Hi, Liu Yun,

How about the final results of your stripline coupler.

Currently, I simulated a stripline coupler at 64MHz with HFSS, and get about
20dB directivity. However, in our application, required directivity more than
30dB!
In ADS, I mean with circuit simulator, more than 40dB can be achieved, however,
with ADS momentum, it decreases to 26dB! Same sturcture is imported into HFSS,
only 20dB.
I have no any other compensation method used in these simulation, just stripline
coupler itself.

Looking forward to your feedback.

Thanks
LiuYu

MR RF Engineer
GE HEALTHCARE

T +86 10 580 68050
F +86 10 678 03267
M +86 13426330039
E YuLiu@...

2 YongChang North Road
Beijing, 100176 China

GE imagination at work

--- In mtt-8@yahoogroups.com, "amicloudly" <amicloudly@...> wrote:
>
> Dear Dr.Levy,
>   My recent simulation of this coupler have achieved directivity of
> 27dB. I adopted several tunable capacitors at the 90degree corner of
> the coupler line. Each capacitor is realized with a tuning screw
> above or under the stripline corner(To do this, I must drill a hole
> on the substrate). I found the discontinuity of the coupler line is
> essential to the directivity and the capacitors can compensate these
> effects.
>   Thanks!
>
> Liu Yun
> --- In mtt-8@yahoogroups.com, Ralph Levy <r.levy@j...> wrote:
> >
> > Dear Liu Yun,
> >
> > I think that your problems are due to the unequal lengths of the
> even and
> > odd mode circuits in your coupler. Yes, the dielectric is the same
> for
> > both modes, but the end effects are not, making the effective
> coupling
> > lengths different. There is some discussion on this in the book
> chapter
> > "Directional Couplers", Advances in Microwaves, vol. 1, Academic
> Press,
> > 1966.  See pp. 186-188. In order to investigate this it would be
> > interesting to analyze the individual even and odd modes on HFSS,
> > converting the results into equivalent circuits, each consisting
> mainly
> > of a transmission line of some length. I expect that the lengths
> will not
> > be identical. The book chapter referenced above indicates that the
> > isolated wave is proportional to the phase difference between the
> two
> > lengths, and the effect on directivity is much worse for loose
> than for
> > tight coupling values.
> >
> > Ralph Levy
> >
> >
> > On Fri, 16 Sep 2005 01:44:10 -0000 "amicloudly" <amicloudly@y...>
> > writes:
> > > Dear Experts,
> > >    sorry that I didnt conduct manual mesh operation in my past
> > > design.After the mesh configuration, I found the results change
> > > continuously and stably according to size chage.
> > >    But I cant find a better directivity than 20dB. I have changed
> > > the width of strip and the space between strips and the coupling
> > > length for many times.
> > >    the return loss on each port is about -30dB,but the
> directivity
> > > is only 19dB.
> > >    does the 1/4lamda  stripline coupler have the potential to
> acheive
> > > 25dB directivity? Any other mistakes in my optimization?Any
> other
> > > coupler form can I adopt in my design?
> > >
> > > Thanks and best regards,
> > > Liu Yun
> > >
> > > --- In mtt-8@yahoogroups.com, "amicloudly" <amicloudly@y...>
> wrote:
> > > > Coupler is always needed in systems.I have read through the
> > microwave
> > > > bible about coupler design.and have the successful experiences
> in
> > > > designing 2 section 800-2500MHz couplers. But these days I
> really get
> >
> > > > puzzled in my  work.
> > > > I think the narrow band coupler should be more simpleer,but
> > > recently I found it is not easy job.Perhaps I am weak in
> designing.
> > > > Recently I am designing a 40dB coupler of 1.8-2.2GHz. Firstly,
> I
> > > used  edge coupled stripline to realize.  I used the formulas,
> and get
> > > the  initial size.
> > > > Eps=2.7
> > > > Height=4mm
> > > > Width of strip=2.8mm
> > > > Space=4mm
> > > > Len=Effective Lamda/4=22.8mm
> > > >
> > > > The simulation in Serenade is good.(This is not a Fullwave
> simulation
> > > > ,but only use some formulas published in related papers,such
> as
> > > Cohn's papers) But the simulation in HFSS is very poor. The VSWR
> and
> > > > Directivity is very poor.So I cant believe in the formulas.
> > > >
> > > > Then I began my tour of HFSS optimization.
> > > > Because of loose coupling,the Zoe and Zoo is very close to Z0
> > (50Ohm).
> > > > So Firstly I optimize the Width to get stripline of 50Ohm.
> with  this
> > > > optimization,I found Width=2.1mm will yield good VSWR.
> > > Terminations use waveports with the Voltage from the strip to
> the uper
> > cover.
> > > > I then use Width=2.1mm and optimize the space to get the
> needed
> > > > coupling. the lenth is 22.8mm. I found that the coupling be
> 40dB,and
> > the
> > > > Space=4.8mm. but the directivity is very poor. I expanded the
> > simulation
> > > > freq range from 1GHz to 4GHz and found the performance is good
> at>
> > > > 1.4GHz. So I increase the length to 30mm. And found the
> performance
> > > > is good.
> > > >
> > > > But my design need the coaxial connectors vertical to the
> dielectric
> > > > plane.It mean there should be a transition from the stripline
> to the
> > > > vertical coaxial connector. So I then optimize a single
> stripline
> > > > connected by two vertical connectors to get good impedance
> matching.
> > > > And get the S11>24dB.
> > > >
> > > > After I adopt the connectors inside the coupler, the
> simulation  show
> > > > very bad performance.Directivity is 0.
> > > >
> > > > I really get very upset and tired.So I ask every experts:
> > > >
> > > > 1, Does formulas make sense?
> > > > 2, What should be a practical and useful procedure for coupler
> > design?
> > > > only for narrow band design.
> > > > 3, how does the length effect performance of the couper? As
> mentioned
> >
> > > > in Young's classic book, 1/4 Lamda is suitable.But practice
> doesnt
> > say
> > > > like that.
> > > > 4, How to realize high directivity? In theory, Stripline has
> equal
> > > > phase velocity of even mode and odd mode, but why the
> simulation is
> > so
> > > > poor?
> > > >
> > > > Thanks and best regards,
> > > >
> > > > Liu Yun
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > ------------------------ Yahoo! Groups Sponsor
> > > --------------------~-->
> > > Get fast access to your favorite Yahoo! Groups. Make Yahoo! your
> > > home page
> > > http://us.click.yahoo.com/dpRU5A/wUILAA/yQLSAA/0CfwlB/TM
> > > -----------------------------------------------------------------
> ---~->
> >
> > >
> > >
> > > To Post a message, send it to:   mtt-8@e...
> > > To Unsubscribe, send a blank message to:
> > > mtt-8-unsubscribe@e...
> > > Yahoo! Groups Links
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> >
>

