F28388D gets hot and short circuited between VDDIO pin and VSS (GND) pin

Longpt,

Were you observing ESD precautions while assembling this board? Glancing at your circuit I don't see anything that would clearly destroy the F28388D device. If you made your schematic symbol I would recommend checking that the pin definitions are correct. Otherwise I suspect that the device was damaged from ESD.

1. Please use the same power source for VDDIO and VDDFL. VDD_MCU_3V3 is recommended.
2. VREFHI pins must be driven by a clean reference, simply supplying regulated 3.3V will not provide good analog performance. Recommend using internal reference if you are not going to use a buffered precision reference as shown of TMDSCNCD28388D.

Regards,
Cody 

Dear Cody,

Thanks for reply,
I have read the power sequencing in the datasheet of F28388D, do you think with the value of filter capacitors, ferrite beads for 3.3V , 1.2V  in my schematic may make the voltage is not good enough for power sequencing and this may be the reason which make the IC died?

Another that I forgot, when soldered the board and tested with above described sequence , I have not yet soldered the R52 - 56 ohm (between VDD_MCU_1.2V and GND) , Was that affect to IC?

Brs,

Longpt

Dear Cody, 
I update  more information about the result I tested the voltage signal on the board.  I measured the  time response of VDD_3.3V , VDD_1.2V ,VDD_MCU_3.3V, and VDD_MCU_1.2V signal (ramp up).
With the VDD_3.3V, and VDD_1.2V picture below: the blue line -VDD_3.3V and yellow - VDD_1.2V

I see that the time response of VDD_3.3V is slower about 14ms than VDD_1.2V. 
I also measured the VDD_MCU_3.3V and VDD__MCU_1.2V so result is quite similar to VDD_3.3V and VDD_1.2V.  I tried to change the value of some filter capacitor of VDD_3.3V and VDD_MCU_3.3V but not come up.

1, Whether "VDD_3.3V is slower than VDD_1.2V. "  this make not true with power sequence? Because I have read  power sequencing in the datasheet of F28388D and also in datasheet of F28379D (quite similar to F28388D), in datasheet of F28379D mentions (not in F28388D ):

" VDDOSC and VDD must be powered on and off at the same time" .

2, Some colleagues notice me about the  peak solder temperature also may damage the chip, can you give me more information about this? 

It is very important that the supplies ramp together, I don't know if this could cause your issue so quickly, typically it takes some sustained time in these invalid conditions to cause a complete failure.

As for the reflow profile please follow this standard. https://www.jedec.org/document_search?search_api_views_fulltext=JSTD020

Regards,
Cody 

 Cody,

I again tried to solder with another board.

At this time, firstly I supply 3.3V and 1.2V from reference supply outside, It  seemed like ok, both led on 3.3V and led on ERRORSTS pin of MCU was lighted out. I used XDS 100V2 programed for the chip and that was ok, MCU was programed and running.

Then , I tried supply 5V for TPS62420DRCR on the board. On the board in the filter for 3.3V I have just soldered cap C9, not soldered cap C10 for the first time ( you can see C9 C10 on schemactic above). And  It also seemed like ok, both led light out, and can program for chip by XDS 100V2. I see the consumption current for 5V  is about 0.6A.

And I think everything may be ok, so I next soldered cap C10 into the board, and re - supply 5V.  But the phenomena was similar to previous board, 2 led was off, and there also was short circuited between 3.3V and GND, I measured the resistance between 3.3V Pin and GND Pin , the value was about 28 ohm. I measured output voltage of TPS62420DRCR, the output 1.2V is still ok, the output 3.3V is only 0.6V. 

Finally, I re - supply 3.3V and 1.2V from outside. It was again ok, the chip was still programmed and running, but there was still short - circuited between 3.3V and GND (28 ohm ).

I am really confused, and don't know why? Can you give me some comment?

longpt

Longpt,

Can you confirm that it is the F28388D on the PCB that is shorted between 3.3V and GND?

If you can supply 3.3V and 1.2V externally and the C2000 device will connect then I would expect that the F28388D device is OK. Are you  again seeing a large current on your 3.3V supply?

 I know the first time the F28388D was hot, in your second experiment are any components hot? 

Please note that soldering in C9 or C10 should not cause any issues assuming that only the correct low voltage DC is applied to the device(3.3V and 1.2V).

Finally, please try to adjust your circuit such that your supplies ramp as closely together as possible. If not this will cause premature circuit ageing.

Regards,
Cody 

Cody , 

1, "Can you confirm that it is the F28388D on the PCB that is shorted between 3.3V and GND?"

 I have not taken out the MCU  in the second board, I measured resistance on the board between 2 pins  of decoupling capacitor for 3.3V ( VDDIO)  - GND and 1.2V (VDD) -GND, the result was about 28 ohm between 2 pins  of decoupling capacitor 3.3V  and GND,  and about 55 ohm between 2 pins  of decoupling capacitor 3.3V  and GND.

the phenomena was similar to the first board which I took the MCU out of the board, for more information: I measured each pin on the chip (taken out of the first board), in reality there is not short circuited in all pin VDDIO and VSS, just only pin VDDIO: 3,11, 15, 20, 26, 138, 147, 152, 159, 168 get short with VSS, other VDDIO pins are not (and also VDD and VDDA, VDDOSC are not either).

