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

Re: [softrock40] Re: Packaging external filters

Expand Messages
  • Kees & Sandy
    Hello Brad and the gang, I figured you would jump in with some good techie advice and/or experience. Very good point. That, I did not think about and look
    Message 1 of 27 , May 7, 2013
    • 0 Attachment
      Hello Brad and the gang,

      I figured you would jump in with some good techie advice and/or experience.

      Very good point. That, I did not think about and look into. The spec for the G6H OMRON relay does say "low magnetic interference enables high-density mounting", but gives no additional details. I took one apart looking for shielding but can't tell if there is any u metal ? in there. The cross-bar Au contacts look nice though and they were really inexpensive in the 2F package.

      73 Kees K5BCQ


      ---------- Original Message ----------
      From: Brad Thompson <brad.thompson@...>
      To: softrock40@yahoogroups.com
      Cc: Kees & Sandy <windy10605@...>
      Subject: Re: [softrock40] Re: Packaging external filters
      Date: Tue, 07 May 2013 15:46:25 -0400

      On 5/7/2013 11:36 AM, Kees & Sandy wrote:
      <snip>

      > The relays all have bifurcated, gold covered, contacts which are also
      > paralleled for reliability (DPDT contacts). The other point is to only
      > have one common connection between the antenna/earth ground and the
      > logic ground for controlling the relays.
      <snip>

      Hello, Kees and the group--

      One point to keep in mind centers on magnetic interactions
      among the relays. IIRC, the plastic-housed box-shaped little
      relays offered by Nais and other manufacturers may require
      at least a minimum separation between the relays, so check
      the data sheets for whatever relays you plan to use.

      73--

      Brad AA1IP
    • Kees & Sandy
      Hi Dave and the group. Great data, Dave. This is FR-4 (Advanced Circuits) so 70mils may be a little too narrow. I haven t measured it and just took data I
      Message 2 of 27 , May 7, 2013
      • 0 Attachment
        Hi Dave and the group.
         
        Great data, Dave. 
         
        This is FR-4 (Advanced Circuits) so 70mils may be a little too narrow. I haven't measured it and just took data I found on the Internet (has to be correct, right ?). I defer to your experience and will widen the traces to 100mils.
         
        73 Kees K5BCQ


        ---------- Original Message ----------
        From: Dave <dave@...>
        To: softrock40@yahoogroups.com
        Subject: Re: [softrock40] Re: Packaging external filters
        Date: Tue, 07 May 2013 13:09:15 -0700

         

        Kees and all:
          Don't know what substrate your pcb is.  But, FR4 epoxy glass, .062" thick, 1 oz copper, 50 ohms is real close to .100" wide. 

        Dave - WB6DHW
        <http://wb6dhw.com>

        On 5/7/2013 11:58 AM, Graeme wrote:
         

        Hi Kees,

        Looks like you are already using "best practice" construction and yes the printed transmission lines are just fine. Only reason I avoided them is they are a bit tricky for home made pcb. Yes you get a stub on the unused filters so I make the through filter the last in the chain them 6 M, 10 M etc. to keep the stub short. On 160 M with the longest stub it shows a a little bit of capacitance on the input and output but does not affect the response. My relays were recovered from an old telephone exchange and probably are good contacts but really unknown although they work just fine.

        I am using 1206 smt ceramic NPO capacitors rated at 200 Volts and have not had a failure even when testing into open circuit. The RD16HHF1's survived also. I am very impressed with them.

        73 Graeme ZL2APV

        --- In softrock40@yahoogroups.com, "Kees & Sandy" <windy10605@...> wrote:
        >
        > Very good info, Graeme. I ran board traces down the two sides of the board, one for input and one for output and made the traces 60-70mil wide. I've found that on a 62mil board, using 1oz copper, over a ground plane it comes pretty close to 50 ohms. It would still be nice to not have the unterminated stubs (to the unused filters), but I don't know what to do about that. The relays all have bifurcated, gold covered, contacts which are also paralleled for reliability (DPDT contacts). The other point is to only have one common connection between the antenna/earth ground and the logic ground for controlling the relays. Right now I'm using SMA connectors but may go back to BNC. Roger, on the bypass capacitor (and diode) across each coil. Always good to have bypass caps on DC circuits in an RF environment. The only "tricky" layout is that I'm putting a single RD16HHF1 right in the middle of the board as an Amp (couldn't see all that space going to waste). Have you ever used NPO, high voltage, ceramic smt capacitors in the filters ? Have you found any reason to use 5V relays vs 12V relays ? .....I got it in my mind that the lower 5V coils (less impedance) would be less suceptible to RF pickup. Probably a DNW (Does Nothing What-so-ever). 73 Kees K5BCQ
        > ---------- Original Message ----------
        > From: Graeme Jury <gvjury@...>
        > To: softrock40@yahoogroups.com
        > Subject: [softrock40] Re: Packaging external filters
        > Date: Tue, 07 May 2013 15:07:23 +1200
        >
        >
        > <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd">
        > There is a couple of things which may be useful when designing switched
        > banks of filters.
        >
        > 1. Use a transmission line down the relay chain. In my case I used
        > RG-179 with the cover stripped back 5 mm or so and the braid tinned. Two
        > knife cuts 2 mm apart were rolled around the centre of the tinned braid
        > and then bent to snap at the knife cuts. The 2 mm piece of braid was
        > pried away and the centre of the coax bared. This was done opposite each
        > relay. The braids were soldered to the earth plane and the centre
        > connected to the relay contact. It did not take long to prepare the coax
        > and avoids the tricky and less reliable design of a printed transmission
        > line.
        >
        > 2. Bypass the coils of each relay. I ran my filter bank through my VNA
        > and could not understand why the roll off at the HF end of the filter
        > went down to about 40 dB and then flattened. They were mesh Chebyshev
        > filters and should have kept going down. In the end I found that a 0.1
        > smt cap on the coil on the 12 Volt common side of the relays fixed the
        > problem. As I added caps to more of the coils it got better and when I
        > had done half of them it was improved to the noise floor of the VNA.
        > Naturally for the negligible cost I did them all and also the switched
        > grounded side of the coils too. It is an obscure fault and maybe if I
        > had used RF relays it would not have happened but I used standard 2 pole
        > changeover miniature ones which switch 30 watts from my RD16HHF1 linear
        > on CW just fine.
        >
        > 73 Graeme ZL2APV
        >

