ADC/DAC QUERY

Started by omprakash alone June 30, 2010
hi,

I am phasing problem with my ADC of controller which is mounted on
BlueBoard.

As my ADC is i/p, i am not able to find out howto make it bipolar.

As my ADC is o/p, i can make it bipolar by connecting its o/p to some
ckt(Op-Amp).

Omprakash Alone

An Engineer's Guide to the LPC2100 Series

What is your question?

Howard
omprakash alone wrote:
>
>
> hi,
>
> I am phasing problem with my ADC of controller which is mounted on
> BlueBoard.
>
> As my ADC is i/p, i am not able to find out howto make it bipolar.
>
> As my ADC is o/p, i can make it bipolar by connecting its o/p to some
> ckt(Op-Amp).
>
> Omprakash Alone
>

--- In l..., omprakash alone wrote:
>
> hi,
>
> I am phasing problem with my ADC of controller which is mounted on
> BlueBoard.
>
> As my ADC is i/p, i am not able to find out howto make it bipolar.
>
> As my ADC is o/p, i can make it bipolar by connecting its o/p to some
> ckt(Op-Amp).
>
> Omprakash Alone
>
You have asked this question before and it has been answered.

You use an op-amp on the input signal to add a constant sufficient to shift the most negative input voltage up to 0V. You will probably also need to rescale the signal such that the most positive value will be <= 5V. The signal -Vn..0..+Vp is now in the range of 0..5V regardless of the input values.

Your DAC has the same issue. It can only deal with digital values of 0..1023 and output voltages of 0..5V. If you want a bipolar result, you first need to do the math internally to represent the signal in a suitable range and then, on the output, you need to use an op-amp to shift the signal down and probably rescale it. Your real 0..5V output signal gets converted to the range of -Vn..0..+Vp.

I pointed you to Chapter 4 of "Op Amps For Everyone" (use Google to search for it).

Chapter 4 deals with single supply rail-to-rail op amps. This will work for the input but not for the output. You will need to do a similar design but with a dual supply op-amp. In the end, it's just an adder (subtracting in this case) with the gain set to fill the desired range.

We've been discussing this project for 3 months. You really need to find an electronics engineer to help with the input and output conversions.

Richard

Assuming that your problem is to convert the bipolar input to a unipolar
input so that the ADC can read it without clipping it.

Here's what you can do, even without an opamp:

Understand first of all that your ADC input can correctly convert all inputs
in the 0V - 3.3V range. I assume that your incoming bipolar signal does not
span more than 3.3V (from its negative peak to its positive peak).

Block the d.c. component of your incoming signal with a series capacitor.
This simply means you place a capacitor in the path between the signal
source and your ADC pin.

Next you need to bias the ADC pin to the median point of its swing. This
will be half of 3.3V, or 1.65V. To do this, construct a resistive divider
with two resistors of 10K each. 1/4 watt is fine, and actually any wattage
will be OK. One resistor should be connected between the ADC pin and
ground. The other resistor should be connected between the ADC pin and
3.3V. This will present a d.c. bias of 1.65V at the ADC pin.

Now feed this point with your bipolar signal, through the capacitor.

That's it!

You will need to select the capacitor based on the lowest frequency of
interest in your incoming signal. Without going into the RC time constant,
and assuming that your signal of interest is above 50 Hz, you could use a 1
micro-farad capacitor. There are many different types of capacitors, but
just go to a shop and ask for 1uF. If he gives you an electrolytic
capacitor, connect the positive terminal of the capacitor to the ADC pin,
and the negative to the signal source.

See if this gives you the results you want. If you want to use much lower
frequencies, you may need to change the capacitor values, and perhaps the
resistor values too.

If you see that the signal amplitude as measured at the signals source
reduces by an unacceptable amount, it may be necessary to introduce an
opamp.

But in most cases this solution should work. Two resistors, one capacitor.
And 10 minutes to solder them in place.

Good luck, you shouldn't need it, but just in case...

