Reply by Ray Andraka January 16, 20042004-01-16
You might start with Skolnik.  It is pretty much the radar bible.

The point with the down conversion is that you are typically after a narrow bandwidth
compared to the carrier.  If you try to work at the carrier directly, you unnecessarily
make the job a bunch harder than it has to be.  The down-conversion basically takes the
bandwidth of interest and moves it down to baseband where it can be sampled with much less
speed.  Typically, you'll be working with complex baseband in modern radar systems.  Until
about 20 years ago, they were analog all the way to complex baseband because the
converters couldn't keep up with much faster.  Going to digital for the IF to baseband has
the advantage of giving you perfect quadrature without painstaking component matching and
calibration, which is generally a good thing (analog systems can still win out on power).
From digital baseband, the subsequent processing depends on what kind of radar and what
parameters you are after.  For distance measurement, you may want to look at FM-CW, as the
processing load is pretty light.  IIRC, Skolnik has a whole chapter on FM-CW.

Rune Allnor wrote:

> Ray Andraka <ray@andraka.com> wrote in message news:<400736F5.73C12887@andraka.com>... > > Rune, > > > > Most of the time we down convert to a workable IF first. You'll need to > > know the bandwidth of your echo. Most of the radar processors I've done > > take a real-only IF somewhere between 20 and 200 MHz and then do a digital > > mix to complex baseband. > > That's what I suspected. You basically say that to get that 10 cm > resolution one has to use some signal processing scheme that is quite > a bit more sophisticated than the crude stuff I'm used to from seismics. > > OK, I have to read up on radar implementations and radar signal > processing. Does anyone have suggestions to good reading material? > > Rune
-- --Ray Andraka, P.E. President, the Andraka Consulting Group, Inc. 401/884-7930 Fax 401/884-7950 email ray@andraka.com http://www.andraka.com "They that give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety." -Benjamin Franklin, 1759
Reply by Rune Allnor January 16, 20042004-01-16
Ray Andraka <ray@andraka.com> wrote in message news:<400736F5.73C12887@andraka.com>...
> Rune, > > Most of the time we down convert to a workable IF first. You'll need to > know the bandwidth of your echo. Most of the radar processors I've done > take a real-only IF somewhere between 20 and 200 MHz and then do a digital > mix to complex baseband.
That's what I suspected. You basically say that to get that 10 cm resolution one has to use some signal processing scheme that is quite a bit more sophisticated than the crude stuff I'm used to from seismics. OK, I have to read up on radar implementations and radar signal processing. Does anyone have suggestions to good reading material? Rune
Reply by Ray Andraka January 15, 20042004-01-15
Rune,

Most of the time we down convert to a workable IF first.  You'll need to
know the bandwidth of your echo.  Most of the radar processors I've done
take a real-only IF somewhere between 20 and 200 MHz and then do a digital
mix to complex baseband.  There are ADCs that work at 1GHz sample rates,
but they are typically limited to about 8 bits, with maybe 6 bits ENOB.
That may not be enough amplitude resolution for anything reasonably
sensitive without trading bandwidth for more bits through downconversion.

Rune Allnor wrote:

> Hi all. > > For some reason I posted this question on sci.engr.radar+sonar without > crossposting to comp.dsp. I am sure there are somebody here who have > some advice. > > I have been invited to comment on a non-mainstream radar application. > From what I can see of the basic prospect, there are some assumptions > on the recieved data that I find a bit dubious. These assumptions > are not, however, obviously right or wrong, so I would like to make > some further inquiries on the raw data the reciever measures and does > its computations on, before expressing any further opinions to > those who approached me. > > To the best of my knowledge, this radar's carrier frequency is > somewhere in the 2-5 GHz area (I'm sorry I can't be more specific, > this is really back-of-the-envelope stuff). > > So, is it possible to digitize raw data at such frequencies? > We are talking about a device that is intended to measure distances > on the ~20 m scale and with a ~10 cm accuracy in a very complex > environment. I would like to digitize and examine the recieved waveform > from one transmitted pulse, from a slant range of 0 m to, say, 50 m. > > Rune
-- --Ray Andraka, P.E. President, the Andraka Consulting Group, Inc. 401/884-7930 Fax 401/884-7950 email ray@andraka.com http://www.andraka.com "They that give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety." -Benjamin Franklin, 1759
Reply by One Usenet Poster January 15, 20042004-01-15
Rune Allnor wrote:
> Hi all. > > For some reason I posted this question on sci.engr.radar+sonar without > crossposting to comp.dsp. I am sure there are somebody here who have > some advice. > > I have been invited to comment on a non-mainstream radar application. > From what I can see of the basic prospect, there are some assumptions > on the recieved data that I find a bit dubious. These assumptions > are not, however, obviously right or wrong, so I would like to make > some further inquiries on the raw data the reciever measures and does > its computations on, before expressing any further opinions to > those who approached me. > > To the best of my knowledge, this radar's carrier frequency is > somewhere in the 2-5 GHz area (I'm sorry I can't be more specific, > this is really back-of-the-envelope stuff).
Most radar receivers down-convert the RF signal to a much lower IF signal, then do the analog-to-digital conversion. I don't think the typical ADC is fast enough for the 2 to 5 GHz band.
> So, is it possible to digitize raw data at such frequencies? > We are talking about a device that is intended to measure distances > on the ~20 m scale and with a ~10 cm accuracy in a very complex > environment. I would like to digitize and examine the recieved waveform > from one transmitted pulse, from a slant range of 0 m to, say, 50 m.
Also, a pulse type radar will have a minimum slant range, since you have to wait for the pulse to travel to the target and return. OUP
Reply by Rune Allnor January 15, 20042004-01-15
Hi all. 

For some reason I posted this question on sci.engr.radar+sonar without
crossposting to comp.dsp. I am sure there are somebody here who have
some advice.

I have been invited to comment on a non-mainstream radar application.
From what I can see of the basic prospect, there are some assumptions
on the recieved data that I find a bit dubious. These assumptions
are not, however, obviously right or wrong, so I would like to make 
some further inquiries on the raw data the reciever measures and does 
its computations on, before expressing any further opinions to
those who approached me. 

To the best of my knowledge, this radar's carrier frequency is 
somewhere in the 2-5 GHz area (I'm sorry I can't be more specific, 
this is really back-of-the-envelope stuff).

So, is it possible to digitize raw data at such frequencies?
We are talking about a device that is intended to measure distances
on the ~20 m scale and with a ~10 cm accuracy in a very complex 
environment. I would like to digitize and examine the recieved waveform 
from one transmitted pulse, from a slant range of 0 m to, say, 50 m.

Rune