[hpsdr] Perseus is a reality

[Nico Palermo]

> You could think of that as the IF sampling rate.  The receiver
> downconverts
> to baseband (effectively, it's a direct-conversion receiver with an
> image-reject mixer.)

Right, John.

> The only question I'd have is, why is it the IF rate limited to 500 kHz?
> I
> know you can get a lot more data than that through USB 2.0 bulk transfers.
> I'd be more interested if it could give me, say, 2.5 MSPS at 16-bit
> resolution.  (I'm not sure what 24-bit resolution buys you with 16-bit
> ADCs.)

Well, there are many factors limiting the so-called IF rate (I'd rather
call it DDC output sampling rate).
First of all, I'm preferring a conservative value just because perseus has
been designed to be a performant communication receiver and not a more
general purpouse platform like that of Matt Ettus. I've already done some
tests at 1 MSPS and I'll probably release them to help the needs of SWL
who are willing to record on their herd-disk as much bandwidth as
possible, but I'm not stressing this point.
>From a more technical perspective it should be said that the DDC output
sampling rate is strictly connected to the computational power through
another important parameter which is the alias rejection.
In the Perseus DDC the last stage of the decimation is equipped with a
decimation filter which has a 130 dB alias rejection. Quite easy to do if
you need just 3 KHz of alias-free bandwidth working at 500 KSPS, but
rather consuming in terms of MMAC/s if you need a 400 KHz alias-free
bandwidth with the same sample rate.
In the DDC I developed the last decimation filter is a complex-input
256-taps FIR MAC unit (I actually use 128 taps at 500 KSPS) with two 18x35
multipliers units and two 52 bits accumulators. At 1 MSPS (output) the
number of taps you can process with it in real-time is less but still
useful to reject the alias to an acceptable value (say 110 dB) over a wide
band. At faster sample rates you need something more than two serial MAC
units to do the job. That's all.
USB 2.0 is not the limiting factor, of course.

> - Does it include the Verilog or VHDL source for the FPGA?

No

> - I assume the data-acquisition and tuning API is fully-documented, and
> can
> be called by user-written Win32 C++ code, right?  Is the interface DLL
> open-source, by any chance?

The interface DLL API will be documented as soon as I find the time to do
it. It is written in Win32 C++ and does the very basic tasks like
configuring the DDC, tuning the DDC NCO, controlling the preselection
filters bank and attenuators and of course passing the data samples to a
user callback. Not planned (yet) to put its source codes in the public
domain.

> - What limits the RF coverage to 10 kHz at the low end, and can it be
> overridden to cover down to 0 Hz?

A transformer. The ADC driver power supply is single-ended and the driver
inputs are biased to half the power supply voltage so the coverage cannot
be overridden.

> It's a nice package, overall, and the software sure looks good.  But other
> implementations like the USRP are able to support much wider output
> bandwidths.  That's pretty important for me, since I'm interested in
> something I can use for instantaneous baseband acquisition from 0 Hz to 1
> MHz in a phase-noise measurement rig.  Closed-source drivers and FPGA code
> would also be a major drawback compared to the USRP, but I'm not sure if
> that's an issue here.

I understand it, John. But since - thanks to Matt - there's already an
USRP available all over the world, I started focusing my attention towards
more specific issues.

73s
Nico iv3nwv



> You could think of that as the IF sampling rate.  The receiver
> downconverts
> to baseband (effectively, it's a direct-conversion receiver with an
> image-reject mixer.)
>
> The only question I'd have is, why is it the IF rate limited to 500 kHz?
> I
> know you can get a lot more data than that through USB 2.0 bulk transfers.
> I'd be more interested if it could give me, say, 2.5 MSPS at 16-bit
> resolution.  (I'm not sure what 24-bit resolution buys you with 16-bit
> ADCs.)
>
> Some other questions:
>
> - Does it include the Verilog or VHDL source for the FPGA?
>
> - I assume the data-acquisition and tuning API is fully-documented, and
> can
> be called by user-written Win32 C++ code, right?  Is the interface DLL
> open-source, by any chance?
>
> - What limits the RF coverage to 10 kHz at the low end, and can it be
> overridden to cover down to 0 Hz?
>
> It's a nice package, overall, and the software sure looks good.  But other
> implementations like the USRP are able to support much wider output
> bandwidths.  That's pretty important for me, since I'm interested in
> something I can use for instantaneous baseband acquisition from 0 Hz to 1
> MHz in a phase-noise measurement rig.  Closed-source drivers and FPGA code
> would also be a major drawback compared to the USRP, but I'm not sure if
> that's an issue here.
>
> -- john, KE5FX
>
>
>
>> -----Original Message-----
>> From: hpsdr-bounces at lists.hpsdr.org
>> [mailto:hpsdr-bounces at lists.hpsdr.org]On Behalf Of wbr_wf4r
>> Sent: Tuesday, October 09, 2007 4:29 PM
>> To: Nico Palermo; Philip Covington
>> Cc: hpsdr at lists.hpsdr.org
>> Subject: Re: [hpsdr] Perseus is a reality
>>
>>
>> ***** High Performance Software Defined Radio Discussion List *****
>>
>> Hello All,
>>               I'm not  a math whiz by any stretch of the imagination,
>> but
>> I'm wondering why in the beautiful panel layout that there are buttons
>> to
>> select the sampling rates of 125K 250K and 500K when the sdr only covers
>> 10KHz to 30MHz?  What does this do for the user?  Should the sampling
>> rate
>> be changed depending on the *displayed*  bandwidth?  Or is this a
>> given.  I
>> assume that the overall sampling rate will be approximately 60 MHz which
>> would be the Nyquist rate.  Maybe I just don't understand the labeling
>> "Sampling Rate"
>>
>> 73, Bill, wf4r
>> ----- Original Message -----
>> From: "Nico Palermo" <nicopal at microtelecom.it>
>> To: "Philip Covington" <p.covington at gmail.com>
>> Cc: <hpsdr at lists.hpsdr.org>
>> Sent: Tuesday, October 09, 2007 5:54 PM
>> Subject: Re: [hpsdr] Perseus is a reality
>>
>>
>> > ***** High Performance Software Defined Radio Discussion List *****
>> >
>> > Thanks, Phil.
>> > I trust that the HPSDR group will be able to do the same and/or even
>> better.
>> >
>> > 73s,
>> > Nico Palermo iv3nwv
>> >
>> >
>> > > ***** High Performance Software Defined Radio Discussion List *****
>> > >
>> > > On 10/9/07, Alberto I2PHD <i2phd at weaksignals.com> wrote:
>> > >> ***** High Performance Software Defined Radio Discussion List *****
>> > >>
>> > >> The direct sampling 10 kHz to 30 MHz receiver Perseus is now
>> a reality.
>> > >>
>> > >> See some pictures here :
>> > >>
>> > >> http://sundry.i2phd.com/perseus
>> > >>
>> > >> 73  Alberto  I2PHD
>> > >>
>> > >
>> > > Wow, very nice!  The software looks great also.
>> > >
>> > > Phil N8VB
>> > > _______________________________________________
>> > > HPSDR Discussion List
>> > > To post msg: hpsdr at hpsdr.org
>> > > Subscription help:
>> http://lists.hpsdr.org/listinfo.cgi/hpsdr-hpsdr.org
>> > > HPSDR web page: http://hpsdr.org
>> > > Archives: http://lists.hpsdr.org/pipermail/hpsdr-hpsdr.org/
>> > >
>> >
>> >



 1192004773.0