Loading ...
Sorry, an error occurred while loading the content.
 

Re: [softrock40] Re: Fractured Rock

Expand Messages
  • FRANCIS CARCIA
    Ian, I was looking at your design and it looks lik eyou are using the VCO in the 4046. I would think an external VCO would be a lot cleaner??? frank wa1gfz
    Message 1 of 14 , Oct 1, 2006
      Ian,
      I was looking at your design and it looks lik eyou are using the VCO in the 4046.
      I would think an external VCO would be a lot cleaner??? frank wa1gfz

      imitchel4you <ian.mitchell@...> wrote:
      Jean, right now there is no kit. I've had some PCBs made to do the
      initial testing but I still need to do some more testing, particularly
      with sub-harmics and the AM broadcast band (in which there are local
      large signal levels).

      Also the microcontroller code needs a little more work to make it
      smart enough to set the up/down signal on the QSD to reverse I/Q based
      on the frequency. It would also be useful to get it to program the
      EEPROM to remember a specific frequency that it would default to at
      start up. Then it could be used as a kind of programmable softrock.

      No plans for a kit, but I would be happy to provide a PCB and
      programmed microcontroller a little later down the track. All the
      other parts are easy to get.

      Cheers, Ian

      --- In softrock40@yahoogro ups.com, "ptapon217" <ptapon217@. ..> wrote:
      >
      > Hello,
      >
      > Is there (will be) a kit available? If yes, I'd like one.
      > Just let me know.
      >
      > JEAN
      >
      >
      > --- In softrock40@yahoogro ups.com, "imitchel4you" <ian.mitchell@ >
      > wrote:
      > >
      > > Hi John, please see:
      > > http://groups. yahoo.com/ group/softrock40 /files/fractured rock/fr80. pdf
      > >
      > > cheers
      > >
      >


    • imitchel4you
      Hi Frank, I m not sure what the phase noise is for the 4046 but the linearity is very good and varies by less than 0.4% across the tuning range. It would be
      Message 2 of 14 , Oct 1, 2006
        Hi Frank,

        I'm not sure what the phase noise is for the 4046 but the linearity is
        very good and varies by less than 0.4% across the tuning range. It
        would be hard to design a varicap based VCO as good as that.

        To a large extent the phase noise will be determined by the noise on
        the VCO input. With a sigma-delta modulator this noise is pushed well
        away from the reference frequency and the loop filter is all that's
        required to do a good job of filtering it. Theoretically the SD
        modulator should push the phase noise below 100db.

        As I've said, I'm still testing, but so far things are looking good.
        My Tektronics TDS200 only shows noise above 50db below the fundamental
        and the oscillator noise is below this. I'll be building a small test
        oscillator like the XG2 to do some more testing. It'll be fun!

        --- In softrock40@yahoogroups.com, FRANCIS CARCIA <carcia@...> wrote:
        >
        > Ian,
        > I was looking at your design and it looks lik eyou are using the
        VCO in the 4046.
        > I would think an external VCO would be a lot cleaner??? frank wa1gfz
        >
        > imitchel4you <ian.mitchell@...> wrote:
        > Jean, right now there is no kit. I've had some PCBs made
        to do the
        > initial testing but I still need to do some more testing, particularly
        > with sub-harmics and the AM broadcast band (in which there are local
        > large signal levels).
        >
        > Also the microcontroller code needs a little more work to make it
        > smart enough to set the up/down signal on the QSD to reverse I/Q based
        > on the frequency. It would also be useful to get it to program the
        > EEPROM to remember a specific frequency that it would default to at
        > start up. Then it could be used as a kind of programmable softrock.
        >
        > No plans for a kit, but I would be happy to provide a PCB and
        > programmed microcontroller a little later down the track. All the
        > other parts are easy to get.
        >
        > Cheers, Ian
        >
        > --- In softrock40@yahoogroups.com, "ptapon217" <ptapon217@> wrote:
        > >
        > > Hello,
        > >
        > > Is there (will be) a kit available? If yes, I'd like one.
        > > Just let me know.
        > >
        > > JEAN
        > >
        > >
        > > --- In softrock40@yahoogroups.com, "imitchel4you" <ian.mitchell@>
        > > wrote:
        > > >
        > > > Hi John, please see:
        > > >
        http://groups.yahoo.com/group/softrock40/files/fracturedrock/fr80.pdf
        > > >
        > > > cheers
        > > >
        > >
        >
      • imitchel4you
        Just one other general comment. I m really appreciating the comments received on this design. One simple change that I think should be made is to the resistors
        Message 3 of 14 , Oct 1, 2006
          Just one other general comment. I'm really appreciating the comments
          received on this design. One simple change that I think should be made
          is to the resistors that set the opamp gain. I think the 1K/100K
          combination is too high and will contribute to noise. They should be
          reduced to 100/10K.