         

         

      • Dave
        Kees: I haven t actually measured the impedance of my traces. But, there are a number of calculators on the internet that calculate either the impedance or
        Message 3 of 27 , May 7, 2013
        • 0 Attachment
          Kees:
            I haven't actually measured the impedance of my traces.  But, there are a number of calculators on the internet that calculate either the impedance or the width.  Note also, the .100" is if there is not a ground near the trace.  The protoype board for the LF2uW, I put ground on the top .01" away from the trace.  According to the calculator, this is 50 ohms.  If the ground is .008", as .040 trace is 50 ohms.

          Dave - WB6DHW
          <http://wb6dhw.com>

          On 5/7/2013 2:19 PM, Kees & Sandy wrote:
           
          Hi Dave and the group.
           
          Great data, Dave. 
           
          This is FR-4 (Advanced Circuits) so 70mils may be a little too narrow. I haven't measured it and just took data I found on the Internet (has to be correct, right ?). I defer to your experience and will widen the traces to 100mils.
           
          73 Kees K5BCQ


          ---------- Original Message ----------
          From: Dave <dave@...>
          To: softrock40@yahoogroups.com
          Subject: Re: [softrock40] Re: Packaging external filters
          Date: Tue, 07 May 2013 13:09:15 -0700

           

          Kees and all:
            Don't know what substrate your pcb is.  But, FR4 epoxy glass, .062" thick, 1 oz copper, 50 ohms is real close to .100" wide. 

          Dave - WB6DHW
          <http://wb6dhw.com>

          On 5/7/2013 11:58 AM, Graeme wrote:
           

          Hi Kees,

          Looks like you are already using "best practice" construction and yes the printed transmission lines are just fine. Only reason I avoided them is they are a bit tricky for home made pcb. Yes you get a stub on the unused filters so I make the through filter the last in the chain them 6 M, 10 M etc. to keep the stub short. On 160 M with the longest stub it shows a a little bit of capacitance on the input and output but does not affect the response. My relays were recovered from an old telephone exchange and probably are good contacts but really unknown although they work just fine.

          I am using 1206 smt ceramic NPO capacitors rated at 200 Volts and have not had a failure even when testing into open circuit. The RD16HHF1's survived also. I am very impressed with them.

          73 Graeme ZL2APV

          --- In softrock40@yahoogroups.com, "Kees & Sandy" <windy10605@...> wrote:
          >
          > Very good info, Graeme. I ran board traces down the two sides of the board, one for input and one for output and made the traces 60-70mil wide. I've found that on a 62mil board, using 1oz copper, over a ground plane it comes pretty close to 50 ohms. It would still be nice to not have the unterminated stubs (to the unused filters), but I don't know what to do about that. The relays all have bifurcated, gold covered, contacts which are also paralleled for reliability (DPDT contacts). The other point is to only have one common connection between the antenna/earth ground and the logic ground for controlling the relays. Right now I'm using SMA connectors but may go back to BNC. Roger, on the bypass capacitor (and diode) across each coil. Always good to have bypass caps on DC circuits in an RF environment. The only "tricky" layout is that I'm putting a single RD16HHF1 right in the middle of the board as an Amp (couldn't see all that space going to waste). Have you ever used NPO, high voltage, ceramic smt capacitors in the filters ? Have you found any reason to use 5V relays vs 12V relays ? .....I got it in my mind that the lower 5V coils (less impedance) would be less suceptible to RF pickup. Probably a DNW (Does Nothing What-so-ever). 73 Kees K5BCQ
          > ---------- Original Message ----------
          > From: Graeme Jury <gvjury@...>
          > To: softrock40@yahoogroups.com
          > Subject: [softrock40] Re: Packaging external filters
          > Date: Tue, 07 May 2013 15:07:23 +1200
          >
          >
          > <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd">
          > There is a couple of things which may be useful when designing switched
          > banks of filters.
          >
          > 1. Use a transmission line down the relay chain. In my case I used
          > RG-179 with the cover stripped back 5 mm or so and the braid tinned. Two
          > knife cuts 2 mm apart were rolled around the centre of the tinned braid
          > and then bent to snap at the knife cuts. The 2 mm piece of braid was
          > pried away and the centre of the coax bared. This was done opposite each
          > relay. The braids were soldered to the earth plane and the centre
          > connected to the relay contact. It did not take long to prepare the coax
          > and avoids the tricky and less reliable design of a printed transmission
          > line.
          >
          > 2. Bypass the coils of each relay. I ran my filter bank through my VNA
          > and could not understand why the roll off at the HF end of the filter
          > went down to about 40 dB and then flattened. They were mesh Chebyshev
          > filters and should have kept going down. In the end I found that a 0.1
          > smt cap on the coil on the 12 Volt common side of the relays fixed the
          > problem. As I added caps to more of the coils it got better and when I
          > had done half of them it was improved to the noise floor of the VNA.
          > Naturally for the negligible cost I did them all and also the switched
          > grounded side of the coils too. It is an obscure fault and maybe if I
          > had used RF relays it would not have happened but I used standard 2 pole
          > changeover miniature ones which switch 30 watts from my RD16HHF1 linear
          > on CW just fine.
          >
          > 73 Graeme ZL2APV
          >

