formatting link
That would be a real bargain for $25.unidirectional attenuator
Jul 28, 2025
Last reply: 11 months ago
9 Replies
We optics folk use them all the time.
The circuit versions (mostly YIG-based) work down to nearly DC, like a gigahertz or something. ;)
Cheers
Phil Hobbs
Heh.
The telephone folk mastered unidirectional with the invention of the hybrid transformer. There was one in every telephone from the 1950s on.
.
formatting link
Joe
It's not really unidirectional. It attenuates -20dB in either direction. The high power end can handle 10 watts (40dBm) in one direction. The low power end maybe 20dBm (100 mW) in the opposite direction. This might be what you mean by "unidirectional".
"Understanding RF Attenuators"
formatting link
Oh, OK, cheap construction. I'll tell my guys about it. We'll be testing 700 volt pulses.
Ummm... maybe a back of the envelope calculation might be useful to check if the Amazon attenuator is suitable for your purposes.
The 20dB Amazon RF attenuator is designed to operate at 10 watts into
50 ohms. At 100% duty cycle, 700 V peak into 50 ohms is: P = E^2 / R = 700^2 / 50 = 9,800 watts If this is how you are using this RF attenuator at 100% duty cycle, you might consider having a fire extinguisher handy. Otherwise, I suggest that you limit your 700 V pulses to: 10 / 9,800 = 0.001 duty cycle. The manufacturers specs probably include a maximum duty cycle, but I couldn't find anything for the Amazon RF attenuator.Also, I don't know at what voltage the internal attenuator resistors will arc over. If you're operating at the rated 10 watts maximum into
50 ohms, the rated maximum voltage would be: E^2 = P * R E^2 = 10 watts * 50 ohms E = sqrt(500) = 22.4 V RMS If I assume the attenuator might need to handle a 2:1 mismatch, the impedance will change to 100 ohms: E^2 = 10 watts * 100 ohms E = sqrt(1000) = 31.6 V RMS = 44.6 V peak If that's the highest RMS voltage at which this attenuator was designed to handle, it might arc over if you apply 700 V peak. I suggest wearing welding goggles during testing. Fortunately, 44.6 V is not sufficient to strike an arc, which usually requires 50 to 100 V.Good luck.
We'll of course keep the pulse widths down, maybe 100 ns, an keep the duty cycle down too. We're evaluating fast GaN fets as pulse generators. The Bertan HV bench power supply is only good for 5 mA.
Fortunately the attenuators are cheap and we can afford to blow some up. They TDR pretty well, plenty good enough for characterizing our pulsers, GaN fets and txline transformers.
I've just got a couple of interns learning about this stuff. Interns are cheap and disposable too.
Most of the Amazon power attenuators are horrible, ring like a church bell.
This one is great.
formatting link
Clean step response with about a 32 ps rise time.
My local surplus house seems to have a lot of various types of broadband circulators in stock for cheap:
formatting link
of course the downside is it seems hard to find data on these products, so you'll have to see what kind of mystery meat they are exactly
This one is not:
formatting link
Join the Discussion
Have something to add? Share your thoughts — no account required.
Didn't find your answer?
Ask the community — no account required