IQ Data from ADC

>On 9/30/2015 3:30 AM, b2508 wrote:
>>> On Tue, 29 Sep 2015 09:47:06 -0500, b2508 wrote:
>>>
>>>> Can I somehow obtain IQ Data by using two ADCs on same input real
>> signal
>>>> and are there some limitations to this?
>>>>
>>>> I see Texas Instrument ADCs have two channels named I and Q but I do
>> not
>>>> understand if I have two inputs (for two analog signals representing
I
>>>> and Q) or they form I and Q from real signal.
>>>>
>>>> Thank you very much in advance.
>>>
>>>
>>> Does this help?
>>> http://dspguru.com/dsp/tricks/complex-downconverters-at-fs-over-4
>>>
>>> Allan
>>
>> I am not sure if this is what I need.
>>
>> I receive IF data on one analog channel (real data) and then this goes
>> through ADC. I get real data at the output and this goes to DDC and I
get
>> IQ Data.
>> My question is, can I get IQ data from ADC and then instead of feeding
>> real data to DDC, feed IQ data to it?
>>
>> I am not sure what is the benefit of this and what is the reason but
it
>is
>> a requirement.
>
>Such a strange requirement.
>
>The DDC is generating I/Q data by digitally multiplying (mixing) your 
>real data stream with a complex carrier (quadrature).  You can do the 
>same mixing in the analog domain and sample the two I/Q analog signals 
>with the ADC chip you have chosen.  Is that what you are looking for?
>
>Seems a lot easier to do in the digital domain using your DDC chip.  Why

>do you need to feed the DDC chip I/Q data?
>
>-- 
>
>Rick

Thank you :) It is a relief that I am not the only one thinking it is
strange requirement. I do not know the reason behind it. 

I also understood at the beginning that requirement is having 2 separate
inputs, one for I and one for Q both obtained through analog down
conversion from RF to IF but later on , I realized it is one real signal
and now I am just confused. If it is real signal why is it not enough to
obtain IQ data after DDC to baseband.

Anyhow, thank you very much to all of you.
---------------------------------------
Posted through http://www.DSPRelated.com

On Tuesday, September 29, 2015 at 10:47:12 AM UTC-4, b2508 wrote:
> Can I somehow obtain IQ Data by using two ADCs on same input real signal
> and are there some limitations to this?
> 
> I see Texas Instrument ADCs have two channels named I and Q but I do not
> understand if I have two inputs (for two analog signals representing I and
> Q) or they form I and Q from real signal.
> 
> Thank you very much in advance.
> ---------------------------------------
> Posted through http://www.DSPRelated.com

One can obtain an approximation to the analytic signal with a technique called "quadrature" or "quarter-wave" sampling. The basic idea is that by sampling with two channels at sub-multiples of the period at the signal center frequency one can effectively filter the RF signals with, e.g., (1/2) [-j, 2, j]. This is a rather crude complex bandpass filter centered at 0.25 (hence the "quarter-wave"). With faster sampling, one can do better ("multi-order" sampling). It's a clever idea and works provided your signal isn't very wide-band.

Here's a reference from the ultrasound literature:
J. E. Powers, et. al., "Ultrasound phased array delay lines based on quadrature sampling techniques," IEEE Transactions on Sonics and Ultrasonics, 1980.

Unfortunately, I do not have access to that article.

I am actually interested in why Texas Instruments calls their ADC channels
I and Q? This  would make sense only if I actually had two signals coming
in and these two be IQ from somewhere else, right? But in graphics where
they explain interleaved sampling with two channels sampling one analog
input they say that at the output you get IQIQIQ... I don't know if they
are just trying to denote that these are interleaved data from two
channels (named I and Q) and not that these are actually I and Q data.
---------------------------------------
Posted through http://www.DSPRelated.com

On Wed, 30 Sep 2015 09:38:01 -0500, "b2508" <108118@DSPRelated> wrote:

>Unfortunately, I do not have access to that article.
>
>I am actually interested in why Texas Instruments calls their ADC channels
>I and Q? This  would make sense only if I actually had two signals coming
>in and these two be IQ from somewhere else, right? But in graphics where
>they explain interleaved sampling with two channels sampling one analog
>input they say that at the output you get IQIQIQ... I don't know if they
>are just trying to denote that these are interleaved data from two
>channels (named I and Q) and not that these are actually I and Q data.

The part is designed to be able to either sample a real stream by
interleaving the two ADCs and providing a sample rate of up to
1.6Gsps, or using the two ADCs to sample I and Q channels of a mixer
output at up to 800Msps.   You can use it either way.