          Thanks Frank, comments much appreciated!


          --- In softrock40@yahoogroups.com, FRANCIS CARCIA <carcia@...> wrote:
          >
          > Ian,
          > I was looking at your design and it looks lik eyou are using the
          VCO in the 4046.
          > I would think an external VCO would be a lot cleaner??? frank wa1gfz
          >
          > imitchel4you <ian.mitchell@...> wrote:
          > Jean, right now there is no kit. I've had some PCBs made
          to do the
          > initial testing but I still need to do some more testing, particularly
          > with sub-harmics and the AM broadcast band (in which there are local
          > large signal levels).
          >
          > Also the microcontroller code needs a little more work to make it
          > smart enough to set the up/down signal on the QSD to reverse I/Q based
          > on the frequency. It would also be useful to get it to program the
          > EEPROM to remember a specific frequency that it would default to at
          > start up. Then it could be used as a kind of programmable softrock.
          >
          > No plans for a kit, but I would be happy to provide a PCB and
          > programmed microcontroller a little later down the track. All the
          > other parts are easy to get.
          >
          > Cheers, Ian
          >
          > --- In softrock40@yahoogroups.com, "ptapon217" <ptapon217@> wrote:
          > >
          > > Hello,
          > >
          > > Is there (will be) a kit available? If yes, I'd like one.
          > > Just let me know.
          > >
          > > JEAN
          > >
          > >
          > > --- In softrock40@yahoogroups.com, "imitchel4you" <ian.mitchell@>
          > > wrote:
          > > >
          > > > Hi John, please see:
          > > >
          http://groups.yahoo.com/group/softrock40/files/fracturedrock/fr80.pdf
          > > >
          > > > cheers
          > > >
          > >
          >
        • olivier_maillard2000
          Hello Ian, Two small questions : As I understand only 2 resistors of the group R23-26 must be inserted depending of the required division coefficient. The
          Message 4 of 14 , Oct 3, 2006
            Hello Ian,

            Two small questions :
            As I understand only 2 resistors of the group R23-26 must be inserted
            depending of the required division coefficient. The UpDown input of
            the '191 is used to reverse the I and Q signal in the case where R24
            and 25 are insterted.

            As I see the AT2313 is used to provide the reference frequency to the
            PLL and as frequency divider ? Is that correct ?
            As I know on the Atmel uC the external counter clock input is gated
            by the system clock. So the half period of the T1 input must be
            longer than a clock period.
            So the abolute max frequency possible on the Q1 output of the hc191
            is 10Mhz.
            Now the stupid question, whats is the consequence of the delay
            introduced by this delay on the phase noise.
            Is this eliminated by the loop filter ?

            I have see somewhere else an interssant trick to overcome this
            problem. The signal to divide was used to clock the CPU. So we have a
            full synchrone divider.

            Other question :
            Is the source code for the AtTiny aviable ?

