Ideality Factor

Discussion in 'General Electronics Chat' started by DC_Kid, May 22, 2009.

  1. DC_Kid

    Thread Starter Distinguished Member

    Feb 25, 2008
    638
    9
    i'm working on a circuit that calls for a PNP bipolar that has ideality factor n=1.01

    i understand what this factor is, but i cant seem to find it in any spec. where do i find this value? or is it the norm that most modern day bi-polars have n real close to 1 ??
     
  2. Wendy

    Moderator

    Mar 24, 2008
    20,765
    2,535
    I've never heard of it myself, so I googled it. I suspect it is a specification that doesn't have much real world use.
     
  3. steveb

    Senior Member

    Jul 3, 2008
    2,433
    469
    I'm confused by this too. The only thing that comes to mind is the thermal voltage factor n*Vt. This number is used in transistor models to better match the transistor curve to the exponential shape assumed. This value can be anywhere in the range of 1 to 2, but is rarely equal to 1 in real devices as far as I know.

    Ic=Io*(exp(Vbe/(n*VT)-1)

    I think we need to know the application and the circuit to understand. Perhaps certain log-amps and multiplier circuits that depend on the exponential shape would benefit from transistors optimized for this application.

    Since the OP knows what the "n" factor is, please explain it and give as much information about the circuit and application so that we can have a chance to help.
     
  4. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
  5. DC_Kid

    Thread Starter Distinguished Member

    Feb 25, 2008
    638
    9
    i suspect in this application context it is a thermal related spec (although it really describes how well the device behaves compared to a perfect NP junction). the application calls for a PNP used as a temp sensing diode for remote temp sensing.

    check out the Maxim 6643 datasheet (page 8)
    http://datasheets.maxim-ic.com/en/ds/MAX6643-MAX6645.pdf

    they suggest a few. and they even suggest that if the device has current gains in the +50 to +150 range it suggests the manufacturer "has good process control and that the devices have consistent characteristics". wondering if this suggests good ideality factors??

    i did manage to find a Fairchild MMT3906. although its not on the list in the 6643 datasheet, the MMT3904 is suggested in the app notes below, so i would suspect the MMT3906 has about the same exact properties as the MMT3904, etc.

    i was just curious to know more about this spec factor "n" (aka "ideality" factor) as it is not a something i could find on any transistor spec sheet, etc. ... and which is why i was wondering if its the norm for most modern day transistors made by reputable manufacturers to have a n real close to 1?


    also more info here:
    http://www.maxim-ic.com/appnotes.cfm/an_pk/1944
    and
    http://www.maxim-ic.com/appnotes.cfm/an_pk/1057
     
    Last edited: May 22, 2009
  6. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
    Perhaps I was talking to myself. It's a diode thing. The factor modifies the PN junction current equation.
     
  7. steveb

    Senior Member

    Jul 3, 2008
    2,433
    469
    I'm wondering why you need a PNP transistor (2N3906 type) instead of the NPN (2N3904 type). It really shouldn't matter which type of tranistor is used if you are going to wire it up as a diode. Once you tie the base and collector together, you are all set. If they are recommending a particular manufacturer, just use that one for the best results.

    Typically, NPN devices are a little better than PNP, so stick with the NPN device. If you are doing something different than typical and really do need a PNP transistor, please explain the details.
     
  8. DC_Kid

    Thread Starter Distinguished Member

    Feb 25, 2008
    638
    9
    the datasheet recommends the PNP. i have asked Maxim if the Fairchild MMBT3906 is acceptable.


    and the datasheet does show a 6643 circuit using a NPN, so not really sure why they suggest (tested) only with the stated 3906's.
     
    Last edited: May 22, 2009
  9. steveb

    Senior Member

    Jul 3, 2008
    2,433
    469
    OK, that's fine too. It really shouldn't matter much which you use. The data sheet shows how to compensate for transistors that don't perfectly match the assumptions anyway.
     
  10. DC_Kid

    Thread Starter Distinguished Member

    Feb 25, 2008
    638
    9
    for Fairchild (and i suspect others) it looks like the MMBT vs the 2N of either NPN or PNP have about the same exact specs.

    however, no such spec of 'n' exists. i suspect this factor is hidden within other specs that are more important in most other applications (such as hfe, etc). 'n' seems to identify how well a manufacturer can make the pn junction.

    i'll also ask Fairchild if they manufacture with a minimum 'n' value and/or if they have 'n' spec for each of their products.

    as far as compensation, its hard to do. n values less than 1.01 will create a inherent negative offset in Tmeasured. forward diode resistence will create a inherent positive offset in Tmeasured. so we certainly dont want n greater than 1.01 because the offsets are additive and this would yield a uncorrectable inherent positive offset in Tmeasured. with n lower than 1.01 and very low forward diode resistence its possible to make offset=0 by adding some resistence to the diode.

    however, in my application i dont think it matters much.

    at minimum, this 'n' is a interesting factor item...
     
Loading...