开云体育

I have been testing LPFs and found that I had filter characteristics that did not make sense. I had much more ripple in the filters then design and simulation showed. I added a 20 dB attenuator to the CH0 output and the problem disappeared. It seems that the drive capability maybe having problems driving the filter input capacitance?

Howard,
The problem could be match at the CH0 port. I checked my -H4 and -H and they both had 25dB or better return loss to 900MHz. Can you measure the Port 0 return loss on your NanoVNA? Note that the RL will probably be different with the power ON or OFF.
--John Gord

On 9/20/20 10:57 AM, Howard Fidel wrote:

I have been testing LPFs and found that I had filter characteristics that did not make sense. I had much more ripple in the filters then design and simulation showed. I added a 20 dB attenuator to the CH0 output and the problem disappeared. It seems that the drive capability maybe having problems driving the filter input capacitance?

what did the S11 look like without the pad?
What happens with a 10dB pad instead of the 20?

It was almost the same with the 10 dB as the 20 dB pads. SWR in the passband was 1.0 to 1.0 without a pad. With the pad is meaningless.

Using high-quality RF capacitors? Or common MLCCs? Voltage rating? I just ask because cheap capacitors can be *very* non-linear - like a effective capacitance ratio of 1:5 across the allowed voltage range. Great recipe for intermodulation, of course.

Anything claiming to be C0G or NP0 should be fairly good.

Caps with X5R and X7R dielectric can vary considerably with voltage as Clifford says,
especially if trying to stuff the maximum possible capacitance into the smallest possible volume.
It is assumed they will be used as bypass caps, not in resonant circuits.

I choose C0G caps over X7R caps, and X7R caps over X5R caps.
In addition to tolerating higher temps than X5R caps, my understanding is that
X7R are less likely to absorb moisture.

Though as a digital design guy, I never used many C0G caps.
Perhaps other folks here know their shortcomings better than I.

Jerry, KE7ER

NPO 1KV parts LPF for my 1 KW linear amp. Anyway you won't see the voltage problem with the VNA, the signal levels are too low.