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

Miller Popper

Expand Messages
  • scott.lindroth
    First of all, I ve just discovered BEAM robots and enjoy them a great deal. I m a professional musician by trade, but a novice in electronics. I ve built 3
    Message 1 of 9 , Aug 25, 2006
      First of all, I've just discovered BEAM robots and enjoy them a great deal. I'm a
      professional musician by trade, but a novice in electronics. I've built 3 solar rollers using
      the Miller engine and am now attempting the Miller Popper from the solarbotics.net
      website.

      The circuit is more or less working on a breadboard -- shine a light on one photodiode,
      and the opposite motor fires. Since I'm supplying about 4.5v from the solar cells and am
      using fairly large motors from standard cassette decks, I am using 1381 L voltage triggers
      with a .33F 5.5v supercap to add more umph. The motors respond with a lot of zip and
      fire about once every 1.25 seconds.

      The problem I _think_ I'm having is that the 1381 L voltage triggers seem highly variable in
      their response, and I can't find two that respond equally. Right now, one 1381 triggers
      much more easily than the other when the light is evenly distributed across the
      photodiodes. When I swap the position of 1381s the motor that never triggered
      before wakes up and fires more readily (as expected). I'm suspicious of this diagnosis
      since the problem seems to occur with every pair of 1381s I try.

      What's more, the 100K? trim pot does not have any noticeable effect; it does not
      compensate for the variable trigger thresholds.

      I hope this description of the problem is useful. Thanks for any advice.

      SL
    • wilf_nv
      Hello Scott, It s been a while since I reviewed that design. I posted a revision which should make the circuit more adjustable albeit at increased leakage
      Message 2 of 9 , Aug 25, 2006
        Hello Scott,

        It's been a while since I reviewed that design. I posted a revision
        which should make the circuit more adjustable albeit at increased
        leakage current during charging (~10uA)

        The 1M resistors together with the 100K pot form two voltage dividers
        between +V and 0V. Adjusting the pot will raise and lower the
        respective reference voltage of each 1381, raising the trigger level
        of of one 1381 while lowering the other. This should allow you to
        trim any difference in thresholds between 1381 pairs.

        wilf

        --- In beam@yahoogroups.com, "scott.lindroth" <scott.lindroth@...>
        wrote:
        >
        > First of all, I've just discovered BEAM robots and enjoy them a
        great deal. I'm a
        > professional musician by trade, but a novice in electronics. I've
        built 3 solar rollers using
        > the Miller engine and am now attempting the Miller Popper from the
        solarbotics.net
        > website.
        >
        > The circuit is more or less working on a breadboard -- shine a
        light on one photodiode,
        > and the opposite motor fires. Since I'm supplying about 4.5v from
        the solar cells and am
        > using fairly large motors from standard cassette decks, I am using
        1381 L voltage triggers
        > with a .33F 5.5v supercap to add more umph. The motors respond
        with a lot of zip and
        > fire about once every 1.25 seconds.
        >
        > The problem I _think_ I'm having is that the 1381 L voltage
        triggers seem highly variable in
        > their response, and I can't find two that respond equally. Right
        now, one 1381 triggers
        > much more easily than the other when the light is evenly
        distributed across the
        > photodiodes. When I swap the position of 1381s the motor that
        never triggered
        > before wakes up and fires more readily (as expected). I'm
        suspicious of this diagnosis
        > since the problem seems to occur with every pair of 1381s I try.
        >
        > What's more, the 100K? trim pot does not have any noticeable
        effect; it does not
        > compensate for the variable trigger thresholds.
        >
        > I hope this description of the problem is useful. Thanks for any
        advice.
        >
        > SL
        >
      • Scott Lindroth
        Thanks, Wilf, but the 100K trim pot still doesn t do anything with the 1M resistors and the 1381 L triggers. The circuit works very well with 1381 G triggers
        Message 3 of 9 , Aug 26, 2006
          Thanks, Wilf, but the 100K trim pot still doesn't do anything with the 1M resistors and the
          1381 L triggers. The circuit works very well with 1381 G triggers with (and without) the
          1M resistors.

          SL

          --- In beam@yahoogroups.com, "wilf_nv" <wrigter@...> wrote:
          >
          >
          >
          >
          > Hello Scott,
          >
          > It's been a while since I reviewed that design. I posted a revision
          > which should make the circuit more adjustable albeit at increased
          > leakage current during charging (~10uA)
          >
          > The 1M resistors together with the 100K pot form two voltage dividers
          > between +V and 0V. Adjusting the pot will raise and lower the
          > respective reference voltage of each 1381, raising the trigger level
          > of of one 1381 while lowering the other. This should allow you to
          > trim any difference in thresholds between 1381 pairs.
          >
          > wilf
          >
          > --- In beam@yahoogroups.com, "scott.lindroth" <scott.lindroth@>
          > wrote:
          > >
          > > First of all, I've just discovered BEAM robots and enjoy them a
          > great deal. I'm a
          > > professional musician by trade, but a novice in electronics. I've
          > built 3 solar rollers using
          > > the Miller engine and am now attempting the Miller Popper from the
          > solarbotics.net
          > > website.
          > >
          > > The circuit is more or less working on a breadboard -- shine a
          > light on one photodiode,
          > > and the opposite motor fires. Since I'm supplying about 4.5v from
          > the solar cells and am
          > > using fairly large motors from standard cassette decks, I am using
          > 1381 L voltage triggers
          > > with a .33F 5.5v supercap to add more umph. The motors respond
          > with a lot of zip and
          > > fire about once every 1.25 seconds.
          > >
          > > The problem I _think_ I'm having is that the 1381 L voltage
          > triggers seem highly variable in
          > > their response, and I can't find two that respond equally. Right
          > now, one 1381 triggers
          > > much more easily than the other when the light is evenly
          > distributed across the
          > > photodiodes. When I swap the position of 1381s the motor that
          > never triggered
          > > before wakes up and fires more readily (as expected). I'm
          > suspicious of this diagnosis
          > > since the problem seems to occur with every pair of 1381s I try.
          > >
          > > What's more, the 100K? trim pot does not have any noticeable
          > effect; it does not
          > > compensate for the variable trigger thresholds.
          > >
          > > I hope this description of the problem is useful. Thanks for any
          > advice.
          > >
          > > SL
          > >
          >
        • wilf_nv
          Curiouser and curiouser! It is not immediately apparent what causes the different behaviors of the 1381G and 1381L but while analyzing that problem, I
          Message 4 of 9 , Aug 27, 2006
            Curiouser and curiouser!