           

           


        • Brad Thompson
          ... Hello, Kees-- Thanks for your comments. I looked at the G6E data sheet... http://www.omron.com/ecb/products/pdf/en-g6e.pdf ...And found a drawing on page 4
          Message 4 of 27 , May 7, 2013
          • 0 Attachment
            On 5/7/2013 5:01 PM, Kees & Sandy wrote:
            > Hello Brad and the gang,
            >
            > I figured you would jump in with some good techie advice and/or experience.
            >
            > Very good point. That, I did not think about and look into. The spec for the G6H OMRON relay does say "low magnetic interference enables high-density mounting", but gives no additional details. I took one apart looking for shielding but can't tell if there is any u metal ? in there. The cross-bar Au contacts look nice though and they were really inexpensive in the 2F package.
            >
            > 73 Kees K5BCQ
            >

            Hello, Kees--

            Thanks for your comments. I looked at the G6E data sheet...

            http://www.omron.com/ecb/products/pdf/en-g6e.pdf

            ...And found a drawing on page 4 that shows effects
            of nearby relays; If my interpretation is correct,
            packing the relays closer together than in the drawing will
            increase the effects. With the spacings shown, operate and
            release voltages change by something less than 1 percent.

            Also, note that the G6E is a relative of the G6H, so we
            may be contemplating apples and oranges<g>.

            73--

            Brad AA1IP
          • Kees & Sandy
            Dave, I tried several of those and found them confusing because the focus seems to be on GHz stripline impedances where you have one stripline on one side and
            Message 5 of 27 , May 7, 2013
            • 0 Attachment
              Dave,
               
              I tried several of those and found them confusing because the focus seems to be on GHz stripline impedances where you have one stripline on one side and a ground plane on the other and no soldermask. ....and yes, one stripline over a ground plane 62 mills away is around 50 ohms if it is about 110 mills wide. This is QRP power at HF frequencies so I was also using floodfill ground on the trace side....separated by say, 20 mils on one or two sides and 62 mills away from a full ground plane on the other side. Assume 1oz Cu is 1.37 mills and permeablity varies from 1 for air to 4.3 for the Advanced Circuits FR-4 board. Guess the only way to really know is to build it and measure it (but that costs) ......got any ideas ? 
               
              Or just light'er up and check for smoke.
               
              73 kees K5BCQ
               
              --------- Original Message ----------
              From: Dave <dave@...>
              To: softrock40@yahoogroups.com
              Subject: Re: [softrock40] Re: Packaging external filters
              Date: Tue, 07 May 2013 17:08:49 -0700

               
              Kees:
              � I haven't actually measured the impedance of my traces.� But, there are a number of calculators on the internet that calculate either the impedance or the width.� Note also, the .100" is if there is not a ground near the trace.� The protoype board for the LF2uW, I put ground on the top .01" away from the trace.� According to the calculator, this is 50 ohms.� If the ground is .008", as .040 trace is 50 ohms.

              Dave - WB6DHW
              <http://wb6dhw.com>

              On 5/7/2013 2:19 PM, Kees & Sandy wrote:
              Hi Dave and the group.
              Great data, Dave.�
              This is FR-4 (Advanced Circuits) so 70mils may be a little too narrow. I haven't measured it and just took data I found on the Internet (has to be correct, right ?). I defer to your experience and will widen the traces to 100mils.
              73 Kees K5BCQ


              ---------- Original Message ----------
              From: Dave <dave@...>
              To: softrock40@yahoogroups.com
              Subject: Re: [softrock40] Re: Packaging external filters
              Date: Tue, 07 May 2013 13:09:15 -0700

              Kees and all:
              � Don't know what substrate your pcb is.� But, FR4 epoxy glass, .062" thick, 1 oz copper, 50 ohms is real close to .100" wide.�

              Dave - WB6DHW
              <http://wb6dhw.com>

              On 5/7/2013 11:58 AM, Graeme wrote:

              Hi Kees,

              Looks like you are already using "best practice" construction and yes the printed transmission lines are just fine. Only reason I avoided them is they are a bit tricky for home made pcb. Yes you get a stub on the unused filters so I make the through filter the last in the chain them 6 M, 10 M etc. to keep the stub short. On 160 M with the longest stub it shows a a little bit of capacitance on the input and output but does not affect the response. My relays were recovered from an old telephone exchange and probably are good contacts but really unknown although they work just fine.

              I am using 1206 smt ceramic NPO capacitors rated at 200 Volts and have not had a failure even when testing into open circuit. The RD16HHF1's survived also. I am very impressed with them.

