TPS2388: TPS2388 does not deliver power at -34C

Part Number: TPS2388


While running some quality testing on a product we are prototyping, I noticed that my PSE would not give power to my attached PoE devices if the PSE was powered on at -34C.

If I started the PSE at room temperature and then lowered it to -34C the devices would remain powered on, but if I started at -34C my attached devices would not power on. Next test I did was power on the PSE at -34C and then raised board to room temperature and left it in the chamber over the weekend. When I got back my PoE devices still had not powered on. 

Any thoughts as to what might be happening? The datasheet says that it should be able to go down as low as -40C. 

Thanks!

  • Hi Josh, 

    Have you noticed this with one single PSE device or multiple devices? Are there any other functional issues you have observed with the device?

    Let me see if I can test this internally and get back to you. 

    Best,

    Anagha

  • Multiple devices and individual. As the product is still in development and this is a public forum I won't say how many ports we are using, but I tried individual ports, multiple combinations of ports at the same time, and all of the ports that I had wired up at the same time. Nothing would power on at -34C. However if I started at room temperature and then brought the temperature down the attached PoE devices would remain on. 

  • Hi Josh, 

    Thank you for the info. Let me test this, I can get back to you by 10/31. 

    Best,

    Anagha

  • Hi Josh, 

    My apologies for the delay in response. 

    I am unable to recreate this issue on our end at this time. Can you provide your test environment? 

    Additionally, have you verified that other components on the board (such as resistors, caps, etc.) are still functional at -34C? Even if these components are not functional, it will still appear as though the PSE is not turning on. I am also checking with our validation team to see if they have any additional data on TPS2388 operation at low temperatures. 

    Best,

    Anagha

  • Hi Josh, 

    Just a quick update - looks like data from the validation team shows normal operation for TPS2388 at temperatures as low as -40C, as mentioned in the datasheet. 

    How are you measuring the temperature? Please ensure that recorded value is the junction temperature of the device. 

    Best,

    Anagha

  • Hey Anagha, 

    Testing was done inside of a temperature controlled environmental chamber. Temperature measurements were taken with a Flir thermal camera. 

    All power rails on the board were operational at -34C as I was also doing power supply load testing in conjunction with functional tests. I ran an additional test by doubling the amount of bypass capacitance to the 3.3V rail (was 1.0uF with a single 0402, I doubled capacitor amount to 2.0uF with two 0402's in parallel) and it still wasn't powering on at low temperature.

    Other supporting components such as resistors, FETs, diodes, and capacitors are rated to -50C and below. 

    I can run some more tests on my end if you have suggestions on what to look for. I will also try with other boards as I have more PSE's on hand. 

    Thanks!

  • Hi Josh,

    Thanks for info. Would you be able to test and check the lowest temperature at which the PSE is operational? Additionally, can you provide the following: 

    • Schematic and layout of board
    • Power on sequence of VDD, VPWR, and /RESET pins

    Thanks,

    Anagha

  • Here is a snippet of some of the ports in the design. Like I said, I have to be incredibly careful as this is a public forum, so information I can pass on is limited. 

    That being said, our DUT is powered off of the 54V, so that comes on first. VDD comes on second. Our /RESET pin is tied directly to the processor, so that could be a potential culprit as I don't have any pull up or pull down resistors on that line and the processor currently does not have firmware on it. 

  • Hi Josh, 

    That might be the issue. /RESET must be held low until VPWR and VDD have hit their UVLO thresholds. This is why the datasheet recommends that some sort of RC delay is added at the /RESET pin. Can you try this and see if it fixes the issue? 

    I don't see any other concerns in the screenshot you shared, but I would recommend adding a port fuse for added protection to the load. 

    Thanks,

    Anagha