In this next part I'm going to look at the output on a static load.


Here's my output load:





For 10A at 1.2V I need a load of 120 mOhms (1.2V/10A) and it has to be able to dissipate 12W (1.2V x 10A).

I made this up from cheap 1 ohm 0.25W resistors. Three resistors in series is 3 ohms; twenty-five of those in

parallel is then 120 mOhms. Each resistor only has to dissipate 160mW, so they're reasonably comfortable. It

measured 123 mOhms, so that will be an output current of something like 9.7A [it will move a bit when it heats up].


Here's the output voltage running the load. [Being a bit contrary, I'm running against the excellence-in-

photography tide.] That's fine - it's just 0.3% high.





Next let's look at the output voltage ripple on load. The following two traces are of exactly the same signal - the first is using the probe's flying-lead ground and the second is using the spring clip ground. Both traces are ac coupled.








Next is the ripple with the bandwidth limited to 20MHz and accumulated traces. This nicely shows the underlying form of

the ripple and we can immediately see that there are four phases which presumably correspond to the phases that the

converter goes through, the spikes coming at the transitions which we would kind of expect.




Finally here's one of the spikes. Unfortunately I'm right up against the limits of my scope here (200MHz,

1Gsps and passive probes). The rise time of the scope is going to be 1.7nS, or so, and that's what I'm seeing

here, so the spike itself is much faster and far taller. The peak on the trace is where the scope is scrabbling to catch up with the spike

when it's on its way down again. Essentially, I'm testing the response of the scope with an impulse waveform.