              73 Graeme ZL2APV

              --- In softrock40@yahoogroups.com, "Kees & Sandy" <windy10605@...> wrote:
              >
              > Very good info, Graeme. I ran board traces down the two sides of the board, one for input and one for output and made the traces 60-70mil wide. I've found that on a 62mil board, using 1oz copper, over a ground plane it comes pretty close to 50 ohms. It would still be nice to not have the unterminated stubs (to the unused filters), but I don't know what to do about that. The relays all have bifurcated, gold covered, contacts which are also paralleled for reliability (DPDT contacts). The other point is to only have one common connection between the antenna/earth ground and the logic ground for controlling the relays. Right now I'm using SMA connectors but may go back to BNC. Roger, on the bypass capacitor (and diode) across each coil. Always good to have bypass caps on DC circuits in an RF environment. The only "tricky" layout is that I'm putting a single RD16HHF1 right in the middle of the board as an Amp (couldn't see all that space going to waste). Have you ever used NPO, high voltage, ceramic smt capacitors in the filters ? Have you found any reason to use 5V relays vs 12V relays ? .....I got it in my mind that the lower 5V coils (less impedance) would be less suceptible to RF pickup. Probably a DNW (Does Nothing What-so-ever). 73 Kees K5BCQ
              > ---------- Original Message ----------
              > From: Graeme Jury <gvjury@...>
              > To: softrock40@yahoogroups.com
              > Subject: [softrock40] Re: Packaging external filters
              > Date: Tue, 07 May 2013 15:07:23 +1200
              >
              >
              > <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd">
              > There is a couple of things which may be useful when designing switched
              > banks of filters.
              >
              > 1. Use a transmission line down the relay chain. In my case I used
              > RG-179 with the cover stripped back 5 mm or so and the braid tinned. Two
              > knife cuts 2 mm apart were rolled around the centre of the tinned braid
              > and then bent to snap at the knife cuts. The 2 mm piece of braid was
              > pried away and the centre of the coax bared. This was done opposite each
              > relay. The braids were soldered to the earth plane and the centre
              > connected to the relay contact. It did not take long to prepare the coax
              > and avoids the tricky and less reliable design of a printed transmission
              > line.
              >
              > 2. Bypass the coils of each relay. I ran my filter bank through my VNA
              > and could not understand why the roll off at the HF end of the filter
              > went down to about 40 dB and then flattened. They were mesh Chebyshev
              > filters and should have kept going down. In the end I found that a 0.1
              > smt cap on the coil on the 12 Volt common side of the relays fixed the
              > problem. As I added caps to more of the coils it got better and when I
              > had done half of them it was improved to the noise floor of the VNA.
              > Naturally for the negligible cost I did them all and also the switched
              > grounded side of the coils too. It is an obscure fault and maybe if I
              > had used RF relays it would not have happened but I used standard 2 pole
              > changeover miniature ones which switch 30 watts from my RD16HHF1 linear
              > on CW just fine.
              >
              > 73 Graeme ZL2APV
              >

               

            • Kees & Sandy
              Ahhhh. My OMRON G6H documentation has no such drawing on page 4 ....but it does have a Discontinued stamp on page 1. those always make the best and least
              Message 6 of 27 , May 7, 2013
              • 0 Attachment
                Ahhhh. My OMRON G6H documentation has no such drawing on page 4 ....but it does have a "Discontinued" stamp on page 1. those always make the best and least expensive surplus parts.

                73 Kees K5BCQ

                ---------- Original Message ----------
                From: Brad Thompson <brad.thompson@...>
                To: softrock40@yahoogroups.com
                Cc: Kees & Sandy <windy10605@...>
                Subject: Re: [softrock40] Re: Packaging external filters
                Date: Tue, 07 May 2013 21:10:30 -0400

                On 5/7/2013 5:01 PM, Kees & Sandy wrote:
                > Hello Brad and the gang,
                >
                > I figured you would jump in with some good techie advice and/or experience.
                >
                > Very good point. That, I did not think about and look into. The spec for the G6H OMRON relay does say "low magnetic interference enables high-density mounting", but gives no additional details. I took one apart looking for shielding but can't tell if there is any u metal ? in there. The cross-bar Au contacts look nice though and they were really inexpensive in the 2F package.
                >
                > 73 Kees K5BCQ
                >

                Hello, Kees--

                Thanks for your comments. I looked at the G6E data sheet...

                http://www.omron.com/ecb/products/pdf/en-g6e.pdf

                ...And found a drawing on page 4 that shows effects
                of nearby relays; If my interpretation is correct,
                packing the relays closer together than in the drawing will
                increase the effects. With the spacings shown, operate and
                release voltages change by something less than 1 percent.

                Also, note that the G6E is a relative of the G6H, so we
                may be contemplating apples and oranges<g>.

                73--

                Brad AA1IP
              • Dave
                Kees: As you bring the floodfill ground closer to the trace, the impedance goes down. Some of the calculators will figure that out. At .020 , you width of
                Message 7 of 27 , May 7, 2013
                • 0 Attachment
                  Kees:
                    As you bring the floodfill ground closer to the trace, the impedance goes down.  Some of the calculators will figure that out.  At .020", you width of .06 to .07 should be close to 50 ohms.

                  Dave -wb6dhw.com
                  <http://wb6dhw.com>

                  On 5/7/2013 6:47 PM, Kees & Sandy wrote:
                   
                  Dave,
                   
                  I tried several of those and found them confusing because the focus seems to be on GHz stripline impedances where you have one stripline on one side and a ground plane on the other and no soldermask. ....and yes, one stripline over a ground plane 62 mills away is around 50 ohms if it is about 110 mills wide. This is QRP power at HF frequencies so I was also using floodfill ground on the trace side....separated by say, 20 mils on one or two sides and 62 mills away from a full ground plane on the other side. Assume 1oz Cu is 1.37 mills and permeablity varies from 1 for air to 4.3 for the Advanced Circuits FR-4 board. Guess the only way to really know is to build it and measure it (but that costs) ......got any ideas ? 
                   
                  Or just light'er up and check for smoke.
                   
