• State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 62 subscribers
  • Views 541 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LP2951-N: LP2951-N Thermal Shutdown

Michael Ellis1
Prodigy 30 points
Part Number: LP2951-N
Other Parts Discussed in Thread: LP2951

Hello,

This is a question relating to Thermal Shutdown of an LP2951-N LDO regulator.

We are trying to debug an issue we have with site failures of a number of our devices. Another manufacturer’s IC is unexpectedly going into a short-duration shutdown mode and then recovering. We think that this may be because of disturbance on the Vcc (+5V0) of the IC, although we are not certain of this.

The Vcc is supplied by an LP2951-N part (markings on the IC are: 3A0J L0DB). Our devices have been in service for 7 years and have only recently failed. The sites where the devices are operated are typically not air-conditioned and ambient could reach +40Deg C, we think. Our devices do not have any forced-air cooling and we estimate that the temperature inside our boxes could get to around 60Deg C.

The LP2951 is dropping +12V to +5V0 whilst supplying approximately 65mA. I’ve measured the temperature on top of the package with a K-type thermocouple and the temperature is around 72Deg C. This is at an ambient of about 22Deg C in our office. The board is 4 layer (with internal GND and Power planes) and the LP2951 routing is reasonable. See below for part of our schematic.

I have read the IC Package Thermal Metrics (SPRA953C) app note and think I understand most of it. I have estimated that the Tj could be about 76Deg C using an appropriate equation. I have applied localised heat to the IC using a Steinel HG2310 heat gun set to 60 and 70Deg C at about a distance of 1-2” for a few minutes. I am not sure whether this is valid or not.

I set our oscilloscope to trigger on a Vcc downward excursion and include the following waveform. See below for the waveform data.

Note that the record length is 5sec. The lowest visible spike is about 3.50V and approx. 5ms wide.

Questions:

  1. Is 72Deg C typical of what you would expect to measure on the top of the IC for 12V to 5V0 regulation @65mA, ambient being around 22Deg C? I’m wondering if this part is running too hot as it feels very hot to the touch?
  2. Are my oscilloscope traces showing the thermal shutdown mode in operation, or are we seeing another phenomenon?
  3. The other manufacturer has said that Vcc dropping to 2V5 might put their part into difficulties and be the cause of our site problems.
  4. There is scant information about Thermal Shutdown in your datasheets. Could the output go as low as 2V5 or even 0V?

Thanks, and Regards

Mike Ellis

Hello,This is a question relating to Thermal Shutdown of an LP2951-N LDO regulator.We are trying to debug an issue we have with site failures of a number of our devices. Another manufacturer’s IC is unexpectedly going into a short-duration shutdown mode and then recovering. We think that this may be because of disturbance on the Vcc (+5V0) of the IC, although we are not certain of this.The Vcc is supplied by an LP2951-N part (markings on the IC are: 3A0J L0DB). Our devices have been in service for 7 years and have only recently failed. The sites where the devices are operated are typically not air-conditioned and ambient could reach +40Deg C, we think. Our devices do not have any forced-air cooling and we estimate that the temperature inside our boxes could get to around 60Deg C.The LP2951 is dropping +12V to +5V0 whilst supplying approximately 65mA. I’ve measured the temperature on top of the package with a K-type thermocouple and the temperature is around 72Deg C. This is at an ambient of about 22Deg C in our office. The board is 4 layer (with internal GND and Power planes) and the LP2951 routing is reasonable. I have read the IC Package Thermal Metrics (SPRA953C) app note and think I understand most of it. I have estimated that the Tj could be about 76Deg C using an appropriate equation. I have applied localised heat to the IC using a Steinel HG2310 heat gun set to 60 and 70Deg C at about a distance of 1-2” for a few minutes. I am not sure whether this is valid or not. I set our oscilloscope to trigger on a Vcc downward excursion and include the following waveforms: The record length is 5sec. The lowest visible spike is about 3.50V and approx. 5ms wide.Questions:1. Is 72Deg C typical of what you would expect to measure on the top of the IC for 12V to 5V0 regulation @65mA, ambient being around 22Deg C? I’m wondering if this part is running too hot as it feels very hot to the touch?2. Are my oscilloscope traces showing the thermal shutdown mode in operation, or are we seeing another phenomenon?3. The other manufacturer has said that Vcc dropping to 2V5 might put their part into difficulties and be the cause of our site problems.4. There is scant information about Thermal Shutdown in your datasheets. Could the output go as low as 2V5 or even 0V?Thanks, and RegardsMike Ellis