            Best regards

            Olivier HB9TOB

            --- In softrock40@yahoogroups.com, "imitchel4you" <ian.mitchell@...>
            wrote:
            >
            > Just one other general comment. I'm really appreciating the comments
            > received on this design. One simple change that I think should be
            made
            > is to the resistors that set the opamp gain. I think the 1K/100K
            > combination is too high and will contribute to noise. They should be
            > reduced to 100/10K.
            >
            > Thanks Frank, comments much appreciated!
            >
            >
            > --- In softrock40@yahoogroups.com, FRANCIS CARCIA <carcia@> wrote:
            > >
            > > Ian,
            > > I was looking at your design and it looks lik eyou are using the
            > VCO in the 4046.
            > > I would think an external VCO would be a lot cleaner??? frank
            wa1gfz
            > >
            > > imitchel4you <ian.mitchell@> wrote:
            > > Jean, right now there is no kit. I've had some PCBs made
            > to do the
            > > initial testing but I still need to do some more testing,
            particularly
            > > with sub-harmics and the AM broadcast band (in which there are
            local
            > > large signal levels).
            > >
            > > Also the microcontroller code needs a little more work to make it
            > > smart enough to set the up/down signal on the QSD to reverse I/Q
            based
            > > on the frequency. It would also be useful to get it to program the
            > > EEPROM to remember a specific frequency that it would default to
            at
            > > start up. Then it could be used as a kind of programmable
            softrock.
            > >
            > > No plans for a kit, but I would be happy to provide a PCB and
            > > programmed microcontroller a little later down the track. All the
            > > other parts are easy to get.
            > >
            > > Cheers, Ian
            > >
            > > --- In softrock40@yahoogroups.com, "ptapon217" <ptapon217@> wrote:
            > > >
            > > > Hello,
            > > >
            > > > Is there (will be) a kit available? If yes, I'd like one.
            > > > Just let me know.
            > > >
            > > > JEAN
            > > >
            > > >
            > > > --- In softrock40@yahoogroups.com, "imitchel4you"
            <ian.mitchell@>
            > > > wrote:
            > > > >
            > > > > Hi John, please see:
            > > > >
            >
            http://groups.yahoo.com/group/softrock40/files/fracturedrock/fr80.pdf
            > > > >
            > > > > cheers
            > > > >
            > > >
            > >
            >
          • imitchel4you
            Hi Olivier sorry for taking so long to respond... discovered that at 100KHz ref the synthsizer was not working so well...read on ... yep you can configure to
            Message 5 of 14 , Oct 4, 2006
              Hi Olivier

              sorry for taking so long to respond... discovered that at 100KHz ref
              the synthsizer was not working so well...read on

              --- In softrock40@yahoogroups.com, "olivier_maillard2000"
              <maillard@...> wrote:
              >
              > Hello Ian,
              >
              > Two small questions :
              > As I understand only 2 resistors of the group R23-26 must be inserted
              > depending of the required division coefficient.

              yep you can configure to divide by an extra 2 or 4 for the lower bands

              > The UpDown input of
              > the '191 is used to reverse the I and Q signal in the case where R24
              > and 25 are insterted.

              or for 3rd harmonic sampling. a side effect of harmonic sampling is
              that I and Q are reversed - the microcontroller can check the
              frequency and count down instead of up so you don't have to change the
              setting in the software

              >
              > As I see the AT2313 is used to provide the reference frequency to the
              > PLL and as frequency divider ? Is that correct ?

              yes, both, a fixed reference frequency and a programmable divider

              > As I know on the Atmel uC the external counter clock input is gated
              > by the system clock. So the half period of the T1 input must be
              > longer than a clock period.
              > So the abolute max frequency possible on the Q1 output of the hc191
              > is 10Mhz.

              atmel recommend clk/2.5 so the max is 8MHz

              > Now the stupid question, whats is the consequence of the delay
              > introduced by this delay on the phase noise.
              > Is this eliminated by the loop filter ?

              not stupid at all and it was a big concern for me. initially I was
              using a reference of 100KHz. the LO locked great on the scope, but in
              my listening peoples voices sounded somewhat distored. i then had the
              idea of listing to the LO on my DX934 receiver and I could hear the
              variation in the tone. a kind a warbling effect.

              i'm not sure why but i don't think it is related to the uC essentially
              sampling the signal but might be more due to the fact that the
              variation of the programmable divider as a consequence of the sigma
              delta modulator is a significat proportion of the integer value. eg if
              the LO is 3.6MHz the integer divider will be 36, but the variation to
              this is about +-7 which is a +-20% variation. it might also be
              integer-N boundary spurs which occur when the fractional part is very
              small and the output frequency is close to a multiple of the
              comparison frequency - which can also happen on submultiples of the
              reference frequency

              as an experiment I set the reference frequency to 10KHz and removed
              the fractional-N code so that it operated in the standard PLL
              arrangment (frequency steps of 10KHz only). then the tone sounded fine
              in the DX394 (the warbling had gone). band conditions were poor last
              night so I didn't get much of a chance to listen to SSB signals

              anyway I replaced the fraction-N code but with a 10KHz reference and
              the warbling had gone (it may have been slightly distorted but it's
              hard to tell) - but I'm still not convinced - I need to do more
              listening (at 10KHz the integer divider is 360 and the variation is
              now just 2%)

              one thing to try is including an extra order (5th) in the modulator
              and see what that sounds like

