Photodiode Recieving circuit - Troubleshooting - Jittery output

[[email protected]]

Hi all,
Built a photodiode circuit (refer schematic Figure 1 - Link below) to
recieve pulses - square wave on "on" period 10us with varyin duty
cycles incident on the photodiode(BP104).

On testing the circuit, found the output at the trans-resistance
amplifier had some ringing at the edges (refer 2nd figure - 1st
response). The further the distance of the incident signal from the
transmitter, the more pronounced the ringing.

Tried to do some trouble shooting. Found the base of the transistor in
the gyrator circuit(R1, Q1, C1, R2) was not stable. Tried increasing C1
to 470nf to help fix this.. but still have the same problem,.
Also noticed C2 had a time constant associated with it.. with the same
response as at C1 (refer figure 3- 2nd response). So tried increasing
C2 to 1uf (non-polarised C), but still of no avail.

SO when the signal is near- clean square wave - can see C2 charging and
discharging. ( figure3)
As signal gets further - ringin at edges - response at C2 is almost a
flat line ( figure 2)
any ideas as to wot is causing this problem?


* please click on link for all figures and detailed explanantion:
http://generous.boy.googlepages.com/photodioderecievingcircuit
was unable to upload onto this page -sorry for the inconvenience

 

[colin]

Hi all,
Built a photodiode circuit (refer schematic Figure 1 - Link below) to
recieve pulses - square wave on "on" period 10us with varyin duty
cycles incident on the photodiode(BP104).

On testing the circuit, found the output at the trans-resistance
amplifier had some ringing at the edges (refer 2nd figure - 1st
response). The further the distance of the incident signal from the
transmitter, the more pronounced the ringing.

Tried to do some trouble shooting. Found the base of the transistor in
the gyrator circuit(R1, Q1, C1, R2) was not stable. Tried increasing C1
to 470nf to help fix this.. but still have the same problem,.
Also noticed C2 had a time constant associated with it.. with the same
response as at C1 (refer figure 3- 2nd response). So tried increasing
C2 to 1uf (non-polarised C), but still of no avail.

SO when the signal is near- clean square wave - can see C2 charging and
discharging. ( figure3)
As signal gets further - ringin at edges - response at C2 is almost a
flat line ( figure 2)
any ideas as to wot is causing this problem?


* please click on link for all figures and detailed explanantion:
http://generous.boy.googlepages.com/photodioderecievingcircuit
was unable to upload onto this page -sorry for the inconvenience

Assuming your light source is a clean squarewave ...
R2 seems a bit low I would try 1k if the gyrator is cuasing trhe ringing,
and maybe a small capacitor in the + input to the op amp might help.

the other possible problem is op amp instability cuased by the phase shift
in the feedback path cuased by R5 and the PD capacitance,
a solution is to put a small capacitor accros R5 to compensate although this
reduces the high frequency response,
another sugestion made here a while ago is to put a resistor in series with
the PD,
if these arent attractive a better alternative might be a faster op amp, or
if it works might be to put a damping network on the output of the op amp
.... resistor load (in series with a capacitor) this just reduces the op amp
gain and so gives a bit more phase margin.


Colin =^.^=

 

[Ban]

Hi all,
Built a photodiode circuit (refer schematic Figure 1 - Link below) to
recieve pulses - square wave on "on" period 10us with varyin duty
cycles incident on the photodiode(BP104).

On testing the circuit, found the output at the trans-resistance
amplifier had some ringing at the edges (refer 2nd figure - 1st
response). The further the distance of the incident signal from the
transmitter, the more pronounced the ringing.

Tried to do some trouble shooting. Found the base of the transistor in
the gyrator circuit(R1, Q1, C1, R2) was not stable. Tried increasing
C1 to 470nf to help fix this.. but still have the same problem,.
Also noticed C2 had a time constant associated with it.. with the same
response as at C1 (refer figure 3- 2nd response). So tried increasing
C2 to 1uf (non-polarised C), but still of no avail.

SO when the signal is near- clean square wave - can see C2 charging
and discharging. ( figure3)
As signal gets further - ringin at edges - response at C2 is almost a
flat line ( figure 2)
any ideas as to wot is causing this problem?


