• State TI Thinks Resolved
  • +1 person also asked this people also asked this
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 49 subscribers
  • Views 116 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

AMC3301-Q1: Stability issues with new design

Frank LaRosa
Prodigy 30 points
Part Number: AMC3301-Q1
Other Parts Discussed in Thread: AMC3301

Tool/software:

Hello,

In testing several new PCBs using this device, a fraction of the circuits exhibit instability, which is shown in the attached video. The trace is DCDC_OUT (high side) with respect to HGND. and it alternates between a stable 3.6V and the oscillation that appears to range form 200uS to 400uS, and has a P-P of ~1V (vertical scale is 2V/division). Whenever this oscillating waveform is observed, the outputs OUTP and OUTN flop aroiund randomly. This occurs even when INP = INN = HGND.

We already know that our design has two deviations from the recommendations in the data sheet:

1. We missed the 100nF capacitor on HLDO_IN (C1 in Figure 10-1 in the data sheet

2. Bypass capacitors are on the opposite side of the board and vias are used.

3. Bypass capacitors are not the recommended parts. One of them on DCDC_OUT is rated for 10V, and they are all 0402.

I have added a 100nF capacitor between pins 3 and 2 right at the device, and the instability continues, though I think it takes less time now to stabilize.

So how bad are these "violations?" Has this been observed in other designs, and are there any recommendations, other than a full re-spin with layout per the data sheet?

Regards, Frank LaRosa

  • 0
    TI__Mastermind 48961 points

    Hi Frank,

    Can you please share the schematic and layout? 

    Are you supplying any other circuitry from the internal DC/DC? 

    What is the low side supply voltage? 

  • 0 in reply to Alexander Smith
    Prodigy 30 points

    Hi Alexander,

    Thanks for responding! We're not powering any other circuitry from the internal DC/DC, nor the HLDO. The low side supply is +3.3V. Attaching PDF of the relevant page. Also side-by-side images of a previous version (on left) that works fine, and the current version that has this problem (on right). Let me know if you'd like the KiCad project. - Frank

    HV_Daughter2_Schematic_Extractor_Current_Sense.pdf

  • 0 in reply to Frank LaRosa
    TI__Mastermind 48961 points

    Hi Frank,

    The main issue I see is the layout. In many instances, the decoupling capacitors are on the other side from the pin they are meant to stabilize. While the connection is made correctly in terms of net, I do not think the capacitors provide much benefit to the switching loads inside the device. 

    CXB7/8 for example. The 3.3V and GND are routed directly to the pins, which is the load. Then the capacitors are on the other side of the board through vias which have an inductive component. The capacitors would be much more effective if placed before the 3.3V and GND traces connected to the device pins. 

    I recommend following the layout as shown in the device datasheet. 

  • 0 in reply to Alexander Smith
    Prodigy 30 points

    Hi Alexander,

    I wish I had the option of doing another spin with the proper layout and the recommended capacitors, but management won't allow it. The mystery here is that on a previous version with a very similar layout ("EDU" above, I have 4 boards containg 4 AMC3301s each, and all 16 of those devices are perfectly stable. The only differences I can see are 1) the source and shunt resistor are on a motherboard, and the AMC circuit is on a daughter board, with a few inches of traces, and 2) different capacitor part numbers were used on the current board, possibly with different ESR ratings. Let me know if you have any other thoughts on possible causes. I am doing some tests to understand this better. If I can't solve it, I'll have to remove the AMC3301 and re-work in a less optimal solution, which an associate is working on as a backup plan. Thanks, Frank

  • 0 in reply to Frank LaRosa
    TI__Mastermind 48961 points

    Hi Frank,

    I certainly think it is related to the capacitors, whether it be the ESR ratings or layout - it is hard to say. You can try probing LDO_OUT and DCDC_Out to see if voltages are stable at certain points compared to others. 

  • 0 in reply to Alexander Smith
    Prodigy 30 points

    Hi Alexander,

    I'll continue testing to see if I can get around this instability issue, and let you know if I get anywhere.

    Thanks,

    Frank

  • 0
    Prodigy 140 points

    Hello Frank,

    Looks like an RF interference from your mobile phone...

    Patches of localised "ground planes" under the high and low side pins, connected obviously to the high and low side "common" respectively, might help (might, not "will"), but I understand, new layout is not an option.

    Regards, Alex