Analog RF to DSP

Hello all,
I have doubt as to what will happen if down converted RF signal of around
16MHz frequency is given directly to DSP. Will there be issues with regard
to signal integrity?

Thanks.

On 6/14/12 9:03 PM, rash412 wrote:
> Hello all,
> I have doubt as to what will happen if down converted RF signal of around
> 16MHz frequency is given directly to DSP. Will there be issues with regard
> to signal integrity?

not if you have lotsa DSP MIPS and a good A/D converter running at about 
35 MSPS.  sounds expensive.

what's the bandwidth of this down-converted RF signal?  if it's still 
decently narrow bandwidth, can you bump it further down than 16 MHz?


> Thanks.

FWIW.

-- 

r b-j                  rbj@audioimagination.com

"Imagination is more important than knowledge."

On Fri, 15 Jun 2012 01:02:09 -0400, robert bristow-johnson
<rbj@audioimagination.com> wrote:

>not if you have lotsa DSP MIPS and a good A/D converter running at about 
>35 MSPS.  sounds expensive.

Example of FM IF (10.7 MHz) handled directly by DSP:
http://ham-radio.com/k6sti/xdr-f1hd.htm

"The SAF7730 provides 16 IF filters of various bandwidths. It dynamically
selects one based on interference and modulation level. The 64-tap FIR filters
use 16-bit coefficients. "

"Adaptive digital IF filters eliminate adjacent-channel interference in nearly
all cases. The filters have flat passbands with extremely steep skirts. The DSP
rapidly swaps filter coefficients while retaining state variables to dynamically
adapt the filter bandwidth to modulation level and interference without
introducing switching artifacts. The digital filters are much more effective
than conventional analog IF filters. In addition, they do not have the
unit-to-unit variation that necessitates ceramic filter selection or tuned
compensation for optimal performance. Their symmetrical finite impulse response
eliminates the group delay error that causes audio distortion in analog
filters."

robert bristow-johnson <rbj@audioimagination.com> writes:

> On 6/14/12 9:03 PM, rash412 wrote:
>> Hello all,
>> I have doubt as to what will happen if down converted RF signal of around
>> 16MHz frequency is given directly to DSP. Will there be issues with regard
>> to signal integrity?
>
> not if you have lotsa DSP MIPS and a good A/D converter running at
> about 35 MSPS.  sounds expensive.

This is within the realm of reason for a processor like the TI 64x+
series, which can run up to 500+ MHz and do 4 16-bit MACs per cycle.

As far as the ADC goes, you can get 12 bits at that rate or better
for 9USD (1k qty): http://www.ti.com/product/adc12040  Not cheap,
but not all that extreme either. 
-- 
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

Randy Yates <yates@digitalsignallabs.com> writes:

> robert bristow-johnson <rbj@audioimagination.com> writes:
>
>> On 6/14/12 9:03 PM, rash412 wrote:
>>> Hello all,
>>> I have doubt as to what will happen if down converted RF signal of around
>>> 16MHz frequency is given directly to DSP. Will there be issues with regard
>>> to signal integrity?
>>
>> not if you have lotsa DSP MIPS and a good A/D converter running at
>> about 35 MSPS.  sounds expensive.
>
> This is within the realm of reason for a processor like the TI 64x+
> series, which can run up to 500+ MHz and do 4 16-bit MACs per cycle.

Ummm, let me modify that. This would be pushing it, but possibly doable,
depending on how good you want your decimation filter.

Also (as we all know here, but for the OP's sake), front-end processing
such as mixing and decimation is much more effectively done in a piece
of hardware such as an FPGA, then fed to the DSP at a much more sane
rate.
-- 
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

Greg Berchin <gjberchin@chatter.net.invalid> wrote:

>   "Adaptive digital IF filters eliminate adjacent-channel 
>    interference in nearly all cases. The filters have flat 
>    passbands with extremely steep skirts. The DSP rapidly 
>    swaps filter coefficients while retaining state variables 
>    to dynamically adapt the filter bandwidth to modulation 
>    level and interference without introducing switching 
>    artifacts. The digital filters are much more effective 
>    than conventional analog IF filters. In addition, 
>    they do not have the unit-to-unit variation that 
>    necessitates ceramic filter selection or tuned 
>    compensation for optimal performance. Their symmetrical 
>    finite impulse response eliminates the group delay error 
>    that causes audio distortion in analog filters."

Sounds pretty neat. I do wonder, though, how many more years
we will have analog radio. With the conversion to digital TV,
and at least a start (I haven't bought one yet) on digital
radio, how long before the complete conversion of broadcast
to digital?

Now, you likely still need IF filters, but it seems to me that
the needs are somewhat different from analog radios, so the
above might not apply. I suppose it depends on the modulation
system in use, though.

-- glen

glen herrmannsfeldt <gah@ugcs.caltech.edu> writes:
> [...]
> and at least a start (I haven't bought one yet) on digital
> radio, how long before the complete conversion of broadcast
> to digital?