              >
              > I have see somewhere else an interssant trick to overcome this
              > problem. The signal to divide was used to clock the CPU. So we have a
              > full synchrone divider.

              yes, but i was worried the uC would crash because the specs say the
              clock should not vary by more that 4% (or something like that) over
              short periods so that scared me off :-)

              >
              > Other question :
              > Is the source code for the AtTiny aviable ?

              yep, just posted it in the fracturedrock folder

              >
              > Best regards
              >
              > Olivier HB9TOB
              >
            • olivier_maillard2000
              Hello, This random delay between the last pulse provided by the 191 and the arrival of the information on the phase comparator sound very bad to me. It can be
              Message 6 of 14 , Oct 6, 2006
                Hello,

                This random delay between the last pulse provided by the '191 and the
                arrival of the information on the phase comparator sound very bad to me.
                It can be on the worst case a whole period of the 20Mhz oscilltor.
                Witch can be 40% of the clock of 8Mhz.

                My sugestion is to add a D Flip-Flop (edge trigerred) between the CT1
                output of the Atmel and the input of the phase comparator.
                CT1 conected on the D input.
                The D Flip-Flop will be clocked by the Output of the '191.

                In the Atmel you divide by n-1.
                After n-1 pulse from the '191 the output of CT1 will change with the
                random delay.
                At the next pulse comming from the '191 the latch will IMMADIATLY send
                the information to the phase comparator.

                The delay is out !

                My second sugestion is for your reference clock.
                My did you not use the same trick used in the dual module PLL.

                Make the divisor between the system colock and the reference CLOCK of
                the PLL variable.
                The divisor of course stay also variable.

                From the RS232 interface you send booth information of the Atmel.
                In the main computer you have to do some math to find the ratio with
                the smallest round divisor possible to obtian the desired frequency.

                But all this is only my point of you and I have almost no time to
                experiment this.
                But building a low cost variable oscillator with common parts for the
                soft rock in definittly a good idea.

                73's de Olivier HB9TOB
              • Ian Mitchell
                Olivier, I really appreciate your suggestions. I d like to try the D-FF idea but that would require an additional IC. I d previously built a 32bit DDS driven
                Message 7 of 14 , Oct 8, 2006
                  Olivier, I really appreciate your suggestions. I'd like to try the D-FF
                  idea but that would require an additional IC.

                  I'd previously built a 32bit DDS driven PLL using AN AVR and a fixed
                  divisor of 256 (but not incorporated into a receiver). The DDS has an
                  equivalent clock of 2.2MHz (with the AVR at 20MHz) and generates a
                  reference frequency of between 39KHz and 78KHz via an R-2R D/A
                  converter. It also has a programmable divisor that the AVR can set to
                  divide the 10-20MHz output. So, I was able to set the output to 3.6MHz
                  and do the listening test with the DX394. And, no warbling, a very pure
                  tone it seemed.

                  I designed this circuit for use in a 0-30MHz SDR that I was hoping to
                  develop after the FR80.

                  With the current FR80 circuit I could use the second half of the 4053 to
                  get a fixed division of 64 and add the components for the A/D converter
                  to the AVR - no extra IC required. The reference frequency would need to
                  be between 156KHz and 312KHz and the equivalent DDS clock would be 2MHz
                  (one extra cycle required because the original circuit had enough SRAM
                  to move the sine table into and access that - with 2313 the table needs
                  to be program memory which takes an extra cycle access).

                  In this case there are 4x fewer samples per cycle but if it works... :-)

