• State TI Thinks Resolved
  • Locked Locked
  • Replies 12 replies
  • Answers 4 answers
  • Subscribers 62 subscribers
  • Views 592 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.

Original question:

LM3478: Unstable PWM-signal in boost design

LM3478: Unstable PWM-signal in boost design, part 2

user6072412
Prodigy 125 points
Part Number: LM3478

Hi!

Vacation is over and I have now done some measurements related to the original question (which was locked).

Vin=3.5V

Ch1, yellow, 5V/div: Gate of M1, used for triggering (during this measure, a 22Ohm resistor was connected between DR and the gate)

Ch2, green, 50mV/div: Isense pin

Ch3, red, 2V/div: Drain of M1

Ch4, blue, 50mV/div and 1.35V offset: Comp pin

Oscilloscope BW = 500MHz

Zoom on short pulse:

There are glitches on Isense and Comp (partly due to bad probing I guess...), but as far as I can see, they come after DR switches off.

The same glitches can be seen on the correct pulse, with same relation to DR goes low:

As I mentioned in the original question, for this test I have used a separate supply for the LM3478 Vin-pin only. For the scope shot above, both supplies were 3.5V. If I change the LM3478 supply to 3.6V (keeping 3.5V to L1), there will be no short pulses.

Any idea why I get this short pulses?

BR

/P-O

  • 0
    TI__Guru 68495 points

    Hi user,

    Thank you for posting.  It sounds like a noise issue.  Can you add an 1 uF local capacitor across VIN pin and GND (placed very close to the IC), and add a 10 Ohm between the input line and the capacitor?  

    Thanks,

    Youhao Xi, Applications Engineering

  • 0 in reply to Youhao Xi
    TI__Guru 68495 points

    Can I assume the issue is resolved?  Let me close this thread here, and you can always re-open it by adding a new post.

    Thanks,

    Youhao

  • 0 in reply to Youhao Xi
    Prodigy 125 points

    Hi Youhao,

    Thanks for your respons! Sorry for the delay from my side, but it seems to be "magic".

    I agree with you that it could be noise related.

    Even if I've tried different options to filter the Vin (see part 1), I decided to try your suggested 10Ohm/1uF. It seems to work, great! But after removing it, I can't see the problem any more...

    Testing another board -> it has the problem. Looking for any soldering issue -> nothing found.

    By accident I found that if I have my Weller solder iron close to the package between pin 4 & 5, I'l get a lot of short pulses. Is the LM3478 known for being sensitive to interference?

    On the scope I can see some 100Hz spikes from the iron, but I'm surprised they cause problems.

    Anyhow, I've ordered a eval board for the LM3478 to eliminate if my design has any layout issues.

    BR,

    P-O

  • 0 in reply to user6072412
    TI__Guru 68495 points

    Apparently your solder iron radiates which may be picked up by the circuit and affect the operation. Usually we don't hold the solder iron to be close the circuit if the circuit is operating.  A drop of solder may cause short circuit and irreversible damage to the IC.  It is sounds like you injected the noise into the ground, which should be clean at the IC side to prevent malfunction. 

  • 0 in reply to Youhao Xi
    Prodigy 125 points

    Hi!

    An update:

    I have modify the LM3478 EVAL-board to my spec and I see a stable PWM-signal.

    If I replace the LM3478 on the EVAL board with one from our product, I have the short pulses in the PWM-signal!

    Some more test gives me this:

    Trace lot code
    8882 unstable
    8882 (2:nd chip) unstable
    7CL4 stable
    95DY stable
    53RT stable

    Putting a LM3478 from another lot on our product makes it stable.

    So something in that lot (and maybe in combination with our design) makes it unstable.

    BR P-O

  • 0 in reply to user6072412
    TI__Guru 68495 points

    Hi P-O,

    Could you test more ICs, fresh ICs from the lot code 8882?  

    Thanks,

    Youhao

  • 0 in reply to Youhao Xi
    Prodigy 125 points

    Hi Youhao,

    Yes, I hope we can do that. We had the same idea, so I have already asked our manufacturing partner to see if they some ICs from that lot in the excess material.

    BR

    P-O

  • 0 in reply to user6072412
    Prodigy 125 points

    Hi Youhao,

    I have got two fresh ICs from the 8882 batch and tested these. The result is that one has the same unstable behavior and one is stable.

    What do you recommend to do next?

    BR,

    P-O

  • 0 in reply to user6072412
    TI__Guru 68495 points

    How many piece do you have that are belonging to the 8882 batch?

  • 0 in reply to Youhao Xi
    TI__Guru 68495 points

    Hope the issue has been resolved, given no further update from your side.  You are welcome to re-open this thread by adding a new post.

    Thanks,

    Youhao

  • 0 in reply to Youhao Xi
    Prodigy 125 points

    Hi Youhao,

    I have not got the info yet regarding how many we have left of the 8882 batch, but I think there is small amount (~10 pcs).

    The issue is not resolved. We think that is a rather strong correlation between the batch and the issue, but how to ensure we will not face the issue with another batch? I can give you some more details, is it possible to do it offline?

    BR

    /P-O

  • 0 in reply to user6072412
    TI__Guru 68495 points

    I just send you a connect request through the PMS. Lets close the thread here and continue the discussion through that channel.

    Thanks,

    Youhao