Hi, I’m using the QPL9547 on a 300 ohm balanced UHF antenna. The amp is connected to the antenna by 6" of 300 ohm twin-lead with a balun on the amp PCB. I noticed that if I touched specific points along the 6" twin-lead, oscillation would start. The spacing of these points was about 1/2λ of the oscillating frequency. (1.5Ghz and 1.8Ghz) I also noticed I could stop oscillation if I touched half way between the points that would usually start it. Touching the ground on the amp’s output coax connector or the coax itself near the amp PCB could invoke oscillation as well.
I realized that the 6" twin-lead must be picking up some of the output signal in the coax as a feedback path. I added a 1/2" copper tube around the 6" twin-lead and I haven’t been able to invoke oscillation since…as long as the tube is grounded at a specific location. When grounded directly to the ground layer on the PCB it is absolutely stable. If however I ground the 1/2" copper tube to the output coax connect the amp is 100% guaranteed to always oscillate. Both results are repeatable over multiple samples.
I don’t understand why the different tube grounding locations cause such a difference. It would be nice to use the coax ground connection as it is more easily manufactured for production.
Thanks for all the information. What is the frequency range you need the LNA to operate over? It may be possible to make some BOM or circuit adjustments to improve stability of the LNA. I guess the tube grounding is just causing the tube to radiate in different ways, you get a better grounding when it is direct to the PCB ground. It’s possible there is coupling between the inductor L1 (and/or L2) and the balun or antenna on the input. You could try changing the orientation of L1 and L2 and see if that helps.
Thanks. You could try adding some high pass filtering to reduce the gain above 600MHz. This may help a lot. Another thing that can help with stability would be to add a 200-500R resistor across L1. This also reduces the gain, in and out of band.
I tested the following with no copper shielding tube installed,
Using a 3 order Chebyshev 600Mhz LP filter on the output (between L2 and coax connector - pi 15nH series and 2x 5.6pF parallel) made it a lot harder to oscillate. The oscillation was lower in frequency and amplitude.
Foil over L1 and L2 area grounded to bottom PCB ground copper in multiple places - had no effect
Removing the 300 ohm twin lead did not stop oscillation. Oscillation did not stop until I removed the balun.
Using balun with a physically smaller core size (both K1 material) shifted up the oscillation frequency
I also noticed lengthening the shield tube’s connection to the PCB ground beyond 3/4" would make it prone to oscillation.
