Go to our main attenuator page.
Go to our page on variable attenuators.
New for June 2019: mechanically variable attenuators are often used in lab equipment, so that one unit can serve many purposes (instead of a drawer full of fixed attenuators). They are continuously adjustable (using a thumb wheel or a screw head for examples), and consume zero DC power. They are available in coax or waveguide. They are passive, reciprocal networks (either side can serve as input). As always there are several different means of accomplishing the same thing, for now we will focus on one particular device as we found a nice video about it!
If you're familiar with lumped element attenuators, you might think that inside a mechanical attenuator, three resistors are being adjusted at the same time while satisfying impedance match conditions. That's not how a mechanical attenuator works. Think of it as a transmission line where some lossy dielectric is added/subtracted to control the attenuation. When a design is done well, the "zero-loss" state will be less than 1 dB, attenuation will be predictable from a dial indicator and will be flat with frequency, and both ports will remain well-matched at all values of attenuation.
In the following video, Dr. Shahriah Shahramian from The Signal Path takes apart a Pasternak 12.4GHz-18GHz (X-band) continuously variable 20dB coaxial attenuator, to show us how it works. That's the easy part... putting it back together before the lab manager sees what you did is always the hard part!
He also goes over a couple of other interesting things:
11:02 - Ultra-high efficiency triple jacket glass distillation dewar.
14:59 - AWG PLC 8-way WDM splitter on 50GHz channels.
24:15 - PCB embedded D-Band waveguide with Vivaldi launch end-points.
28:45 - Hidden security features of the Canadian currency.
Thanks, Shahriar!
One thing to consider with mechanically adjustable attenuators is their power rating. The slab of lossy material will heat up, and can melt at high power. We've seen this happen, when a PhD engineer put a waveguide "vane" attenuator on the output of a 20W TWT and started calibration... no problem at zero dB, one dB, two dB... the power meter kept going down. Eventually at 10 dB or so, the power meter indicated that full power was back. Waht the heck? Hey, what's that smell? The vane had turned into a puddle of something that smelled like New Jersey!