            It is not immediately apparent what causes the different behaviors of
            the 1381G and 1381L but while analyzing that problem, I discovered
            another design flaw that reduces the effectiveness of this circuit.
            The problem is that adding components in series with the 1381 ground
            pin raises the 1381 output voltage while charging caused by the
            voltage drop of the 1381 ground pin current (~1uA). If the 1381
            output pin voltage, which is connected to the base of the output
            transistor, exceeds 500mV it may cause the transistor to partially
            turn on and leak solar cell current to ground.

            The original Miller Miller uses Si diode in series with the ground
            pin which raises the output by 100mV. The Miller Popper V1 adds a PD
            (300mV) in series which raises the transistor base voltage while
            charging to ~400mV. When the 1M resistors are added, this offset will
            be even higher so Miller Popper V2 is out the window.

            The likely cause of the different behavior of the 1381G and 1381L in
            the Miller Popper circuits is that the higher threshold of the 1381L
            is accompanied by higher ground pin current and raises the output
            voltage closer to 500mV.

            There are three possible solutions (Miller Popper V3, V4 and V5) to
            reduce the 1381 output offset during charging which I will post
            shortly in the file section :

            1) reverse the PDs so that their photovoltaic voltage substracts from
            the series voltage drop.

            2) connect the PDs in series between the 1381 ground pins so that the
            difference of the PD photovoltaic voltages and currents will modify
            the two diode drops.

            3) place the PDs in series with the 1381 Vcc pin while the 1uF timing
            cap remains directly connected to the 1381 Vcc and ground pins.


            The last solution is interesting for increasing the Miller Engine
            threshold. For example connecting a LED between Vcc and the positive
            side of the timing) allows the trigger level of the ME to be
            increased by 2V without side effects

            No time for me to test these circuits so they are experimental but
            you would do me a big favour by trying them out and reporting the
            results.

            good luck

            wilf


            --- In beam@yahoogroups.com, "Scott Lindroth" <scott.lindroth@...>
            wrote:
            >
            > Thanks, Wilf, but the 100K trim pot still doesn't do anything with
            the 1M resistors and the
            > 1381 L triggers. The circuit works very well with 1381 G triggers
            with (and without) the
            > 1M resistors.
            >
            > SL
            >
            > --- In beam@yahoogroups.com, "wilf_nv" <wrigter@> wrote:
            > >
            > >
            > >
            > >
            > > Hello Scott,
            > >
            > > It's been a while since I reviewed that design. I posted a
            revision
            > > which should make the circuit more adjustable albeit at increased
            > > leakage current during charging (~10uA)
            > >
            > > The 1M resistors together with the 100K pot form two voltage
            dividers
            > > between +V and 0V. Adjusting the pot will raise and lower the
            > > respective reference voltage of each 1381, raising the trigger
            level
            > > of of one 1381 while lowering the other. This should allow you to
            > > trim any difference in thresholds between 1381 pairs.
            > >
            > > wilf
            > >
            > > --- In beam@yahoogroups.com, "scott.lindroth" <scott.lindroth@>
            > > wrote:
            > > >
            > > > First of all, I've just discovered BEAM robots and enjoy them a
            > > great deal. I'm a
            > > > professional musician by trade, but a novice in electronics.
            I've
            > > built 3 solar rollers using
            > > > the Miller engine and am now attempting the Miller Popper from
            the
            > > solarbotics.net
            > > > website.
            > > >
            > > > The circuit is more or less working on a breadboard -- shine a
            > > light on one photodiode,
            > > > and the opposite motor fires. Since I'm supplying about 4.5v
            from
            > > the solar cells and am
            > > > using fairly large motors from standard cassette decks, I am
            using
            > > 1381 L voltage triggers
            > > > with a .33F 5.5v supercap to add more umph. The motors respond
            > > with a lot of zip and
            > > > fire about once every 1.25 seconds.
            > > >
            > > > The problem I _think_ I'm having is that the 1381 L voltage
            > > triggers seem highly variable in
            > > > their response, and I can't find two that respond equally.
            Right
            > > now, one 1381 triggers
            > > > much more easily than the other when the light is evenly
            > > distributed across the
            > > > photodiodes. When I swap the position of 1381s the motor that
            > > never triggered
            > > > before wakes up and fires more readily (as expected). I'm
            > > suspicious of this diagnosis
            > > > since the problem seems to occur with every pair of 1381s I
            try.
            > > >
            > > > What's more, the 100K? trim pot does not have any noticeable
            > > effect; it does not
            > > > compensate for the variable trigger thresholds.
            > > >
            > > > I hope this description of the problem is useful. Thanks for
            any
            > > advice.
            > > >
            > > > SL
            > > >
            > >
            >
          • wilf rigter
            I posted yet another circuit: Miller Popper V4a, which I think will work better than V4 wilf ... From: wilf_nv To: beam@yahoogroups.com Sent: Sunday, August
            Message 5 of 9 , Aug 27, 2006
              I posted yet another circuit:  Miller Popper V4a, which I think will work better than V4
               
              wilf
               
              ----- Original Message -----
              From: wilf_nv
              Sent: Sunday, August 27, 2006 6:50 PM
              Subject: [beam] Re: Miller Popper

              Curiouser and curiouser!

