Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations waross on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Making your own 20 db attenuator

Status
Not open for further replies.

1capybara

Industrial
Aug 3, 2008
40
GB
I want to make my own attenuator - would this work or what should I change?

/-------100 ohm resistor-------ground
/
input----/
\
\
\-------1 ohm resistor--------output

i will make each lead length = wavelength/2

is it do-able like this?
:)


 
Replies continue below

Recommended for you

That's not going to work. Assuming that you're using 50-ohm circuits (?), then the 1-ohm is almost not there at all (almost no effect), and the 100-ohm effectively-parallel load just creates a slight mismatch.

You should look up Pi and T attenuator circuits. There are plenty of on-line calculators. You'll need to specify your input and output Z.


 
By the way, if this is for 2-4 GHz (clue extracted from your other post), then you'd better just find a commercial product.

Or perhaps use a length of lossy cable (if your application can accept the delay).

 
Two antennas back to back can make a variable attenuator.

Strip two cable ends having the center conductors poke out of the cables, and place cables near each other. If you make them about 1/4 wavelength center conductors you may get good VSWR on your attenuator.

Put them in a box to stop leakage. you can have from about 6 dB up to 40 dB attenuator ranges. You could probably put them in a metal box and make some slide mechanism or tuning screw arrangement to move them apart and set your attenuation level if you're mechanically clever.

One long cable with loss is another way to make an attenuator. It has very good VSWR too.


kch

PS: Attenuators with resitors always require 3 of them to make a X dB attenuator and maintain good VSWR. Two in series with one to ground (T type), or one in series and two to ground (Pi type). Search Resistive T or Resistive Pi attenuator and you'll find good details. I don't suggest trying this though with normal attenuators, usually companies make a tiny package that you insert into a microstrip package.
 
PI attenuators as TDEW has shown in his link are used frequently in RF circuits up to about 500 MHz if made from SMT resistors. You can do it to over 1GHz if you have a network analyzer to check and tweek your results. Of course, this all depends upon how accurate and how flat (in frequency response) you want your attenuator. Layout and skin effect are issues - standard thick-film resistors are no longer accurate due to skin effect.

You can buy thin-film PI resistor attenuators on a chip for reasonable cost. Check out the Panasonic EXB24A series and the Susumu PAT series at Digikey's web site.
 
I assumed he was trying to operate below 100 MHz because above that even using SMT can be hectic, but buying an attenuator is really the best solution especially for wide-band applications.

5-1000MHz antenuators can be had for as little as 3 bucks (Note these are ultra cheap and I would not use them for accuracy) for coax applications with 250 buying a set of good quality attenuators.
 
ive gotta give Higgler the gold and bronze medals here for his two suggestions, using a pair of antennas as a variable attenuator and using a lossy coax.
let see....58' of rg-58 loses 20 db!
thanks Higgler, IOU one!
 
Me (first): "...use a length of lossy cable..."

Sniff.


[!!!! ;-) !!!!]

 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Top