Connector to microstrip discontinuity

[Mikko Kiviranta]

Dear Colleagues,

I wonder if any of you can point me to data
on the amount of reflection at a microstrip-to-SMA
or (even better) at a microstrip-to-MCX connection?
A few hours of googleing didn't help in finding
a good reference.

I know something about the SMA microstrip edge
launchers (e.g. from Hoffman's MMIC design handbook),
but I'd be more interested about the connector type
which is assembled in through holes and which
launches the coax cable perpendicular to the PCB
surface. The fact that such SMA and MCX connectors
are fabricated at all suggests actually that the
discontinuity shouldn't be that bad. One would
imagine that the VSWR specified to the particular
connector type would include the microstrip-to-connector
transition, but as this depends delicately on
the size and shape of the transition I'd expect those
to be specified along with the VSWR data.

As side information: I'd like to connect the MCX to a
0.5mm wide microstrip, which yields a 50-ohm line on a
multilayer PCB. The goal frequency is 2.45 GHz.

Thanks for any info / pointers.

Regards,
Mikko Kiviranta

 

[Don Pearce]

On Fri, 14 May 2004 16:40:19 +0300, Mikko Kiviranta
<Okkim.Atnarivik@iki.fi.invalid> wrote:

Dear Colleagues,

I wonder if any of you can point me to data
on the amount of reflection at a microstrip-to-SMA
or (even better) at a microstrip-to-MCX connection?
A few hours of googleing didn't help in finding
a good reference.

I know something about the SMA microstrip edge
launchers (e.g. from Hoffman's MMIC design handbook),
but I'd be more interested about the connector type
which is assembled in through holes and which
launches the coax cable perpendicular to the PCB
surface. The fact that such SMA and MCX connectors
are fabricated at all suggests actually that the
discontinuity shouldn't be that bad. One would
imagine that the VSWR specified to the particular
connector type would include the microstrip-to-connector
transition, but as this depends delicately on
the size and shape of the transition I'd expect those
to be specified along with the VSWR data.

As side information: I'd like to connect the MCX to a
0.5mm wide microstrip, which yields a 50-ohm line on a
multilayer PCB. The goal frequency is 2.45 GHz.

Thanks for any info / pointers.

Regards,
Mikko Kiviranta

At that sort of frequency, if you are prepared to spend a little
effort, you should be seeing a return loss better than 20dB. If you
have a TDR and a scalpel, perhaps even better still. Unfortunately
this parameter is not easily specified because of the huge number of
board thicknesses and Er's you can expect to find.

d

--
http://www.pearce.uk.com

 

[Tim Wescott]

Mikko Kiviranta wrote:

Dear Colleagues,

I wonder if any of you can point me to data
on the amount of reflection at a microstrip-to-SMA
or (even better) at a microstrip-to-MCX connection?
A few hours of googleing didn't help in finding
a good reference.

I know something about the SMA microstrip edge
launchers (e.g. from Hoffman's MMIC design handbook),
but I'd be more interested about the connector type
which is assembled in through holes and which
launches the coax cable perpendicular to the PCB
surface. The fact that such SMA and MCX connectors
are fabricated at all suggests actually that the
discontinuity shouldn't be that bad. One would
imagine that the VSWR specified to the particular
connector type would include the microstrip-to-connector
transition, but as this depends delicately on
the size and shape of the transition I'd expect those
to be specified along with the VSWR data.

As side information: I'd like to connect the MCX to a
0.5mm wide microstrip, which yields a 50-ohm line on a
multilayer PCB. The goal frequency is 2.45 GHz.

Thanks for any info / pointers.

Regards,
Mikko Kiviranta

You can't really specify it with just one VSWR. Since the
discontinutity has a 50-ohm line on either side the SWR will change with
frequency and transition techniques.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

 

[John Larkin]

On Fri, 14 May 2004 16:40:19 +0300, Mikko Kiviranta
<Okkim.Atnarivik@iki.fi.invalid> wrote:

Dear Colleagues,

I wonder if any of you can point me to data
on the amount of reflection at a microstrip-to-SMA
or (even better) at a microstrip-to-MCX connection?
A few hours of googleing didn't help in finding
a good reference.

I know something about the SMA microstrip edge
launchers (e.g. from Hoffman's MMIC design handbook),
but I'd be more interested about the connector type
which is assembled in through holes and which
launches the coax cable perpendicular to the PCB
surface. The fact that such SMA and MCX connectors
are fabricated at all suggests actually that the
discontinuity shouldn't be that bad. One would
imagine that the VSWR specified to the particular
connector type would include the microstrip-to-connector
transition, but as this depends delicately on
the size and shape of the transition I'd expect those
to be specified along with the VSWR data.

As side information: I'd like to connect the MCX to a
0.5mm wide microstrip, which yields a 50-ohm line on a
multilayer PCB. The goal frequency is 2.45 GHz.

Thanks for any info / pointers.

Regards,
Mikko Kiviranta

I'll post some pics to a.b.s.e.

John

 

[Mikko Kiviranta]

Don Pearce wrote:

Okkim.Atnarivik@iki.fi.invalid> wrote:
I wonder if any of you can point me to data
on the amount of reflection at a microstrip-to-SMA
or (even better) at a microstrip-to-MCX connection?

As side information: I'd like to connect the MCX to a
0.5mm wide microstrip, which yields a 50-ohm line on a
multilayer PCB. The goal frequency is 2.45 GHz.

At that sort of frequency, if you are prepared to spend a little
effort, you should be seeing a return loss better than 20dB. If you

Thanks, an educated order-of-magnitude guess is already useful.
I had estimated that the point where the microstrip dives
underneath the MCX flange introduces ~30 fF shunting capacitance
for a 0.5mm-wide ustrip, and the 1.4 x 1.4 mm pad means ~300 fF
more. This alone implies ~18dB RL. But the effect of the bend and
redistribution of the ground plane currents into the coax is
hard to estimate. I can't use the microstrip bend formulas
(from Waddell for instance) because the ground geometry is
different.

have a TDR and a scalpel, perhaps even better still. Unfortunately

Isn't trimming a 0.5mm strip a bit impractical? Besides, I'll
have a number of such junctions, and would prefer to compensate
for the discontinuity on the PCB layout already. Thanks for the
suggestion anyway.

this parameter is not easily specified because of the huge number of
board thicknesses and Er's you can expect to find.

Yes. I was wondering whether there is available theory similar to the
Majewski's IEEE Tran MTT vol 29 no 8 paper, which could be applied to a
number of different conditions.

Regards,
Mikko Kiviranta