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.

  • TI Thinks Resolved

LM22675: LM22675 startup burn

Prodigy 50 points

Replies: 9

Views: 2641

Part Number: LM22675

I am in trouble with LM22675MR-ADJ.
I use this device since years on many boards, one type (T1006) of them don't works as well as I expected.
Many of T1006 electronic boards delivered during years work fine until, during a startup,
the chip burns, fusing its pin n°7 (Vin on SOIC8 package), actually this happens frequently.

I'm testing the device with stressing both in input then output without problems
(applying 38Vdc on input , with a load of more than 1A on output).

Searching on website, I found people with the same problem (see:,
the solutions suggested are made by me yet.

The device is used to power some internal circuitery and provide power supply to external optical encoders (total current max 600mA).

Is it possible the chip or some lot of them is bugged ?
Do you think it's better to select another ones if so, please suggest me the part number.

device is U118

Elitron - Divisione Elettronica
Via del Chiesino, 8
56025 - Pontedera (Pisa) - Italy
Cell. +39.380.6334480 Sciarrini Massimo
Cell. +39.380.6334479 Giannessi Roberto
  • Your schematic looks OK.  I notice that you have two instances of LM22675.  Are they both having trouble?

    You highlighted the one connected to your input connector; I assume that is the one giving trouble?

    Can you send wave-forms of your startup.  Vin, EN, Vout, and SW would be helpful to see during start-up and normal running.

    What is your expected input voltage and normal load current?

    I would look for the possibility of an over voltage on the input.

    Be sure that you inductor is not saturating during start-up.  It should be rated for the current limit of the device.

    I have a few comments on the layout:

    1.  I would try to get a better ground connection between D35 and C1027/28.  It looks like it is going down to a ground plane???

    2.  I do not see C1027.  This should be close to the device; close to C1028

    3.  I would make the trace to C1024 and pin 1 fatter.

  • In reply to Frank De Stasi:

    No I got problem on U118 only.

    All images refer to start up ,unless last one  image, this refers to normally run.

    All wave form are made with a load of 800mA and Vin 24 Vdc.

    The normal input  voltage is 24Vdc and with 600mA load.

    Yes this is 24 vdc of industrial machines. I put TVS 40 V working, and 400uF electrolitic capacitor with some 100nF x7r to reduce disturbs in input

    Looking waveforms I think  inductor isn't saturated, however i am not able to measure in this circuit. (suggestions?)

    The only differnce between this circuit and all others (many hundreds working fine) is the enable pin used, which is left floating

    track blue:enable

    track yellow: switch




    track blue:Vin

    track yellow: switch


    track blue:Vin

    track yellow: enable


    track yellow: first pulse


    track yellow: normal run

  • In reply to Roberto Giannessi3:

    I do not see anything wrong with the wave forms.

    Can you indicate the location of C1027.

    It may be that the device is overheating in some cases, even though the load is small.

    This might be if the DAP is not soldered down well on some of the builds.

    If you have a current probe you can lift up the inductor and put a small wire loop from the inductor

    to the pad and clip on the current probe.

  • In reply to Frank De Stasi:


    I've just received an board with the problem (burned chip),
    this case has burned chip with only 200mA load so I think it isn't a inductance problem.
    The DAP is soldered correctly and as picture shows the pin burned is the same yet; on this picture you can see the C1027 too.
    I haven't a current probe so I made a measure through a 0.5Ohm resistor,
    the aquired waveforms don't show any saturation, however I'll substitute the
    inductor with another one (code Wurth 744066470 1.85 A sat.).

    To remark the difference between this application and all others,which
    functioning correct:
    this application use the 'enable' so the Vin value is already present at startup,
    all others the Vin rises during all startup, is this a trouble ?
    The output drives some incremental encoders externally via wiring connection,
    may some disturbance on this wires cause a problem to the chip ?

    If the problems will be come again even inductor change, the only soluction
    will be to create a new power circuit with a large input voltage and load characteristics, in this case can you suggest me a correct device ?

    Thank you, regards.

  • In reply to Roberto Giannessi3:

    C1027 seems a little far away.  I am still worried that some kind of over voltage or spike is occurring on the input.

    If your load is inductive (such as long wires) then a bad connection between load and the device may cause the output

    voltage at the device to go negative.  This can cause a run-away condition in the regulator.

    I do not see any obvious issue that could be causing this trouble.

    You can return the damaged units to TI and we can try to do some FA.

  • In reply to Frank De Stasi:

    thank you a lot for your answers,

    right now, I haven't any broken devices, the one arrived is already repaired.
    I'm waiting for another one, what would you like to receive, the burned device or
    only the chip ?
    If you prefer I can provide you both the entire burned device and a device
    correctly functioning.
    Please let me know your choice and the shipping address.
  • In reply to Roberto Giannessi3:

    Let me pass this along to the responsible party.

  • In reply to Frank De Stasi:

    Please return just the damaged IC.  Please fill out the attached form with shipping address.


  • In reply to Frank De Stasi:

    I sended burned chip on monday

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.