  • 0
    TI__Mastermind 37945 points

    Hi Mike,

    Using the datasheet thermal metrics for the D package, I calculate the junction temperature rise due to power dissipation to be 44C. With an ambient temperature of 70C, the virtual junction temperature would be 114C. This is well within specs, as the actual thermal shutdown temperature is normally above 150C , and it is typical for it to be around 170C for many parts. So to me, this does not seem like a junction temperature violation. Furthermore, the scope shot does not show thermal shutdown. During thermal shutdown the output voltage and load current typically head towards zero. If the thermal shutdown is prolonged, the output voltage and load current will both stay near zero, and you would not see prolonged shallow valleys like you are seeing. 

    Have you monitored the 12V rail when this even is happening? I do not know why this would be happening if the 12V rail is behaving normally.

    Best regards,

    Nick

  • 0 in reply to Nick Butts
    Prodigy 30 points

    Hi Nick,

    Thanks for the quick response, it is appreciated.

    Unfortunately, whilst debugging the board and removing various ICs, I managed to lift the pads of the regulator IC whilst cleaning them with de-soldering braid. They are so very fragile! So yesterday, I got another failed device and tried heating the regulator IC with the heat gun. This one did not produce the same disrupted 5V0 output, even though I went as high as 80DegC This is good news, I guess?

    Regarding your comment on the health of the 12V rail. I'm fairly certain that it is good as it's produced by a very expensive, proprietary isolated DC-DC converter made for us by Murata. However, I understand your point and will try to get the original board and IC working again, by some micro-surgery. My eyesight isn't great these days, so wish me luck!

    I'll let you know how I get on and hopefully will have some scope traces showing healthy 12V0 and disrupted 5V0 to show you. This is likely to be sometime next week, as we are not fully back to work due to the Covid crisis.

    With best regards,

    Mike

  • 0 in reply to Michael Ellis1
    TI__Mastermind 37945 points

    Hi Mike,

    I suppose that may be good news as it adds to the evidence that this is likely not a thermal shutdown issue. 

    Have you tested to be sure that it is not another IC that is being supplied by the 5V0 line that may be causing a load transient effect? Is it possible that the heat generated by the LDO is too much for nearby ICs?

    I look forward to seeing your scope shots.

    Best regards,

    Nick

  • 0 in reply to Nick Butts
    Prodigy 30 points

    Hi Nick,

    Unfortunately, I haven't been able to get the first board back up and running again. I established all of the pin connections of the part to the correct nets (there were only 3 missing pads) and turned it on, but it is not now producing the correct output voltage. I think that the part may be damaged. It produced an output of 4.3V with a regular drop to 0.6V for 40ms in approx 1 sec intervals. I could not get anything else out of it, even after re-working all of the the joints.

    Next, I removed the part and put it into a known good board where this new board's LP2951 regulator had just been removed. It then produced an output of 2.9V, with similar mark-space?? (See attached diagram). Hopefully, this might make some sense to you? I have written this part off as being damaged and now will concentrate on other returned devices. (I stupidly forgot to record the 12V0 supply to this part whilst it was doing this, but it was a healthy 12V0).

    We have received one more device back from site and I set about heating the part up and checking the 5V0. This part was heated to 90Deg C and this didn't produce any errors on the output. We now have had 3 devices returned and only one has shown this output voltage error (at 60Deg C), and this just happened to be the first one I looked at.

    I'm convinced that we are not seeing Thermal Shutdown in operation and really that was what I had wanted to know. The other IC manufacturer has confirmed that a short duration drop to 2V5 would cause the effect that we see in our site failures. So, I will continue my search!

    As for your points: I will be concentrating on looking for the source of this transient effect. My suspicion is pointing to the other manufacturers part, which may be consuming 20mA more than I am expecting. I doubt that the heat from the LDO could cause a problem with other IC's as it is the top most IC in a vertically mounted PCB - its' unlikely, I think.

    I'll leave this thread open, if that's okay with you, as I may well have more questions to ask?

    With best regards,

    Mike