                  olivier_maillard2000 wrote:
                  >
                  > Hello,
                  >
                  > This random delay between the last pulse provided by the '191 and the
                  > arrival of the information on the phase comparator sound very bad to me.
                  > It can be on the worst case a whole period of the 20Mhz oscilltor.
                  > Witch can be 40% of the clock of 8Mhz.
                  >
                  > My sugestion is to add a D Flip-Flop (edge trigerred) between the CT1
                  > output of the Atmel and the input of the phase comparator.
                  > CT1 conected on the D input.
                  > The D Flip-Flop will be clocked by the Output of the '191.
                  >
                  > In the Atmel you divide by n-1.
                  > After n-1 pulse from the '191 the output of CT1 will change with the
                  > random delay.
                  > At the next pulse comming from the '191 the latch will IMMADIATLY send
                  > the information to the phase comparator.
                  >
                  > The delay is out !
                  >
                  > My second sugestion is for your reference clock.
                  > My did you not use the same trick used in the dual module PLL.
                  >
                  > Make the divisor between the system colock and the reference CLOCK of
                  > the PLL variable.
                  > The divisor of course stay also variable.
                  >
                  > >From the RS232 interface you send booth information of the Atmel.
                  > In the main computer you have to do some math to find the ratio with
                  > the smallest round divisor possible to obtian the desired frequency.
                  >
                  > But all this is only my point of you and I have almost no time to
                  > experiment this.
                  > But building a low cost variable oscillator with common parts for the
                  > soft rock in definittly a good idea.
                  >
                  > 73's de Olivier HB9TOB
                  >
                  >

                  --
                  Ian Mitchell
                  Research Services
                  University of Tasmania
                  Private Bag 1
                  Hobart, Tasmania, Australia, 7001

                  Phone (03) 6226 7457
                  Fax (03) 6226 2765

                  Intl Phone +61 3 6226 7457
                  Intl Fax +61 3 6226 2765

                  http://www.research.utas.edu.au/
                  "...for who can escape what he desires."
                  Genesis, And Then There Were Three, 1978
                • old_newcomer
                  ... Where is the file FR80.PDF just? The FRACTUREDROCK directory is not existing. Regards
                  Message 8 of 14 , Apr 6, 2009
                    --- In softrock40@yahoogroups.com, "imitchel4you" <ian.mitchell@...> wrote:
                    >
                    > Hi John, please see:
                    > http://groups.yahoo.com/group/softrock40/files/fracturedrock/fr80.pdf
                    >
                    > cheers
                    >

                    Where is the file FR80.PDF just?

                    The FRACTUREDROCK directory is not existing.

                    Regards

                    >
                    > --- In softrock40@yahoogroups.com, john sakellakis <yiannis_w8gxu@>
                    > wrote:
                    > >
                    > > Hello Ian..
                    > >
                    > > Your posting about the "fractured rock" is very
                    > > interesting. Do you have any schematics to pass on?
                    > > If you could email me a set, I would greatly
                    > > appreciate it... 73... John
                    > >
                    > > yiannis_w8gxu@
                    > > ***********************************************
                    > >
                    > > --- imitchel4you <ian.mitchell@> wrote:
                    > >
                    > > > Hi, here's just a bit of info on the fractured rock:
                    > > >
                    > > > * tunes 80M band but can be configured for 160, 40,
                    > > > 20 or 20/40 etc
                    > > > and AM broadcast using either fundamental or
                    > > > harmonic sampling
                    > > > * 7 pole bandpass filter using moulded inductors
                    > > > * preamp of about 12db (but you can change it) - in
                    > > > my area signals
                    > > > are few and low level - also it provides the correct
                    > > > impedance to the
                    > > > filter , and sets the bias voltage and low impedance
                    > > > to the QSD
                    > > > * the LO is based on a 32bit 4th order sigma-delta
                    > > > modulator feeding a
                    > > > PLL with a reference frequency of 100KHz (I didn't
                    > > > notice any spurs
                    > > > but i've only been testing for a couple of nights) -
                    > > > frequency range
                    > > > is 10-20MHz
                    > > > * the QSD divides the LO by four and is a four-way
                    > > > commutating switch
                    > > > * the LM837 has a GBW of 25MHz and noise of less
                    > > > than 4nV/rtHz and is
                    > > > powered at 9V to improve dynamic range - it is a
                    > > > quad op-amp
                    > > > configured to offer a high impedance to all four
                    > > > switches
                    > > > * the LO is set by sending the decimal ascii
                    > > > frequency in Hz to a COM port
                    > > >
                    > > > Cheers, Ian
                    > > >
                    > > >
                    > > >
                    > > >
                    > >
                    > >
                    > > __________________________________________________
                    > > Do You Yahoo!?
                    > > Tired of spam? Yahoo! Mail has the best spam protection around
                    > > http://mail.yahoo.com
                    > >
                    >
                  Your message has been successfully submitted and would be delivered to recipients shortly.