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.

TLV3691: START-UP GLITCH IN OUTPUT

Part Number: TLV3691
Other Parts Discussed in Thread: TLV7031

Hello,

I am looking into the potential use of the TLV3691 in an autonomous power supply design where there is only one voltage source.  I have simulated the comparator in LTspice as shown below.  I have not reproduced the power generation side because of IP issues.  Instead I have simply used a voltage supply to simulate one power source.  The smaller capacitor C7 is supposed to store energy to power the comparator and the larger capacitor is supposed to store energy for the load.  The TLV3691 is supposed to turn ON a MOSFET once the comparator is operational at 0.9V.

Why do we see the glitch at the comparator output between 0.5Ks and 1.0Ks?  Should the pull-down resistor R5 not ensure that node "Vgate" is held LOW until switching occurs?  I have read one thread in the forum closely related to what I am talking about here.  The prescribed pull-down does not seem to work in a reliable way when the comparator supply is less than 0.9V.  Any comments?

Chundra

  • Chundra

    I will simulate in TINA and see what I get.  I will get back to you by tomorrow.

    Chuck

  • Chundra

    I confirmed the same thing in TiNA.  Apparently there is something in the model that is causing the output to have this strange behavior.  From experience on the bench, this result is not expected and the output should remain low.  Are you able to obtain some samples to prototype the circuit?  Circuit startup is something worth investigating on the bench if you have ability to build on a breadboard.

    Chuck

  • Hi Chuck,

    I am still at the simulation stage.  Because of Covid-19, I have been working from home since mid-March.  Access to the lab is reserved for researchers who absolutely have to be present in their offices/labs.  I could have prototyped it at home but the IC size is unfortunately too small and I do not have the required tools.

    The presence of the glitch is a major issue because the autonomous power supply can be stuck at very low voltages for quite a while.  It will drain away the miniscule amount of generated power which would otherwise slowly accumulate in a capacitor.

    Chundra

  • I presume you are referring to power being wasted in the load resistor R5?

    The voltage of the spike should not be high enough to turn on the gate, so no power should be wasted there.

    My experience testing the POR is that a spike is only experienced if there is something to pull the output up externally.

    In your case, there isn't anything, so I don't that will be a problem.

  • Yes, I was referring to power being wasted in R5.

    If you are correct, do you think TI will review its spice model to have the glitch removed?  The TLV3691 is an amazing device and an improved model would be good I feel.

    Chundra

  • I’ll see what can be done with the model. I estimate it will take a couple days but I’ll report back.

    Chuck

  • Chundra

    I noticed that we haven't been able to repair the model yet.  I will check again on Monday to see what the status is. 

    One thought is to try the TLV7031 model which has the same architecture for the power on reset circuit.  I don't believe that model has this problem.  Our assumption is that it has something to do with how we are modeling the output stage.

    Chuck

  • Hi Chuck,

    Thank you for the update and suggestion.  I will try it and let you know how it works out.

    Chundra

  • Thanks Chundra

    If we find out that this model performs better in this test condition, we can modify the specs to match the performance of the TLV3691.

    Chuck

  • Hi Chundra,

    For the sake of time, I made the simulation slightly faster. Please let me know if this model works with your setup. From what I can tell, my simulation shows that this PSpice model does not have the glitch you were experiencing. I hope this resolved your issue.

    tlv3691.zip

    Thanks,

    Joe

  • Hi Joe,

    Sorry for the delay in responding... got caught up with other things.  I created a new symbol using the .zip file you attached previously.  I ran the same simulation as I had when I first came across the issue with the model (new symbol created).  Unfortunately the problem persists.  I really wanted the TLV3691 model fixed but... My next step is to try what Chuck had suggested earlier, the TLV7031.

    I am attaching the simulation circuit in LTspice as well as the resulting waveforms.  You will also see the steps in creating the new symbol and its location on my laptop..CR - 21Nov2020.zip

    Chundra

  • Hi Chuck,

    I tried the TLV7031 and found a similar issue, i.e. There is a power up glitch when the supply voltage of the chip is below the specified minimum.  I would have thought that the 1meg pull down would be enough to hold the output close to 0V.   Please take a look at the attached waveforms.

  • Hi Chundra,

    Thank you for your detailed document. Using the model that I sent you, I am not seeing the same behavior in the output (green). The output remains low with all of your component values. Instead of having a 500 second delay I shortened it to 1 second and the simulation window is 750 seconds as opposed to 5Ks.

    I think the issue lies in the conversion from PSpice for TI to LTSpice or your simulation settings. I used a maximum step size of 7.5s for this simulation. I recommend requesting access to PSpice for TI or TINA for TI. Both programs don’t show the glitch you are experiencing.

    Regards,

    Joe