Electronics > Projects, Designs, and Technical Stuff

DC-10MHz isolation design

(1/6) > >>

Has anyone designed an isolation circuit for measuring switching power supplies using an oscilloscope? The frequency range is DC-10MHz or so and the isolation needs to be at least 400-600V CAT II.

I have thought about amplitude modulating a 100MHz or so (or optical) carrier, sending it through a transformer, and then synchronous rectifying that to get back the original signal. But that would require careful RF design to work properly.

Another idea is to use a transformer to couple the high frequencies and a some other isolation method to couple the DC and low frequencies, but would that require careful adjustments to ensure a flat frequency response? (Curves are for models, not linear isolation circuits!)

The other big problem I see is measuring such a high voltage signal (= using high impedance) up to that frequency, since, for example, at 10MHz cutting frequency, 100kohm input resistance means 0.16pF in parallel.

However, maybe I'm saying a stupid thing, but... what about acquiring data on the measure side and transferring it digitally to the other?

Switching power supplies are not going to be affected by a few pF of loading.

Digitizing the data would basically mean building an oscilloscope from scratch, not exactly practical. That would indeed make the isolation part easy.

How about modifying a 2.4 GHz video link to accept your desired signal voltage range? Well, maybe not up to 10 MHz, but might go near.


I'm pretty sure video is AC coupled with automatic zeroing so that wouldn't work.


[0] Message Index

[#] Next page

There was an error while thanking
Go to full version