Photodiode amplifier

Is the amplifier saturating?

What purpose does the current/voltage converter serve? Wouldn't V proportional to logarithm-of-current output be more useful? Alternately, if you're going to mix down a modulated signal, can you just steer the photodiode current with a Gilbert cell mixer?

Well I was working with the Seattle Robotics Society on this too, and the person said I am driving it into saturation. He never said on how to fix it. Because the photodiode has high impedance, it'd be hard to interface it directly to the stages of amplification. I've never heard nor I have any ideas on "logarithm-of-current" (Could someone please tell me what it is?).

And one more thing, I am a middle school student working on my project.

Ray Xu wrote in news:ef831cf3-db5f-413c-9473- snipped-for-privacy@z66g2000hsc.googlegroups.com:

Think of the photodiode as a 'light to current' transducer. You probably want a 'light to voltage' transducer, so you need to feed the current to a resistance or, alternatively, to a 'transimpedance' amplifier. Of course, like most current sources, photodiodes have their limitations.

The first limitation to overcome is the fact that photodiodes have a semiconductor junction in parallel with the current source. If you allow much voltage to develop across the junction in its forward direction, most of the photocurrent will shunted through the junction rather than into your amplifier. If you maintain a reverse voltage across the photodiode, you're in 'photoconductive' mode (which also has the benefit of reducing photodiode capacitance, more later). However, reverse voltage also gives rise to leakage currents, which can be a problem in some systems.

So, another way to keep the junction from conducting is to keep the voltage across the photodiode low, preferably near zero. For obscure reasons, you are then said to be operating in 'photovoltaic' mode. The classic negative feedback 'transimpedance amplifier' is often employed for that purpose, and I'm guessing that's your circuit. The circuit can be as simple as a photodiode, an opamp, and a feedback resistor (plus power supply, bypass capacitors). This circuit is 'driven into saturation' (as some people would say) when the output voltage of the opamp reaches its maximum or minimum level permitted by its internal design and the supply voltages. Since the output voltage of a simple transimpedance amplifier is given by V = Rf * I, where Rf is the feedback resistance and I is the input current, you can 'bring a transimpedance amp out of saturation' by reducing the feedback resistance. In your case, this might involve reducing the 100K resistor to 10K.

However, I've left out a very important part of amplifier design: stability. Simple transimpedance amps are generally unstable at high frequencies, especially when the input source is a large area photodiode operating in photovoltaic mode. The instability has to do with the way that the capacitance of the photodiode slows down the negative feedback. Oscillation of the transimpedance amplifier is often mistaken for some other kinds of problems, and you should be sure that all of your amplifier stages are stable before going much further. The simplest way to tame oscillations in this case is to add a parallel capacitance across the feedback resistance. Usually, a capacitance that is a small fraction of the photodiode capacitance will settle things down.

It's best to use an oscilloscope and a proper probe to sniff around your circuit. In the dark, every point in your circuit should be 'quiet', with only dc levels on any opamp pin. If you let your photodiode 'see' some fluorescent light fixtures, you might find some 120 hertz 'ripple' showing up. Any other kind of oscillations mean you have stability problems.

After you've quieted things down (if necessary), tell us more about what you're trying to do and come back with a better description of the optics. At that point, we'll be in a better position to suggest things like feedback resistances, number of stages, lensing, frequency response, etc.

Paul Mathews

From this post alone i already expect that your problem area is optics not electronics. Try working on the optics please.

The basic diode has a log transfer function, and the transfer function is voltage as log of current. Make of it what you will.

Sounds like something to pass on to S.E.K ;-)

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
  My simulation, struggling to converge, is telling me something
    Maybe my idea of a ring of current mirrors is flawed ?:-)

Thanks Paul for the reply. I'll go look for some transimpedance opamps on Monday. Just yesterday, I re-looked on the concept of photoconductive and re-designed the current to voltage converter to be in photoconductive mode (I need it to be like this because I am making a laser range finder; this portion of the circuit needs to be precise and fast). For the stability part, I haven't seen this problem at all since I started building this, but I added on some 10pF caps just in case it ever does happen, for some reason. There's also a 50Hz lowpass filter I designed before and I'm planning to add it to the receiver's end. For now, I'm not going to be modulating the laser above 20Hz. And to avoid saturation in the opamp stages, would it be best to have more stages of opamps and to cascade them, or to try and put all the gain in 1 or 2 stages of amplification?

Anyways, I'm building a phase shift-based laser range finder. On the receiver side, I have a 555 built oscillating at about 10Hz (eventually going to replace it with a MCU). That 10 Hz is used to drive a LA modulation-capable laser diode module from Meredith Instruments. For the phase shifting part, I am using a THS3202 (mounted on a adapter board) as a high-speed comparator. I choose this because of it's high slew rate, bandwidth, and that the rise and fall times is 480 pS. The THS3202 will compare the sent modulation against the received modulation, and the received modulation should have some lag proportional to the distance from the object. On the comparator's output will be a pulse train, and (if I remember correctly) a DPRG

said to feed it through a RC network which will turn it into a analog voltage and where a ordinary MCU can read it and determine the distance. And for tuning, I may add a programmable delay chip in between the laser modulation wire and the output of the 555 (or MCU) to "cancel out" the circuit delay. For the hardware, I'm going to mount the laser and photodiode parallel to each other but on the same "level", and the object is going to be placed perpendicular to the laser and photodiode. The convex lens will be placed on top of the photodiode. I hope this helps.

Ray Xu wrote in news:59b3ca8c-a90a-40a6-8c63- snipped-for-privacy@m36g2000hse.googlegroups.com:

Although you can buy 'transimpedance amps', the usual approach is to simply use a low-noise opamp.

Just yesterday, I re-looked on the concept of

Your modulation frequency, if I understand, is just 10 Hz, so it's not clear to me why anything needs to be fast. On the other hand, if you're planning to have a ranging resolution on the order oa centimeter (for example), your time resolution needs to be on the order of 60 picoseconds. In other words, for such a low modulation frequency, you need incredibly high phase resolution.

. For the stability part, I haven't seen this problem at all

You haven't said anything yet about the size of your photodiode and/or its capacitance. If you were getting away with no feedback capacitance, I'm guessing your photodiode is small. Large lens plus small photodiode means very precise aiming is required. Maybe that's why you got no signal.

Plus a comparatively huge lag produced by the slow receiver.....

On the

From what you've told us so far, I'm guessing that you'll be needing a modulation frequency in the 10s megahertz, with correspondingly fast receiver, but there are many unanswered questions. Paul Mathews

d e

It needs to be fast so there is the smallest possible lag in the circuit. And yes, the PNZ334 has a pretty small junction.

Two questions: what comprised the "seperate test" and what do you mean by fast?

Hul

Lay you dollars to donuts that the amplifier string is railed, thus "it is not sensitive enough".

This thread belongs on the kindergarten group ;-)

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

Obama, Questioned about the Presidency, it\'s "above my pay grade"