Hi, Liu Yun,

How about the final results of your stripline coupler.

Currently, I simulated a stripline coupler at 64MHz with HFSS, and get about
20dB directivity. However, in our application, required directivity more than
30dB!
In ADS, I mean with circuit simulator, more than 40dB can be achieved, however,
with ADS momentum, it decreases to 26dB! Same sturcture is imported into HFSS,
only 20dB.
I have no any other compensation method used in these simulation, just stripline
coupler itself.

Looking forward to your feedback.

Thanks
LiuYu

MR RF Engineer
GE HEALTHCARE

T +86 10 580 68050
F +86 10 678 03267
M +86 13426330039
E YuLiu@...

2 YongChang North Road
Beijing, 100176 China

GE imagination at work

--- In mtt-8@yahoogroups.com, "amicloudly" <amicloudly@...> wrote:
>
> Dear Dr.Levy,
>   My recent simulation of this coupler have achieved directivity of
> 27dB. I adopted several tunable capacitors at the 90degree corner of
> the coupler line. Each capacitor is realized with a tuning screw
> above or under the stripline corner(To do this, I must drill a hole
> on the substrate). I found the discontinuity of the coupler line is
> essential to the directivity and the capacitors can compensate these
> effects.
>   Thanks!
>
> Liu Yun
> --- In mtt-8@yahoogroups.com, Ralph Levy <r.levy@j...> wrote:
> >
> > Dear Liu Yun,
> >
> > I think that your problems are due to the unequal lengths of the
> even and
> > odd mode circuits in your coupler. Yes, the dielectric is the same
> for
> > both modes, but the end effects are not, making the effective
> coupling
> > lengths different. There is some discussion on this in the book
> chapter
> > "Directional Couplers", Advances in Microwaves, vol. 1, Academic
> Press,
> > 1966.  See pp. 186-188. In order to investigate this it would be
> > interesting to analyze the individual even and odd modes on HFSS,
> > converting the results into equivalent circuits, each consisting
> mainly
> > of a transmission line of some length. I expect that the lengths
> will not
> > be identical. The book chapter referenced above indicates that the
> > isolated wave is proportional to the phase difference between the
> two
> > lengths, and the effect on directivity is much worse for loose
> than for
> > tight coupling values.
> >
> > Ralph Levy
> >
> >
> > On Fri, 16 Sep 2005 01:44:10 -0000 "amicloudly" <amicloudly@y...>
> > writes:
> > > Dear Experts,
> > >    sorry that I didnt conduct manual mesh operation in my past
> > > design.After the mesh configuration, I found the results change
> > > continuously and stably according to size chage.
> > >    But I cant find a better directivity than 20dB. I have changed
> > > the width of strip and the space between strips and the coupling
> > > length for many times.
> > >    the return loss on each port is about -30dB,but the
> directivity
> > > is only 19dB.
> > >    does the 1/4lamda  stripline coupler have the potential to
> acheive
> > > 25dB directivity? Any other mistakes in my optimization?Any
> other
> > > coupler form can I adopt in my design?
> > >
> > > Thanks and best regards,
> > > Liu Yun
> > >
> > > --- In mtt-8@yahoogroups.com, "amicloudly" <amicloudly@y...>
> wrote:
> > > > Coupler is always needed in systems.I have read through the
> > microwave
> > > > bible about coupler design.and have the successful experiences
> in
> > > > designing 2 section 800-2500MHz couplers. But these days I
> really get
> >
> > > > puzzled in my  work.
> > > > I think the narrow band coupler should be more simpleer,but
> > > recently I found it is not easy job.Perhaps I am weak in
> designing.
> > > > Recently I am designing a 40dB coupler of 1.8-2.2GHz. Firstly,
> I
> > > used  edge coupled stripline to realize.  I used the formulas,
> and get
> > > the  initial size.
> > > > Eps=2.7
> > > > Height=4mm
> > > > Width of strip=2.8mm
> > > > Space=4mm
> > > > Len=Effective Lamda/4=22.8mm
> > > >
> > > > The simulation in Serenade is good.(This is not a Fullwave
> simulation
> > > > ,but only use some formulas published in related papers,such
> as
