Discussion in 'General Technical Questions and Answers' started by KK5JY, Oct 19, 2017.
The test room is shielded too.
I didn't say a variable was needed. Where did you get that?
Variable (or step) attenuator isn't needed, although handy to have.
The photo in post #38 shows why it's important for the attenuator to be well shielded, with shields between sections, to achieve any reasonable attenuation between input and output ports once you go beyond about 10 or 20 dB.
I don't think an "80 dB" resistive attenuator using three components can be built that will work at radio frequencies.
Fair enough. I think I have enough of the right kinds of resistors handy to build a multi-stage attenuator that will bring the power level down into the range of the SDR, but still be able to dissipate the power without overheating.
I was going to spend tomorrow doing the 160m contest from a state park, but since the weather is supposed to be very bad here (storms all day), I may spend tomorrow building an attenuator and retrying my measurement experiment.
By the way, thanks to everybody for the feedback. This is all very helpful.
Remember if you make an attenuator "directional" (with a specific input and a specific output, as opposed to bi-directional so components are all the same looking into either port), you only need "power" components at one end, the end connected to the transmitter. Once you knock that down 6-10 dB or so, the following components can be really small and probably all 1/4W or 1/8W resistors. Makes building easier and cheaper.
Just remember to label which end is which!
All very good advice. I'm not sure if I can make a directionally symmetric attenuator or not, from the parts I have. I can swipe the card at Mouser and get more resistors, but I was hoping to tinker on this over the weekend.
Thankfully, this is just a one-off. Once I can verify the proper spurious suppression, I won't need this circuit again.
For checking harmonics up to 900MHz I would probably choose a MiniCircuits TFM-2+ mixer. These give a very flat response to about 900MHz with a decent VSWR spec on the RF port. I used to use the TFM2-LH version a lot but that was over 25 years ago and this may have been a custom part with a higher LO drive level. But the (level 7) TFM-2 should be fine as long as you don't drive it too hard on the RF port. But the TFM-2+ is quite expensive at $20. Or you could use the classic SBL-1 as Karl-Arne suggested. This will be a lot cheaper although it will only be flat to about 500MHz. Still good enough I think...
Note that your AD9850 LO will have lots of spurs on it at -65dBc typical and maybe a few at -55dBc. So this will never be a clean LO. But still good enough to meet the requirements I guess.
I'm okay with -65dBc, or even -50dBc. That's still much better than the square wave that I have seen others use, and it's that much less stuff I have to filter out later. And -50dBc is FCC-compliant. It may not be great, but it's street-legal.
I'm sure the amplifier stages will put back some of the harmonic spurs just from nonlinearity, but I try to start as good as I can.
Thankfully, harmonics aren't called "spurs" much (although technically, they are) because they're really easy to identify and also really easy to filter using bandswitched low pass filters as found in all solid state rigs. With a good LPF designed to cut off just above the band you're using, even a terribly filthy transmitter can meet -45 dBc on harmonics.
There can still be spurs below that frequency, of course, but they won't be harmonics.
This is my hope. The LPFs were designed by others as part of a kit, so I like to think they knew what they were doing, and it will all just work out. The plots they posted on a real spectrum analyzer were promising.