* please click on link for all figures and detailed explanantion:
http://generous.boy.googlepages.com/photodioderecievingcircuit
was unable to upload onto this page -sorry for the inconvenience

What is the purpose of this "gyrator"? Did you do some calculations
regarding the expected photocurrent and -bandwidths.
IMHO this circuit just amplifies the ripple on the supply(which is injected
into the opamp input via R1 and C1 and amplified through the emitter) just
the opposite of your intention. If the circuit is suppüsed to filter the
supply, C11 must be grounded instead of going to the input. You rather use a
polarized 100u in this case, and increase R2 to 10k-100k, depending on your
desired sensitivity. You can even just use the resistor alone.

 

[colin]

What is the purpose of this "gyrator"? Did you do some calculations
regarding the expected photocurrent and -bandwidths.
IMHO this circuit just amplifies the ripple on the supply(which is injected
into the opamp input via R1 and C1 and amplified through the emitter) just
the opposite of your intention. If the circuit is suppüsed to filter the
supply, C11 must be grounded instead of going to the input. You rather use a
polarized 100u in this case, and increase R2 to 10k-100k, depending on your
desired sensitivity. You can even just use the resistor alone.

The cap needs to be accross the vbe so that it provides a constant current
source above some freq at wich it cant amplify supply noise, below that freq
the current adapts to the light on the PD and passes any supply noise.

whatever is used to bias the PD, supply decoupling needs to be good enough.

Colin =^.^=

 

[[email protected]]

What is the purpose of this "gyrator"?
Did you do some calculations
regarding the expected photocurrent and -bandwidths.

the gyrator acts as a constant current source. It has a time constant
such that it is not fast enough to react to the incident pulse of IR
(10us), hence giving me a low (square wave) signal at the photodiode.
It serves to filter out ambient light or any other form of IR
interference.

IMHO this circuit just amplifies the ripple on the supply(which is injected
into the opamp input via R1 and C1 and amplified through the emitter) just
the opposite of your intention.

i wouldnt think it would not amplify the noise. The noise would have to
get through R1 and the noise would be filtered out by the large
capacitance C1

 

[Robert Baer]

the gyrator acts as a constant current source. It has a time constant
such that it is not fast enough to react to the incident pulse of IR
(10us), hence giving me a low (square wave) signal at the photodiode.
It serves to filter out ambient light or any other form of IR
interference.





i wouldnt think it would not amplify the noise. The noise would have to
get through R1 and the noise would be filtered out by the large
capacitance C1

Above some frequency, C1 acts like a short; then R1 is fully active,
injecting PS noise int U1; it is not balanced by the attenuation of that
noise via R3, R4.
Furthermore keeping Vbe constant is a rather poor and unreliable way
to have "constant current", as transistor current can change due to
small temperature changes.
The circuit needs major surgery and perhaps radiation therapy to get
rid of the cancer.

 

[[email protected]]

Furthermore keeping Vbe constant is a rather poor and unreliable way
to have "constant current", as transistor current can change due to
small temperature changes.

Hmm checked on that too.. 2.5mV/degree celsius..
Im only planning to use the circuit at ambient temperature, so dont
think that would be a problem.

The circuit needs major surgery and perhaps radiation therapy to get
rid of the cancer.

any suggestions?

The range i require is only 5 cm.

I am transmitting unique ID's:
10us on, 10 us off - ID 1
10us on, 20 us off- ID 2
and so on..

At the moment i am just simply turnin the transmitter on and off with a
fixed "on" period an duty cycle. It is controlled by a simple program
toggling the pins of a micro which is controlling the mosfet to the
transmitter LED.
Photoreciever modules work on carrier frequencies, which would mean i
would have to modulate my transmitted signal with a carrier frequency.

so using a module would mean i wld have to shift away from the workings
i have until now.
Or is it an easy shift?

 

[colin]

Hmm checked on that too.. 2.5mV/degree celsius..
Im only planning to use the circuit at ambient temperature, so dont
think that would be a problem.

The temperature changes arnt a problem anyway
as the average current through the PD sets the bias point it is not effected
by change in vbe.