              It is not immediately apparent what causes the different behaviors of
              the 1381G and 1381L but while analyzing that problem, I discovered
              another design flaw that reduces the effectiveness of this circuit.
              The problem is that adding components in series with the 1381 ground
              pin raises the 1381 output voltage while charging caused by the
              voltage drop of the 1381 ground pin current (~1uA). If the 1381
              output pin voltage, which is connected to the base of the output
              transistor, exceeds 500mV it may cause the transistor to partially
              turn on and leak solar cell current to ground.

              The original Miller Miller uses Si diode in series with the ground
              pin which raises the output by 100mV. The Miller Popper V1 adds a PD
              (300mV) in series which raises the transistor base voltage while
              charging to ~400mV. When the 1M resistors are added, this offset will
              be even higher so Miller Popper V2 is out the window.

              The likely cause of the different behavior of the 1381G and 1381L in
              the Miller Popper circuits is that the higher threshold of the 1381L
              is accompanied by higher ground pin current and raises the output
              voltage closer to 500mV.

              There are three possible solutions (Miller Popper V3, V4 and V5) to
              reduce the 1381 output offset during charging which I will post
              shortly in the file section :

              1) reverse the PDs so that their photovoltaic voltage substracts from
              the series voltage drop.

              2) connect the PDs in series between the 1381 ground pins so that the
              difference of the PD photovoltaic voltages and currents will modify
              the two diode drops.

              3) place the PDs in series with the 1381 Vcc pin while the 1uF timing
              cap remains directly connected to the 1381 Vcc and ground pins.

              The last solution is interesting for increasing the Miller Engine
              threshold. For example connecting a LED between Vcc and the positive
              side of the timing) allows the trigger level of the ME to be
              increased by 2V without side effects

              No time for me to test these circuits so they are experimental but
              you would do me a big favour by trying them out and reporting the
              results.

              good luck

              wilf

              --- In beam@yahoogroups. com, "Scott Lindroth" <scott.lindroth@ ...>
              wrote:
              >
              > Thanks, Wilf, but the 100K trim pot still doesn't do anything with
              the 1M resistors and the
              > 1381 L triggers. The circuit works very well with 1381 G triggers
              with (and without) the
              > 1M resistors.
              >
              > SL
              >
              > --- In beam@yahoogroups. com, "wilf_nv" <wrigter@> wrote:
              > >
              > >
              > >
              > >
              > > Hello Scott,
              > >
              > > It's been a while since I reviewed that design. I posted a
              revision
              > > which should make the circuit more adjustable albeit at increased
              > > leakage current during charging (~10uA)
              > >
              > > The 1M resistors together with the 100K pot form two voltage
              dividers
              > > between +V and 0V. Adjusting the pot will raise and lower the
              > > respective reference voltage of each 1381, raising the trigger
              level
              > > of of one 1381 while lowering the other. This should allow you to
              > > trim any difference in thresholds between 1381 pairs.
              > >
              > > wilf
              > >
              > > --- In beam@yahoogroups. com, "scott.lindroth" <scott.lindroth@ >
              > > wrote:
              > > >
              > > > First of all, I've just discovered BEAM robots and enjoy them a
              > > great deal. I'm a
              > > > professional musician by trade, but a novice in electronics.
              I've
              > > built 3 solar rollers using
              > > > the Miller engine and am now attempting the Miller Popper from
              the
              > > solarbotics. net
              > > > website.
              > > >
              > > > The circuit is more or less working on a breadboard -- shine a
              > > light on one photodiode,
              > > > and the opposite motor fires. Since I'm supplying about 4.5v
              from
              > > the solar cells and am
              > > > using fairly large motors from standard cassette decks, I am
              using
              > > 1381 L voltage triggers
              > > > with a .33F 5.5v supercap to add more umph. The motors respond
              > > with a lot of zip and
              > > > fire about once every 1.25 seconds.
              > > >
              > > > The problem I _think_ I'm having is that the 1381 L voltage
              > > triggers seem highly variable in
              > > > their response, and I can't find two that respond equally.
              Right
              > > now, one 1381 triggers
              > > > much more easily than the other when the light is evenly
              > > distributed across the
              > > > photodiodes. When I swap the position of 1381s the motor that
              > > never triggered
              > > > before wakes up and fires more readily (as expected). I'm
              > > suspicious of this diagnosis
              > > > since the problem seems to occur with every pair of 1381s I
              try.
              > > >
              > > > What's more, the 100K? trim pot does not have any noticeable
              > > effect; it does not
              > > > compensate for the variable trigger thresholds.
              > > >
              > > > I hope this description of the problem is useful. Thanks for
              any
              > > advice.
              > > >
              > > > SL
              > > >
              > >
              >

            • Scott Lindroth
              Thanks so much, Wilf. I ve got my work cut out for me! I ll get to these revisions later this week and report back. Thanks again! SL
              Message 6 of 9 , Aug 28, 2006
                Thanks so much, Wilf. I've got my work cut out for me! I'll get to these revisions later this
                week and report back.

                Thanks again!

