TPS630250: input side shorted and output is partially shorted as well

Hi  Rong,

could you share a picture including more info, espcially the silk laryer?

Regards

Tao

Our application has a BLDC motor involved. Three motor wires run through the top of this chip. I captured the Feedback signals between motor wires  close or a little far away from TPS630250.

A little far away

Closed to TPS chip

Chan 1: 5v chan 2: IC203 input, chan3: IC203 enable, chan 4: 3.3v output

Also, we got this design from WBENCH tool. The voltage divider resistance is pretty large. We are wondering if we can change them to 4.7K and 1.5K? Any risk to do it. Our application needs 1.5A in the most of time. Thank you.

Rong 

Sorry a type on picture above, Chan 3 is IC203 feedback 

Hi Rong,

The only concern you change the resister divider to smaller one should be the high consumption, if it is acceptable, i think it is should be ok.

i am afraid the root cause is not the disturbed FB, what is the actual input voltage in your APP? i suspect the damaged device is caused by the high volltage at node Vin, Vin+spikes.

Regards

Tao

Tao, thanks for your reply.  For the smaller resister divider, you mentioned it would introduce high consumption. Is there any thermal concern?  For our application, the normal input voltage is 4.0v or  5V. I checked the input voltage, no very spike found, under 5.5V. The 5v is from another DC-DC buck converter and 4V is from a big charged capacitor. 

Tao, for this chip, the switch between buck and boost converter is based on input voltage and output voltage, right? if so, for case below

Chan 1: 5v chan 2: IC203 input, chan3: IC203 feedback, chan 4: 3.3v output

Was this chip switching between buck and boost mode? 

I guess the gate control ( in green circle)  is the one that decides to run as buck or boost. 

On one of our failure board, I removed L203 and check impedance for L1 and L2. The L2 to ground is zero, and L1 to ground is 283.5K. Does it mean that the chip was running as boost mode sometime?

Thanks.

Hi  Rong,

if the input voltage is 4-5V and output is 3.3V, i think device will always work in buck mode. although there is a diode in series between power input and your board. Device should not work in pure boost mode. so I am confused about the attached picture. 

Mode transition of this device is not only depend on the voltage, also the output current. But, yes, you can almost think it is depend on voltage only.

i still think the high risk should be the voltage out of spec in you Vin and L1 pin.

Regards

Tao

Hi Tao,

You said you were confused about the picture. Could you please let me know which picture confuse you? Thanks. Also, you mentioned 

"Mode transition of this device is not only depend on the voltage, also the output current" 

Hi  Rong,

Rong Liu said:

Chan 1: 5v chan 2: IC203 input, chan3: IC203 feedback, chan 4: 3.3v output

In  the begin of this picture, there are some switching noise coupling, but the noise gone later. seems no switching any more?

Yes, because always have different voltage drop at the FET internal with different Iout.

Regards

Tao

At the beginning, the motor was spinning. then motor stopped, so the noise/switch was gone. The motor wires were on the top of the IC203 (this buck-boost coveter

)

I found an interesting during I switched battery. We have a big cap (charged to 4.5v) in place which functions as an temp power supply so the system is not shut down for 15 seconds. The operator can change battery during the 15s. There are a quick transient on 3.3v output line that is higher than 4v. I am wondering if such high voltage transient will kill the internal output MOSFET?

Chan 1:  5v Chan:2 IC203 input  ; Chan 3: IC203 Feedback, Chan 4: 3.3V output

Another capture , the input side dropped below 3.3V which may push the chip to work on boost mode?

Hi  Rong,

Do you know why the output transient to higher than 4V? i think it should be stable at 3.3V because the feedback divider is a fixed one and will not change during this power switching, right?

For 3.3V  output, the mode transition point should be similar as below. if the ouput is higher than 4V, yes, the device will work in boost mode.

Regards

Tao

I guess the input side jump between 2.93 to 5.17 v, so the chip jumped between buck and boost modes.

Hi Rong,

Ok,  i still think the root cause of damaged device is caused by high spikes because your maximum input voltage is high and layout is not good.

if you want to re-design your layout, please take the example of layout in datasheet as ref. it will help much to decrease the spikes.

Regards

Tao

Thanks for providing this feedback Tao!

Rong has shared the design files with me and I have just sent it to you directly for further review in order to comment on improvements that can be made in the layout and schematic.

Feel free to share your design review feedback to me and I will be sure to pass it along to Rong.

Best regards,

Matt Calvo

Hi  Matt,

I got your mail. will feedback ASAP.

Regards

Tao