• Not Answered

LMR16006 shutdown hysteresis not working

Hi,

I'm using a LMR16006 simple switcher and I'm using the !Shutdown input to get an undervoltage lockout.

Im using a voltage divider with 330k and 22k. This results in a Vstart = 19.7V and Vstop = 18.7V. Vhysteresis is about 1V.

When I tested the circuit I found that the switcher doesn't stay shutdown. It comes back up only to shutdown again after a couple of milliseconds.

See attached scope (Ch1 = open; Ch2 = LMR16006 output voltage; Ch3 = Input voltage; Ch4 = voltage at !shutdown pin of simple switcher)

First I thought that this is probably due to the noise on the shutdown pin. But at t = 0 (where it seems that the shutdown is triggered) the voltage at the shutdown pin should drop by about 66mV (3µA * 22k || 330k) but it doesn't. Which makes me think that the shutdown logic may not be responsible for the first dip in the output voltage. But what is?

Thanks in advance!

Regards,

Konstantin

8 Replies

  • Some additional information:

    So when the voltage at the shutdown pin is below the Vshdn threshold the Ishdn_hys current source is enabled which results in a change of the voltage on the shdn pin. With the voltage divider that i have i should see a change of about 66mV. But I don't - not at t = 0.

  • We have forwarded your query to the right engineer. We will get back to you shortly.
  • Hi Konstantin,

    1. Do you add a capacitor parallel with the 22k resistor? If yes, remove the capacitor and re-test.
    2. I find the the Vin drop slew rate is quite slow. If the Vin is around the system UVLO threshold, and its slew rate is flat, then noise will enable/disable the IC by mistaken. Have you tried fast Vin slew rate? Does this issue also happen?

    BR
    Luke

    BR

    Luke

  • In reply to Luke Pan:

    Hi Luke,

    I tried it with and without the capacitor in parallel to the 22k resistor. The capacitor reduces the noise a little but the problem stays the same.

    The slew rate of the Vin voltage is that slow because I'm reducing it very slowly on purpose. My circuit will be supplied by a supercap. So the voltage will actually be decreasing very slowly. I think that the slow slew rate shouldn't be a problem. Isn't that what the hysteresis is there for?


    With faster slew rate this issue doesn't happen.

    Best Regards,

    Konstantin

  • In reply to Konstantin W:

    Hi,

    I looked at the scope of my initial post again today and i still think that there is something wrong with the behaviour of the LMR16006.

    The first time the DCDC output voltage (green) drops the voltage at the !SHDN input doesn't seem to change much. But it should drop by about 66mV!

    To me it looks like the I_SHDN_HYS current source doesn't get enabled. It only gets enabled about 10milliseconds later when the DC/DC properly shuts down and stays that way.

    I think i found a solution though:

    I increased the hystereses voltage. And because i did't want to incease the total resistance of the voltage divider I added an additional 1MOhm resistor connecting the 12V output voltage of the DC/DC to the !SHDN pin. By doing that I get about 5V of hysteresis

    With that resistor the DC/DC disables itself and stays disabled even with a very slowly decreasing input voltage.

    CH2: 12V DC/DC output voltage; Ch3: Input voltage; Ch4: Voltage at !SHDN pin of LMR16006

    Best Regards,

    Konstantin

  • In reply to Konstantin W:

    Hi Konstantin,

    Will the battery charge the output cap through the 1Mohm resistor even if the IC is disabled?

    BR

    Luke

    BR

    Luke

  • In reply to Luke Pan:

    Hi Luke,

    well, its not a battery but a super capacitor. But yes, a small current is flowing through the 1Mohm resistor to the output cap (and connected circuitry) even if the IC is disabled. In my application I see about 5mV at the output capacitor.

    Regards,

    Konstantin

  • In reply to Konstantin W:

    Konstantin W

    To me it looks like the I_SHDN_HYS current source doesn't get enabled. It only gets enabled about 10 milliseconds later when the DC/DC properly shuts down and stays that way.

    I have similar problem here. I want to use the I_SHDN_HYS current source for bi-stable power hold. With about 520 kOhm to GND it holds either on or off. But I have to pull down more than 300ms to switch the hysteresis from 4µA to 1µA.

    Is there any more documentation how the hysteresis is timed or calculated?