selenium cell current density

[BetterSense]

I have an old GE lightmeter that has a selenium cell hooked to an analog needle-style current meter, and computer dial on the front for calculating settings. The selenium cell is too old to be accurate but the meter itself is very cool.

I want to gut the meter and put in digital guts and a silicon sensor. Using a microcontroller I can read the sensor and then drive the current meter so that the needle readout works just like on an unmodified meter. The issue is I'm trying to figure out how much analog electronics I'm going to need to drive the meter. Does anyone know how much current these types of analog current meters range over, and whether they are linear or log? Or maybe you know how much current selenium cells put out? My microcontroller has a 0-5v analog output which I suppose I can just run through a resistor for a current source, but are we talking microamps here or mililamps?

 

[ic-racer]

Does you existing galvometer work? If you are going to buy a new selenium cell anyway, why not try to hook it up to the exiting galvometer.

If you really want to make it electronic, for some reason I think you will be much better off with a CDS cell and a battery to make your voltage. You'll need a battery anyway for the electronic voltage metering.

 

[BetterSense]

Like I said, the existing galvanometer works. I intend to reuse it. That's the whole point.

Like I said, I'm not going to use a selenium cell. I'm going to use a silicon cell.

I suppose the thing to do is take the meter into the electronics lab and work and hook it up to one of the power supplies and just start turning up the current until the needle moves and plot its motion.

 

[tkamiya]

Typically speaking, many of analog meters are in hundreds of micro amp range full scale. It won't take much to full-scale it. If you hook it up to a power supply, you can easily burn up the coil, unless you are talking about lab grade power supply with fine control. Almost all meters I've seen are linear if you are talking about deflection in degrees verses current. I'd put some series resister of known value to be safe. If you hook up 1kohm in series, 1 volt will give you 1mA. It may full scale it but not enough to burn it up...

I have an old meter similar to yours but unfortunately, the meter part is far too damaged to be any reference.

 

[BetterSense]

unless you are talking about lab grade power supply with fine control.

Oh yeah. I have access to some $5k+ lab test equipment. Unless I mess up with the knobs I should be fine. I should just be able to set it source microamps and turn it up till the needle moves. Then I can plot current versus needle deflection. If it's linear, I don't really see why I can't just run the PWM output of my microcontroller through a resistor into the meter. Then the fun comes of writing the program to sync the needle reading with the mechanical computer wheel. I have already determined that by adding an artificial pointer to the mechanical computer I can get 14 stops out of it.

 

[ath]

Get a good multimeter, place it in series to the analog meter and determine the currents for the scale markings.
For the frontend you want to meter the current of a silicon photo diode. It is proportional to the light level and small. You have to cover quite a bit of dynamic range (i.e. the current varies over 2^(number of stops)). You will need additional circuitry with a good OPAMP.

 

[unclemack]

Surely easier to recalibrate for a photovoltaic from a $2 calculator if the selenium has died?
Works for Westons.

 

[BetterSense]

The rest of the meter is not a problem. I have already made several light meters using silicon light-to-frequency chips and I'm just going to use the same circuit again. This one will be easier because I don't need an LCD.

chazmiller.com/projects/lightmeters.html