I'm intermittently doing some final FAT (first-article tests) on my little toy pulse generator, but I'm frying 20 dB attenuators. This seems to help:
It doesn't help that the box's output SMA connector gets hot all by itself.
Probably some ham has worked out optimum stacking of attenuators.
I wonder how it's possible to make a multi-GHz, hundreds of watts attenuator. Big and fast don't usually go together.
torsdag den 27. august 2020 kl. 20.28.33 UTC+2 skrev John Larkin:
I just ordered a couple of SMA 10-watt 3 GHz attenuators from Amazon, about $20 each. It's amazing how much stuff Amazon has.
Engineering is now officially forbidden from going downstairs and pulling parts from stock; we're supposed to make an email request and wait to have them delivered. So we're getting a lot of resistor and cap kits from Amazon, and lots of hardware and connectors too.
For my microwave uses, it just takes a borosilicate measuring cup of water, to keep the cavity from completely overpowering the fluorescent-lamp-under-test.
Attenuator is easy; measurement is harder.
How many octave bandwidth do you need ?
I have used a 100 m reel of RG-58 coax as dummy load for frequencies above 1 GHz. It doesn't matter if the other end is open, shorted or terminated by 50 ohms :-).
If you are going to feed several watts, you might uncoil a few meters of the reel, so the first meters of the coax will cool and not melt due to the power losses.
If you need say 20 dB attenuation, use some coax and then use some low power BNC/TNC/SMA attenuators after that.
A load is much easier to make than a wideband attenuator. Skin-effect loss makes the bandwidth of the line go like 1/length**2.
I don't do high-power RF stuff, but for the few-watt range, I often use
1 dB, 1 dB, 2 dB, 3 dB, then whatever more I need. That's good for about four times the power rating of the individual pads. (I have a couple of Mini Circuits 6 GHz SMA attenuator sets that I use a fair amount.)Cheers
Phil Hobbs
I really need to see the waveforms accurately. My rise times are only about 1 ns, so I don't need a super wideband attenuator.
My 500 MHz Rigol scope only tolerates 5v inputs when set to 50 ohms, and my 20 GHz Tek only 2 volts. The little pulse generator can output
44v peak.So I really need an accurate 40 dB attenuator that can soak up a few watts. I can keep the duty cycle down some, but I do want to load the DUT.
use a 50R dummy load and probe made with a 2.5K series resistor and coax to a 50R terminated scope?
Am 27.08.20 um 23:56 schrieb Lasse Langwadt Christensen:
<I use the big black one when there comes more than I want.
In principle, I also have a really large KW size Bird dummy load with -40 dB output, but one should never lend that stuff away.
And there are directional couplers.
Gerhard
I ordered a couple of these, 20 and 40 dB. I'll test them with my 20 GHz TDR/TDT scope and see how good they are.
Bad idea. Use the same binned parts that paperwork and labor determine will go in the product, or be surprised later.
RL
There's not a lot of variation in 0805 1% resistors. I always have a DVM and an LC meter in reach of my bench too.
The parts in a bin downstairs can come from various manufacturers too.
I'm an engineer. I can deal with this sort of thing. The gain is low. Do the math. Change the resistor. Now the gain is right.
Sample kits are great to have around. Amazon has some amazing stuff.
I keep a local stock of common parts, including ICs, in my office in plastic bins, in coin envelopes. That saves time too.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required