The psu noise injected into the op amp via R1 C1 is limited in frequency and
is half canceled out by R3/R4
this could be more completly canceled out by a capacitor accross R3
carefull choice of the right value is needed to cancel at all frequencies,
you might need a resistor in series with that and another smaller capacitor
(30pf) acrros R4 (this is probaly needed anyway)
or more simply just decouple the supply to the input with suitable RC
network.

any suggestions?

Its realy not that bad, just needs to pay attention to the op amp feedback
response to get rid of the ringing by taking acount of the PD capacitance.

The range i require is only 5 cm.

I am transmitting unique ID's:
10us on, 10 us off - ID 1
10us on, 20 us off- ID 2
and so on..

At the moment i am just simply turnin the transmitter on and off with a
fixed "on" period an duty cycle. It is controlled by a simple program
toggling the pins of a micro which is controlling the mosfet to the
transmitter LED.
Photoreciever modules work on carrier frequencies, which would mean i
would have to modulate my transmitted signal with a carrier frequency.

so using a module would mean i wld have to shift away from the workings
i have until now.
Or is it an easy shift?

You might want to have a more robust encoding sequence to ignore errors due
to noise/glitches (ie some redundancy) or filter out the glitches in
software.

Colin =^.^=

 

[[email protected]]

The psu noise injected into the op amp via R1 C1 is limited in frequency and
is half canceled out by R3/R4

Could u please explain as to how that is? i dont think i quite
understand as to how.

this could be more completly canceled out by a capacitor accross R3
carefull choice of the right value is needed to cancel at all frequencies, you might need a resistor in series with that and

Do u mean a de-coupling capacitor from supply to ground just before R3?

another smaller capacitor
(30pf) acrros R4 (this is probaly needed anyway)

Y would i need that?

or more simply just decouple the supply to the input with suitable RC
network.

In the meantime i was doing some reading and am going to try a reserve
capacitor (47uf), from the point of supply to the circuit to ground. it
would serve to filter out any noise coming in from the supply.
does this sound like a good idea?

You might want to have a more robust encoding sequence to ignore errors due
to noise/glitches (ie some redundancy) or filter out the glitches in
software.

Yea i have written some simple noise detection software.. Using Pulse
Position Modulation for transmission.

GB

 

[Jamie]

the gyrator acts as a constant current source. It has a time constant
such that it is not fast enough to react to the incident pulse of IR
(10us), hence giving me a low (square wave) signal at the photodiode.
It serves to filter out ambient light or any other form of IR
interference.





i wouldnt think it would not amplify the noise. The noise would have to
get through R1 and the noise would be filtered out by the large
capacitance C1

maybe it won't help, but did you try a little hysteresis in the circuit?
that would be a small feed back from the OP-AMP out to the + input of
the OP-amp.

 

[colin]

Could u please explain as to how that is? i dont think i quite
understand as to how.

Sure ... the voltage on the PD is held just below vcc so any vcc supply
noise apears here too,
although it is cutoff at LF by the coupling capacitor,
and at hf by the PD capacitance.

The R3/R4 divides the vcc by two and hence the supply noise too, therefore
the op amp sees all the supply noise on one input and half the noise on the
other and the difference is half the noise.
frequencies, you might need a resistor in series with that and

Do u mean a de-coupling capacitor from supply to ground just before R3?

no, a capacitor accros R3 would couple all the supply noise into the op amp
input so it sees the same noise on both inputs wich cancels out with the
noise on the other input.

as the noise drops of with increasing frequency due to the capacitance of
the PD so the noise coupled into the other input needs to be reduced to
balance it.

Y would i need that?

To improve the high frequency response, this may be the cuase of the ringing
anyway.
with a high impedance on the + input at hf any capacitive coupling is a
potential problem.

In the meantime i was doing some reading and am going to try a reserve
capacitor (47uf), from the point of supply to the circuit to ground. it
would serve to filter out any noise coming in from the supply.
does this sound like a good idea?

could help, but a resistor in series with the supply to the PD and R3/R4 as
well as a cap down to ground will block more noise.


