Forums

Hey Joe,

I think I am going to need a little more information about the set-up that is causing this issue to be able help you. I have personally never seen any of these devices blow up.

Different pins on the device have different protection and are rated for different voltages. For example the CANH and CANL pins can withstand DC voltages from -36 to +36 volts, whereas the supply voltage pin is only able to take up to 7 volts.

Which pins are seeing the 12 volts? Is it a specific side of the device that is burning up (bus pin side or TX RX pin side)? Can you share with me a schematic? Are you running a specific test when you see this issue? Has this happened on multiple devices? Once I know some of these things I think I will be able to help you more.

Thanks,

John

Joe:

We use that same CAN bus transceiver and have never had a failure in 20K pieces.  We often short CAN_H and CAN_L together as one of our tests.  If you want to post that section of your schematic, I'll review it for you.

Good luck finding your issue.

Thanks,

JR

http://simmasoftware.com

Hi John,

What we are seeing is that pin 1 is dead and there appears to be some burning on one side of the housing.  We are doing HASS testing and shaking them at 20 gRMS for 30 minutes and temp cycling from -40°C to 70°C.

Thanks for your help,

Joe

Hi Joe,

Sorry for the delay in getting back to you, I was out on vacation.

During your HASS testing what voltages are applying to all the pins on the device? Are you applying 12 volts to the D pin (Pin 1) during your stress screening? The absolute max voltage for this pin is 7.0 volts. We do not gaurentee device performance if this rating is exceeded.

Could you share with me a circuit schematic of how the device is being tested?

John

Hi John,

I'll have to look deeper to see if we are getting unwanted voltage on the lines.  Even if we are I'm not certain how to verify that.  We have sent two failed units out for X-ray removal of the lid.  They replied"

External Visual Examination/Documentation: Failed devices contained a clear conformal coating on the device body. S/N: 2 exhibited broken/missing “Pin 1” lead (See Figures 1 and 2).
Physical Dimensions/Documentation: All of the samples met the manufacturers drawing requirement (See attached data).
X-Ray Inspection/Documentation: No damage observed (See Figures 3 thru 5). Device de-capsulation/Documentation: Failed devices and good sample showed the same chip die part markings, no
damage observed (See Figures 6 thru 10). No further analysis performed.
Conclusion: The failed devices did not exhibit any signs of mechanical or electrical overstress damage. S/N: 2 broken device lead could have been induced during extraction of PCB/assembly. The results of this analysis are inconclusive.

I'll have to check into whether I can release the schematic beyond they pins of your part.

Joe,

Could you try attaching the pictures you are referring to again? You can also email me at johngriffith@ti.com.

John

John, earlier today I emailed you the failure report with the TI part number in the subject line.  Hopefully it made it past your spam filter.

Hi John, did you receive the test report email?

Hi John,

Can you say more about what we could observe or test for if failures where caused by getting 12V on the D pin?

Hello,

Sorry for the delay, I was out of the office for a week on travel. I did receive your test report. The issue with over voltage stresses is that it is a race condition on what fails within the device. I cannot make any suggestion on tests that would verify what caused the failure.

I am going to need schematic / test procedure to be of more help. Also if you let me know your company and location I can help get you in touch with a local ditributuion FAE or field sales team.

Thanks.

John

John, a coworker has suggested another potential cause.  Because we have a slip ring as part of the bus and we are testing with vibration he felt that a failure he had seen in other transceiver chips my apply.  He stated that the some designs are such that if ground opens there is a diode that will allow voltage to go to pin 1.  Is that the case with this chip?

Hi Joe,

Unfortunately, I do not beleive that loss of ground would cause any issue like you are describing. When the device looses ground, all of the nodes on the device will float up to Vcc. I do not see how this could cause a the amoun of current needed to burn a pin out.

In the past when I have seen device loose I have seen the bus voltage float up to VSUP. That is the worse thing I have seen with loss of ground. Also the datasheet specifically states "Designed for operation in especially harsh environments, the devices feature cross-wire, overvoltage and loss of ground protection to ±36 V..."

Can you share with me a schematic / test procedure? Also what is your company and location?

Thanks,

John

John, I have emailed you a portion of the schematic.  The test procedure involves rapid temp cycling within the design specs of the transceiver and fairly intense random vibration for 30 minutes.  We are in Southern California.

Hi Joe,

I did receive your schematic. Everything looks good to me. With 10 Ohms to ground you should still be in high speed mode. I went in the lab and measured thevoltage on the RS pin with 10 ohms to ground and it was less than 10mV.  Is there a reason you choose not to connect it to ground without a resistor?

I do not suspect the temp cycling to be an issue, but I do not have any information on levels of stress our devices can handle in terms of G rms on a vibration table. If this is a concern I can talk with our packaging team to find out more information.

Thanks,

John