TPS2660:Damage occurs around EFUSE GND (PIN17) when surge is applied.

Kwangyoung Lee

Part Number: TPS2660
Other Parts Discussed in Thread: TPS2640,

1. During a surge test, EFUSE was damaged when -3kV was applied to line to line or when -4kV from N to PE is applied, Efuse is damaged.

Waveform explanation below) CH1 is the EFuse input voltage, CH2 is the EFuse output voltage, and M is the voltage of CH1-CH2.

Related circuit)

If you proceed with various countermeasures, for reference, check the surge countermeasure data for TPS2640 to add a TVS diode to the input/output terminal, and no EFuse damage occurred during surge evaluation. However, if TVS was added, Efuse damage occurred for some reason when a surge was applied. I don't understand how it was prevented, so please explain.

---EFUSE input/output terminal voltage when surge is applied after adding TVS---

Inquiry details
1. EFuse damage mechanism when surge is applied
2. Mechanism that prevents EFuse from being damaged when surge is applied when adding TVS on top of EFuse after damage.
3. Why is the TVS missing from the input/output terminal of the TPS26601 when a surge is applied? Is this an unnecessary part? I know you are busy, but please reply.

7 months ago

0 Rakesh 7 months ago

TI__Guru** 102111 points

Hi Lee,

The reason for damage could be due to violation of VIN to gnd max voltage spec.

We need to use bidirectional tvs diode SMCJ28CA at the input for 3kv surge. Please add it and test again

Best regards

Rakesh

0 Kwangyoung Lee 7 months ago in reply to Rakesh

Prodigy 80 points

Thank you for answer.
Although it is not in the attached circuit diagram, a bi-directional TVS diode (8.0SMDJ33CA) is also installed in the input terminal. This is the measured waveform when the input voltage is up to 36V and the SMDJ33CA diode is configured.

0 Kwangyoung Lee 7 months ago in reply to Rakesh

Prodigy 80 points

The rated specification of TPS2660 is Max 62V. In the above waveform, the Efuse input terminal waveform is yellow and the Efuse output terminal waveform is blue. Please explain at what point the VIN and GND ratings were exceeded.

0 Kunal Goel 7 months ago in reply to Kwangyoung Lee

TI__Guru 56860 points

Hi Kwangyoung,

I will get back by tomorrow EOD on this.

Regards

Kunal Goel

0 Kunal Goel 7 months ago in reply to Kunal Goel

TI__Guru 56860 points

Hi Kwangyoung,

1. You are using a TVS (8.0SMDJ33CA) from VIN to GND and also a TVS from VIN to VOUT?

2. Waveforms you shared are measured on fail device or good device?

3. In waveforms is it VIN,VOUT with respect to GND or RTN pin?

4. Is this fail repeatable?

Regards

Kunal Goel

0 Kwangyoung Lee 7 months ago in reply to Kunal Goel

Prodigy 80 points

1. You are using a TVS (8.0SMDJ33CA) from VIN to GND and also a TVS from VIN to VOUT?

->EFuse is damaged, so TVS is configured and evaluated on Vin and Vout.

2. Waveforms you shared are measured on fail device or good device?

->Above is the failure waveform, and below is the normal waveform with the TVS installed between VIn and Vout.

3. In waveforms is it VIN,VOUT with respect to GND or RTN pin?

->It's GND

4. Is this fail repeatable?

->In the absence of one TVS, it is often repeated.

Add Q)EFUSE damage appears to be caused by the IN-OUT voltage difference exceeding the spec. When a TVS is installed between VIN and VOUT and a surge is applied, the VOUT voltage drops to almost 0V and a power reset occurs. Is there perhaps another way?

0 Kunal Goel 7 months ago in reply to Kwangyoung Lee

TI__Guru 56860 points

Hi Kwangyoung,

Looks like when you use TVS from VIN to GND and VOUT to GND fail does not happen. So what is the concern? Sorry for inconvenience.