                  73 kees K5BCQ
                   
                  --------- Original Message ----------
                  From: Dave <dave@...>
                  To: softrock40@yahoogroups.com
                  Subject: Re: [softrock40] Re: Packaging external filters
                  Date: Tue, 07 May 2013 17:08:49 -0700

                   
                  Kees:
                  � I haven't actually measured the impedance of my traces.� But, there are a number of calculators on the internet that calculate either the impedance or the width.� Note also, the .100" is if there is not a ground near the trace.� The protoype board for the LF2uW, I put ground on the top .01" away from the trace.� According to the calculator, this is 50 ohms.� If the ground is .008", as .040 trace is 50 ohms.

                  Dave - WB6DHW
                  <http://wb6dhw.com>

                  On 5/7/2013 2:19 PM, Kees & Sandy wrote:
                  Hi Dave and the group.
                  Great data, Dave.�
                  This is FR-4 (Advanced Circuits) so 70mils may be a little too narrow. I haven't measured it and just took data I found on the Internet (has to be correct, right ?). I defer to your experience and will widen the traces to 100mils.
                  73 Kees K5BCQ


                  ---------- Original Message ----------
                  From: Dave <dave@...>
                  To: softrock40@yahoogroups.com
                  Subject: Re: [softrock40] Re: Packaging external filters
                  Date: Tue, 07 May 2013 13:09:15 -0700

                  Kees and all:
                  � Don't know what substrate your pcb is.� But, FR4 epoxy glass, .062" thick, 1 oz copper, 50 ohms is real close to .100" wide.�

                  Dave - WB6DHW
                  <http://wb6dhw.com>

                  On 5/7/2013 11:58 AM, Graeme wrote:

                  Hi Kees,

                  Looks like you are already using "best practice" construction and yes the printed transmission lines are just fine. Only reason I avoided them is they are a bit tricky for home made pcb. Yes you get a stub on the unused filters so I make the through filter the last in the chain them 6 M, 10 M etc. to keep the stub short. On 160 M with the longest stub it shows a a little bit of capacitance on the input and output but does not affect the response. My relays were recovered from an old telephone exchange and probably are good contacts but really unknown although they work just fine.

                  I am using 1206 smt ceramic NPO capacitors rated at 200 Volts and have not had a failure even when testing into open circuit. The RD16HHF1's survived also. I am very impressed with them.

                  73 Graeme ZL2APV

                  --- In softrock40@yahoogroups.com, "Kees & Sandy" <windy10605@...> wrote:
                  >
                  > Very good info, Graeme. I ran board traces down the two sides of the board, one for input and one for output and made the traces 60-70mil wide. I've found that on a 62mil board, using 1oz copper, over a ground plane it comes pretty close to 50 ohms. It would still be nice to not have the unterminated stubs (to the unused filters), but I don't know what to do about that. The relays all have bifurcated, gold covered, contacts which are also paralleled for reliability (DPDT contacts). The other point is to only have one common connection between the antenna/earth ground and the logic ground for controlling the relays. Right now I'm using SMA connectors but may go back to BNC. Roger, on the bypass capacitor (and diode) across each coil. Always good to have bypass caps on DC circuits in an RF environment. The only "tricky" layout is that I'm putting a single RD16HHF1 right in the middle of the board as an Amp (couldn't see all that space going to waste). Have you ever used NPO, high voltage, ceramic smt capacitors in the filters ? Have you found any reason to use 5V relays vs 12V relays ? .....I got it in my mind that the lower 5V coils (less impedance) would be less suceptible to RF pickup. Probably a DNW (Does Nothing What-so-ever). 73 Kees K5BCQ
                  > ---------- Original Message ----------
                  > From: Graeme Jury <gvjury@...>
                  > To: softrock40@yahoogroups.com
                  > Subject: [softrock40] Re: Packaging external filters
                  > Date: Tue, 07 May 2013 15:07:23 +1200
                  >
                  >
                  > <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd">
                  > There is a couple of things which may be useful when designing switched
                  > banks of filters.
                  >
                  > 1. Use a transmission line down the relay chain. In my case I used
                  > RG-179 with the cover stripped back 5 mm or so and the braid tinned. Two
                  > knife cuts 2 mm apart were rolled around the centre of the tinned braid
                  > and then bent to snap at the knife cuts. The 2 mm piece of braid was
                  > pried away and the centre of the coax bared. This was done opposite each
                  > relay. The braids were soldered to the earth plane and the centre
                  > connected to the relay contact. It did not take long to prepare the coax
                  > and avoids the tricky and less reliable design of a printed transmission
                  > line.
                  >
                  > 2. Bypass the coils of each relay. I ran my filter bank through my VNA
                  > and could not understand why the roll off at the HF end of the filter
                  > went down to about 40 dB and then flattened. They were mesh Chebyshev
                  > filters and should have kept going down. In the end I found that a 0.1
                  > smt cap on the coil on the 12 Volt common side of the relays fixed the
                  > problem. As I added caps to more of the coils it got better and when I
                  > had done half of them it was improved to the noise floor of the VNA.
                  > Naturally for the negligible cost I did them all and also the switched
                  > grounded side of the coils too. It is an obscure fault and maybe if I
                  > had used RF relays it would not have happened but I used standard 2 pole
                  > changeover miniature ones which switch 30 watts from my RD16HHF1 linear
                  > on CW just fine.
                  >
                  > 73 Graeme ZL2APV
                  >

                   

                  _

                • David Turnbull
                  ... http://oshpark.com/pricing a 2 square inch board would cost $10 and you’d get three 73 David AE9RB http://AE9RB.com/ On Tue, May 7, 2013 at 6:47 PM,
                  Message 8 of 27 , May 7, 2013
                  • 0 Attachment
                    On Tue, May 7, 2013 at 6:47 PM, Kees & Sandy <windy10605@...> wrote:
                    Guess the only way to really know is to build it and measure it (but that costs) ......got any ideas ? 

