Making a quater wave connection and 2 GHz source

[reasonableman]

Apologies for what may be a basic question, I am unfamiliar with RF systems but much more familiar with optics.

I need to connect some standard coax to some microcoax; they do not have exactly matching impedances so there are reflections. It seems to me I should be able to make an 'anti-reflection coating' out of some coax with a different impedance. How do I go about doing this, in particular what components are available and where can I buy the components? Obviously the frequency is important which brings me to my next question.

How difficult is it to obtain a 2 GHz source? I only need +/- a few volts, is it a 'voltage controlled oscillator' that I'm looking for?

 

[the_emi_guy]

Can you be more specific about your "standard coax" and "microcoax". There is a big difference between trying to match 75 ohm RG-59 to 50 ohm RG-402, vs. trying to get 40dB return loss across a pair of 50 ohm connectors at 2GHz.

For a single frequency source you may be better off buying a component like the Crystek CRFS75-2000 vs. purchasing an instrument.

 

[reasonableman]

I have nominally 50 ohm coax connecting to nominally 50 ohm micro-coax. I have a data sheet for the micro-coax (http://www.micro-coax.com/pages/pro...oaxial/Semi-RigidDetails.asp?ID=UT-034-TP-M17), I don't know the details of the 'standard coax' off the top of my head. Probably bought from farnell or RS...

 

[davenn]

LOL there's no such thing as standard coax
there's just so many variations of makes and models
Is it not written along the length ?
what diameter is it 1/2 inch, eg RG213, RG-8U, LMR400, etc etc

1/4 inch ? RG58, RG59 etc

the_emi_guy's Crystek CRFS75-2000 is an excellent choice, failing that you have companies like Minicircuits or Zcomm that produce a wide range of fixed oscillators and VCO's ( just in case you want to be able to do a freq sweep

Dave

 

[reasonableman]

Well, RF people might not think there is no such thing as 'standard coax' but I think most physics labs just have some 50 ohm coax, with female BNC connectors that they use most of the time. Anyway, moving on...I checked and the 'standard' coax is RG58.

 

[reasonableman]

Apologies for double posting but I'm confused. I have an email saying the_emi_guy has responded but it's not visible on the thread.

Anyway in answer to your questions I need to clarify some things. I am trying to replicate some results from McCall et al 1985 Microwave interferometer... In this paper it gives equations for the signal one would expect to see from a microcoax based interferometer with a mis-match. Based on comparison of my results with this model it appears the join (which is a SMA-BNC coupler) has a reflection of ~10%. This seems quite high to me as both components are nominally 50 ohm, but the results seem to indicate that. If this is correct reducing the reflection to 1% or less would be great.

 

[the_emi_guy]

The normal way to couple your semi-rigid to a generic BNC RG-58 would be to mount an SMA connector onto the semi-rigid, then use SMA to BNC adapter.

You are worried about reflection from this interface? Do you have a particular requirement for max reflection (VSWR or return loss)? Can you use RG-58 with SMA connectors?

The normal way to obtain lower reflections, if you need it, is to use better quality, precision connectors and keep them clean. There are metrology grade connectors that provide very good performance but are quite delicate and easily damaged if mated with a low quality connector.

I have never heard of anyone compensating for reflections from ordinary connectors using narrowband impedance matching techniques.

 

[the_emi_guy]

1% is a pretty tall order. You will not be able to use BNC, or even SMA. Use good quality 3.5mm connectors on both your semi-rigid and your RG-58. You are looking for VSWR of 1.01.

 

[f95toli]

Also, you simply shouldn't use RG58 at 2 GHz, I doubt its impedance will be 50 ohms at 2 GHz, which explains why you get reflections. It will also be quite lossy.

That said, 10% isn't really that much for a microwave system.

 

[reasonableman]

All of a sudden I feel spoilt by fibre optics! -60 dB reflection at connectors is an everyday occurence.

I was thinking I must have made a mistake to be getting 10%.

With regard to RG58, most of the system is made out of mini circuits 18 GHz cable with N connectors (http://217.34.103.131/pdfs/CBL-0.5M-NMNM+.pdf). I'll look at replacing that section.

Thanks to everyone for their help so far.

 

[Baluncore]

A quarter wave matching transformer would ideally have an impedance that is the geometric mean of the two transmission lines, but Sqrt(50*50) = 50, which suggests it is not that simple. So you could use a quarter wavelength of 50 ohm coax with two adjustable tuning stubs, one at each end. Two GHz is a low enough frequency that you should be able to do it with standard 50 ohm connectors, maybe not BNC but SNC, Ntype or SMA should be no problem.

 

[Mordred]

if you working with a specific frequency, you might want to calculate the 1/4 wavelength then multiply that by the velocity factor of the cable your using. This avoids relective wave issues. In RF line length is an important aspect as per frequency used. Wrong lengths can increase reflective power and reduce the forward power.

this link has a handy calculator and the formula on how to calculate

http://www.arcticpeak.com/antennapa...htm#Calculation_of_Quarter-wave_stub_lenght:_