                SL


                --- In beam@yahoogroups.com, "wilf rigter" <wrigter@...> wrote:
                >
                > I posted yet another circuit: Miller Popper V4a, which I think will work better than V4
                >
                > wilf
                >
                > ----- Original Message -----
                > From: wilf_nv
                > To: beam@yahoogroups.com
                > Sent: Sunday, August 27, 2006 6:50 PM
                > Subject: [beam] Re: Miller Popper
                >
                >
                > Curiouser and curiouser!
                >
                > It is not immediately apparent what causes the different behaviors of
                > the 1381G and 1381L but while analyzing that problem, I discovered
                > another design flaw that reduces the effectiveness of this circuit.
                > The problem is that adding components in series with the 1381 ground
                > pin raises the 1381 output voltage while charging caused by the
                > voltage drop of the 1381 ground pin current (~1uA). If the 1381
                > output pin voltage, which is connected to the base of the output
                > transistor, exceeds 500mV it may cause the transistor to partially
                > turn on and leak solar cell current to ground.
                >
                > The original Miller Miller uses Si diode in series with the ground
                > pin which raises the output by 100mV. The Miller Popper V1 adds a PD
                > (300mV) in series which raises the transistor base voltage while
                > charging to ~400mV. When the 1M resistors are added, this offset will
                > be even higher so Miller Popper V2 is out the window.
                >
                > The likely cause of the different behavior of the 1381G and 1381L in
                > the Miller Popper circuits is that the higher threshold of the 1381L
                > is accompanied by higher ground pin current and raises the output
                > voltage closer to 500mV.
                >
                > There are three possible solutions (Miller Popper V3, V4 and V5) to
                > reduce the 1381 output offset during charging which I will post
                > shortly in the file section :
                >
                > 1) reverse the PDs so that their photovoltaic voltage substracts from
                > the series voltage drop.
                >
                > 2) connect the PDs in series between the 1381 ground pins so that the
                > difference of the PD photovoltaic voltages and currents will modify
                > the two diode drops.
                >
                > 3) place the PDs in series with the 1381 Vcc pin while the 1uF timing
                > cap remains directly connected to the 1381 Vcc and ground pins.
                >
                > The last solution is interesting for increasing the Miller Engine
                > threshold. For example connecting a LED between Vcc and the positive
                > side of the timing) allows the trigger level of the ME to be
                > increased by 2V without side effects
                >
                > No time for me to test these circuits so they are experimental but
                > you would do me a big favour by trying them out and reporting the
                > results.
                >
                > good luck
                >
                > wilf
                >
                > --- In beam@yahoogroups.com, "Scott Lindroth" <scott.lindroth@>
                > wrote:
                > >
                > > Thanks, Wilf, but the 100K trim pot still doesn't do anything with
                > the 1M resistors and the
                > > 1381 L triggers. The circuit works very well with 1381 G triggers
                > with (and without) the
                > > 1M resistors.
                > >
                > > SL
                > >
                > > --- In beam@yahoogroups.com, "wilf_nv" <wrigter@> wrote:
                > > >
                > > >
                > > >
                > > >
                > > > Hello Scott,
                > > >
                > > > It's been a while since I reviewed that design. I posted a
                > revision
                > > > which should make the circuit more adjustable albeit at increased
                > > > leakage current during charging (~10uA)
                > > >
                > > > The 1M resistors together with the 100K pot form two voltage
                > dividers
                > > > between +V and 0V. Adjusting the pot will raise and lower the
                > > > respective reference voltage of each 1381, raising the trigger
                > level
                > > > of of one 1381 while lowering the other. This should allow you to
                > > > trim any difference in thresholds between 1381 pairs.
                > > >
                > > > wilf
                > > >
                > > > --- In beam@yahoogroups.com, "scott.lindroth" <scott.lindroth@>
                > > > wrote:
                > > > >
                > > > > First of all, I've just discovered BEAM robots and enjoy them a
                > > > great deal. I'm a
                > > > > professional musician by trade, but a novice in electronics.
                > I've
                > > > built 3 solar rollers using
                > > > > the Miller engine and am now attempting the Miller Popper from
                > the
                > > > solarbotics.net
                > > > > website.
                > > > >
                > > > > The circuit is more or less working on a breadboard -- shine a
                > > > light on one photodiode,
                > > > > and the opposite motor fires. Since I'm supplying about 4.5v
                > from
                > > > the solar cells and am
                > > > > using fairly large motors from standard cassette decks, I am
                > using
                > > > 1381 L voltage triggers
                > > > > with a .33F 5.5v supercap to add more umph. The motors respond
                > > > with a lot of zip and
                > > > > fire about once every 1.25 seconds.
                > > > >
                > > > > The problem I _think_ I'm having is that the 1381 L voltage
                > > > triggers seem highly variable in
                > > > > their response, and I can't find two that respond equally.
                > Right
                > > > now, one 1381 triggers
                > > > > much more easily than the other when the light is evenly
                > > > distributed across the
                > > > > photodiodes. When I swap the position of 1381s the motor that
                > > > never triggered
                > > > > before wakes up and fires more readily (as expected). I'm
                > > > suspicious of this diagnosis
                > > > > since the problem seems to occur with every pair of 1381s I
                > try.
                > > > >
                > > > > What's more, the 100K? trim pot does not have any noticeable
                > > > effect; it does not
                > > > > compensate for the variable trigger thresholds.
                > > > >
                > > > > I hope this description of the problem is useful. Thanks for
                > any
                > > > advice.
                > > > >
                > > > > SL
                > > > >
                > > >
                > >
                >
              • Scott Lindroth
                So far Miller Popper 4a is the winner, and I succumbed to the temptation to build a photopopper with that circuit. It works well! The circuits seem more
                Message 7 of 9 , Sep 2, 2006
                  So far Miller Popper 4a is the winner, and I succumbed to the temptation to build a
                  photopopper with that circuit. It works well!