If you use it to sample the output of a mixer, say a downconverter
that is moving a broadband multi-access channel to baseband, then what
else would you call the I and Q channels?   That's the normal
terminology for that application.


Eric Jacobsen
Anchor Hill Communications
http://www.anchorhill.com

On Wednesday, September 30, 2015 at 10:38:06 AM UTC-4, b2508 wrote:

> I am actually interested in why Texas Instruments calls their ADC channels
> I and Q?

Sigh... Then why did you ask this question in your first post?

> Can I somehow obtain IQ Data by using two ADCs on same input real signal 
> and are there some limitations to this?

On 9/30/2015 10:38 AM, b2508 wrote:
> Unfortunately, I do not have access to that article.
>
> I am actually interested in why Texas Instruments calls their ADC channels
> I and Q? This  would make sense only if I actually had two signals coming
> in and these two be IQ from somewhere else, right? But in graphics where
> they explain interleaved sampling with two channels sampling one analog
> input they say that at the output you get IQIQIQ... I don't know if they
> are just trying to denote that these are interleaved data from two
> channels (named I and Q) and not that these are actually I and Q data.

I think you understand the part.  The naming I and Q is just for 
convenience even though it is not being very convenient for you. 
Normally the I and Q inputs would be used to sample two analog inputs 
which are I and Q from an analog mixer, or they can be used to sample 
two arbitrary inputs just as easily in which case the I and Q 
designations are meaningless.

The interleaved sampling is a way to achieve higher sample rates of one 
signal.  The two ADCs operate on opposite edges of the clock (they even 
talk about adjusting the clock edges so the phasing is very close to 180 
degrees) so it becomes effectively a 2x clock with each ADC working on 
alternate clock cycles.  The merged data output is then one stream of 
the same signal, no I and Q involved.

-- 

Rick

On 9/30/2015 2:18 PM, rigby@crd.ge.com wrote:
> On Wednesday, September 30, 2015 at 10:38:06 AM UTC-4, b2508 wrote:
>
>> I am actually interested in why Texas Instruments calls their ADC channels
>> I and Q?
>
> Sigh... Then why did you ask this question in your first post?

Sighing, but no answer.  What was the point of your post?

-- 

Rick

On Wednesday, September 30, 2015 at 2:45:40 PM UTC-4, rickman wrote:
> On 9/30/2015 2:18 PM, rigby wrote:
> > On Wednesday, September 30, 2015 at 10:38:06 AM UTC-4, b2508 wrote:
> >
> >> I am actually interested in why Texas Instruments calls their ADC channels
> >> I and Q?
> >
> > Sigh... Then why did you ask this question in your first post?
> 
> Sighing, but no answer.  What was the point of your post?
> 
> -- 
> 
> Rick

You're mistaken. I did answer the OP's original question, "Can I somehow obtain IQ data by using two ADCs on the same real signal." (The answer is, yes, approximately.) He replied that what he actually wanted was the answer to a completely different question.

On 9/30/2015 5:31 PM, rigby@crd.ge.com wrote:
> On Wednesday, September 30, 2015 at 2:45:40 PM UTC-4, rickman wrote:
>> On 9/30/2015 2:18 PM, rigby wrote:
>>> On Wednesday, September 30, 2015 at 10:38:06 AM UTC-4, b2508 wrote:
>>>
>>>> I am actually interested in why Texas Instruments calls their ADC channels
>>>> I and Q?
>>>
>>> Sigh... Then why did you ask this question in your first post?
>>
>> Sighing, but no answer.  What was the point of your post?
>>
>> --
>>
>> Rick
>
> You're mistaken. I did answer the OP's original question, "Can I somehow obtain IQ data by using two ADCs on the same real signal." (The answer is, yes, approximately.) He replied that what he actually wanted was the answer to a completely different question.

He asked for additional info.  My point is you don't need to be rude. 
Reply or don't reply.  Why be rude?

-- 

Rick

On Wed, 30 Sep 2015 02:15:09 -0500, b2508 wrote:

> I am using ADC12D800RF. I don't understand if I can, by using one single
> analog input obtain iq data with this ADC.
> This ADC basically has 2 channels named I and Q but from what I've
> understood, it can use both of them to sample the same input signal
> with,
> say, fs and obtain 2*fs rate that way. But is this IQ data and if it is,
> how is that the case?

It looks like the two built-in track and hold amplifiers hold on every 
other clock, out of phase, in a way that makes it easy to digitally 
extract an inphase and quadrature channel.

But -- I'd just read the data sheet carefully and try to cipher it out.  
Maybe see if you can find an app note, or if you're working on a product, 
the attention of an applications engineer?

-- 

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com