> > > Cohn's papers) But the simulation in HFSS is very poor. The VSWR
> and
> > > > Directivity is very poor.So I cant believe in the formulas.
> > > >
> > > > Then I began my tour of HFSS optimization.
> > > > Because of loose coupling,the Zoe and Zoo is very close to Z0
> > (50Ohm).
> > > > So Firstly I optimize the Width to get stripline of 50Ohm.
> with  this
> > > > optimization,I found Width=2.1mm will yield good VSWR.
> > > Terminations use waveports with the Voltage from the strip to
> the uper
> > cover.
> > > > I then use Width=2.1mm and optimize the space to get the
> needed
> > > > coupling. the lenth is 22.8mm. I found that the coupling be
> 40dB,and
> > the
> > > > Space=4.8mm. but the directivity is very poor. I expanded the
> > simulation
> > > > freq range from 1GHz to 4GHz and found the performance is good
> at>
> > > > 1.4GHz. So I increase the length to 30mm. And found the
> performance
> > > > is good.
> > > >
> > > > But my design need the coaxial connectors vertical to the
> dielectric
> > > > plane.It mean there should be a transition from the stripline
> to the
> > > > vertical coaxial connector. So I then optimize a single
> stripline
> > > > connected by two vertical connectors to get good impedance
> matching.
> > > > And get the S11>24dB.
> > > >
> > > > After I adopt the connectors inside the coupler, the
> simulation  show
> > > > very bad performance.Directivity is 0.
> > > >
> > > > I really get very upset and tired.So I ask every experts:
> > > >
> > > > 1, Does formulas make sense?
> > > > 2, What should be a practical and useful procedure for coupler
> > design?
> > > > only for narrow band design.
> > > > 3, how does the length effect performance of the couper? As
> mentioned
> >
> > > > in Young's classic book, 1/4 Lamda is suitable.But practice
> doesnt
> > say
> > > > like that.
> > > > 4, How to realize high directivity? In theory, Stripline has
> equal
> > > > phase velocity of even mode and odd mode, but why the
> simulation is
> > so
> > > > poor?
> > > >
> > > > Thanks and best regards,
> > > >
> > > > Liu Yun
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > ------------------------ Yahoo! Groups Sponsor
> > > --------------------~-->
> > > Get fast access to your favorite Yahoo! Groups. Make Yahoo! your
> > > home page
> > > http://us.click.yahoo.com/dpRU5A/wUILAA/yQLSAA/0CfwlB/TM
> > > -----------------------------------------------------------------
> ---~->
> >
> > >
> > >
> > > To Post a message, send it to:   mtt-8@e...
> > > To Unsubscribe, send a blank message to:
> > > mtt-8-unsubscribe@e...
> > > Yahoo! Groups Links
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> >
>

Dear Colleagues,

We have an open Post-doc/Engineer position at our lab on the topics of
multiplexer synthesis. See following link for details:
http://www-sop.inria.fr/teams/apics/joomla/index.php?option=com_content&view=sec\
tion&layout=blog&id=5&Itemid=6

Please forward this message to people possibly interested,

Regards,

F.Seyfert

 

(The workshop registration is free)

 

Title: International Workshop on Microwave Filters

Start date: 16 November 2009

End date:  - 18 November 2009

Organizer: CNES and ESA

Location: CNES, Toulouse, France

Description: 

Satellite telecommunication systems continue to play a major role in broadcasting services and provide important point to point communication. The continued demand to improve the traffic flow capacity and efficiency of telecommunication satellite payloads is triggering many technological improvements. In the area of microwave filters, standard hardware such as input multiplexers, output multiplexers and planar filters are experiencing a growing maturity, whereas surface acoustic wave, film bulk acoustic and micromachined filters have just started to emerge for the space market. Satellite payload hardware needs to withstand severe mechanical and thermal conditions while maintaining all electrical specifications. This requires accurate CAD tools and methodologies for analysis, design and optimisation to reduce time to market and production cost.