Colin =^.^=

 

[[email protected]]

Just found that adding a capacitor from pin 3 (biasing - non-inverting
input) of the op-amp to ground, almost halved the noise.
I am happy it works, but any ideas as to why it worked?

thanks

 

[Mike Monett]

The photodiode capacitance makes the noninverting gain of the stage rise
to a huge peak at high frequency, so any junk on the noninverting input
gets amplified enormously, along of course with the amplifier voltage
noise--which is the primary limitation on TIA performance at low signal
levels.

Cheers,

Phil Hobbs

This also shows up in SPICE. What is your recommendation to cure the
problem?

Regards,

Mike Monett

Antiviral, Antibacterial Silver Solution:
http://silversol.freewebpage.org/index.htm
SPICE Analysis of Crystal Oscillators:
http://silversol.freewebpage.org/spice/xtal/clapp.htm
Noise-Rejecting Wideband Sampler:
http://www3.sympatico.ca/add.automation/sampler/intro.htm

 

[Phil Hobbs]

Just found that adding a capacitor from pin 3 (biasing - non-inverting
input) of the op-amp to ground, almost halved the noise.
I am happy it works, but any ideas as to why it worked?

thanks

The photodiode capacitance makes the noninverting gain of the stage rise
to a huge peak at high frequency, so any junk on the noninverting input
gets amplified enormously, along of course with the amplifier voltage
noise--which is the primary limitation on TIA performance at low signal
levels.

Cheers,

Phil Hobbs

 

[Phil Hobbs]

This also shows up in SPICE. What is your recommendation to cure the
problem?

Compared with a simple load resistor/noninverting buffer setup, a TIA
improves the frequency response but does nothing to improve the SNR--the
noise physics is stil purely RC. In the TIA, the op amp just jiggles
the other end of the resistor to make the summing junction stay
reasonably still, so the signal current doesn't disappear into the
photodiode capacitance.

Another way of looking at it is that with the load resistor approach,
the signal rolls off while the noise floor stays flat, whereas with the
TIA, the signal stays flat and the noise floor rises with frequency.
The SNR is the same--there is no free lunch.

At moderate frequencies, the three main things you can do are:

1. Use a lower capacitance photodiode, e.g. reduce the area, choose a
PIN and not a PN device, and use lots of reverse bias. (The old wives'
tale that you get lower noise at zero bias is an outrageous fib--unless
you have a really stinky photodiode, the amplifier noise always
dominates the leakage current shot noise.)

2. Use a quieter amplifier, and bypass the daylights out of it.

3. Use a really quiet bootstrap transistor, e.g. a superbeta MPSA18.

4. Use a common-base input stage (which I loosely call a cascode though
it isn't really).

There's an article of mine on all this stuff (including accurate
cookbook rules for choosing amplifiers) at http://www.electrooptical.net
about halfway down the (long) page.

Cheers,

Phil Hobbs

 

[Phil Hobbs]

At moderate frequencies, the three main things you can do are:
<snip>
1. <snip>
2. <snip>
3. <snip>
4. <snip>

5. ruthless efficiency
6. an almost fanatical devotion to the Pope.

Amongst our weaponry are such diverse elements as...

(can't count this morning)

Cheers,
Phil Hobbs

 

[Mike Monett]

[... snip good info]

There's an article of mine on all this stuff (including accurate
cookbook rules for choosing amplifiers) at
http://www.electrooptical.net about halfway down the (long) page.

Cheers,

Phil Hobbs

Sorry for the delay - been offline. Thanks for the excellent info. I tried
looking for your article at electrooptical.net but it wouldn't respond.
I'll try again later.

I also found some very good stuff on your web site. There seems to be two
identical (or very similar pages):

http://users.bestweb.net/~hobbs/pergamos.htm

and

http://users.bestweb.net/~hobbs/

You should put one in your sig

Regards,

Mike Monett

Antiviral, Antibacterial Silver Solution:
http://silversol.freewebpage.org/index.htm
SPICE Analysis of Crystal Oscillators:
http://silversol.freewebpage.org/spice/xtal/clapp.htm
Noise-Rejecting Wideband Sampler:
http://www3.sympatico.ca/add.automation/sampler/intro.htm