low noise, real analog SPDT switch..

Hi:

I am trying to automate an LED to work as a light emitter as well as a light detector by electronic control. I tried using MAX4053A to act as a SPDT switch, but it does not work. On one of the inputs of the SPDT switch, I am trying to give the constant current input for the LED. Other input will have a detector connected to it. The COM output will have LED connected to it. So I am trying to solve this issue. Is there a proper SPDT switch which will solve this problem. I am looking at a very low noise switch, because the current to be detected is in the nano-amp range. Other ideas to solve this issue is greatly appreciated..

Thanks..

Provide a block diagram (or schematic of what you tried) if you can.

your questions is also very vague?

I tried using MAX4053A to act as a SPDT switch, but it does not work.

Click to expand...

You should specify what doesnot work exactly
is it the switch is not working (then you should check the select inputs of the switch)
or is it your LED is not working after switching???
so as suggested by Ron H provide the schematic and block diagram and also specify your exact problem.

cjoseph2 said:

Hi:

I am trying to automate an LED to work as a light emitter as well as a light detector by electronic control. I tried using MAX4053A to act as a SPDT switch, but it does not work. On one of the inputs of the SPDT switch, I am trying to give the constant current input for the LED. Other input will have a detector connected to it. The COM output will have LED connected to it. So I am trying to solve this issue. Is there a proper SPDT switch which will solve this problem. I am looking at a very low noise switch, because the current to be detected is in the nano-amp range. Other ideas to solve this issue is greatly appreciated..

Thanks..

Click to expand...

You have to choose your switch carefully. You can easily get nA leakage currents ruining your detection. Can you use a reed relay switch?

Perhaps you need a debouncing circuit?

If you look at a switch being switched on an oscilloscope, you can see that it creates a huge amount of noise for a very short time. A debouncing circuit gives you a very clean switch. Look it up on google.

instruite said:

your questions is also very vague?

I tried using MAX4053A to act as a SPDT switch, but it does not work.

Click to expand...

You should specify what doesnot work exactly
is it the switch is not working (then you should check the select inputs of the switch)
or is it your LED is not working after switching???
so as suggested by Ron H provide the schematic and block diagram and also specify your exact problem.

Click to expand...

My issue is I would like to make an LED work both as a light emitter and as a photo-detector. And I want to do that automatically (electronically). I checked this concept manually and it is working. LED in a photoconductive mode acts as a photo-detector.

Being said all these, now I tried to get an SPDT switch to solve my issue. I selected MAX4053A, which has a very low leakage current of less than 100pico-amps in the OFF state. MAX4053A has 3 SPDT switches. I am trying to use only one switch (say switch A). I connected the LED to the switch-A output (COMA) of the 4053A and then to GND. One of the switch-A inputs, NOA is connected to the 20mA input. Other input, NCA is connected to the input of a nano-ammeter.

Now when I give '0' as input to the ADDA input, the LED is supposed to act as a photo-detector, by having current flow in the reverse direction and the nano-ammeter should show some nano-amp reading. But there is current flowing in the positive direction of the LED and the LED lights. And during this test, the nano-ammeter is overloaded. But when I reduce the current below 7mA to the input NOA, and repeat the same test described above, I am able to measure some value of nano-amp current inthe nano-ammeter.

So my conclusion was, when I increase the value of current above 7mA at NOA, there is some leakage into the COMA output. Therefore all these problems.

I hope you all understand what I am trying to say. Please advise. Sorry, I dont have a schematic..

I am not sure what a reed relay is.. Could you please tell, if that will help me solve the issue. Will that totally isolate the input and the output.


Thanks..

How are you generating the 20ma source? What supply voltage are you providing to the MAX4053A? You really need to post a schematic if you want serious help.

Ron H said:

How are you generating the 20ma source? What supply voltage are you providing to the MAX4053A? You really need to post a schematic if you want serious help.

Click to expand...

I am using LM317 to generate 20mA..
15V at Vin of LM317..
5V at V+ and 0V for V- of Max4053A..

I have added a schematic drawn in MSWord. Hope this helps..

I can't figure out a mechanism to explain your results, but consider this:
The 20ma has no intentional alternate path when it is deselected, so it flows through the ESD protection diode on the MAX4053A to +5V. If your +5V supply can't sink 20ma (most positive regulators can't sink current), then +5V will go up to around +12 to +15V. This may raise your input logic threshold above your logic control voltage. It will also raise leakage currents on the other switch pins, although I wouldn't think it would be high enough to overload your nanoammeter.
Try this: Use another section of your switch to ground the output of your current source when it is not selected to be flowing through the LED.
Keep in mind that the analog switch pin voltages must not exceed the supply voltage. This part has an on resistance as high as 225 ohms (typical=125 ohms) with a 5 volt supply. 20ma*225=4.5V. Add this to the LED drop, and the voltage on your current source pin is exceeding the supply voltage. Even at 125 ohms, you may have a problem, depending on the LED forward voltage.
You can extract yourself from this can of worms with the reed relay idea if you don't need fast switching. If you don't know what they are, Google. You can buy them from a multitude of online suppliers.

use a mercury wetted switch there is NO contact bounce and is very fast swiching