 

-- Chris Radcliffe
Phase2 Microwave Ltd
Office +44 (0) 1438313502
Mobile +44 (0) 7711613926
www.phase2mw.co.uk
Unit 1A Boulton Rd
Pin Green Ind Estate
Stevenage SG1 4QX
England

No virus found in this outgoing message.
Checked by AVG - www.avg.com
Version: 9.0.704 / Virus Database: 270.14.60/2495 - Release Date: 11/10/09
19:56:00

Dear fellow coaxial friends

Anyone knows where I can find the information regarding the followings
equipment, reading the internet seems confusing at times, as all semms to be
rather similar.

Taps
Directional couplers
Combiner
Splitter
Separator

Really appreciate and a thank you to all who help.

Signing off as Patrick Lim of Singapore

Dear experts,

Now I am puzzled with a filter problem. I want to design a varactor tunable
bandpass filter which operates 100MHz-200MHz.The insert loss and fractional
bandwidth of the filter are 2dB and 3% respectively.

The two-order top-L coupled lumped bandpass filter model is selected. To get the
smallest Shape Factor (BW30dB/BW30dB), how can I select the appropriate
susceptance slope paremeter?  The unloader Q of the varactor is high. 
Considering the finite Q of lumped inductor, what's the smallest Shape Factor I
can reached actually? How can I deal with the  Shape Factor probelm?

I am looking forward for your response.

Thanks a lot!

----- Original Message -----

From: lycloud1978

To: mtt-8

Sent: Sunday, November 01, 2009 5:52 AM

Subject: [mtt-8] Narrow band Notch filter

 

Dear experts,

      Can anybody give some guidance to the design of narrow band notch filters(for example,

the notch band is 2000+/-0.5MHz? I guess

there need to realize weak couplings between mainline and resonators. Design should well

solve the tuning of both couplings and resonating frequencies, and high Q resonators should

be adopted to give lower loss.Thanks if any reply.

 

Best regards,

Yun Liu

The small differences that you are observing are probably due to rounding
errors.  When the TX zeros were optimized to give the required lobe levels, the
calculations were done using double precision arithmetic. The final TZ positions
quoted in Table 8.3 were rounded down to 4 decimal places.

When I use the 4-decimal place TZ positions, I get the same coefficients for the
P(s) polynomial as you have.

Don't worry about it.

Richard J Cameron



--- In mtt-8@yahoogroups.com, 정학 <agileahn@...> wrote:
>
> Thank you for your e-mail.
>
> Could you explain why there is small differece for P?
>
> At the page 312, at the table 8.3
>
> The value of P
>
> j1.0
> +2.2128
> +j26.5826
> +1.4870
> j65.6671
>
> but my result is that
>
> Ptr =
> > >
> > >         0 + 1.0000i
> > >    2.2127
> > >         0 +26.5831i
> > >    1.4865
> > >         0 +65.6698i
>
> BR
>
> Junghak Ahn
>
>
>
>
>
> --- In mtt-8@yahoogroups.com, "RJC_1947" <richard.cameron@> wrote:
> >
> > There was a small error in the formula for epsilon in equation (8.38)on page
304 of the book 'Microwave Filters ...' :-
> > F(s) should read F(s)/epsR
> > This equation is the s-plane equivalent of the w-plane equation (6.34) on
page 220. This error might account for the small differences in the values of
eps, epsR etc. that you are seeing.
> >
> > To obtain the values for eps and epsR, proceed as follows:
> > 1. Knowing the monic polynomial P(w) which contains the finite-position
transmission zeros, calculate the S11 numerator polynomial F(w) using the method
outlined in Section 6.3.
> > 2. Now knowing the return loss (RL), the ratio eps/epsR (eq.6.34) may be
calculated, and then the monic polynomial E(w) constructed using the alternating
pole method (see procedure after eq. 6.53 on p.227).
> > 3. Now eps may be evaluated at a convenient value of w, usually w=+/-1.0 for
a Chebyshev characteristic:
> > eps=P(w)/(E(w)*SQRT(1.0-10**(-RL/10))
> > 4. Finally epsR may be calculated from eq. 6.26
> >
> > Hope that this helps.
> > Richard J Cameron
> >
> >
> > --- In mtt-8@yahoogroups.com, &#65533;��&#65533;�� <agileahn@> wrote:
> > >
> > > Dear experts,
> > >
> > > I'm now studying for polynomial synthesis from the book
> > > Microwave filters for communication systems.
> > >
> > > At the page 312, in the table 8.3
> > > I got little different value for eps and polynomial for P(s)
> > > as below.
> > >
> > > eps =
> > >
> > >    33.1422
> > >
> > >
> > > epsr =
> > >
> > >     1.0005
> > >
> > >
> > > Etr =
> > >
> > >    1.0000
> > >    2.2463 - 0.0047i
> > >    3.6055 - 0.0032i
> > >    3.2888 - 0.0489i
> > >    1.9868 - 0.0025i
> > >
> > >
> > > Ptr =
> > >
> > >         0 + 1.0000i
> > >    2.2127
> > >         0 +26.5831i
> > >    1.4865
> > >         0 +65.6698i
> > >
> > >
> > > Ftr =
> > >
> > >    1.0000
> > >         0 - 0.0026i
> > >    1.0615
> > >         0 - 0.0009i
> > >    0.1580
> > >
> > >
> > > m1 =
> > >
> > >          0    0.3646   -0.6539    0.6680   -0.3438    0.0151
> > >     0.3646    1.3140         0         0         0    0.3639
> > >    -0.6539         0    0.7829         0         0    0.6536
> > >     0.6680         0         0   -0.8040         0    0.6677
> > >    -0.3438         0         0         0   -1.2966    0.3431
> > >     0.0151    0.3639    0.6536    0.6677    0.3431         0
> > >
> > > >>
> > >
> > > Please help me why there is little differece for eps and Ps
> > >
> > > BR
> > >
> > > Junghak Ahn
> > >
> >
>

