UCC28780: ucc28780- exposed pad contact in prototype

robin gangopadhya18

Part Number: UCC28780

Ulrich

Is there any problem if the exposed pad meant for thermal reasons is not soldered to the pcb pad in this prototype phase?

Our last attempt(Friday) to get NTC pin to be as in the Eval board had the controller fail quickly.

Our sense is that it was due to excessive heat during attempts to solder the exposed pad. We cannot reflow it with all the parts in place. So we used hot air - thermocouple in vicinity showed controlled hot air temp to be 230 deg C. But it lasted several minutes & still did not flow the solder. Heating it further did it.

Then of course we had to hand solder the pads also at high temp.

When power up to 20 V or so...it quickly failed.

We are going to reinstall now. With conductive epoxy dot at the center just to have easy alignment of the chip to pads. & Will solder at leat temp recommended for good solder flow.

- Also: seems NTC pin may not be used for EN/disbale by an Opto . I am suspecting the plastic pkg if transparent & the phototransistor is conducting ...( I have experimental verification of this in a previous life ..& vendor had to change the material!)....of course I cannot be sure of this: why NTC pin will tend to be low...just before the controller failed Friday.

-so removing the Opto from NTC pin from now on.

robin

over 4 years ago

0 Ulrich Goerke over 4 years ago

TI__Mastermind 31130 points

Hello Robin,

That is a good question, and I posed it to the IC designer, but have not heard back yet. I'll let you know when I do.

I think it is better to have the pad grounded, but I don't know what could happen if it is floating. The IC substrate is internally bonded to the pad, so if it floats, the substrate floats and maybe certain P-N isolation wells may or may not be fully back-biased? I'm just guessing here.

I think it is a good idea to use the conductive epoxy as substitute for soldering.

The NTC pin has a pull-up current source for NTC biasing purposes. However, to minimize stand-by power much of the IC functions are shut down during the wait-state between bursts. This includes the NTC current source and the circuit that monitors the pin voltage. During the wait-state, these are shut off and the voltage falls to zero, then turned on again only briefly during each burst interval. In general, 0V on NTC is not a harbinger of imminent failure.
However, external shutdown on NTC may not work during light load because the shutdown requires 3 consecutive switching pulses with NTC < 1V, but the wait-state between 2-pulse bursts (in SBP, LPM, and lowest ABM) also shuts off and clears the counter that counts 3 pulses.

I think the opto-coupled on/off function can be restored for all loads if a small capacitor is used to couple the collector voltage to the base of a PNP, whose emitter goes to REF and collector goes to GND. When the opto pulls down on NTC, it also briefly drives the PNP to overload REF which will reset the IC. Then the low voltage on NTC should prevent it from restarting until the opto is turned off. The cap on the PNP base keeps the REF overload brief, not constant.

Anyway, refining opto on/off control should be deferred until the controller runs satisfactorily.

Regards,
Ulrich

0 robin gangopadhya18 over 4 years ago in reply to Ulrich Goerke

Mastermind 6055 points

Ulrich:

I suspect the pad needs grounding. We will see what the IC designers say for sure...In real product, it will be grounded for sure.

As to ON/OFF, yes, we disconnected all that. NTC now is connected to 30.1k( some value we had) alone.

A new proto pcb is being assembled this morning.

Hope it kicks off nicely.

As always, thnx a lot for insights into the design issues.

-robin

Power management

Power management forum

UCC28780: ucc28780- exposed pad contact in prototype