• State Resolved
  • Locked Locked
  • Replies 13 replies
  • Answers 1 answer
  • Subscribers 33 subscribers
  • Views 4102 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.

TMP421, remote thermal diode(BJT) beta value?

vijay koganti
Prodigy 40 points
Other Parts Discussed in Thread: TMP421, TMP441, DXP, LM95235

Hi,

I have a question, regarding the recommended value of beta for pnp substrate diode (BJT) used.

The recommended value in the data sheet is as below

"Tight control of VBE characteristics indicated by noisy environment. Layout should adhere to the small variations in hFE (that is, 50 to 150)."

Does the value of beta require to be in this range, as many of the substrate PNP diodes(BJT) have beta of even less than 1.?

please respond ASAP.

thanks and regards

Vijay

  • 0
    TI__Mastermind 23140 points

    Vijay,

    I am currently looking into this and will get back to you shortly. 

  • 0
    TI__Mastermind 23140 points

    Vijay,

    Due to the dependence of Vbe on beta we recommend that this criteria is applied regardless of the type of discrete transistor used. The datasheet is referring to the 2N3906 (PNP) which has an hfe (beta) between 30 and 100 depending on the test condition as shown below.

    As for other types of transistors including the lateral pnp and substrate pnp, the forward current gain (beta) can range from 19 to around 75 as shown below:

  • 0 in reply to Chris Featherstone
    Prodigy 40 points

    Hi Chris,

    Thanks for the reply, but i'm unable to see the snapshots you have posted,

    can you please point out where can i find these?

    And more over, according to the equation  [8] addendum , as pointed in the article http://eetimes.com/design/power-management-design/4011557/How-to-measure-temperature-in-integrated-systems,

    the temperature sensing would be dependent on the variation of beta over the temperature range, but not on the absolute value of beta.

    can you please explain me, in more detail on how the TMP421 is sensing the temperature and the dependency on the value of beta?

    thankyou verymuch for your reply

    regards

    Vijay

  • 0 in reply to vijay koganti
    TI__Mastermind 23140 points

    Vijay,

    Try using another browser. I checked and had the same problem using Google Chrome. If I use Firefox I can see the images. I will get back to you shortly on your other questions.

  • 0 in reply to Chris Featherstone
    Prodigy 40 points

    Hi Chris,

    Any update on the question I had?

    regards

    Vijay

  • 0 in reply to vijay koganti
    TI__Mastermind 23140 points

    Vijay,

    On digital processes 40nm and below the PNP transistors can have a beta of less than 1. Ti designed the TMP441 which is a footprint and pin for pin compatible device to the TMP421. The only difference is that we have added a Beta Compensation register that provides the option of automatic or manual beta compensation as seen below.

    To address your main question the answer is yes. With the TMP421 the preferred beta will be 50 or above as given by the recommended discrete transistors and or analog processes. With a digital process and a beta of any value the TMP441 will provide beta compensation to accommodate the user.

  • 0 in reply to Chris Featherstone
    Prodigy 50 points

    Hi Chris,

    This is Murali and I work with Vijay.

    Could you please provide your answers on below questions.

    1. Is there any possibility (programability) to have intermediate points between 1.042759 and 1.747977 (just not listed in the n-factor range table of TMP441), like 1.06XXX

    2. Does auto beta correction is applicable to diode connected (base and collector shorted) PNP transistor?

    Best Regards,

    murali

  • 0 in reply to Murali Krishna1
    Prodigy 50 points

    HI Chris,

    We are looking for your reply . Could you please answer the query.

    Best Regards,

    Murali

  • 0 in reply to Murali Krishna1
    TI__Mastermind 23140 points

    Murali,

    Question:

    If an n-Factor of let’s say 1.05 is needed how do you determine the corresponding binary value to program the register?


    Answer:

    Using equation 3 above plug in nEFF = 1.05

    NADJUST = 300 – 300*1.008/1.05 = 12

    12 = 0xC = 1100

    Programming the register to 0000 1100 will provide the desired result of nEFF = 1.05.


    I will look into your second question and get back to you shortly.

  • 0 in reply to Murali Krishna1
    TI__Mastermind 23140 points

    Murali,

    Yes auto beta correction is applicable to a diode connected sensor. 

    On page 14 of the datasheet you will find the following:

    "If the Beta compensation Configuration Register is set to '1xxx' (beta compensation enabled) for a given channel at the beginning of each temperature conversion the TMP441/42 automatically detects if the sensor is diode-connected or GND collector connected, selects the proper beta range, and measures the sensor temperature appropriately."

    You will see in Table 8 that setting the value to '1xxx' in the Beta Compensation Register corresponds to the automatic detection while setting the value to '0xxx' not including 0111 corresponds to manual selection. 

  • 0 in reply to Chris Featherstone
    Prodigy 40 points

     

    HI Chris,

    Could you please provide me the information on TMP441 leakage current compensation?

    ESD diodes in the path of Anode and Cathode of Thermal Diode will have the leakage (through these diodes) especially at high temperatures and impact the accuracy of measurement.

    Does TMP441 has compensation method of leakage current for accurate measurement?

    Best Regards,

    Murali

  • 0 in reply to vijay koganti
    TI__Mastermind 23140 points

    Murali,

    There is no compensation other than having very low leakage ESD protection on the device. There isn't a way for the user to compensate in the device for leakage currents. Your main concern is with accuracy and we do guarantee the accuracy over temperature as shown below which accounts for any leakage within the device itself.

    Ti doesn't have control over the leakage currents due to the remote sensor itself or from flux leakage between DXP and DXN.

  • 0 in reply to vijay koganti
    TI__Expert 4905 points

    Parts that support TruTherm do actually compensate for some of the leakage current. You could try one of them like the LMP95235 and see what you get. This part supports Beta Correction but does not support ideality calibration.

    Also look for a sensor that biases the negative side of the thermal diode lower. The LM95235 biases it around 400mV. The higher the voltage the worse the affect of leackage as you are biasing the ESD closer to on. This is especially true as CPU voltages drop. Don't go for ground biasing though - I have another posting that describes the issues with that (basically noise coupling from the CPU/ASIC ground).

    What ever you decide I highly recommend that you also check out the performance on the bench with the target diode in a well controlled temperature environment. I have found that many folks have gotten the ideality wrong when calculated.

     

    Take care,