Hi,

I am wondering if someboby of you might recommend me reference (paper or book)
about the fundamental definition of active scattering parameters for passive
devices.

Marco

Thank you for your e-mail.

Could you explain why there is small differece for P?

At the page 312, at the table 8.3

The value of P

j1.0
+2.2128
+j26.5826
+1.4870
j65.6671

but my result is that

Ptr =
> >
> >         0 + 1.0000i
> >    2.2127
> >         0 +26.5831i
> >    1.4865
> >         0 +65.6698i

BR

Junghak Ahn





--- In mtt-8@yahoogroups.com, "RJC_1947" <richard.cameron@...> wrote:
>
> There was a small error in the formula for epsilon in equation (8.38)on page
304 of the book 'Microwave Filters ...' :-
> F(s) should read F(s)/epsR
> This equation is the s-plane equivalent of the w-plane equation (6.34) on page
220. This error might account for the small differences in the values of eps,
epsR etc. that you are seeing.
>
> To obtain the values for eps and epsR, proceed as follows:
> 1. Knowing the monic polynomial P(w) which contains the finite-position
transmission zeros, calculate the S11 numerator polynomial F(w) using the method
outlined in Section 6.3.
> 2. Now knowing the return loss (RL), the ratio eps/epsR (eq.6.34) may be
calculated, and then the monic polynomial E(w) constructed using the alternating
pole method (see procedure after eq. 6.53 on p.227).
> 3. Now eps may be evaluated at a convenient value of w, usually w=+/-1.0 for a
Chebyshev characteristic:
> eps=P(w)/(E(w)*SQRT(1.0-10**(-RL/10))
> 4. Finally epsR may be calculated from eq. 6.26
>
> Hope that this helps.
> Richard J Cameron
>
>
> --- In mtt-8@yahoogroups.com, &#65533;��&#65533;�� <agileahn@> wrote:
> >
> > Dear experts,
> >
> > I'm now studying for polynomial synthesis from the book
> > Microwave filters for communication systems.
> >
> > At the page 312, in the table 8.3
> > I got little different value for eps and polynomial for P(s)
> > as below.
> >
> > eps =
> >
> >    33.1422
> >
> >
> > epsr =
> >
> >     1.0005
> >
> >
> > Etr =
> >
> >    1.0000
> >    2.2463 - 0.0047i
> >    3.6055 - 0.0032i
> >    3.2888 - 0.0489i
> >    1.9868 - 0.0025i
> >
> >
> > Ptr =
> >
> >         0 + 1.0000i
> >    2.2127
> >         0 +26.5831i
> >    1.4865
> >         0 +65.6698i
> >
> >
> > Ftr =
> >
> >    1.0000
> >         0 - 0.0026i
> >    1.0615
> >         0 - 0.0009i
> >    0.1580
> >
> >
> > m1 =
> >
> >          0    0.3646   -0.6539    0.6680   -0.3438    0.0151
> >     0.3646    1.3140         0         0         0    0.3639
> >    -0.6539         0    0.7829         0         0    0.6536
> >     0.6680         0         0   -0.8040         0    0.6677
> >    -0.3438         0         0         0   -1.2966    0.3431
> >     0.0151    0.3639    0.6536    0.6677    0.3431         0
> >
> > >>
> >
> > Please help me why there is little differece for eps and Ps
> >
> > BR
> >
> > Junghak Ahn
> >
>

There was a small error in the formula for epsilon in equation (8.38)on page 304
of the book 'Microwave Filters ...' :-
F(s) should read F(s)/epsR
This equation is the s-plane equivalent of the w-plane equation (6.34) on page
220. This error might account for the small differences in the values of eps,
epsR etc. that you are seeing.

