LTC4269-2 has an internal MOSFET which is used to pull down VNEG ("ground" for active operation) to VPORTN ("ground" of power supply). This MOSFET is kept off when the SHDN signal is asserted. Then VNEG is essentially pulled up to the positive rail VPORTP.
I am wondering if pulling "externally" in this state VNEG down to VPORTN could damage the chip. Would it make any difference whether it is pulled down strongly ("external" short to VPORTN) or through some resistance (even as high as 200kOhm)?
That's good. SHDN should be either below 450mV for normal PoE operation or above 3V for auxiliary operation. SHDN and /EN are both consider inputs, so tying them together only makes sense…
Can you explain what you're tying to do a bit more? This is during auxiliary operation? VPORTN, AUX-, and VNEG (Power Ground) will be at about the same potential when the auxiliary supply is connected and the PSE disconnects the port. In this condition, nothing happens if you connect VNEG to VPORTN. While the PSE and Auxiliary supply are both connected, VPORTN will equal to Auxiliary voltage minus the PoE voltage until the PSE disconnects the port.
in this case I'm trying to understand what could caused explosion of the chip. The relevant portion of the schematics looked like this:
In these conditions the chip was kept in shutdown, because the internal pullup in LT4321 raised voltage on SDHN above the shutdown threshold.
I was measuring the voltage on SHDN line and apparently I've made a short to VNEG from the node between R4 and R5. In the same moment the chip exploded, the crater was 1.5mm from VNEG leads, if that means anything.
When the chip is held in shutdown, VNEG is not pulled to VPORTN - the internal MOSFET is off. If a MOSFET is off you can basically apply to its drain any voltage below its maximal ratings, further, here the voltage was decreased, not increased, so that should be safe. From the node behind R4 there is still quite high resistance on any path to POENEG. Is there anything inside the chip that could get damaged so violently in such circumstances?
What was the voltage of the SHDN pin? Was the PD running with an auxiliary supply? Is this issue a similar to the other post you made today? It is difficult to see the whole picture with these snips of schematics.
It is difficult to see the whole picture with these snips of schematics. If you have more questions, could you share full schematics or send them to me directly? This will make it easier for me to help you.
this issue is not related to the other one.
The measured voltage on SHDN was near 650mV, it was due to internal pullup in LT4321. The datasheet says the logic low level threshold is below 450mV, so the chip could legitimately not start operation. I've recalculated and replaced the resistors and then it was working fine.
No auxiliary power supply was connected. Just POE.
The only short which I could have made was from between R4 and R5 to VNEG. I think no rapid discharge could have occured here, because we get at least near 100k ohms from VNEG to VPORTN anyway. That would have limited the current intensity to value near 50V/100kohm = 0.5mA.
What I cannot preclude is that there was a problem related to assembly, but I can't tell what could that be.
After I replaced the damaged chip the prototype is working.
I've only the snippets here not to obscure the problem with irrelevant (I think) details. The whole schematics is rather big anyway, too big for a post on the forum. How could I reach you to send it directly to you?
That's good. SHDN should be either below 450mV for normal PoE operation or above 3V for auxiliary operation. SHDN and /EN are both consider inputs, so tying them together only makes sense if they are both controlled by another signal. Usually /EN would be connected to /PWRGD, or EN to PWRGD.
I cannot see how a VNEG to R4/R5 short would cause damage. You can send the whole schematic or just the PoE pages to me, either by direct message or to my listed email.
where can I find the option to send a direct message, or your email?Please, help me with that.
I'm still figuring out EngineerZone, if you don't see an option then just send it to firstname.lastname@example.org