inverting op-amp circuit. (fet input op-amp, and gain of 5 in my current design) In this case, the output voltage is proportional to the log of the incident optical power giving useful outputs from pitch black to full sun. (i.e. massive dynamic range)
Lawson
In general, all the really fast circuits force the voltage across the diode to 0V. Just as that in your circuit example. While the capacitance is higher the op-amp forces the the voltage to be 0V. Since the voltage doesn't change the large capacitance doesn't hinder fast rise times, assuming the op-amp is up to snuff.
Of course reduced capacitance is always desirable. There are specialized photodiodes called PIN, (P-Intrinsic-N), diodes. These have the P and N junction doping regions separate by an Intrinsic or un-doped region. This greatly reduces the capacitance without materially reducing the photon to electron conversion efficiency. These can have pS current rise times.
I agree, this is a very nice device.
However, since its silicon it can't be used in photovoltaic mode to directly drive a silicon transistor in the simple circuits without an op-amp.
In my other life with "Kansas City Space Pirates" we used a 4 quadrant PIN diode sensor to track a 6.4 KHz LED beacon at 1 KM distance in bright sunlight. The light we were sensing was vanishing small. This used tracking quadrature detection amplifiers at 6.4KHz with a bandwith of 0.1Hz to 1Hz.
Duane J
Hey Bruce, you sure you are not living in Az or Guam?
If you really want sensitivity and fast response, go old school and use a PMT. Google photomultiplier tube and response. Multiple hits including Hamamatsu with a manual of their offerings.
FF
I will try your 3 transistor stage for reading a phototransistor. Actually, just with a 10k resistor, I get 50mV when dark, and 100 to 250mV when lit.
and want to connect it to a propeller MCU.
But when I am near a light bulb, I get full 3.0V. So, if I emplify with your circuitry, what will happen if I allow the detector see daylight?, Will something fry?
So, I was asking myself if I need to limit it. I am not an expert at all, but I was thinking about an 555 IC driving a transistor that would block when voltage reach lets say 2.8v.
Does it sould good or there is better simpler? (I do not understand well zenner diodes or maybe there is a component that would do the protections by itself)
Thanks in advance for advices.
Marc (EMHmark7)
Duane J
I am planning to make a bb bullet detector which uses bpw34 with 5mw red laser.
First of all we need to figure out the fastest detecing speed of circuit. The 6mm bb would flies as fast as 140m/s. Anyone knows if we can calculate the period which passes the sensor?
My concern is if I can make the circuit as simple as possible for example the 3904x2 circuit.
Any idea?
Thanks.
This will have skirts, so be _/=\_ in nature, as the round BB eclipses the source.
43 microseconds is getting fast, and to hit best speeds, you need to avoid voltage swings across the sensor device.
that usually means no BPT, but instead a fast Opamp as shown above.
To get good skirt slicing, you may want to also average the not-blocked linear value, and set a slice to (say) 50% of that.
(that can be a simple RC +2:1 divider)
That compensates for Temperature and sensor gain variations.
You would still need to carefully calibrate some 'apparent size' into the speed calculation, but that might be possible with a spinning test target.
For the laser receiving windows I opened some 2mm holes and hide the sensor 2cm behide the shell to avoid lights from outside. The circuit is drove by a lipo battery for sensor circuit with 100uf cap, arduino and lasers so I believe voltage swing is very low. At this moment circuit is powered by arduino and arduino is powered by computer. $ is not a big concern so which Opamp should I use? (of course don't suggest me to use pricy audio Opamp)
Moreover, I used 74hc4067 to receive that many of sensor signals, for example 32x32. With the all in one sensor I tested IS0103 from china market, when laser is blocked the 3 pin chip output gives a digital high signal. From datasheet which rise time is only 0.1us. But my circuit is only fine to detect a bb throw by hand. Programming I used 74hc4067 libary with arduino uno. 8 sensor would use 5 pins on arduino. If connect all 64 sensors I would need 2 uno connect by 3wire.
IS0103
https://img.alicdn.com/imgextra/i2/430959460/TB2.k3dcFXXXXbAXXXXXXXXXXXX-430959460.png
It is kind of success to detect slower objects pass though the sensors but it is too slow for a shooting target. First of all I want to change the sensor to bpw34s. Since I am not sure if the IS0103 speed in datasheet is real and if so, bpw34s speed is faster than that very much. And for the 74hc4067 mux, anyone thinks it is suit for this application or not? Finally the programming, I tried 74hc4067 libary with uno. Should I use teensy instead of uno for a faster speed??
Rise time is only one spec, and probably the least important.
the Tplh,tphl and skews matter more, and those in the typical column of 2us and 3us are ok for a 43us transit.
Did you try that on a storage scope ?
I'd place an opamp+Slicer right next to each sensor, and then you should be ok to MUX the digital outputs.
This does mean you miss signals on non-selected channels.
Is that tolerable, or would you prefer to capture many at the same time ?
If you know the trigger order & gaps, you could do a ripple-scan design.
Depends what precision you seek.
The transit looks ~ 43us, so 430ns is 1% and 43ns is 0.1%.
A Parallax Propeller can time to 12.5ns to 32bit precision, and do that with a max of 16 timer channels, or more if you use software.
I will try to connect a storage scope for a test within this week.
After discussed with friends they suggested me to add a logic latch on each sensor before mux. It sounds easy to solve the insufficient of speed problem.
http://blog.svenbrauch.de/2017/02/19/homemade-10-mbits-laser-optical-ethernet-transceiver/