                  The circuits seem more responsive when the resistor between the trigger and the base of
                  the transistor is omitted. Not sure if this is kosher. I've also gone back to using a 4700 uf
                  storage cap. The motors trigger with more punch than the supercap.

                  I hope to have time to work more carefully with the other experimental circuits (3 & 4) over
                  the weekend. Thanks again!

                  Scott


                  --- In beam@yahoogroups.com, "wilf rigter" <wrigter@...> wrote:
                  >
                  > I posted yet another circuit: Miller Popper V4a, which I think will work better than V4
                  >
                  > wilf
                  >
                  > ----- Original Message -----
                  > From: wilf_nv
                  > To: beam@yahoogroups.com
                  > Sent: Sunday, August 27, 2006 6:50 PM
                  > Subject: [beam] Re: Miller Popper
                  >
                  >
                  > Curiouser and curiouser!
                  >
                  > It is not immediately apparent what causes the different behaviors of
                  > the 1381G and 1381L but while analyzing that problem, I discovered
                  > another design flaw that reduces the effectiveness of this circuit.
                  > The problem is that adding components in series with the 1381 ground
                  > pin raises the 1381 output voltage while charging caused by the
                  > voltage drop of the 1381 ground pin current (~1uA). If the 1381
                  > output pin voltage, which is connected to the base of the output
                  > transistor, exceeds 500mV it may cause the transistor to partially
                  > turn on and leak solar cell current to ground.
                  >
                  > The original Miller Miller uses Si diode in series with the ground
                  > pin which raises the output by 100mV. The Miller Popper V1 adds a PD
                  > (300mV) in series which raises the transistor base voltage while
                  > charging to ~400mV. When the 1M resistors are added, this offset will
                  > be even higher so Miller Popper V2 is out the window.
                  >
                  > The likely cause of the different behavior of the 1381G and 1381L in
                  > the Miller Popper circuits is that the higher threshold of the 1381L
                  > is accompanied by higher ground pin current and raises the output
                  > voltage closer to 500mV.
                  >
                  > There are three possible solutions (Miller Popper V3, V4 and V5) to
                  > reduce the 1381 output offset during charging which I will post
                  > shortly in the file section :
                  >
                  > 1) reverse the PDs so that their photovoltaic voltage substracts from
                  > the series voltage drop.
                  >
                  > 2) connect the PDs in series between the 1381 ground pins so that the
                  > difference of the PD photovoltaic voltages and currents will modify
                  > the two diode drops.
                  >
                  > 3) place the PDs in series with the 1381 Vcc pin while the 1uF timing
                  > cap remains directly connected to the 1381 Vcc and ground pins.
                  >
                  > The last solution is interesting for increasing the Miller Engine
                  > threshold. For example connecting a LED between Vcc and the positive
                  > side of the timing) allows the trigger level of the ME to be
                  > increased by 2V without side effects
                  >
                  > No time for me to test these circuits so they are experimental but
                  > you would do me a big favour by trying them out and reporting the
                  > results.
                  >
                  > good luck
                  >
                  > wilf
                  >
                  > --- In beam@yahoogroups.com, "Scott Lindroth" <scott.lindroth@>
                  > wrote:
                  > >
                  > > Thanks, Wilf, but the 100K trim pot still doesn't do anything with
                  > the 1M resistors and the
                  > > 1381 L triggers. The circuit works very well with 1381 G triggers
                  > with (and without) the
                  > > 1M resistors.
                  > >
                  > > SL
                  > >
                  > > --- In beam@yahoogroups.com, "wilf_nv" <wrigter@> wrote:
                  > > >
                  > > >
                  > > >
                  > > >
                  > > > Hello Scott,
                  > > >
                  > > > It's been a while since I reviewed that design. I posted a
                  > revision
                  > > > which should make the circuit more adjustable albeit at increased
                  > > > leakage current during charging (~10uA)
                  > > >
                  > > > The 1M resistors together with the 100K pot form two voltage
                  > dividers
                  > > > between +V and 0V. Adjusting the pot will raise and lower the
                  > > > respective reference voltage of each 1381, raising the trigger
                  > level
                  > > > of of one 1381 while lowering the other. This should allow you to
                  > > > trim any difference in thresholds between 1381 pairs.
                  > > >
                  > > > wilf
                  > > >
                  > > > --- In beam@yahoogroups.com, "scott.lindroth" <scott.lindroth@>
                  > > > wrote:
                  > > > >
                  > > > > First of all, I've just discovered BEAM robots and enjoy them a
                  > > > great deal. I'm a
                  > > > > professional musician by trade, but a novice in electronics.
                  > I've
                  > > > built 3 solar rollers using
                  > > > > the Miller engine and am now attempting the Miller Popper from
                  > the
                  > > > solarbotics.net
                  > > > > website.
                  > > > >
                  > > > > The circuit is more or less working on a breadboard -- shine a
                  > > > light on one photodiode,
                  > > > > and the opposite motor fires. Since I'm supplying about 4.5v
                  > from
                  > > > the solar cells and am
                  > > > > using fairly large motors from standard cassette decks, I am
                  > using
                  > > > 1381 L voltage triggers
                  > > > > with a .33F 5.5v supercap to add more umph. The motors respond
                  > > > with a lot of zip and
                  > > > > fire about once every 1.25 seconds.
                  > > > >
                  > > > > The problem I _think_ I'm having is that the 1381 L voltage
                  > > > triggers seem highly variable in
                  > > > > their response, and I can't find two that respond equally.
                  > Right
                  > > > now, one 1381 triggers
                  > > > > much more easily than the other when the light is evenly
                  > > > distributed across the
                  > > > > photodiodes. When I swap the position of 1381s the motor that
                  > > > never triggered
                  > > > > before wakes up and fires more readily (as expected). I'm
                  > > > suspicious of this diagnosis
                  > > > > since the problem seems to occur with every pair of 1381s I
                  > try.
                  > > > >
                  > > > > What's more, the 100K? trim pot does not have any noticeable
                  > > > effect; it does not
                  > > > > compensate for the variable trigger thresholds.
                  > > > >
                  > > > > I hope this description of the problem is useful. Thanks for
                  > any
                  > > > advice.
                  > > > >
                  > > > > SL
                  > > > >
                  > > >
                  > >
                  >
                • wilf_nv
                  Thanks for testing the circuit in the mobile platform unit. That is essential because the mechanics and electronics are interactive. Best photopopper
                  Message 8 of 9 , Sep 3, 2006
                    Thanks for testing the circuit in the mobile platform unit. That is
                    essential because the mechanics and electronics are interactive.