I presume you're talking of Ibiquity, AKA HD Radio?
-- 
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

Randy Yates <yates@digitalsignallabs.com> wrote:

(snip, I wrote)
>> [...]
>> and at least a start (I haven't bought one yet) on digital
>> radio, how long before the complete conversion of broadcast
>> to digital?

> I presume you're talking of Ibiquity, AKA HD Radio?

I believe so. There are both AM and FM versions, with presumably
different technologies. And there is also Sirius/XM, which I
presume is also digital.

I suppose before it gets very popular, a new technology might
come along, and everything shift over.

There is the story about the beginning of FM radio, they were
in the 30MHz or so band, and selling (slowly) radios.

Then when the standard came out at 88-108MHz (as I understand it,
influenced by AM radio broadcasters) all the old radios were useless.

-- glen

On Fri, 15 Jun 2012 22:55:48 +0000 (UTC), glen herrmannsfeldt
<gah@ugcs.caltech.edu> wrote:

>Randy Yates <yates@digitalsignallabs.com> wrote:
>
>(snip, I wrote)
>>> [...]
>>> and at least a start (I haven't bought one yet) on digital
>>> radio, how long before the complete conversion of broadcast
>>> to digital?

Personally, even as one who makes their living from wireless digital
comm, I hope it never happens. 

Analog receivers are dirt simple and consumer very little power.   A
crystal set can receive AM usably with no power supply at all.  This
makes cheap receivers that last a long time on one set of batteries
(or a hand crank or a little solar cell) practical.  This makes public
broadcast of emergency directions easy, cheap, and more practical than
with digital.

A receiver for an all-digital signal is more complex, more expensive,
and consumes more power.   It is less therefore less likely to be
easily accessible and usable for long periods without recharge or
battery replacement.

There's been a move to include FM receivers in cellular handsets
partly for this reason, but there's been some resistance from the
manufacturers.   Personally I think it's a good idea, but since it
doesn't generate revenue for the cellular ecosystem there's not much
support from that industry.

>> I presume you're talking of Ibiquity, AKA HD Radio?
>
>I believe so. There are both AM and FM versions, with presumably
>different technologies. And there is also Sirius/XM, which I
>presume is also digital.
>
>I suppose before it gets very popular, a new technology might
>come along, and everything shift over.
>
>There is the story about the beginning of FM radio, they were
>in the 30MHz or so band, and selling (slowly) radios.
>
>Then when the standard came out at 88-108MHz (as I understand it,
>influenced by AM radio broadcasters) all the old radios were useless.

Just like the shift to digital TV, or from horses to cars.   Such
technology shifts happen a bit more frequently now, and they're always
a little painful.


Eric Jacobsen
Anchor Hill Communications
www.anchorhill.com

Eric Jacobsen <eric.jacobsen@ieee.org> wrote:

(snip, I wrote)
>>>> and at least a start (I haven't bought one yet) on digital
>>>> radio, how long before the complete conversion of broadcast
>>>> to digital?

> Personally, even as one who makes their living from wireless 
> digital comm, I hope it never happens. 

> Analog receivers are dirt simple and consumer very little power.   
> A crystal set can receive AM usably with no power supply at all.  
> This makes cheap receivers that last a long time on one set 
> of batteries (or a hand crank or a little solar cell) practical.
> This makes public broadcast of emergency directions easy, 
> cheap, and more practical than with digital.

> A receiver for an all-digital signal is more complex, more 
> expensive, and consumes more power.   It is less therefore 
> less likely to be easily accessible and usable for long 
> periods without recharge or battery replacement.

Is that required, or just current technology? Since the power
used by digital circuits scales as the transistors shrink, 
the power per unit of computation shrinks. Modern processors
use more power because they do a lot more computing.

But there has to be some analog electronics at the front end,
and it could be that will always need more power. Wouldn't that
change with the power of the broadcast signal? If they stop
analog, and put more power into the digital signal, doesn't
that mean less power needed for the receiver?

Now, sometimes I am in the mood to do things the old way.
I have recently got more into darkroom photography, partly 
for the fun of doing it the old way. For similar reasons,
I some might be interested in analog audio, LP records 
and cassette tapes. But for audio, the future is digital.

For video, digital is even more obvious. High quality video
takes a lot of bandwidth, which is not easy to store as
an analog signal. People lived with amazingly poor quality
VHS for a long time before DVDs came along and got cheap enough.
(Personally, I never used the 6 hour mode of VHS, but many 
people did.)

> There's been a move to include FM receivers in cellular handsets
> partly for this reason, but there's been some resistance from the
> manufacturers.   Personally I think it's a good idea, but since it
> doesn't generate revenue for the cellular ecosystem there's not much
> support from that industry.

I have known phones with build-in MP3 players, I hadn't
noticed FM receivers. (That is, not counting smart phones.)

It seems to me that it depends on the price difference, and what
you can convince someone of the added value in the store.

-- glen