I use two ferrites on each probe, the type that you normally use for power cables, one near the BNC, and the other near the probe. The probe cable is looped once around each of these ferrites for double the inductance
TomC, No I used two ferrites without any looped to cable. I'll do a test like this and reply about this.
Attention the following information is from Adapter Ver. 3.0I upload some measurements captures from the AOZ circuits as I modificated this at the last.
Unfortunately the
initial referenced measurements had limited BW (20MHz) and I can't to compare with these that is full BW.
But, there is one measurement with a full BW that with D127 180°turn modification (this method decrease almost 40% the noise). This measurement has taken at the output of 5.5V after the C45 capacitor. Upper is waveform from 5.5V output, bottom is gnd-noise.
The
5.52V Output.png is the measurement at the same position as before but now is at the last decoupling capacitor after the second 5R6 inductor. There is another timing than before, 200ns vs 100us but the level of noise is significantly lower than previous measurement (28mVp-p with 10X attenuation = 2.8mVp-p with 1X). The previous measurement was 74mVp-p (10X).
I am very sure that if I capture with the 100us the noise will be bigger but sure very lower than 74mVp-p.
The second yellow waveform is the gnd-noise capture.
The
AOZ1094.png is the capture with the probe tip over the chip. As you see there are some signs that related to some spikes at the yellow waveform (gnd-noise). Sure there is need to investigate that but I don't know if that is normal or my AOZ is defected from many soldering-desoldering that I make to all this area of circuits.
The
D127.png and
L3.png is the captures with the probe over of them. They have identical pattern. At this point, I had some measurements with decoupling capacitor or no at the cathode of diode. The decoupling capacitor is minimize at little the ringing of pulse and for that reason I applied it.
At the end there is no any measurement on the last 5R6 Coiltronics inductor. It is shielded and you can't to record any waveform over this.
I think that the modification of this area has reached its limits given the existing circuit topology. There is one option, yet. The replacement of second inductor by ferrite with a very low DC resistance.
Now I'll focus to the crappy LED Circuit Backlight.