To obtain the values for eps and epsR, proceed as follows:
1. Knowing the monic polynomial P(w) which contains the finite-position
transmission zeros, calculate the S11 numerator polynomial F(w) using the method
outlined in Section 6.3.
2. Now knowing the return loss (RL), the ratio eps/epsR (eq.6.34) may be
calculated, and then the monic polynomial E(w) constructed using the alternating
pole method (see procedure after eq. 6.53 on p.227).
3. Now eps may be evaluated at a convenient value of w, usually w=+/-1.0 for a
Chebyshev characteristic:
eps=P(w)/(E(w)*SQRT(1.0-10**(-RL/10))
4. Finally epsR may be calculated from eq. 6.26

Hope that this helps.
Richard J Cameron


--- In mtt-8@yahoogroups.com, 정학 <agileahn@...> wrote:
>
> Dear experts,
>
> I'm now studying for polynomial synthesis from the book
> Microwave filters for communication systems.
>
> At the page 312, in the table 8.3
> I got little different value for eps and polynomial for P(s)
> as below.
>
> eps =
>
>    33.1422
>
>
> epsr =
>
>     1.0005
>
>
> Etr =
>
>    1.0000
>    2.2463 - 0.0047i
>    3.6055 - 0.0032i
>    3.2888 - 0.0489i
>    1.9868 - 0.0025i
>
>
> Ptr =
>
>         0 + 1.0000i
>    2.2127
>         0 +26.5831i
>    1.4865
>         0 +65.6698i
>
>
> Ftr =
>
>    1.0000
>         0 - 0.0026i
>    1.0615
>         0 - 0.0009i
>    0.1580
>
>
> m1 =
>
>          0    0.3646   -0.6539    0.6680   -0.3438    0.0151
>     0.3646    1.3140         0         0         0    0.3639
>    -0.6539         0    0.7829         0         0    0.6536
>     0.6680         0         0   -0.8040         0    0.6677
>    -0.3438         0         0         0   -1.2966    0.3431
>     0.0151    0.3639    0.6536    0.6677    0.3431         0
>
> >>
>
> Please help me why there is little differece for eps and Ps
>
> BR
>
> Junghak Ahn
>

Dear experts,

I'm now studying for polynomial synthesis from the book
Microwave filters for communication systems.

At the page 312, in the table 8.3
I got little different value for eps and polynomial for P(s)
as below.

eps =

    33.1422


epsr =

     1.0005


Etr =

    1.0000
    2.2463 - 0.0047i
    3.6055 - 0.0032i
    3.2888 - 0.0489i
    1.9868 - 0.0025i


Ptr =

         0 + 1.0000i
    2.2127
         0 +26.5831i
    1.4865
         0 +65.6698i


Ftr =

    1.0000
         0 - 0.0026i
    1.0615
         0 - 0.0009i
    0.1580


m1 =

          0    0.3646   -0.6539    0.6680   -0.3438    0.0151
     0.3646    1.3140         0         0         0    0.3639
    -0.6539         0    0.7829         0         0    0.6536
     0.6680         0         0   -0.8040         0    0.6677
    -0.3438         0         0         0   -1.2966    0.3431
     0.0151    0.3639    0.6536    0.6677    0.3431         0

>>

Please help me why there is little differece for eps and Ps

BR

Junghak Ahn

Dear sir,

I tried to follow the paper, Novel Approch to the synthesis of microwave
diplexers.

I'v got same result for the Type 2 example case. But for the First Tx
couping : k0,1tx and k0,1 rx,

I got different values.

f0=1917.4MHz, B=146.5MHz so Bn=B/f0=0.0764, c0=0.3983.

Mtx(0,1)=0.005, Mrx(0,1)=0.0284

k0,1tx=Bn*Mtx(0,1)/sqrt(c0)=0.00061,

k0,1rx=Bn*Mrx(0,1)/sqrt(c0)=0.0034

I don't know why ? Pls help me.



BR

Junghak Ahn

Is there commercial software available that can extract the coupling matrix from
the s-parameter file derived from the HFSS filter model with lumped ports on
each resonator?  My company actually has software that does this but it is
proprietary.

Recently I was asked to comment on a narrow bandpass filter for high power at
low VHF frequencies. A brief analysis revealed that the insertion loss would be
very low for the given unloaded Q of the proposed resonators. However, a filter
with the same resonator Q and identical absolute bandwidth at UHF would have a
much higher insertion loss

This got me thinking again about insertion loss vs fc.
One might simply argue that the relative bandwidth at low VHF would be much more
than at UHF and therefore the insertion loss is much lower at VHF. However, I
tend to think: 'Yes - but why ?' Mathematically it is all very clear. At VHF,
the transfer function poles are shifted to the left in the complex plane by a
smaller amount and hence the lower insertion loss ( epsilon = 2 x PI x fc / (BW
x Qu) ). But filter maths does not seem to tell us the full story.