                    "a 2 square inch board would cost $10 and you’d get three"

                    73 David AE9RB

                  • Kees & Sandy
                    Dave, That s great. Do you have a link to one that takes floodfill into accont. I never found one of those. 73 Kees K5BCQ ... From: Dave To:
                    Message 9 of 27 , May 7, 2013
                    • 0 Attachment
                      Dave,
                       
                      That's great. Do you have a link to one that takes floodfill into accont. I never found one of those.
                       
                      73 Kees K5BCQ


                      ---------- Original Message ----------
                      From: Dave <dave@...>
                      To: softrock40@yahoogroups.com
                      Subject: Re: [softrock40] Re: Packaging external filters
                      Date: Tue, 07 May 2013 19:15:41 -0700

                       

                      Kees:
                      � As you bring the floodfill ground closer to the trace, the impedance goes down.� Some of the calculators will figure that out.� At .020", you width of .06 to .07 should be close to 50 ohms.

                      Dave -wb6dhw.com
                      <http://wb6dhw.com>

                      On 5/7/2013 6:47 PM, Kees & Sandy wrote:
                      Dave,
                      I tried several of those and found them confusing because the focus seems to be on GHz stripline impedances where you have one stripline on one side and a ground plane on the other and no soldermask. ....and yes, one stripline over a ground plane 62 mills away is around 50 ohms if it is about 110 mills wide. This is QRP power at HF frequencies so I was also using floodfill ground on the trace side....separated by say, 20 mils on one or two sides and 62 mills away from a full ground plane on the other side. Assume 1oz Cu is 1.37 mills and permeablity varies from 1 for air to 4.3 for the Advanced Circuits FR-4 board. Guess the only way to really know is to build it�and measure it (but that costs)�......got any ideas ?�
                      Or just light'er up and check for smoke.
                      73 kees K5BCQ
                      --------- Original Message ----------
                      From: Dave <dave@...>
                      To: softrock40@yahoogroups.com
                      Subject: Re: [softrock40] Re: Packaging external filters
                      Date: Tue, 07 May 2013 17:08:49 -0700

                      Kees:
                      � I haven't actually measured the impedance of my traces.� But, there are a number of calculators on the internet that calculate either the impedance or the width.� Note also, the .100" is if there is not a ground near the trace.� The protoype board for the LF2uW, I put ground on the top .01" away from the trace.� According to the calculator, this is 50 ohms.� If the ground is .008", as .040 trace is 50 ohms.

                      Dave - WB6DHW
                      <http://wb6dhw.com>

                      On 5/7/2013 2:19 PM, Kees & Sandy wrote:
                      Hi Dave and the group.
                      Great data, Dave.�
                      This is FR-4 (Advanced Circuits) so 70mils may be a little too narrow. I haven't measured it and just took data I found on the Internet (has to be correct, right ?). I defer to your experience and will widen the traces to 100mils.
                      73 Kees K5BCQ


                      ---------- Original Message ----------
                      From: Dave <dave@...>
                      To: softrock40@yahoogroups.com
                      Subject: Re: [softrock40] Re: Packaging external filters
                      Date: Tue, 07 May 2013 13:09:15 -0700

                      Kees and all:
                      � Don't know what substrate your pcb is.� But, FR4 epoxy glass, .062" thick, 1 oz copper, 50 ohms is real close to .100" wide.�

                      Dave - WB6DHW
                      <http://wb6dhw.com>

                      On 5/7/2013 11:58 AM, Graeme wrote:

                      Hi Kees,

                      Looks like you are already using "best practice" construction and yes the printed transmission lines are just fine. Only reason I avoided them is they are a bit tricky for home made pcb. Yes you get a stub on the unused filters so I make the through filter the last in the chain them 6 M, 10 M etc. to keep the stub short. On 160 M with the longest stub it shows a a little bit of capacitance on the input and output but does not affect the response. My relays were recovered from an old telephone exchange and probably are good contacts but really unknown although they work just fine.

                      I am using 1206 smt ceramic NPO capacitors rated at 200 Volts and have not had a failure even when testing into open circuit. The RD16HHF1's survived also. I am very impressed with them.

