[hpsdr] QSD architecture and image suppression

[Ahti Aintila]

Hi Oleg,

My earlier measurements are about the same that Phil already reported.
Now with the new PowerSDR versions and especially with the new
Flex-5000 hardware(firmware) the situation has changed. If I
understand it correctly, there are new possibilities of saving the
tuning settings in the files rx_image.csv and power.csv.

How that is implemented in practise with the Flex-5000, I don't know.
Theoretically it should be possible to save the receiver AND
transmitter balance settings on every band with 10 kHz intervals.
SDR-1000 would be more problematic due to the variations of the
commercial sound cards and contact instabilities of the connectors
(plugs and jacks).

I agree with you Oleg that the image should ideally be more than 100
dB down, especially in the transmitter. I'm impatiently waiting for
the results of Penelope and Mercury. What do you say, Phil?

73, Ahti OH2RZ


On 21/02/2008, Oleg Skydan <oleg at skydan.in.ua> wrote:
> ***** High Performance Software Defined Radio Discussion List *****
>
> Hi, Bob, Anti and Phil!
>
> Thank you for your comments. And special thanks to
> Phil for the numbers - that was exactly what I needed.
>
> I was interested in those numbers as I have some plans to
> start a new transceiver. So I had three alternatives:
> 1. Use traditional superheterodyne architecture with
> DSP at low IF (just as my current design).
> 2. Use QSD/QSE architecture for the frontend.
> 3. Use DDC/DUC architecture.
>
> I have several goals for the new design. Particulary
> the image rejection should be more then 90-100dB.
> It can be easilly achived with 1st and 3td architecture, but
> I can not use high speed ADC with more then 12 bits
> (export regulations), so DDC architecture can not be
> used.
>
> It looks like I will use the traditional superheterodyne
> with low IF DSP.
>
> You can find my site here http://t03dsp.skydan.in.ua
> The information is a bit outdated - there were many changes
> since the last site update. The new DSP block with faster
> and better ADC (CS5381) and DAC (CS4392) and new
> processor (DSP56367). The FST3125 based first mixer.
> The new control block and front panel (you can see the
> picture of it here http://skydan.in.ua/downloads/NewTrcrv.JPG).
>
> All the best!
> Oleg
> 73 de UR3IQO
>
> ----- Original Message -----
> From: "Phil Harman" <phil at pharman.org>
> To: <hpsdr at lists.hpsdr.org>
> Sent: Thursday, February 21, 2008 7:26 AM
> Subject: Re: [hpsdr] QSD architecture and image suppression
>
>
> > ***** High Performance Software Defined Radio Discussion List *****
> >
> >
> > Hi Oleg,
> >
> > I'd like to add my warm welcome to group as well. I've also followed
> > your work for some time and it would be great to have you participate
> > in the various HPSDR projects.
> >
> > Bob, N4HY, has replied to your question regarding image suppression
> > from a theoretical perspective so I would like to address what is
> > presently being achieved in practice.
> >
> > My own measurements on QSD based receivers indicates that it is
> > possible to achieve >90dB of suppression over a narrow bandwidth if a
> > manual or automatic adjustment is made at a single frequency. If such
> > adjustment is made as say the edge of a ham band then the suppression
> > drops to typically 40-45dB at the band edges.
> >
> > Software that enables correction to be made across a band, such as
> > Rocky, does much better than this. However, since this software is
> > aimed at the SoftRock range of radios, that typically do not have any
> > isolation between the QSD and antenna, then changes in antenna SWR etc
> > can significantly degrade the suppression.  Also whilst this technique
> > is effective over a single band I'm not sure how effective it would be
> > for a multi-band radio or one that covered the entire HF range.
> >
> > There was some work started some time ago to model the impulse response
> > of the receiver by injecting a series of very narrow pulses into the
> > antenna socket.
> >
> > Given the impulse response I gather that it should be possible to
> > accurately compensate for I and Q amplitude and phase errors.
> >
> > I'm not sure what the current progress of this idea is and perhaps
> > others can comment as to how successful this is/was.
> >
> > From my own perspective, since there does not so far seem to be an
> > effective solution, I've been working on DDC and DUC radios since image
> > suppression is usually not an issue with such technologies.
> >
> > 73's Phil...VK6APH
> _______________________________________________
> 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/
>

 1203576661.0