Then I looked at the group delay at fc which is of course the same whether we
are at UHF or VHF for a constant absolute bandwidth. This - I think - gave me
the answer.

Any comment on this subject would be most welcome.


Dieter Pelz

order 7 ,25dB 3 zeros 1.05 1.1504 1.2
[r22,L22,k22]=residue1(m1,n1);
[r21,L21,k21]=residue1(i*Ps1,e*n1);

r21=real(r21);
r22=real(r22);

L21=imag(L21);
L22=imag(L22);
T1k=-0.4113 0.5745  -0.2530 0.2450 -0.1448 0.3772 -0.5277


Tnk=0.4113 0.5745 0.3677 0.3642 0.1448 0.3772 0.5277
L21=1.3939 0.9055  -1.15042  -1.15042  -0.993  -0.7749   -0.079

for second zero:1.1504 closed abs(L21(3)) and abs(L21(4)),the response is error
,can you help me

--- In mtt-8@yahoogroups.com, "fabseyfert" <fabseyfert@...> wrote:
>
> Dear mtt-8 members,
> In the hope to clarify things about the delicate step of the synthesis
> of a coupling matrix when starting from a scattering matrix I have
> written a short tutorial. It is meant as a mathematical resume of
> Caemron's 1999 paper and lists in particular all properties verified
> by the residues, loads etc...together with a justification why it is
> so. I hope this will help people to develop and debug their own
> software. The tutorial is accessible under the Dedale-HF homepage at:
> http://www-sop.inria.fr/apics/Dedale
> under the section "Related Tutorials".
>
> F.Seyfert
>

hi F.Seyfert,can you tell me the extracted methods of your coupling matrix ,the 
methods of Caemron's 1999 or not.

I read about this paper "Novel Approach to the Synthesis
of Microwave Diplexers" many times.I still  have a question unsolved.
In this paper,there are  5 stepes to evaluate the diplexer polynomials. In the
first step, have to get F0tx F0rx E0tx E0x P0tx P0rx . But as far as I could, I
only get the F E P polynomial in the normalized dormain [-1 1]. how to get F0tx 
E0tx  P0tx in the [omega_tx 1] and  F0rx  E0rx  P0rx  in the [-1 omega_rx] .
Please help. thansk in advance.

Dear Yun,

 

I missed the 6 dB coupling requirement in my initial response. Actually there is no problem in realizing waveguide couplers for this case. There are such couplers described in my reference.

 

Ralph Levy 

 

----- Original Message -----

From: lycloud1978

To: mtt-8

Sent: Friday, May 01, 2009 12:10 PM

Subject: Re: Re: [mtt-8] circular waveguide coupler design.

 

Dear Dr.Snyder and Dr.Levy,

     Thanks for your suggestions.

     In fact I designed a 6dB branch line coupler in  microstrip form, and  used

microstrip-circular waveguide transitions at each ports, but I have not got good enough

VSWR and isolation though I have been conducting  parametric optimization with HFSS for

many days. Trasitions add to the complexity and overall dimension.

    I think perhaps a  coupler directly in waveguide form is much simpler in structure 

with fewer parameters to be tuned. And methods, such as Mode matching method

can be used for simulation with much higher efficiency. So I hope to try.

 

Best regards,

Yun Liu 

---

�����ˣ� Dr. Richard V. Snyder

����ʱ�䣺 2009-05-02  01:32:43

�ռ��ˣ� mtt-8

���ͣ�

���⣺ Re: [mtt-8] circular waveguide coupler design.

I have not had to build a waveguide coupler with coupling as strong as -6 dB, but if I did, perhaps I would build a stripline coupler at -6 dB and transition to the waveguide, at input and output.
Dick Snyder


Ralph Levy wrote:

Dear Yun,

 

One method to realize, say, a 10 or 20 dB directional coupler is to couple a rectangular waveguide running in parallel with the circular waveguide and using the modified Bethe small-aperture coupling method to calculate the dimensions of the coupling apertures. The coupling will take place when the circular waveguide, presumably operating in the TE11 mode, has the correct polarization, and it will then not work for other polarization angles. You may also wish to consider using a similar circular waveguide for the secondary arm since then the propagation constants of the two waveguides will be the same and the theory becomes simpler. For further information see my paper  "Improved Single and Multiaperture Waveguide Coupling Theory...." in MTT-28, pp. 331-338, April 1980, along with the references listed therein.

 

Best Regards,

 

Ralph Levy

 

 

----- Original Message -----

From: lycloud1978

To: mtt-8

Sent: Friday, May 01, 2009 8:24 AM

Subject: [mtt-8] circular waveguide coupler design.

Dear experts,

      I hope to have some guidance for circular waveguide coupler design.

     The copling is 6dB, and frequency range is 7-8.5GHz.

      Which structure is more suitable for this case?

 

Thanks if any response on this question.

 

 

Best regards,

Yun Liu

-- Dr. Richard V. Snyder
President
RS Microwave

Hello, all

In making practical coaxial filter or duplexer for BTS, what is the recommended
multiplying factor for skin depth of plating on the aluminum
housing?

Does anyone who knows the relation between pim and plating performance
systematically?

BR

Junghak Ahn

----- Original Message -----

From: lycloud1978

To: mtt-8

Sent: Friday, May 01, 2009 12:10 PM

Subject: Re: Re: [mtt-8] circular waveguide coupler design.

 

Dear Dr.Snyder and Dr.Levy,

     Thanks for your suggestions.

     In fact I designed a 6dB branch line coupler in  microstrip form, and  used

microstrip-circular waveguide transitions at each ports, but I have not got good enough

VSWR and isolation though I have been conducting  parametric optimization with HFSS for

many days. Trasitions add to the complexity and overall dimension.

    I think perhaps a  coupler directly in waveguide form is much simpler in structure 

with fewer parameters to be tuned. And methods, such as Mode matching method

can be used for simulation with much higher efficiency. So I hope to try.

 

Best regards,

Yun Liu 

---

�����ˣ� Dr. Richard V. Snyder

����ʱ�䣺 2009-05-02  01:32:43

�ռ��ˣ� mtt-8

���ͣ�

���⣺ Re: [mtt-8] circular waveguide coupler design.

I have not had to build a waveguide coupler with coupling as strong as -6 dB, but if I did, perhaps I would build a stripline coupler at -6 dB and transition to the waveguide, at input and output.
Dick Snyder


Ralph Levy wrote:

Dear Yun,

 

One method to realize, say, a 10 or 20 dB directional coupler is to couple a rectangular waveguide running in parallel with the circular waveguide and using the modified Bethe small-aperture coupling method to calculate the dimensions of the coupling apertures. The coupling will take place when the circular waveguide, presumably operating in the TE11 mode, has the correct polarization, and it will then not work for other polarization angles. You may also wish to consider using a similar circular waveguide for the secondary arm since then the propagation constants of the two waveguides will be the same and the theory becomes simpler. For further information see my paper  "Improved Single and Multiaperture Waveguide Coupling Theory...." in MTT-28, pp. 331-338, April 1980, along with the references listed therein.

 

Best Regards,

 

Ralph Levy

 

 

----- Original Message -----

From: lycloud1978

To: mtt-8

Sent: Friday, May 01, 2009 8:24 AM

Subject: [mtt-8] circular waveguide coupler design.

Dear experts,

      I hope to have some guidance for circular waveguide coupler design.

     The copling is 6dB, and frequency range is 7-8.5GHz.

      Which structure is more suitable for this case?

 

Thanks if any response on this question.

 

 

Best regards,

Yun Liu

-- Dr. Richard V. Snyder
President
RS Microwave

Dear Yun,

 

One method to realize, say, a 10 or 20 dB directional coupler is to couple a rectangular waveguide running in parallel with the circular waveguide and using the modified Bethe small-aperture coupling method to calculate the dimensions of the coupling apertures. The coupling will take place when the circular waveguide, presumably operating in the TE11 mode, has the correct polarization, and it will then not work for other polarization angles. You may also wish to consider using a similar circular waveguide for the secondary arm since then the propagation constants of the two waveguides will be the same and the theory becomes simpler. For further information see my paper  "Improved Single and Multiaperture Waveguide Coupling Theory...." in MTT-28, pp. 331-338, April 1980, along with the references listed therein.

 

Best Regards,

 

Ralph Levy

 

 

----- Original Message -----

From: lycloud1978

To: mtt-8

Sent: Friday, May 01, 2009 8:24 AM

Subject: [mtt-8] circular waveguide coupler design.

Dear experts,

      I hope to have some guidance for circular waveguide coupler design.

     The copling is 6dB, and frequency range is 7-8.5GHz.

      Which structure is more suitable for this case?

 

Thanks if any response on this question.

 

 

Best regards,

Yun Liu

-- Dr. Richard V. Snyder
President
RS Microwave

----- Original Message -----

From: lycloud1978

To: mtt-8

Sent: Friday, May 01, 2009 8:24 AM

Subject: [mtt-8] circular waveguide coupler design.

Dear experts,

      I hope to have some guidance for circular waveguide coupler design.

     The copling is 6dB, and frequency range is 7-8.5GHz.

      Which structure is more suitable for this case?

 

Thanks if any response on this question.

 

 

Best regards,

Yun Liu