ADC differential input requirements

Go To Last Post
5 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Greetings,

I am using the differential channels on the mega 128 and am having trouble getting a useful result.

I have two solar cells connected + against + and wish to obtain a differential reading from the minus ends to determine the direction of incoming sunlight. The differential signal is about +/- 20 mV, so I thought, great, I'll just plug the minus ends into the differential channels, add 200x gain and get the -512..+512 signal out...not so easy.

There are obviously some requirements for the ADC inputs that I have missed. I have read a lot of information in the forums and manual, but am still having trouble putting it all together.

My questions are in 3 main areas:
1) Input current,
2) Voltage ranges, and
3) Offset error.

1)
At present, the minus ends of the solar cells are connected directly to the pins. i.e. I have practically no current flow an am measuring open circuit voltages.

The manual specifies an input resistance of less than 10k. Does this imply a minimum short-circuit current for the source in order to drive the ADC input?

2)
I would have assumed that a differential voltage signal is a differential voltage signal, however... the input voltage must be between 0 and Vcc. What is the input voltage? Is this at the pins, or after the amplifier, or a bit of both?

More specifically is the relative voltage between differential source and uC important? Do I need some sort of common ground?

3)
The magnitude of the signal (as well as the sign) is also different, depending on which way around I connect my wires. Is there any obvious reason for this apparent offset error?

I have noticed that there is ADMUX setting for measuring channel 0 against channel 0 that "Can be used for offset calibration." Calibration of what? Does the offset refer to the amplifier, or to some sort of parasitic voltage on the pins? Where can I throw this factor in to correct the offset?

I hope I have not missed anything really obvious here.

Grateful for any responses!

Nick

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

> What is the input voltage? Is this at the pins, or after the amplifier, or a bit of both?

At the pins.

> Does the offset refer to the amplifier, or to some sort of parasitic voltage on the pins?

Essentially to the amplifier, yes. You can subtract it from your actual result.

p.s.: You are aware that the voltage of a silicon photo diode has only a flat dependency on
the strength of light, overlaid by a temperature coefficient, are you? Measuring the short-circuit
(or load) current would be much more useful.

Jörg Wunsch

Please don't send me PMs, use email if you want to approach me personally.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi Joerg,

I'm using actual solar cells, for which I do have to look into the electrical characteristics a bit more. I'm pretty sure the o/c voltage does increase with the strength of light though, as I have measured this using a voltmeter. I did try using the load current, which currently works better, but I am still trying to figure out why, so that I make the right decision.

Given your answer to 2), I would assume that the differential voltages both have to be above the 0V uC reference.

Your answer to 3) has also added some clarity.

Thanks

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

To me 'solar cell' means a silicon photovoltaic cell that outputs voltage. To determine the direction of the sun, I'd use a light dependent resistor (cadmium sulphide cell). The differential voltage between two of these gives a + or - signal that can be amplified to drive a gearmotor to turn the panel to null the voltage difference.

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi Bob,

the reason I am using solar cells in this case is because of the integrator effect their larger surface area has. This makes them much less sensitive to reflections and shadows. Light concentration is also involved and the LDRs I have seen all saturate at quite low levels.

At the moment, the signal from the solar cells' negative ends seems fine, but I now understand that both the negative ends must be higher than the uC 0V reference...looks like I should have stuck the negative ends together and measured the positive ends instead.