TMS570LS0714: CPU not booting on reset if > 105*C

We monitored nRST, 3.3V supply, 1.2V supply and power hold signal. As it can be seen in the video, power supply voltages are up to spec at all times.

What happens when PCB temp rises above 80*C is that nRST signal switches between 0.2 to 0.8V with the high signal threshold slowly increasing as the temperatures drops until the reset pin voltage level reaches the minimum necessary voltage to trigger the reset in CPU.

nRST singal is pulled up via external 1k2 resistor.

Here you can view the video: https://files.fm/u/y838c2yx

I added our test case description with results as a word file, accessible using the link from my previous post. As well as a new video for additional information. Please have a look at section: 9 Additional tests – 20181009 in 1006302-B file.

Our PCB employs TI external watchdog IC. Tests were repeated with and without the IC on the board, giving exactly the same results. There is no other functionality on the board besides that, driving CPU nRST pin.

We have a 1.2kohm pullup which is within the spec. Max current sourcing into the CPU through nRST is 2.6mA when pin is low which doesn't exceed the datasheet limit.

We have an older revision of PCB which employs TI TMS570LS1114 CPU and we couldn't reproduce the problem. Reset line didn't sink and/or float after power cycle at temperatures above 80*C and below 125*C.

Our conclusion so far is that the transistors used for reset pin configuration become leaky and can't hold the desired voltage level on startup if CPU temperature goes above 80*C. 

we are looking forward for any response from TI

The latest data sheet for the TMS570LS0714 is available online (SPNS226E –JUNE 2013–REVISED NOVEMBER 2016)

There is a highlighted notice on page 163.

• Section 4.2. Describes connections affected by the PORRST, including the nRST.
• Section 5.4. Recommended operating conditions are -40 to +125oC. (ambient).
• Section 6.4. The warm reset (nRST) function description. T.I. recommends an external resistor is attached to the terminal pin.

EHS drive pcb has a 10K resistor attached to the pin. Sub-section’s of this folder highlight internal causes of a warm rest event and highlight timing requirements, which are met within the design

• Section 4.2.1.13. The nRST i/o pin has an internal 100uA pullup. (-195 to -40uA range)
• Section 4.4. nRST low level output current 4mA max for VL = VOLmax over -40 to +125oC range

The pull-up resistor has been changed from 10K to 1K2, which increases the nRST pull up peak current close to 3mA.

It can be seen, (look at photos attached), that with the 1K2 pull up resistor on the nRST pin the negative voltage transients on the nRST continue to become larger after a power on reset as the temperature increases.  This occurs at the processor initializes as before. The processor being tested will start with a local temperature of 99 oC. before the failure to start situation appears again

Initial testing of the product was carried out using a 1114 variant of the processor. The 0714 was replaced by a 1114 and the test repeated with the TPC3850 \WDO pin not connected. The 1114 will restart with a temperature in excess of 102oC with no sign of voltage reduction during the initialization cycle.

Trace key

CH1 (yellow) - nPORRST

CH2 (blue) – nRST

CH3 (pink) – 1V2

CH4 (green) – 3V3

Our conclusion (for in depth detail of testing, please look at the attached file):

The TMS570LS0714 nRST pin becomes leaky at higher temperatures, requiring a lower value of pull up resistor to enable device configuration after power on reset if the pull up resistor is connected to 3.3V as recommended.

With a current restriction of -4mA associated with the pin, it is not possible to have a low enough pull up resistor to ensure the processor can configure itself and carry out startup tests for the full device temperature range.  It is possible to do this with the TMS570LS1114.

It is not clear if the leakage current variation is typical for devices of different manufacturing dates.

The customer expects the system to operate in a 90oC ambient.  A resistor change will improve the nRST situation and if the processor is too hot, the processor will always start when it cools down.

It is not clear that this situation will be acceptable to the customer and it doesn’t meet their requirement specification.

5102.1006313-A (EHS150-1257 - 1258 (G2 - Wake up & CAN loss at temperature)- 2018-10-12).pdf

Hi Sunil.  My name is Duncan Hamilton - I have been working on the nRST investigation.  Is response to your question, we see it on all the 0714's.  The silicon we have is at revision A.  We did not see this problems (and still don't) when using the 1114 - we originally used a 10K pull up resistor and carried this forward to the 0714 designs.  

We also have the /WDO pin of the TPS3850DRC  watchdog connected to the nRST pin and a diode with another pull up resistor.  We have removed all external influences from the pcb except for the 10K local pull up resistor.  Changing the 10K to a 1K2 improves the situation in that the processor will start just below 100C, but still have difficulty with meeting the customer expectation.

Uvis has attached the report above, which is 1006313-A.  The 2 embedded videos will not be available in the pdf document taken with the external 10K resistor, which think will interest you.  I will attach them below. 

The image file without the watchdog fitted is the more interesting one - the nRST pin is the lowest one on screen at the start of the video.  In both cases, the other traces are the nPORRST, 3.3V and 1.2v supplies.  We have had to make the image files smaller in size - hope they are ok...

Could you please confirm the TMS0714PGE is certified for automotive use?  It is not on the list of approved parts T.I. sent us.

Is it possible to have direct contact with you, either by email, Skype or phone? 

8176.EHS150 G2 pcb INC wdog - PonReset 1v2 3v3 and nRST with air cooling from 80C 20181011_reduced.mp43835.EHS150 G2 pcb no wdog - PonReset 1v2 3v3 and nRST with air cooling from 80C 20181011_reduced.mp4

Is it possible to have direct contact with you, either by email, Skype or phone? 

Two further questions

The  TMS5700714APGEQQ1  is lower cost than the TMS5701114CPGEQQ1  

The TMS5700714APGEQQ1   does not have safety application guidelines  TMS5701114CPGEQQ1 has safety application guidelines including one endorsed by TUV

The TMS5700714APGEQQ1   does not have an ASIL D compatible qualification  the TMS5701114CPGEQQ1 does

So can we conclude that the TMS5700714APGEQQ1  is automotive qualified, but not as a safety processor?