Photoconductive Photocells

[Jacques St-Pierre]

Hi,

I am designing a small system to measure water contaminant using light going
trought the water.

So on one side I use a LED drive by a constant current source, and on the
other side a Cds (Photoconductive Photocells).

Without any water, just in air, I can not adjust my circuit because the
resistive value of the Cds is constantly changing. So I remove all my
component and simply connect the Cds on a ohms meter.

The resistance value of the Cds change with variation in light condition as
is suppose to do, but instead of stabilizing around a value, the resistance
value is slowly going up, and it's doing so as long I wait to check it.

So because of that constant rise in value it's impossible to set any form of
calibration to my device. I need something that give me a resonably stable
value for a constant light source.

Do someone know why that append? Anyway to stabilize the resistance value?
Any other low cost sensor element to suggest?

Bye
Jacques

 

[Phil Hobbs]

Hi,

I am designing a small system to measure water contaminant using light going
trought the water.

So on one side I use a LED drive by a constant current source, and on the
other side a Cds (Photoconductive Photocells).

Without any water, just in air, I can not adjust my circuit because the
resistive value of the Cds is constantly changing. So I remove all my
component and simply connect the Cds on a ohms meter.

The resistance value of the Cds change with variation in light condition as
is suppose to do, but instead of stabilizing around a value, the resistance
value is slowly going up, and it's doing so as long I wait to check it.

So because of that constant rise in value it's impossible to set any form of
calibration to my device. I need something that give me a resonably stable
value for a constant light source.

Do someone know why that append? Anyway to stabilize the resistance value?
Any other low cost sensor element to suggest?

Bye
Jacques

Cad sulphide and cad selenide cells are the pits. You can get a 5:1
resistance hysteresis based only on previous history.

You're far better off with a silicon photodiode or a small solar cell.

Cheers,

Phil Hobbs

 

[GregS]

Hi,

I am designing a small system to measure water contaminant using light going
trought the water.

So on one side I use a LED drive by a constant current source, and on the
other side a Cds (Photoconductive Photocells).

Without any water, just in air, I can not adjust my circuit because the
resistive value of the Cds is constantly changing. So I remove all my
component and simply connect the Cds on a ohms meter.

The resistance value of the Cds change with variation in light condition as
is suppose to do, but instead of stabilizing around a value, the resistance
value is slowly going up, and it's doing so as long I wait to check it.

So because of that constant rise in value it's impossible to set any form of
calibration to my device. I need something that give me a resonably stable
value for a constant light source.

Do someone know why that append? Anyway to stabilize the resistance value?
Any other low cost sensor element to suggest?

How about a bridge circuit to help stabilize.

greg

 

[Ban]

Hi,

I am designing a small system to measure water contaminant using
light going trought the water.

So on one side I use a LED drive by a constant current source, and on
the other side a Cds (Photoconductive Photocells).

Without any water, just in air, I can not adjust my circuit because
the resistive value of the Cds is constantly changing. So I remove
all my component and simply connect the Cds on a ohms meter.

The resistance value of the Cds change with variation in light
condition as is suppose to do, but instead of stabilizing around a
value, the resistance value is slowly going up, and it's doing so as
long I wait to check it.
So because of that constant rise in value it's impossible to set any
form of calibration to my device. I need something that give me a
resonably stable value for a constant light source.

Do someone know why that append? Anyway to stabilize the resistance
value? Any other low cost sensor element to suggest?

Photoresistors are no "cells". Very temperature sensitive so if you hold it
in the hand, the resistance changes. This only happens at lowish
illumination (at 0.1lx the resistance doubles from 20° to 40°). They are
also dependent on the past illumination, the resistance rises. They might
need a whole day in the dark to recover.
Another disadvantage is the long time constant for higher resistance. BUT
they are almost as sensitive as a photomultiplier tube, and can drive loads
like a relais directly. Don't use it as a measuring sensor or use at least
100 lx for 100ms and then keep a longer pause(30s) until the next
measurement.

 

[Adrian Tuddenham]

The resistance value of the Cds change with variation in light condition as
is suppose to do, but instead of stabilizing around a value, the resistance
value is slowly going up, and it's doing so as long I wait to check it.

The CdS cell has different time constants for attack (light increasing)
and decay (light decreasing), so it is extremely difficult to calibrate.
In addition, the decay has two different time constants acting at the
same time, one of a few seconds, another of many minutes.

It would be better if you used a photodiode.

 

[Phil Hobbs]

Ban wrote:

Another disadvantage is the long time constant for higher resistance. BUT
they are almost as sensitive as a photomultiplier tube, and can drive loads
like a relais directly.

When you say "almost as sensitive as a PMT", I don't think you're taking
stability into account adequately. It's true that photoconductors
exhibit a gain of (carrier lifetime)/(transit time), which is large in
CdS, but (a) photoconductors have twice the shot noise of photodiodes,
because the recombination contributes as much noise as generation, and
(b) CdS is so variable with time and history that it's really hard to
make use of all that sensitivity.

PMTs are also a bit noisier than photodiodes, but are much, much more
stable than CdS cells. Cad sulphide and phototransistors are the sorts
of devices that look less and less wonderful the closer you get.

Cheers,

Phil Hobbs

 

[Pooh Bear]

Hi,

I am designing a small system to measure water contaminant using light going
trought the water.

So on one side I use a LED drive by a constant current source, and on the
other side a Cds (Photoconductive Photocells).

Without any water, just in air, I can not adjust my circuit because the
resistive value of the Cds is constantly changing. So I remove all my
component and simply connect the Cds on a ohms meter.

The resistance value of the Cds change with variation in light condition as
is suppose to do, but instead of stabilizing around a value, the resistance
value is slowly going up, and it's doing so as long I wait to check it.

Just try checking the dark resistance if you want to see something interesting !

So because of that constant rise in value it's impossible to set any form of
calibration to my device. I need something that give me a resonably stable
value for a constant light source.

Do someone know why that append? Anyway to stabilize the resistance value?
Any other low cost sensor element to suggest?

A CdS or CdSe cell is totally unsuitable for this application. To be honest, if
you're looking for low levels of contamination I think you're looking at the
wrong method.

Graham

 

[Ban]

Ban wrote:


When you say "almost as sensitive as a PMT", I don't think you're
taking stability into account adequately. It's true that
photoconductors exhibit a gain of (carrier lifetime)/(transit time),
which is large in CdS, but (a) photoconductors have twice the shot
noise of photodiodes, because the recombination contributes as much
noise as generation, and (b) CdS is so variable with time and history
that it's really hard to make use of all that sensitivity.

PMTs are also a bit noisier than photodiodes, but are much, much more
stable than CdS cells. Cad sulphide and phototransistors are the
sorts of devices that look less and less wonderful the closer you get.

I think there are situations where CdS-Photoresistors are of advantage. I
have an old reflex camera with a 22.5V battery and it works very well since
40yrs. The meter is so sensitive in low light situations, no amplifier
needed. And the accuracy is sufficient for this purpose.
Lets see: the current at 10mlx will be around 20 uA. When you compare that
with a diode it will hardly put out more than 1nA, since you need also more
filtering even less. The temp. stability is not much worse. And it seems if
you avoid strong illumination the stability and repeatibility can be good as
well, at least for this app. Well, maybe Hasselblad have researched a bit
how to treat this resistor, but they did well, IMHO
The noise of the diode will be much higher, since the resistance is about 30
times that of the photoconductor. And even the time constant of the diode is
difficult to get lower with all its own + amplifier input capacity.
I totally agree about precision measurements with you though.