                    Best photopopper performance is based adjusting the following
                    variables:

                    A) Electronics

                    1) solar cell
                    2) 1381 trigger voltage
                    3) Power supply capacitor value
                    4) timing capcitor value
                    5) base resistor

                    B) Mechanics

                    1) motor efficiency
                    2) "wheel" diameter
                    3) mass
                    4) PD orientation

                    The motion of the photopopper is a waggle towards a bright light
                    source. When fully optimized, the two 1381s trigger sequentially,
                    dumping a portion of the supply capacitor charge into each motor
                    winding, causing each wheel to rotate and displace that side of the
                    popper by a small distance.

                    Depending on the materials on hand, either the motor or the solar
                    cell sets the requirements for the other components.

                    In terms of overall performance, the motor is the most influential.
                    The preferred motor is designed for 3V operation with a winding
                    resistance of at least 8 ohms and with low starting friction.

                    The recommended solar cell should have a "working" voltsge of 2.5V or
                    4.5V and the 1381 type should be chosen so that the overall trigger
                    voltage matches the solar cell working voltage.

                    The solar cell supplies maximum (contant) current while the output
                    voltage is equal or less than the "working" voltage specification.
                    Therefore, the maximum power and efficiency is highest when the
                    output voltage is relatively constant near the "working" voltage.

                    The main storage capacitor value should be at least 2200uF up to 1F
                    super caps. Large caps have an initial long wait befor reaching the
                    trigger level. However once charged they have only a small voltage
                    drop during each discharge since the pulse duration is set by the
                    timing cap. The advantage is that the main capacitor average volatge
                    remains near the working voltage for optimum charging efficiency.

                    The super caps should be low ESR type to minimize the internal
                    voltage drop. These caps are generally rated at 2.5V so use two in
                    series for 5V applications.

                    The timing cap should be adjusted so that the duration of each motor
                    current pulse causes the popper to pivot around the stalled wheel so
                    that the waggle motion is primarily forward to avoid wasting energy
                    on the side to side motion.

                    If a smaller storage cap is used, make sure that the 1381 resets
                    befor the motor stops turning otherwise you will be wasting time and
                    energy by continuing to discharge the cap when the motor has already
                    stopped.

                    The 1381 trigger voltage adjusts the level to which the power supply
                    cap is charged just before the popper triggers. This voltage should
                    be high enough to cause a good kick from the motor that overcomes the
                    static friction and the eneriat of the mobile platform.

                    The motor specs, wheel diameter and platform mass interact so much
                    that it is difficult to optimize the design by specification alone
                    and the mechanics must be always be evaluated by experiment.

                    In conclusion, overall performance of the photopopper requires a
                    careful mariage of mechanics and electonics and can only be optimized
                    by testing in a real live robot preferably in competion with other
                    designs.

                    May the best tech win!