-- Ahmad

On Wed, Jun 30, 2010 at 10:14 AM, rtstofer wrote:

> --- In l... , omprakash
> alone wrote:
> >
> > hi,
> >
> >
> >
> > I am phasing problem with my ADC of controller which is mounted on
> > BlueBoard.
> >
> >
> >
> > As my ADC is i/p, i am not able to find out howto make it bipolar.
> >
> >
> >
> > As my ADC is o/p, i can make it bipolar by connecting its o/p to some
> > ckt(Op-Amp).
> >
> >
> >
> > Omprakash Alone
> > You have asked this question before and it has been answered.
>
> You use an op-amp on the input signal to add a constant sufficient to shift
> the most negative input voltage up to 0V. You will probably also need to
> rescale the signal such that the most positive value will be <= 5V. The
> signal -Vn..0..+Vp is now in the range of 0..5V regardless of the input
> values.
>
> Your DAC has the same issue. It can only deal with digital values of
> 0..1023 and output voltages of 0..5V. If you want a bipolar result, you
> first need to do the math internally to represent the signal in a suitable
> range and then, on the output, you need to use an op-amp to shift the signal
> down and probably rescale it. Your real 0..5V output signal gets converted
> to the range of -Vn..0..+Vp.
>
> I pointed you to Chapter 4 of "Op Amps For Everyone" (use Google to search
> for it).
>
> Chapter 4 deals with single supply rail-to-rail op amps. This will work for
> the input but not for the output. You will need to do a similar design but
> with a dual supply op-amp. In the end, it's just an adder (subtracting in
> this case) with the gain set to fill the desired range.
>
> We've been discussing this project for 3 months. You really need to find an
> electronics engineer to help with the input and output conversions.
>
> Richard
>
>
>
On Wednesday 30 June 2010 12:09:55 Ahmad wrote:

>
> You will need to select the capacitor based on the lowest frequency
> of interest in your incoming signal. Without going into the RC
> time constant, and assuming that your signal of interest is above
> 50 Hz, you could use a 1 micro-farad capacitor. There are many
> different types of capacitors, but just go to a shop and ask for
> 1uF. If he gives you an electrolytic capacitor, connect the
> positive terminal of the capacitor to the ADC pin, and the negative
> to the signal source.

Most of the time the source side will be higher than 1.65V. You should
reverse the cpacitor. -ve to adc pin and positive to source.

> See if this gives you the results you want. If you want to use
> much lower frequencies, you may need to change the capacitor
> values, and perhaps the resistor values too.
>
> If you see that the signal amplitude as measured at the signals
> source reduces by an unacceptable amount, it may be necessary to
> introduce an opamp.

He will also be pumping all the powersupply noise into the adc. His
small signals will be barely useful.

But annything should be better than the ops current state of affairs.
--
Rgds
JTD
You're right, JTD.

The negative terminal of the capacitor should be connected to the ADC, and
the positive to the signal source (assumed to be bipolar). Thanks for the
correction.

Omprakash, please take note. Which, by the way, is a good illustration of
the fallacy of taking anyone's word for anything until you've thought it
through yourself!

:-)

Ahmad
On Wed, Jun 30, 2010 at 2:20 PM, jtd wrote:

> On Wednesday 30 June 2010 12:09:55 Ahmad wrote:
>
> >
> > You will need to select the capacitor based on the lowest frequency
> > of interest in your incoming signal. Without going into the RC
> > time constant, and assuming that your signal of interest is above
> > 50 Hz, you could use a 1 micro-farad capacitor. There are many
> > different types of capacitors, but just go to a shop and ask for
> > 1uF. If he gives you an electrolytic capacitor, connect the
> > positive terminal of the capacitor to the ADC pin, and the negative
> > to the signal source.
>
> Most of the time the source side will be higher than 1.65V. You should
> reverse the cpacitor. -ve to adc pin and positive to source.
> > See if this gives you the results you want. If you want to use
> > much lower frequencies, you may need to change the capacitor
> > values, and perhaps the resistor values too.
> >
> > If you see that the signal amplitude as measured at the signals
> > source reduces by an unacceptable amount, it may be necessary to
> > introduce an opamp.
>
> He will also be pumping all the powersupply noise into the adc. His
> small signals will be barely useful.
>
> But annything should be better than the ops current state of affairs.
>
> --
> Rgds
> JTD
>
>
>