Other Parts Discussed in Thread: LM5008
We are using a LM5008A circuit to step down 48V to 12V in our design. That 12V is then used by two linear regulators (UA78M05, UA78M33) to generate our 3.3V and 5V rails.
We are running into a couple issues with this circuit which has been implemented following the TI reference design guidelines.
First, the voltage regulation is not very good, with a very unstable PWM output, the frequency is way higher than anticipated, and the Turn On time is a lot shorter than what we'd expect.
With this, we expect ~385kHz switching frequency at constant load and Ton of ~750ns.
What we see is this, a sort of very short double pulse (way shorter than the expected minimum Ton), at a very high frequency.
This issue seems to be coming from the Linear regulator load. I've tried removing them and connecting a pure resistive load and it seems to be regulating fine with a stable PWM frequency (though for some reason, it won't go into continuous conduction mode before ~250mA load, but that is another story). I'm not sure why it's behaving this way with the linear regulators. I've bypassed the LM5008A and the 12V rail was using ~180mA, so a lot less than the rated current.
The other issue with this circuit is the switching noise. On every rising edge of the PWM pulse, we see a pretty big noise spike in the rest of the circuitry. I've tried adding RC snubbers on the inductor and on the diode, but it didn't help. This is messing with our current measurement circuit.
Thanks for your help.
