Author Topic: LTC4368 failing when shorting output  (Read 1200 times)

0 Members and 1 Guest are viewing this topic.

Offline tfranco_apexar

  • Contributor
  • Posts: 13
  • Country: ar
LTC4368 failing when shorting output
« on: April 19, 2018, 12:26:22 pm »
Hello,
I'm using a LTC4368 to protect some automotive boards but in testing have found that the IC can protect from UV and OV but has issues with the overcurrent protection, specially when you short the output to ground.
Rather than shut-down it "explodes". The IC seems to burn, the most damage being near the VIN pin.
I'm curious if anyone has had similar issues with these LTC protections and perhaps an idea on how to not have itself blow up when it should be shutting down to protect the board.

I'm going to try adding some current limiting resistors in the gate and vin paths and seeing if that helps somewhat. Aside from the charge pump being damaged by the load step, I don't see anything in the functional diagram that could explain how the IC itself is dying. Perhaps I need to add a snubber or something. At the output I have a boost converter which being enabled or not doesn't affect the IC dying (it's basically just a 10uH inductor in series with 300uF worth of caps to ground). MOS:https://www.vishay.com/docs/62634/si7252dp.pdf

http://www.analog.com/media/en/technical-documentation/data-sheets/4368f.pdf

Thanks


Edit: adding a 1K current limiting resistor in Vin seemed to work.
« Last Edit: April 19, 2018, 01:50:58 pm by tfranco_apexar »
 

Offline capt bullshot

  • Super Contributor
  • ***
  • Posts: 1668
  • Country: de
    • Mostly useless stuff, but nice to have: wunderkis.de
Re: LTC4368 failing when shorting output
« Reply #1 on: April 19, 2018, 01:50:48 pm »
It's usually not mentioned in their datasheet - for use cases like yours, you need some additional protection on the IC's input side. Try a suitable TVS. I vastly remember having seen an app note adressing this.

What's happening: Due to the short, the IC will turn off the MOSFETs rapidly. This causes an inductive kickback from the input line (there's just a beefy current beeing interrupted, and this current won't stop in an instant), causing a serious ringing or overvoltage transient at the input side. This easily exceeds the IC's abs max ratings, et voila: it burns.
Safety devices hinder evolution
 
The following users thanked this post: tfranco_apexar

Online SiliconWizard

  • Super Contributor
  • ***
  • Posts: 3436
  • Country: fr
Re: LTC4368 failing when shorting output
« Reply #2 on: April 19, 2018, 02:37:53 pm »
@capt bullshot: That's probably the case. Funnily enough, the datasheet claims that this part can be used to mitigate problems such as "The load or the input is shorted to ground or to another
supply". :-BROKE

It does mention the mitigation of "Transients During OV Fault" on page 19, with the use of a bidirectional clamping diode at the input.

 

Offline tfranco_apexar

  • Contributor
  • Posts: 13
  • Country: ar
Re: LTC4368 failing when shorting output
« Reply #3 on: April 19, 2018, 03:22:50 pm »
You're right. It's the voltage spike being generated. For some reason I wasn't seeing it before. I saw as high as 114V, over the 100V the IC can take.
I guess the TVS isn't really optional.
Thanks
 

Offline tfranco_apexar

  • Contributor
  • Posts: 13
  • Country: ar
Re: LTC4368 failing when shorting output
« Reply #4 on: April 20, 2018, 03:05:32 pm »
I added in a TVS and that made the MOS burn itself out when I shorted the output. I might need to look into a different TVS.
So far having a 550 Ohm resistor in Vin seems to be enough to protect the IC.
 

Offline CopperCone

  • Super Contributor
  • ***
  • Posts: 1415
  • Country: us
  • *knock knock*
Re: LTC4368 failing when shorting output
« Reply #5 on: April 20, 2018, 03:49:02 pm »
What kind of tvs diode did you use? That sounds really suspicious.

If the mosfet burned that would imply a current problem. Are you sure the ic is driving the pass transistors correctly? Maybe the driver was damaged and they are being turnrd off slow spending too much time in the linear region.

The tvs does not really have much of a response time.

Why dont you plug in a resistive load and experiment with controlled shorting so you can run the experiment more then once and also have a triggered scope setup with it to capture the various waveforms.

This situation sounds off. And a 500 ohm pass transistor on a rail is huge. I could see a few ohms to dampen things but with 500 ohm why even bother usi g the ic?

You need to measure the transistors and see if soa is being violated during the switching.
« Last Edit: April 20, 2018, 03:57:08 pm by CopperCone »
 

Offline tfranco_apexar

  • Contributor
  • Posts: 13
  • Country: ar
Re: LTC4368 failing when shorting output
« Reply #6 on: April 20, 2018, 05:26:31 pm »
The 500 ohms are just in the input line to the IC, not in the way of the high current path, so considering the IC consumes less than 3mA that should be inconsequential.
I have done a few different tests shifting the components around, resulting in 3 successful configurations.
Having either the resistor or the TVS on its own is enough to protect the IC, as is having the TVS before the resistor. What seems to destroy the MOS is having the TVS next to the IC's Vin pin while having the resistor in the way.

I've done controlled load steps with the layouts that worked, not with the one that doesn't, and I have some waveforms measured but nothing worth sharing on the one that fails. I'm guessing that somehow, having the TVS after the resistor makes it harder for the MOS to shutdown.

TVS: SMAJ48CA, just for testing the positive side, I'd move to a more appropriate one once I knew that a TVS was able to keep the IC safe.

Edit: The board so far worked best with a 100 Ohm resistor to protect the Vin pin and the bidirectional input TVS. Attempting to add a separate TVS to ensure VGS didn't burn the MOS was unfruitful. I also increased the inrush cap and added the hotswap cap, in hopes it made a difference.
« Last Edit: April 24, 2018, 12:38:52 pm by tfranco_apexar »
 

Online SiliconWizard

  • Super Contributor
  • ***
  • Posts: 3436
  • Country: fr
Re: LTC4368 failing when shorting output
« Reply #7 on: April 20, 2018, 06:41:27 pm »
TVS: SMAJ48CA, just for testing the positive side, I'd move to a more appropriate one once I knew that a TVS was able to keep the IC safe.

They have unidirectional and bidirectional with this reference, which one is it?
LT recommends a bidirectional diode.
 

Online Yansi

  • Super Contributor
  • ***
  • Posts: 2643
  • Country: 00
  • STM32, STM8, AVR, 8051
Re: LTC4368 failing when shorting output
« Reply #8 on: April 20, 2018, 07:16:14 pm »
SMAJ48CA
TVS: SMAJ48CA, just for testing the positive side, I'd move to a more appropriate one once I knew that a TVS was able to keep the IC safe.

They have unidirectional and bidirectional with this reference, which one is it?
LT recommends a bidirectional diode.


SMAJ48CA - that one is bidirectional. It is explicitly specified by the C in the type. A types are uni, CA are always bi.

Unidirectional should be enough to eat the inductive kickback.  But should not be used because it would burn in case of a reverse polarity fault.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf