balun designing steps

[aniakhan]

hi,
i need some help in balun designing.i have simualted a planar antenna on quartz substrate with 100 ohms lumped port now to connect it with 50 ohm feed line i will be in need of balun design,but i have no idea how to design a balun??plz anyone who have design it plz tell me what are designing considration??should i think about frequency??how to design it plz basic steps are requird if some one has any example plz do share with me????plz i need help

 

[Ow@i$]

I am attaching some docs that will surely help you out, do go through them thoroughly ask for any further query !!!

 

[WimRFP]

What is the center frequency, required bandwidth (given certain VSWR), available space, maximum attenuation, input power, phase delay, preferred technology, etc?

 

[aniakhan]

What is the center frequency, required bandwidth (given certain VSWR), available space, maximum attenuation, input power, phase delay, preferred technology, etc?

thanks for answering.
my center frequency is 300 GHz,band width is 6 i-e 100-600GHz. my antenna's total dimensions are lembda/2*lemba/2 and substrate is quartz of total length 2*2 mm.i have no idea when you asked phase delay what does it mean,balun is quite new for me,can you plz explain it to me???i just read about it in books but on practical grounds its new for me.i.plz guide me about it.thanks

 

[WimRFP]

This makes sense! When your unbalanced feed is a microstrip, without further info, I would recommend you a tapered microstrip balun. You may also include klopfenstein into the search. The "disadvantage" is that you need large space for it. The physical length is several electrical wavelengths as this type of balun uses a gradual taper to enable 6:1 bandwidth.

The phase behavior could be of importance for very wide band signals and/or beam forming.

 

[aniakhan]

hi
wimf,
can you plz help me to understand one thing in balun designing,i have found a paper where microstip line to double slot line transiotion is used as balun ,now i am only trying to simualte same paper structure in hfss,but my problem is that i dont understand impedance of both lines.if i am using slot line as balanced structure and for impednace matching then i should put its impedance value equal to mine antenna i-e 100 ohms or i should put same values of 50 ohms for both ports one for micrstrip line and other for slot???i am confused i think i should put 100 ohm port to slot line that will be connected to feed in fabrication and 50 ohm port to microstrip but i am not sure.plz help me in this

This makes sense! When your unbalnced feed is a microstrip, without further info, I would recommend you a tapered microstrip balun. You may also include klopfenstein into the search. The "disadvantage" is that you need large space for it. The physical length is several electrical wavelengths as this type of balun uses a gradual taper to enable 6:1 bandwidth.

The phase behavior could be of importance for very wide band signals and/or beam forming.

 

[WimRFP]

You are right, when your balanced antenna was designed for 100 Ohms, you also have to set the reference/port impedance for the balanced side of the balun to 100 Ohms. The unbalanced input side remains 50 Ohms.

Check hfss documentation for what is best for you: two unbalanced ports for the balanced side and using post processing to convert this to a balanced/differential port port, or a lumped/localized port scheme (that is balanced by itself). I can't guide you in that as I don't have hfss.

 

[aniakhan]

thanks wimf,
one more thing about electric length calculation,for microstrip calcultion i use Tx line cal but they also asked for electric length,is this mean phase delay in antenna and line???i saw some threads about it so if i have wavelength of 1 then my electrical length will be 360 degree??? am i rigth??thanks for reply it really helped me

You are right, when your balanced antenna was designed for 100 Ohms, you also have to set the reference/port impedance for the balanced side of the balun to 100 Ohms. The unbalanced input side remains 50 Ohms.

Check hfss documentation for what is best for you: two unbalanced ports for the balanced side and using post processing to convert this to a balanced/differential port port, or a lumped/localized port scheme (that is balanced by itself). I can't guide you in that as I don't have hfss.

 

[WimRFP]

When you have some dielectric between the conductors, the propagation speed is less then c0. So a certain length of transmission line gives more phase delay then based on c0 (velocity in vacuo). In formulas for calculating transmission line parameters they may call it velocity factor. Velocity factor = v = 1/(rt(epsr.eff) ). epsr.eff = effective relative epsilon. c = v*c0. where c = propagation velocity.

Electrical length = (physical length/v). For example if you need a quarter wave length at 300 MHz, in air that would be 0.25 m (90 degrees delay). However when c=0.666 (polyethylene coaxial cable), the physical length would be 0.66*0.25 m = 0.167 m.

So if you need 360 degr delay, you need 1 lambda of electrical length. The physical length (the actual length) will be less for microstrip or olther 2 conductor structures with dielectric in between.