                      73 Graeme ZL2APV

                      --- In softrock40@yahoogroups.com, "Kees & Sandy" <windy10605@...> wrote:
                      >
                      > Very good info, Graeme. I ran board traces down the two sides of the board, one for input and one for output and made the traces 60-70mil wide. I've found that on a 62mil board, using 1oz copper, over a ground plane it comes pretty close to 50 ohms. It would still be nice to not have the unterminated stubs (to the unused filters), but I don't know what to do about that. The relays all have bifurcated, gold covered, contacts which are also paralleled for reliability (DPDT contacts). The other point is to only have one common connection between the antenna/earth ground and the logic ground for controlling the relays. Right now I'm using SMA connectors but may go back to BNC. Roger, on the bypass capacitor (and diode) across each coil. Always good to have bypass caps on DC circuits in an RF environment. The only "tricky" layout is that I'm putting a single RD16HHF1 right in the middle of the board as an Amp (couldn't see all that space going to waste). Have you ever used NPO, high voltage, ceramic smt capacitors in the filters ? Have you found any reason to use 5V relays vs 12V relays ? .....I got it in my mind that the lower 5V coils (less impedance) would be less suceptible to RF pickup. Probably a DNW (Does Nothing What-so-ever). 73 Kees K5BCQ
                      > ---------- Original Message ----------
                      > From: Graeme Jury <gvjury@...>
                      > To: softrock40@yahoogroups.com
                      > Subject: [softrock40] Re: Packaging external filters
                      > Date: Tue, 07 May 2013 15:07:23 +1200
                      >
                      >
                      > <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd">
                      > There is a couple of things which may be useful when designing switched
                      > banks of filters.
                      >
                      > 1. Use a transmission line down the relay chain. In my case I used
                      > RG-179 with the cover stripped back 5 mm or so and the braid tinned. Two
                      > knife cuts 2 mm apart were rolled around the centre of the tinned braid
                      > and then bent to snap at the knife cuts. The 2 mm piece of braid was
                      > pried away and the centre of the coax bared. This was done opposite each
                      > relay. The braids were soldered to the earth plane and the centre
                      > connected to the relay contact. It did not take long to prepare the coax
                      > and avoids the tricky and less reliable design of a printed transmission
                      > line.
                      >
                      > 2. Bypass the coils of each relay. I ran my filter bank through my VNA
                      > and could not understand why the roll off at the HF end of the filter
                      > went down to about 40 dB and then flattened. They were mesh Chebyshev
                      > filters and should have kept going down. In the end I found that a 0.1
                      > smt cap on the coil on the 12 Volt common side of the relays fixed the
                      > problem. As I added caps to more of the coils it got better and when I
                      > had done half of them it was improved to the noise floor of the VNA.
                      > Naturally for the negligible cost I did them all and also the switched
                      > grounded side of the coils too. It is an obscure fault and maybe if I
                      > had used RF relays it would not have happened but I used standard 2 pole
                      > changeover miniature ones which switch 30 watts from my RD16HHF1 linear
                      > on CW just fine.
                      >
                      > 73 Graeme ZL2APV
                      >

                      _

                       

                    • Kees & Sandy
                      Cool ! Is that what you use for your Peaberry boards ? what files are required ? gerber? 73 Kees K5BCQ ... From: David Turnbull To:
                      Message 10 of 27 , May 7, 2013
                      • 0 Attachment
                        Cool !  Is that what you use for your Peaberry boards ? what files are required ? gerber?
                         
                        73 Kees K5BCQ

                        ---------- Original Message ----------
                        From: David Turnbull <dturnbull@...>
                        To: softrock40@yahoogroups.com
                        Subject: Re: [softrock40] Re: Packaging external filters
                        Date: Tue, 7 May 2013 19:25:30 -0700

                         

                        On Tue, May 7, 2013 at 6:47 PM, Kees & Sandy <windy10605@...> wrote:
                        Guess the only way to really know is to build it�and measure it (but that costs)�......got any ideas ?�
                         
                         
                        "a 2 square inch board would cost $10 and you�d get three"
                         
                        73 David AE9RB
                         

                         

                      • Dave
                        Kees: The latest version of KICAD includes a pcb calculator. The grounded coplanar calculates the impedance with a trace over a groundplane and ground areas
                        Message 11 of 27 , May 7, 2013
                        • 0 Attachment
                           Kees:
                            The latest version of KICAD includes a pcb calculator.  The "grounded coplanar" calculates the impedance with a trace over a groundplane and ground areas on the sides.

                          Dave - WB6DHW
                          <http://wb6dhw.com>

                          On 5/7/2013 7:31 PM, Kees & Sandy wrote:
                           
                          Dave,
                           
                          That's great. Do you have a link to one that takes floodfill into accont. I never found one of those.
                           
                          73 Kees K5BCQ


                          ---------- Original Message ----------
                          From: Dave <dave@...>
                          To: softrock40@yahoogroups.com
                          Subject: Re: [softrock40] Re: Packaging external filters
                          Date: Tue, 07 May 2013 19:15:41 -0700

                           

                          Kees:
                          � As you bring the floodfill ground closer to the trace, the impedance goes down.� Some of the calculators will figure that out.� At .020", you width of .06 to .07 should be close to 50 ohms.

                          Dave -wb6dhw.com
                          <http://wb6dhw.com>

                          On 5/7/2013 6:47 PM, Kees & Sandy wrote:
                          Dave,
                          I tried several of those and found them confusing because the focus seems to be on GHz stripline impedances where you have one stripline on one side and a ground plane on the other and no soldermask. ....and yes, one stripline over a ground plane 62 mills away is around 50 ohms if it is about 110 mills wide. This is QRP power at HF frequencies so I was also using floodfill ground on the trace side....separated by say, 20 mils on one or two sides and 62 mills away from a full ground plane on the other side. Assume 1oz Cu is 1.37 mills and permeablity varies from 1 for air to 4.3 for the Advanced Circuits FR-4 board. Guess the only way to really know is to build it�and measure it (but that costs)�......got any ideas ?�
                          Or just light'er up and check for smoke.
                          73 kees K5BCQ
                          --------- Original Message ----------
                          From: Dave <dave@...>
                          To: softrock40@yahoogroups.com
                          Subject: Re: [softrock40] Re: Packaging external filters
                          Date: Tue, 07 May 2013 17:08:49 -0700

                          Kees:
                          � I haven't actually measured the impedance of my traces.� But, there are a number of calculators on the internet that calculate either the impedance or the width.� Note also, the .100" is if there is not a ground near the trace.� The protoype board for the LF2uW, I put ground on the top .01" away from the trace.� According to the calculator, this is 50 ohms.� If the ground is .008", as .040 trace is 50 ohms.