==> with all information above, I and some colleagues have believed that with the second board is similar to first board, the chip also has short in just only some VDDIO Pins (and these short pins may or not the same with short pin in the first board ), It make the resistance between 3.3V and GND become very low - just 28 ohm , and make the current become high and the TPS 62420 does not have enough power to supply, and output 3.3V has decreased. 

2, "If you can supply 3.3V and 1.2V externally and the C2000 device will connect then I would expect that the F28388D device is OK. Are you  again seeing a large current on your 3.3V supply?"

Yes, for the first time when supply 3.3V and 1.2V externally I did not notice, but you ask me that make me remember , yes a large current on 3.3V supply for the first time about 0.6A . I think at that time some VDDIO pins were got short, but some were not so the C2000 still could connect,

3. " I know the first time the F28388D was hot, in your second experiment are any components hot? "

In the second board, the F28388D was hot too, but not hot as the same as the first one (I think because It has less short VDDIO pin than the fisrt one).

So Now I am still investigating the reason that make the chip died in some pins VDDIO, for issue that I am concerning:

 I soldered the board by my hand, with the chip when  soldered it into the board, the pins in the 4 corner of the chip were ok, get through to the pad in the board, but with the VSS pin - 177 in the bottom of the chip , It was not got through - connect to pad GND in the board. whether this could cause my problem?

Cody,

I confirm your question 2, "If you can supply 3.3V and 1.2V externally and the C2000 device will connect then I would expect that the F28388D device is OK. Are you  again seeing a large current on your 3.3V supply?"

Here the picture when I supply 3.3V and 1.2V externally, consumption current of 3.3V up to 0.7A, the C2000 device still can connect and running.

The resistance between 3.3v and GND on the board is about 19 ohm (my previous reply is 28 ohm because I take the TPS6420 out of the board, when I put it back to the board, It is only 19Ohm). The resistance is small, it make the current big,  If I supply 5V, so output 3.3v of TPS6420 does not have enough power to supply, so the voltage has decrease down to 0.6V.

Here are some pic of my board , the MCU in the top and the TPS in the bottom

If possible, can you review my project altium of this board.

Longpt,

Long Pham Tan said:

I soldered the board by my hand, with the chip when  soldered it into the board, the pins in the 4 corner of the chip were ok, get through to the pad in the board, but with the VSS pin - 177 in the bottom of the chip , It was not got through - connect to pad GND in the board. whether this could cause my problem?

Pin177 is the only VSS pin, I would not expect the device to work correctly without this soldered. It is possible that you were able to connect to the device due to some conductivity between VDDIO and some of the other VSSxx pins, but this conduction path is not indented and will certainly damage the device. I cannot say if it would fail exactly as you are seeing, but it is possible. Pin177 "VSS" MUST be connected.

Lets start by connecting VSS. As an additional check, if you made your schematic symbol please check the the pin assignments against the device datasheet.

For hand assembling boards like this I would highly recommend a solder stencil and a small reflow oven. The time savings during assembly of prototypes will be worth more than the cost of the stencil and reflow oven.

Regards,
Cody 

Cody, 

Can you review my altium project of this board (include schematic and layout)?

With all information that we have discussed, I am still really confused and don't know why and what makes the chip died. I also checked pin assignments of chip in PCB It seems there is no problem.

I am not professional in hardware designing, so when designed this board, I leverage schematic mostly  from reference board like launch pad kits and control cards.

I have make 2 chips died without knowing the exact reason, F28388D has some features which are really suitable for my project, but it is quite new. I and my colleagues don't have too much experience with it. If we still don't find the reason exactly, we may change to another kind of chips which we are more familiar. But F28388D is really  strong and has amazing features , so we really don't want to change it. 

So please, help me check the altium project of this board, maybe the problem comes from layout? 

The link download my file here: drive.google.com/.../view

Longpt,

I don't have any conclusive proof this is your only issue, but you failed to connect the primary device ground pin( Vss 177). You had connected specialized ground pins (Vssa, VSSOSC), but these pins are not designed to handle the current from VDD or VDDIO and therefore the internal circuits were likely damaged by this unintended current flow resulting in some pins being shorted to ground.

As for a schematic or layout review I have looked at the images provided above and have provided my high-level comments. If you want an in-depth schematic review please work with your local FAE. If you do not have an FAE a request can be submitted on the TI Support site.

Regards,
Cody 

Cody,

Thanks for your support,

May be I found the error which is more critical in my layout, there is a short circuit in my layout, It makes a 5V line come directly to a GPIO Pin of MCU. I think it is  the main reason cause my problem beside the VSS - 177 not got soldered. 

Brs,