 

[aniakhan]

hi,
wimrfp
i have one question about balun design.i have simulated microstrip to slot line transition as balun now i got these results ,i am not understanding either it is working right,what i believe about balun and reading papers is that after simulation my s parameter of microstrip line should be below -10 dB with real impedance of 50 ohm and similar for S21 of microcstip line but what about s para meter of slot line??that is if i have used port of 100 ohm for slot line then its s para meter should also be lower than -10 db for 100 ohm throughtout frequency band??i my case i am getting same as it should be for microstrip but for slot line real impedance is not constant and is increasing upto 200 ohm ,normally in papers of balun design they just talk about S11 of unbalanced ine but i am confused that i either should not bother about balanced line???can you plz help me to understand it??? i am attaching my results with this post.thanks for your guidance in this regard.

 

[WimRFP]

Regarding RL > 10 dB for the balun, this is not sufficient when your balanced load also have RL>10 dB. When you measure RL of both balun and antenna together, you may get very bad RL, so when you design them seperately, your antenna and balun should have RL > 16 dB (rough guess).

In a lossless transition (such as your balun) RL is same for both sides. your 200 Ohms from the 100 Ohm balanced side is not strange. 200 Ohm load from a 100 Ohms source impedance is VSWR=2 and that is RL = 9.54 dB.

You can be sure when you have good RL at the 50 Ohm port, RL at the balanced port is good also. So you only can consider the 50 Ohms side.

I can't read PPTX. Can you convert them to PNG image format?

 

[aniakhan]

thanks a lot for you reply i got your point,so if it means my RL on slot line should also be less than -10 db for good performance??and what about insertion loss its value should be in what range?? ,i am attaching fig of both RL

Regarding RL > 10 dB for the balun, this is not sufficient when your balanced load also have RL>10 dB. When you measure RL of both balun and antenna together, you may get very bad RL, so when you design them seperately, your antenna and balun should have RL > 16 dB (rough guess).

In a lossless transition (such as your balun) RL is same for both sides. your 200 Ohms from the 100 Ohm balanced side is not strange. 200 Ohm load from a 100 Ohms source impedance is VSWR=2 and that is RL = 9.54 dB.

You can be sure when you have good RL at the 50 Ohm port, RL at the balanced port is good also. So you only can consider the 50 Ohms side.

I can't read PPTX. Can you convert them to PNG image format?

 

[WimRFP]

I have no idea what you are actually simulating. Your microstrip looks like a narrow band network and your slotline seems to have a complete other then designed characteristic impedance. I am completely lost now.

Regarding insertion loss, it depends on your application. To be honest, I don't know what is "state of the art" at 600 GHz for your process.

 

[aniakhan]

thanks for your reply ,actually this simulation was a try to make it for w band slot antenna(which is narrow band),now i am going to make modification in slot width for working in 600 GHz , my confusion was i have first time design for balun and i have confusion for balanced feed side but by your advice you given in your last post i have got idea of analyzing balun for my own frequency range.i am making this antenna for terahertz camera application,thanks a lot for your reply it really helped me alot.i will modify my design i hope it will work ,thanks again

I have no idea what you are actually simulating. Your microstrip looks like a narrow band network and your slotline seems to have a complete other then designed characteristic impedance. I am completely lost now.

Regarding insertion loss, it depends on your application. To be honest, I don't know what is "state of the art" at 600 GHz for your process.

 

[WimRFP]

Do you really need 1:6 bandwidth ratio? If not, you can make a microstrip to slotline transition with less space requirements (such as used in vivaldi antenna designs with a circular aperture behind the feedpoint of the antenna). Such transition also provides your balun function.

If you have lots of time, you may add a second resonating structure to increase the bandwidth of the microstrip to slotline transition.

 

[aniakhan]

Hi,
Wimrfp,
thanks for your answers they have helped me in working with balun first time.now i have made a microstrip line to strip line transition with double slots also do it for w band as i have changed my frequency region for test set up.i have one question about results my return loss for microstrip line side is better with wave port of 50 ohm and on the slot side i attach 100 ohm port.now my results are in attach file now question is that
1. insertion loss curve that i normally saw in paper looks like a straight line curve rather than what i obtained. i am just taking it like S21 from s parameters of software am i right in taking that S21 as insrion loss??
2.secondly my S22 is not exactly below 10 db ans varies -8 -12 db at my frequencny region when i plot magnitude impedance v/s frequencny curve then it value was changing from 100-130 ohm mine antenna is of 100 ohm so when i go for fabrication will it make huge change in results or not????
3.also i am still nconfused about port impednace as according to my thinking i should attach differnet port impedance on different sides but normally papers donot have such port .am i right in doing this???
thanks for your answers.