                          Dave - WB6DHW
                          <http://wb6dhw.com>

                          On 5/7/2013 2:19 PM, Kees & Sandy wrote:
                          Hi Dave and the group.
                          Great data, Dave.�
                          This is FR-4 (Advanced Circuits) so 70mils may be a little too narrow. I haven't measured it and just took data I found on the Internet (has to be correct, right ?). I defer to your experience and will widen the traces to 100mils.
                          73 Kees K5BCQ


                          ---------- Original Message ----------
                          From: Dave <dave@...>
                          To: softrock40@yahoogroups.com
                          Subject: Re: [softrock40] Re: Packaging external filters
                          Date: Tue, 07 May 2013 13:09:15 -0700

                          Kees and all:
                          � Don't know what substrate your pcb is.� But, FR4 epoxy glass, .062" thick, 1 oz copper, 50 ohms is real close to .100" wide.�

                          Dave - WB6DHW
                          <http://wb6dhw.com>

                          On 5/7/2013 11:58 AM, Graeme wrote:

                          Hi Kees,

                          Looks like you are already using "best practice" construction and yes the printed transmission lines are just fine. Only reason I avoided them is they are a bit tricky for home made pcb. Yes you get a stub on the unused filters so I make the through filter the last in the chain them 6 M, 10 M etc. to keep the stub short. On 160 M with the longest stub it shows a a little bit of capacitance on the input and output but does not affect the response. My relays were recovered from an old telephone exchange and probably are good contacts but really unknown although they work just fine.

                          I am using 1206 smt ceramic NPO capacitors rated at 200 Volts and have not had a failure even when testing into open circuit. The RD16HHF1's survived also. I am very impressed with them.

                          73 Graeme ZL2APV

                          --- In softrock40@yahoogroups.com, "Kees & Sandy" <windy10605@...> wrote:
                          >
                          > Very good info, Graeme. I ran board traces down the two sides of the board, one for input and one for output and made the traces 60-70mil wide. I've found that on a 62mil board, using 1oz copper, over a ground plane it comes pretty close to 50 ohms. It would still be nice to not have the unterminated stubs (to the unused filters), but I don't know what to do about that. The relays all have bifurcated, gold covered, contacts which are also paralleled for reliability (DPDT contacts). The other point is to only have one common connection between the antenna/earth ground and the logic ground for controlling the relays. Right now I'm using SMA connectors but may go back to BNC. Roger, on the bypass capacitor (and diode) across each coil. Always good to have bypass caps on DC circuits in an RF environment. The only "tricky" layout is that I'm putting a single RD16HHF1 right in the middle of the board as an Amp (couldn't see all that space going to waste). Have you ever used NPO, high voltage, ceramic smt capacitors in the filters ? Have you found any reason to use 5V relays vs 12V relays ? .....I got it in my mind that the lower 5V coils (less impedance) would be less suceptible to RF pickup. Probably a DNW (Does Nothing What-so-ever). 73 Kees K5BCQ
                          > ---------- Original Message ----------
                          > From: Graeme Jury <gvjury@...>
                          > To: softrock40@yahoogroups.com
                          > Subject: [softrock40] Re: Packaging external filters
                          > Date: Tue, 07 May 2013 15:07:23 +1200
                          >
                          >
                          > <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd">
                          > There is a couple of things which may be useful when designing switched
                          > banks of filters.
                          >
                          > 1. Use a transmission line down the relay chain. In my case I used
                          > RG-179 with the cover stripped back 5 mm or so and the braid tinned. Two
                          > knife cuts 2 mm apart were rolled around the centre of the tinned braid
                          > and then bent to snap at the knife cuts. The 2 mm piece of braid was
                          > pried away and the centre of the coax bared. This was done opposite each
                          > relay. The braids were soldered to the earth plane and the centre
                          > connected to the relay contact. It did not take long to prepare the coax
                          > and avoids the tricky and less reliable design of a printed transmission
                          > line.
                          >
                          > 2. Bypass the coils of each relay. I ran my filter bank through my VNA
                          > and could not understand why the roll off at the HF end of the filter
                          > went down to about 40 dB and then flattened. They were mesh Chebyshev
                          > filters and should have kept going down. In the end I found that a 0.1
                          > smt cap on the coil on the 12 Volt common side of the relays fixed the
                          > problem. As I added caps to more of the coils it got better and when I
                          > had done half of them it was improved to the noise floor of the VNA.
                          > Naturally for the negligible cost I did them all and also the switched
                          > grounded side of the coils too. It is an obscure fault and maybe if I
                          > had used RF relays it would not have happened but I used standard 2 pole
                          > changeover miniature ones which switch 30 watts from my RD16HHF1 linear
                          > on CW just fine.
                          >
                          > 73 Graeme ZL2APV
                          >

                          _

                           


                        • David Turnbull
                          ... I use OSH Park for the prototypes and Advanced Circuits for production. I saw Laen yesterday and showed him the Advanced Circuits boards that resulted from
                          Message 12 of 27 , May 7, 2013
                          • 0 Attachment
                            On Tue, May 7, 2013 at 7:38 PM, Kees & Sandy <windy10605@...> wrote:
                            Cool !  Is that what you use for your Peaberry boards ? what files are required ? gerber?

                            I use OSH Park for the prototypes and Advanced Circuits for production. I saw Laen yesterday and showed him the Advanced Circuits boards that resulted from his prototypes.

                            I send gerbers but he takes other formats too. It's all automated. Uploads are processed immediately so just try what you got and you'll know within minutes if it's good.

                            73 David AE9RB
                            http://AE9RB.com/
                          Your message has been successfully submitted and would be delivered to recipients shortly.