I reach out to you guys in the hope to see something that I can’t see.
I have had a board manufactured with the bellow power supply. It came with the mail a few weeks ago and I finally got around to test it the other night. Big disappointment. Seems like the step-down circuitry is shoring out some way.
The symptoms are:
When applying 14V to 12V_SW and GND the board consumes a few mA and the 5V and 12V LEDs lights up nicely. After about 2 seconds the current limitation on my power supply kicks in and the boards is consuming about 4A at 3V. 5V and 12V LEDs dim down. The ACT4060 regulator gets very hot and eventually lets out smoke.
I have tried two boards; both do the same thing. Once they are bad, they don’t recover.
On a burnt board there is very little resistance between 5V and GND. About 8 ohm. On am unused board there is about 10K ohm between 5V and GND.
Applying 5V to the 5V rail on a fresh unsused board. There was a short high pitch tone for a second, then my bench supply strated to current limit at about 300mA. And the ACT heated up.
Why L1 (there). There is now no buffer cap on the input of the chip.
And that buffer cap must be of high quality (< 0.05 ohm ESR).
Layout also matters when switching at 0.5Mhz.
Leo..
According to the data sheet, the maximum voltage of EN is 6v.
You should, in this case, leave it disconnected instead of holding it at the input voltage which you have said could be 14v.
Also more explicitly from the data sheet:
Note that EN is a low voltage input with a maximum voltage of 6V, it should never be directly connected to IN.
Your input protection circuit also appears to be very elaborate. I’m sure you could rationalize and otherwise improve it’s handling of over voltage.
How 'good' is C4. 470uF sounds like a normal electrolytic cap, and that's not good enough.
The datasheet mentions 22uF ceramic or tantalum, and it needs to be very close to the chip's input AND ground(plane).
If you have a cap there. then I suppose L1 can stay.
Leo..
What is surprising is the datasheet implies that the ACT4060 is quite well protected.
I guess you've checked the obvious things like the resistance between FB to ground and FB to +5V matches the schematic and D3 is not reversed.
One suggestion is to take the latest dead board (where EN has not been subjected to overvoltage) and isolate the preceding protection circuit and apply voltage direct to IN (pin 2). That is just incase the unexpected power drain is due to say a mis-specified (or wrongly supplied) component earlier in the circuit.
What is the application that you require such a protection circuit ? Is it automotive electronics ? I guess you've ruled out the risk of lightning because there is no spark gap device in there.
Problem solved! It was indeed the L1 component that caused the trouble, as Wawa pointed out. As soon as I removed it and bridged the gap all was working very nicely. Big thank you for that.
However I'm still curious about the improvements that can be done on the protection of the input. You mention spark gap device.... ?
Tobbera:
Thank you both for your quick replies. Seem to be absolutely correct about EN-pin. I have removed it on a fresh board, but the same result I'm afraid.
Should I try to short out L1?
You have any suggestions on how to rationalize and improve the input protection?
Use exactly the circuit suggestion from the datasheet, with the correct components, in the correct layout - anything else is not guaranteed to work. In particular the ESR of each capacitor matters, the inductor choice matters, the diode choice matters, the PCB layout is critical.
L1 is completely utterly wrong, it has no place at all there, it will destroy the chip most likely. Why did you add it?