                    wilf


                    --- In beam@yahoogroups.com, "Scott Lindroth" <scott.lindroth@...>
                    wrote:
                    >
                    > So far Miller Popper 4a is the winner, and I succumbed to the
                    temptation to build a
                    > photopopper with that circuit. It works well!
                    >
                    > The circuits seem more responsive when the resistor between the
                    trigger and the base of
                    > the transistor is omitted. Not sure if this is kosher. I've also
                    gone back to using a 4700 uf
                    > storage cap. The motors trigger with more punch than the supercap.
                    >
                    > I hope to have time to work more carefully with the other
                    experimental circuits (3 & 4) over
                    > the weekend. Thanks again!
                    >
                    > Scott
                    >
                    >
                    > --- In beam@yahoogroups.com, "wilf rigter" <wrigter@> wrote:
                    > >
                    > > I posted yet another circuit: Miller Popper V4a, which I think
                    will work better than V4
                    > >
                    > > wilf
                    > >
                    > > ----- Original Message -----
                    > > From: wilf_nv
                    > > To: beam@yahoogroups.com
                    > > Sent: Sunday, August 27, 2006 6:50 PM
                    > > Subject: [beam] Re: Miller Popper
                    > >
                    > >
                    > > Curiouser and curiouser!
                    > >
                    > > It is not immediately apparent what causes the different
                    behaviors of
                    > > the 1381G and 1381L but while analyzing that problem, I
                    discovered
                    > > another design flaw that reduces the effectiveness of this
                    circuit.
                    > > The problem is that adding components in series with the 1381
                    ground
                    > > pin raises the 1381 output voltage while charging caused by the
                    > > voltage drop of the 1381 ground pin current (~1uA). If the 1381
                    > > output pin voltage, which is connected to the base of the
                    output
                    > > transistor, exceeds 500mV it may cause the transistor to
                    partially
                    > > turn on and leak solar cell current to ground.
                    > >
                    > > The original Miller Miller uses Si diode in series with the
                    ground
                    > > pin which raises the output by 100mV. The Miller Popper V1 adds
                    a PD
                    > > (300mV) in series which raises the transistor base voltage
                    while
                    > > charging to ~400mV. When the 1M resistors are added, this
                    offset will
                    > > be even higher so Miller Popper V2 is out the window.
                    > >
                    > > The likely cause of the different behavior of the 1381G and
                    1381L in
                    > > the Miller Popper circuits is that the higher threshold of the
                    1381L
                    > > is accompanied by higher ground pin current and raises the
                    output
                    > > voltage closer to 500mV.
                    > >
                    > > There are three possible solutions (Miller Popper V3, V4 and
                    V5) to
                    > > reduce the 1381 output offset during charging which I will post
                    > > shortly in the file section :
                    > >
                    > > 1) reverse the PDs so that their photovoltaic voltage
                    substracts from
                    > > the series voltage drop.
                    > >
                    > > 2) connect the PDs in series between the 1381 ground pins so
                    that the
                    > > difference of the PD photovoltaic voltages and currents will
                    modify
                    > > the two diode drops.
                    > >
                    > > 3) place the PDs in series with the 1381 Vcc pin while the 1uF
                    timing
                    > > cap remains directly connected to the 1381 Vcc and ground pins.
                    > >
                    > > The last solution is interesting for increasing the Miller
                    Engine
                    > > threshold. For example connecting a LED between Vcc and the
                    positive
                    > > side of the timing) allows the trigger level of the ME to be
                    > > increased by 2V without side effects
                    > >
                    > > No time for me to test these circuits so they are experimental
                    but
                    > > you would do me a big favour by trying them out and reporting
                    the
                    > > results.
                    > >
                    > > good luck
                    > >
                    > > wilf
                    > >
                    > > --- In beam@yahoogroups.com, "Scott Lindroth" <scott.lindroth@>
                    > > wrote:
                    > > >
                    > > > Thanks, Wilf, but the 100K trim pot still doesn't do anything
                    with
                    > > the 1M resistors and the
                    > > > 1381 L triggers. The circuit works very well with 1381 G
                    triggers
                    > > with (and without) the
                    > > > 1M resistors.
                    > > >
                    > > > SL
                    > > >
                    > > > --- In beam@yahoogroups.com, "wilf_nv" <wrigter@> wrote:
                    > > > >
                    > > > >
                    > > > >
                    > > > >
                    > > > > Hello Scott,
                    > > > >
                    > > > > It's been a while since I reviewed that design. I posted a
                    > > revision
                    > > > > which should make the circuit more adjustable albeit at
                    increased
                    > > > > leakage current during charging (~10uA)
                    > > > >
                    > > > > The 1M resistors together with the 100K pot form two
                    voltage
                    > > dividers
                    > > > > between +V and 0V. Adjusting the pot will raise and lower
                    the
                    > > > > respective reference voltage of each 1381, raising the
                    trigger
                    > > level
                    > > > > of of one 1381 while lowering the other. This should allow
                    you to
                    > > > > trim any difference in thresholds between 1381 pairs.
                    > > > >
                    > > > > wilf
                    > > > >
                    > > > > --- In beam@yahoogroups.com, "scott.lindroth"
                    <scott.lindroth@>
                    > > > > wrote:
                    > > > > >
                    > > > > > First of all, I've just discovered BEAM robots and enjoy
                    them a
                    > > > > great deal. I'm a
                    > > > > > professional musician by trade, but a novice in
                    electronics.
                    > > I've
                    > > > > built 3 solar rollers using
                    > > > > > the Miller engine and am now attempting the Miller Popper
                    from
                    > > the
                    > > > > solarbotics.net
                    > > > > > website.
                    > > > > >
                    > > > > > The circuit is more or less working on a breadboard --
                    shine a
                    > > > > light on one photodiode,
                    > > > > > and the opposite motor fires. Since I'm supplying about
                    4.5v
                    > > from
                    > > > > the solar cells and am
                    > > > > > using fairly large motors from standard cassette decks, I
                    am
                    > > using
                    > > > > 1381 L voltage triggers
                    > > > > > with a .33F 5.5v supercap to add more umph. The motors
                    respond
                    > > > > with a lot of zip and
                    > > > > > fire about once every 1.25 seconds.
                    > > > > >
                    > > > > > The problem I _think_ I'm having is that the 1381 L
                    voltage
                    > > > > triggers seem highly variable in
                    > > > > > their response, and I can't find two that respond
                    equally.
                    > > Right
                    > > > > now, one 1381 triggers
                    > > > > > much more easily than the other when the light is evenly
                    > > > > distributed across the
                    > > > > > photodiodes. When I swap the position of 1381s the motor
                    that
                    > > > > never triggered
                    > > > > > before wakes up and fires more readily (as expected). I'm
                    > > > > suspicious of this diagnosis
                    > > > > > since the problem seems to occur with every pair of 1381s
                    I
                    > > try.
                    > > > > >
                    > > > > > What's more, the 100K? trim pot does not have any
                    noticeable
                    > > > > effect; it does not
                    > > > > > compensate for the variable trigger thresholds.
                    > > > > >
                    > > > > > I hope this description of the problem is useful. Thanks
                    for
                    > > any
                    > > > > advice.
                    > > > > >
                    > > > > > SL
                    > > > > >
                    > > > >
                    > > >
                    > >
                    >
                  • wilf rigter
                    I have posted Miller Popper V6 which uses just one 1381 and two darlington transistors capable of driving a 1A load. D2 and D3 should be Schottky or Ge type
                    Message 9 of 9 , Sep 4, 2006
                      I have posted Miller Popper V6 which uses just one 1381 and two darlington transistors capable of driving a 1A load.
                      D2 and D3 should be Schottky or Ge type for best performance.
                       
