Shocking to Halo test story (1): Oscilloscope + standard probe to measure linear power ripple?

Engineers who want to see these stories don't take a shot at their heads and think they've done the same thing. I have been in the field of testing and surveying for more than 10 years and have experienced many interesting stories. Of course, the characters in these stories are my placement, if there is a similar, purely coincidental.

One man, at a cheap price, bought an adjustable linear power supply to power the part under test. The nominal noise value on this power supply specification is very low, only 2mV. And that's what he needs, really. and proudly told me that although the price of this power supply is much cheaper than Agilent's, the noise index is not bad. So I asked him, "Are you sure that this indicator is OK?" A PAT chest: No problem, I measured with an oscilloscope. I asked him again: How did you measure it. I was very confident to answer: I used a 500MHz import oscilloscope measured, almost no ripple. I asked him again, "Can you tell me what probe I am using?" I am very confident to answer: that is the oscilloscope standard 500HMz probe. My last question is: Is the load being added to the measurement? -Answer: No.

I started to get dizzy: first, the power supply noise measurement should be when the power supply in the maximum output state of the indicator, if there is no load, it will not know the actual state of the power supply after the load. More importantly, the Oscilloscope standard probe is 10:1, in other words, the noise signal comes in, has been attenuated to the original amplitude of 10%, even if the power supply ripple and noise indicator is 20mV, after the probe, the amplitude is attenuated to 2mV. The background noise of the general purpose oscilloscope is about 1-2MV, so that the ripple signal is almost completely submerged. So, if you want to verify the noise of your power supply, the first is to load the power supply, the output to the maximum, and then use the following method: Using 1:1 oscilloscope probe, bandwidth needs above 20MHz, such as Agilent's n2870a, 35MHz, 1:1 of the probe. This is suitable for low noise measurements above 5mVpp with a 30MHz wideband differential amplifier, which amplifies the noise 10 times times and then uses the oscilloscope to measure. This is more suitable for smaller amplitude ripple noise measurements. Regarding this measurement method, I have described in detail in the previous "Program control Power Technology and Application (5): ultra-low ripple measurement" blog.

In short, a good instrument is very important to measurement, but the probe is the door. If you neglect the janitor, you will get into trouble. In addition, the ripple indicator of the power supply is measured, and the load is required to measure the value in the maximum output state.