I have designed a 5V@4A boost-buck voltage regulator using the LTC3785 - I found the Excel Spreadsheet calculator tool and Applications Information provided in the datasheet to be most helpful in this process, thank you.
However, I am seeing some unexpected behaviour when operating in Burst Mode, as load current becomes significant, when supplying a nominal input voltage of 6.4V.
At light loads, switching Burst Mode on and off changes the quiescent current draw, as expected.
If Burst Mode is left on and I increase the load current above around 0.5A, it appears that the current limiting action takes effect as the efficiency and output voltage falls and the C&D FETs (Buck) heat up rapidly.
If I disable Burst Mode, the regulator supplies current all the way up to the 4A limit before feedback into the error amp starts to reduce the output voltage.
Is this behaviour expected? I had the impression that I could simply leave Burst Mode on permanently and that it's function would be disabled by the action of the (separate) Undervoltage Protection.
Please note that I have used two separate voltage divider networks (albeit with same resistor values) for the FB and VSENSE control pins.
I could use our host controller to monitor load current and engage Burst Mode as required, but this will complicate the software and I'm not sure how quickly it would need to respond to avoid damaging the controller or FETs in the event of a large and sudden increase in load current.
I'd appreciate any advice you may have in relation to this.
This behavior is not expected. Be sure the input de-coupling caps are close to the input and Vcc de-coupling is close. If you are still having issues, I suggest getting our demo board to see how it should work.
The input caps are close to the switches and the Vcc cap is close to the controller. If their placement was problematic, wouldn't they also affect the regulator in continuous mode?
The demo board uses a common FB & VSENSE connection and so wouldn't represent the circuit I am trying to debug.
Can you think of any other reasons why the C&D FETs (Buck) might heat up?
Please check that you follow the recommended PCB layout. Many unexpected issues show up with bad layout.
I'm still having trouble with this. I took another look at the demo board schematic and separate voltage dividers are indeed used for the FB and VSENSE pins - my mistake and apologies for this.
I followed the layout advice carefully and as I mentioned above, I'd expect that if this was the cause, I would also have trouble when Burst Mode is disabled.
I'd be interested to hear from anyone else that may have the demo board to see if the output voltage falls and the buck switch heats up as they increase the load gradually with Burst Mode enabled. The 'knee' appears to be around half the current limit in my case.
Can you get an ADI demo board through appropriate channels? If not, can you tell me what is Estimate Annual Usage of your product so I can check this for you?