You can check VIN to RTN and VOUT to RTN, maybe that is getting violated.

Regards

Kunal Goel

0 Kwangyoung Lee 7 months ago in reply to Kunal Goel

Prodigy 80 points

Hi Kunal Goel

Thank you for answering.
With TVS configured between Vin and Gnd and Vout and Gnd, EFuse was damaged when a surge was applied. However, when TVS was added between Vin and Vout, no damage occurred up to 4kv. However, when TVS is added, as shown in the waveform above, Vout (ch2) drops to 0V for about 2.5m and a power reset occurs. What is the solution? And is there any other way to deal with the surge without adding a TVS between Vin and Vout?

0 Kwangyoung Lee 7 months ago in reply to Kunal Goel

Prodigy 80 points

Hi Kunal Goel

In the DATABOOK, TVS is connected to power GND, so TVS and RTN are not connected.

0 Kunal Goel 7 months ago in reply to Kwangyoung Lee

TI__Guru 56860 points

Hi Kwangyoung,

We need to zoom in there and see if VOUT is -ve. That can reset device. Also is VIN going below UVP?

Regards

Kunal Goel

0 Kwangyoung Lee 7 months ago in reply to Kunal Goel

Prodigy 80 points

What does -ve represent?

0 Kwangyoung Lee 7 months ago in reply to Kunal Goel

Prodigy 80 points

Hi Kunal Goel

This is a different waveform when a 3kV surge is applied under the same conditions. The Vout output is close to 0V for about 5ms.

I also checked the output voltage reset in the log.

0 Kunal Goel 7 months ago in reply to Kwangyoung Lee

TI__Guru 56860 points

I mean, does VOUT goes negative?

0 Kwangyoung Lee 7 months ago in reply to Kunal Goel

Prodigy 80 points

Hi Kunal Goel

In the waveform above, the Ch2 minimum is -1.6V.

0 Roosbel Guaya 7 months ago in reply to Kwangyoung Lee

Prodigy 40 points

Hi:
Maybe on the PCB you have PPAD connected to GND?
On your schematic your PPAD is connected to RTN, if it is connected to GND too, then when you apply a negative surge you will have problems, even with low negative V.

Best regards

0 Kwangyoung Lee 7 months ago in reply to Roosbel Guaya

Prodigy 80 points

Hi Roosbel Guaya

PPAD and RTN are connected, and GND is handled separately.

Best regards

0 Kunal Goel 7 months ago in reply to Kwangyoung Lee

TI__Guru 56860 points

Hi Kwangyoung,

VOUT going -1.6V will reset device. You will need more cap and schottky diode at output to clamp VOUT to near -0.3 to -0.5V.

Regards

Kunal Goel

0 Kwangyoung Lee 7 months ago in reply to Kunal Goel

Prodigy 80 points

Hi Roosbel Guaya

Looking at the top waveform, if you increase the output stage Cap capacity without adding D2 (TVS between Vin and Vout), there will not be a large voltage change in Vout like the Ch2 (Vout) voltage waveform. However, when a surge is applied, a large voltage difference of over 60V occurs between Vin and Vout, causing damage to the Efuse. When D2 is added, Vin and Vout become closer to 0V, and the voltage difference between Vin and Vout does not reach 60V, so there is no damage to the Efuse. However, as explained above, the problem is that Vout drops and a power reset occurs. Is there a way to solve this? And is D2 absolutely necessary?

0 Roosbel Guaya 7 months ago in reply to Kwangyoung Lee

Prodigy 40 points

Hi:
I think that happens to you when you apply a negative voltage surge. Then it may be the case that the Schottky diode enters in conduction at that time.

About the PPAD(PowerPad), usually you have it connected to GND for heatsinking. The connection is with VIAS. That is what is recommended by the datasheet.

So maybe you can inspect the PCB visually or maybe do a continuity test of the vias at the bottom of the IC with GND?

Best regards.