                      The next version, V7, uses plain NPN transistors and is more effcicient up to 300mA because of the lower output  transistor saturation voltage. Because of that lower voltage drop, D1-3 can all be 1N914s or equal.
                       
                      wilf
                       
                       
                      ----- Original Message -----
                      From: wilf_nv
                      Sent: Sunday, August 27, 2006 6:50 PM
                      Subject: [beam] Re: Miller Popper

                      Curiouser and curiouser!

                      It is not immediately apparent what causes the different behaviors of
                      the 1381G and 1381L but while analyzing that problem, I discovered
                      another design flaw that reduces the effectiveness of this circuit.
                      The problem is that adding components in series with the 1381 ground
                      pin raises the 1381 output voltage while charging caused by the
                      voltage drop of the 1381 ground pin current (~1uA). If the 1381
                      output pin voltage, which is connected to the base of the output
                      transistor, exceeds 500mV it may cause the transistor to partially
                      turn on and leak solar cell current to ground.

                      The original Miller Miller uses Si diode in series with the ground
                      pin which raises the output by 100mV. The Miller Popper V1 adds a PD
                      (300mV) in series which raises the transistor base voltage while
                      charging to ~400mV. When the 1M resistors are added, this offset will
                      be even higher so Miller Popper V2 is out the window.

                      The likely cause of the different behavior of the 1381G and 1381L in
                      the Miller Popper circuits is that the higher threshold of the 1381L
                      is accompanied by higher ground pin current and raises the output
                      voltage closer to 500mV.

                      There are three possible solutions (Miller Popper V3, V4 and V5) to
                      reduce the 1381 output offset during charging which I will post
                      shortly in the file section :

                      1) reverse the PDs so that their photovoltaic voltage substracts from
                      the series voltage drop.

                      2) connect the PDs in series between the 1381 ground pins so that the
                      difference of the PD photovoltaic voltages and currents will modify
                      the two diode drops.

                      3) place the PDs in series with the 1381 Vcc pin while the 1uF timing
                      cap remains directly connected to the 1381 Vcc and ground pins.

                      The last solution is interesting for increasing the Miller Engine
                      threshold. For example connecting a LED between Vcc and the positive
                      side of the timing) allows the trigger level of the ME to be
                      increased by 2V without side effects

                      No time for me to test these circuits so they are experimental but
                      you would do me a big favour by trying them out and reporting the
                      results.

                      good luck

                      wilf

                      --- In beam@yahoogroups. com, "Scott Lindroth" <scott.lindroth@ ...>
                      wrote:
                      >
                      > Thanks, Wilf, but the 100K trim pot still doesn't do anything with
                      the 1M resistors and the
                      > 1381 L triggers. The circuit works very well with 1381 G triggers
                      with (and without) the
                      > 1M resistors.
                      >
                      > SL
                      >
                      > --- In beam@yahoogroups. com, "wilf_nv" <wrigter@> wrote:
                      > >
                      > >
                      > >
                      > >
                      > > Hello Scott,
                      > >
                      > > It's been a while since I reviewed that design. I posted a
                      revision
                      > > which should make the circuit more adjustable albeit at increased
                      > > leakage current during charging (~10uA)
                      > >
                      > > The 1M resistors together with the 100K pot form two voltage
                      dividers
                      > > between +V and 0V. Adjusting the pot will raise and lower the
                      > > respective reference voltage of each 1381, raising the trigger
                      level
                      > > of of one 1381 while lowering the other. This should allow you to
                      > > trim any difference in thresholds between 1381 pairs.
                      > >
                      > > wilf
                      > >
                      > > --- In beam@yahoogroups. com, "scott.lindroth" <scott.lindroth@ >
                      > > wrote:
                      > > >
                      > > > First of all, I've just discovered BEAM robots and enjoy them a
                      > > great deal. I'm a
                      > > > professional musician by trade, but a novice in electronics.
                      I've
                      > > built 3 solar rollers using
                      > > > the Miller engine and am now attempting the Miller Popper from
                      the
                      > > solarbotics. net
                      > > > website.
                      > > >
                      > > > The circuit is more or less working on a breadboard -- shine a
                      > > light on one photodiode,
                      > > > and the opposite motor fires. Since I'm supplying about 4.5v
                      from
                      > > the solar cells and am
                      > > > using fairly large motors from standard cassette decks, I am
                      using
                      > > 1381 L voltage triggers
                      > > > with a .33F 5.5v supercap to add more umph. The motors respond
                      > > with a lot of zip and
                      > > > fire about once every 1.25 seconds.
                      > > >
                      > > > The problem I _think_ I'm having is that the 1381 L voltage
                      > > triggers seem highly variable in
                      > > > their response, and I can't find two that respond equally.
                      Right
                      > > now, one 1381 triggers
                      > > > much more easily than the other when the light is evenly
                      > > distributed across the
                      > > > photodiodes. When I swap the position of 1381s the motor that
                      > > never triggered
                      > > > before wakes up and fires more readily (as expected). I'm
                      > > suspicious of this diagnosis
                      > > > since the problem seems to occur with every pair of 1381s I
                      try.
                      > > >
                      > > > What's more, the 100K? trim pot does not have any noticeable
                      > > effect; it does not
                      > > > compensate for the variable trigger thresholds.
                      > > >
                      > > > I hope this description of the problem is useful. Thanks for
                      any
                      > > advice.
                      > > >
                      > > > SL
                      > > >
                      > >
                      >

                    Your message has been successfully submitted and would be delivered to recipients shortly.