TPS4H000-Q1: Fault recovery after short-to-GND condition does not work

Hello Paolo,

Welcome to E2E! Could you please provide a schematic surrounding the TPS4H000B-Q1 and scope shots of channels 1-3 hitting the current limit? If you could also provide a scope shot of Vs when you introduce the short- it could help narrow things down a bit. The behavior you've described is not expected for the device. A short on one channel should not affect the other channels. 

Regards,

Kalin Burnside

Hello Kalin,

the schematic surrounding the TPS4H000B is that of the TPS4H000EVM I picked up to facilitate the device evaluation and in which the original BOM part TPS4H000AQPWPRQ1 was replaced with a TPS4H000BQPWPRQ1 chip.

I provide the scope shots you've asked for, with the following setup:

• VS = 24 V
• R(CS) = 3200 Ω (R18 variable resistor) => Iout = 100 mA nominal current in the 0 to 4 V current sense range
• R(CL) = 3830 Ω (R21 variable resistor)  => Iout = 62,5 mA expected current limit value
• DIAG_EN = 5V to enable the diagnostic function
• THER pulled to GND to have the termal shutdown operating in the auto-retry mode
• IN1 = IN2 = IN3 = IN4 = 5 V for channels activation
  

The shots reported in the following 1), 2), 3), 4) and 5) show that a short on one channel always affects the other channels and the device does not automatically recover after the fault condition is removed.

1) Short-to-ground on OUT4 while OUT 1 feeds a resistive load - Scope shot of OUT1, OUT4, CS and nFAULT when I introduce the short between OUT4 and GND

Yellow => OUT1 -- connected to a R = 300 Ω resistive load
Green => OUT4 -- no load connected, used for short-to-ground
Orange => CS -- monitoring CH 1 with SEH=0 and SEL=0
Blue => nFAULT

2) Short-to-ground on OUT4 while OUT 2 feeds a resistive load - Scope shot of OU2, OUT4, CS and nFAULT when I introduce the short between OUT4 and GND

Yellow => OUT2 -- connected to a R = 300 Ω resistive load
Green => OUT4 -- no load connected, used for short-to-ground
Orange => CS -- monitoring CH 2 with SEH=0 and SEL=1
Blue => nFAULT

3) Short-to-ground on OUT4 while OUT 3 feeds a resistive load - Scope shot of OUT3, OUT4, CS and nFAULT when I introduce the short between OUT4 and GND

Yellow => OUT3 -- connected to a R = 300 Ω resistive load
Green  => OUT4 -- no load connected, used for short-to-ground
Orange  => CS -- monitoring CH 3 with SEH=1 and SEL=0
Blue => nFAULT

4) Short-to-ground on OUT4 while OUT1 and OUT2 feed a resistive load - Scope shot of OUT1, OUT2, CS and nFAULT when I introduce the short between OUT4 and GND

Yellow => OUT1 -- connected to a R = 340 Ω resistive load
Green => OUT2 -- connected to a R = 300 Ω resistive load
Orange  => CS -- monitoring CH 1 with SEH=0 and SEL=0
Blue => nFAULT

5) Short-to-ground on OUT4 while OUT1 and OUT2 feed a resistive load -- Scope shot of OUT1, VS, CS and nFAULT when I introduce the short between OUT4 and GND

Yellow => OUT1 -- connected to a R = 340 Ω resistive load
no trace provided => OUT2 -- connected to a R = 300 Ω resistive load
Green => VS supply voltage
Orange  => CS -- monitoring CH 1 with SEH=0 and SEL=0
Blue => nFAULT

Thank you for your kind attention.

Regards,

Paolo

Hello Paolo,

Thank you for the detailed description and scope traces. One of the aspects that look suspicious here is the voltage dip on the output of the channels that are not being short circuited. This suggests that there is a possibility the supply could be collapsing or there might be some issue with the upstream power rail.

What is the supply connected to? Is it an external power source like a bench supply or something integrated like a DC/DC converter?

Hello Timothy,

the VS and GND connectors of the TPS4H000EVM board are connected to the positive and negative output terminals, respectively, of a bench supply set to provide a constant supply voltage of 24 V and a maximum current of 1 A.

To rule out the external power source from the potential causes, I even changed the bench supply used to perform the tests. Unfortunately, the behaviour of the device is always the same I described.     

Regards,

Paolo

Paolo,

Understood here and thanks for the additional data point about the supply.

I think at this point the best course of action would be for us to reproduce this in our lab here locally and post the scope shots. The way that we would do this is to use the TPS4H000EVM in our bench setup. Each channel's thermal protections should be independent of each other and should not be showing the behavior that you are seeing. I have reserved some time in the lab and should have results by the end of the week at latest.

As a sanity test could you triple check the soldering work of your TPS4H000-Q1? What we have seen in the past is if the thermal pad is incorrectly soldered or if there is low solder coverage then the device performs very poorly thermally and could potentially explain the behavior you are seeing. 

Hello Timothy,

Timothy Logan said:

I think at this point the best course of action would be for us to reproduce this in our lab here locally and post the scope shots. The way that we would do this is to use the TPS4H000EVM in our bench setup. Each channel's thermal protections should be independent of each other and should not be showing the behavior that you are seeing. I have reserved some time in the lab and should have results by the end of the week at latest.

Thank you for your attention to this issue. I look forward to hearing your results.

Timothy Logan said:

As a sanity test could you triple check the soldering work of your TPS4H000-Q1? What we have seen in the past is if the thermal pad is incorrectly soldered or if there is low solder coverage then the device performs very poorly thermally and could potentially explain the behavior you are seeing.

I inspected the soldering under a microscope and everything seems fine. Unfortunately, I have no X-Ray inspection equipment to check the thermal pad soldering.

Regards,

Paolo

Hello Paolo,

I created your test setup in the lab and was unable to see the behavior you were experiencing.  Specifically what I recreated in your list was condition #5. On the TPS4H000EVM (B silicon populated) I was seeing that when shorting OUT4 to GND the CS when set to OUT1 was unaffected:

When I put the voltage probe to OUT4 and set SEL/SEH to monitor on channel 4 the CS pin goes to the appropriate fault voltage:

A few things to check on your end:

• Are you sure there isn't some way you are increasing your current on OUT1 and partially shorting? It looks like your output voltage dips on VOUT1 when it shouldn't be affected by the short on VOUT4.
• Could you monitor the current (if you have a current probe) on OUT1 during the fail condition?
• Could you try to resolder the device and make sure the thermal pad is soldered correctly? If the thermal pad has poor solder coverage the thermal performance will be unreliable.
• Could you share your layout so that I can double check it for any problems?

Hello Timothy,

thank you for the tests and for sharing your results.

Timothy Logan said:

Specifically what I recreated in your list was condition #5. On the TPS4H000EVM (B silicon populated) I was seeing that when shorting OUT4 to GND the CS when set to OUT1 was unaffected

May I please ask you if, recreating condition #5, the value of current limit resistor R(CL) and the value of the load resistor connected to OUT1 were the same I set for my test, i.e. R(CL) = 3830 Ω and R = 340 Ω respectively? In case it wasn't, then I may have some explanation about why I saw that the CS when set to OUT1 was affected when shorting OUT4 to GND.

Thank you for your kind attention.

Regards,

Paolo