• Resolved

SN74LVC1G123: MODEL ISSUE IN LTspice

Prodigy 190 points

Replies: 7

Views: 77

Part Number: SN74LVC1G123

Hi,

I am using LTspice.  I made a symbol using the "sn74lvc1g123.inc" file and tried testing the model.  Simulation fails because of an error.  Please refer  to the attached document for details.

Chundrasn74lvc1g123 - Model Issue.docx

  • The HSPICE model is encrypted and works only with HSPICE.

  • In reply to Clemens Ladisch:

    Hi Clemens,

    No possibility of getting a Pspice model I guess.  Any other device you might suggest having a Pspice model that would approach the behaviour of the SN74LVC1G123?

    Chundra

  • In reply to Chundra Ramful1:

    Hello Chundra,

    While it may not be identical, the SN74LV123A device has a pspice model that you could use.

    https://www.ti.com/product/SN74LV123A 

    Michael

    Have a question about Logic or Level Translation?
    Find answers through our Logic Frequently Asked Questions

  • In reply to Michael J Schultis:

    Hi Michael,

    Thanks, I created the SN74LV123A component in LTspice and I am getting the expected results except for one detail...  The component has two parameters, Rext and Cext, with default values.  I am not sure how these values are interacting with the actual Rext and Cext component values I am using in my schematic.

    What is the prescribed method in this case to make sure that the output pulse width matches with the component values I am using?  I tried deleting the default parameters but then everything fails.

    Chundra

  • In reply to Chundra Ramful1:

    Hello Chundra,

    I am not sure exactly what values you are referring to, however, I believe these are internal values that are set based on the device parameters in order to try to closely replicate the device behavior based on the R and C that you connect externally.

    This is because internal parasitics and components affect the timing, and the output pulse is not only dependent upon the external values of R and C used, but also other device factors, which is why a 'K' value for a given device is provided.

    See the following FAW for additional information.

    https://e2e.ti.com/support/logic/f/151/t/864007 

    I believe you have to leave those internal components defined in order to have the model operate properly.

    Michael

    Have a question about Logic or Level Translation?
    Find answers through our Logic Frequently Asked Questions

  • In reply to Michael J Schultis:

    Hi Michael,

    I have reproduced part of the "sn74lv123a.lib" in the attached document.  It is only now that I noticed note 4, so I do not think it is related to the parasitics you mentioned above.  If I understand correctly, the model will work properly only if the Rext & Cext values used in the schematic are also entered explicitly as parameters within the model.

    Also, note 1 of the "sn74lv123a.lib" file states that the model is only for a 5V supply.  Its reference to "figure 6" in the latest datasheet seems to be outdated.  I simulated the model with a supply of 2V and then a supply of 5V.  The attached document shows that the difference in pulse width is hardly noticeable.  The ~87us is well below the expected ~96us (5V supply) or ~98us (2V supply).

     Any comments?

    Chundrasn74lv123a - Test.docx

  • In reply to Chundra Ramful1:

    Hello Chundra,

    I would say that 87us vs 96us is not too far off to be honest from a model. 

    A couple of points here:

    1. These components are not intended to be used in highly precise applications, as variation is expected not only from the device, but also from external components across process and temperature. If you are looking for a very precise output pulse, then it is better to use something that is designed to be highly precise. 

    2. While we do have a model available for this part, it is very difficult to model these precisely as there are multiple dependencies, so for it to be exactly accurate to every data point given in the datasheet is exceedingly challenging. Please refer to the following FAQ for more information: https://e2e.ti.com/support/logic/f/151/t/787588

    For these components it is generally much better to test in a system or on an actual board to "tune" the circuit to your needs as opposed to depending on a model to get it exactly accurate.

    I would rely more on the datasheet output than the model output when considering a starting point for design.

    Michael

    Have a question about Logic or Level Translation?
    Find answers through our Logic Frequently Asked Questions