Hi Alexey,
Those are very good questions and I might do a follow up video to clarify those questions.
A year or so ago I approached Bart in implementing an FRA feature into the Cleverscope. He had already been able to do frequency response bode plots on passive components. I asked him to try it on a power supply and I advised him to use an injection transformer to isolate the function generator from the power supply voltage. So he used a Jenson transformer. We got good results at high frequency about 100Hz but the lower frequency around 1 to 10 Hz were not too accurate. Some colleague of mine advice Bart that if he could design an isolated function generator, then we could do away with the injection transformer and greatly improve the low frequency response of the Cleverscope. About 3 months later, Bart came out with the isolated function generator. I bought the first isolated function generator. So this was a great improvement that was implemented.
The second big improvement was adding the ability to reduce the amplitude of the function generator as the frequency is swept to higher frequencies. The problem that arises is that at low frequency the power supplies generally have a lot of gain. And since you are injecting a signal in the feedback path by means of a 10 ohm resistor, the generator is connected to across the 10 ohm resistor which means that you need an isolation transformer or an isolated function generator. The probes of the Cleverscope are then connected across the 10 ohms resistor i.e. Channel 1 to the high side and channel 2 to the low side of the resistor. At low frequency around 1 to 10Hz, because of the high gain in the power supply, the injected signal gets greatly attenuated. So you must inject as much signal as possible without saturating the power supplies error amp. Now as the frequency is swept to higher frequencies, the gain of the power supply will start to drop and you do not need as much injection signal from the function generator. In fact, if you do not lower the signal, it will start to saturate the power supplies' error amp. So the need to be able to reduce the injected signal is important.
Bart implemented this by making a table where you can fill the frequency and the amount of amplitude signal for the function generator to put out. This second improvement really helps in getting a good loop response.
I might do a detailed follow up video and explain that the Cleverscope is the only FRA in the market that I know of that you do not need an injection transformer.
Questions and comments can be posted here or sent to rbola35618@aol.com
Robert