MOSFET datasheet problem with K value

Discussion in 'General Electronics Chat' started by johnnymoha, Apr 2, 2013.

  1. johnnymoha

    Thread Starter New Member

    Apr 2, 2013
    4
    0
    Hello all. Long time reader, first time poster.
    I've used MOSFETs as switches for applications before, but I'm curious as to why the data sheets don't contain a K value for the device. I hate when folks reference "back in school", but we always had the K values of a device when we were trying to work problems or solve the the length and wide etc...
    My question is: Why do MOSFET data sheets only show some generic curves for different Vsd, Vsg values? If I'm using a pmos as a high side switch and want to make sure that I can draw the amount of current needed via Id, how can I be sure that the mosfet is going to be in the currect region and able to do so without being able to calculate Id because K is unknown. I have to be missing something and just wrapped around the axel since I haven't done this in a while. Thanks in advance for the help.
     
  2. blah2222

    Well-Known Member

    May 3, 2010
    554
    33
    What you will find is that mosfet and especially jfet manufacturing companies produce devices with a large spec range for their parameters. This basically allows them to sell more product with more relaxed testing limits. The more expensive the part (primarily military grade devices) usually the more testing it has been through and tighter spec variance.

    In school they refer to these device parameters such as 'K' as constants when in fact they are not. Normally what people do is use the max/min limits (datasheet) with some buffer room to guarantee correct operation of the device.

    It would be helpful if you could provide the name of the device you are talking about.
     
  3. #12

    Expert

    Nov 30, 2010
    16,261
    6,770
    Complete agreement. One of the first rules of design for mass production is to arrange circuits so that any device within the limits (and then some) will work.

    Some of the nastiest problems I've had to work on were TV's with "tight spec" parts that you could only get from the OEM.
     
  4. crutschow

    Expert

    Mar 14, 2008
    13,000
    3,229
    Welcome to the non-academic world. Most transistor courses are based upon analysis of the inner device characteristics that typically are only used when actually designing the transistor or using them in IC design, which involves determining their W/L among other things. Discrete devices never use those parameters since they aren't really that useful is designing the devices into a circuit. Thus they give externally derived (black box) type parameters such as transconductance, threshold voltage, ON resistance, voltage and current ratings, etc. You just have to adjust you thinking to using those parameters, which are generally sufficient to determine how to reliably use the device in a specific circuit.

    For example, when using a MOSFET as a switch, you don't worry about the gain or threshold voltage. To insure it's on you just apply the minimum voltage necessary to insure that the device is fully on for any current up to its maximum. This is typically stated in the spec sheet as 10V Vgs for standard MOSFETs and 3V or 5V for logic level type MOSFETs. No K value needed. ;)
     
  5. Jony130

    AAC Fanatic!

    Feb 17, 2009
    3,957
    1,097
  6. johnnymoha

    Thread Starter New Member

    Apr 2, 2013
    4
    0
    Thanks for the help folks. The part is http://www.digikey.com/product-detail/en/NDP6020P/NDP6020P-ND/1055922. Logic level PMOS from Fairchild Semi.
    I understand that in the academic world we had more information than we get in the real world. The new nomenclature is something I don't understand apparently. Fully on? Is this in saturation, triode or both? If it's saturation region, then the current is supposed to fluctuate mostly with Vsg right?

    For practical design application(if you had a chance to look at the datasheet): How can I determine the maximum current I can draw through the FET at the listed Vsg values? The channel will fluctuate base on the gate voltage and I need to know how the FET will react to the changes.
     
  7. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    Figures 1 and 2 should be enough. What current and voltage do you plan to switch?
    Will it be PWM or plain slow switching?
     
  8. Jony130

    AAC Fanatic!

    Feb 17, 2009
    3,957
    1,097
    Yes figure 1 and 2 and when MOSFET if full ON the MOSFET working in a triode region.
     
  9. johnnymoha

    Thread Starter New Member

    Apr 2, 2013
    4
    0
    Thanks folks. I'm just using as a plain old switch, no PWM. One last question though. When it comes to practical applications for things other than switches, without having the equations that they generate the curves with, how can you design amplifiers or other things where you need to implement a specific transfer function? Without K or un Cox all that stuff.
     
  10. #12

    Expert

    Nov 30, 2010
    16,261
    6,770
    For a single transistor, the surrounding circuitry sets the gain, and it is always a lot less than the alleged parameter on the datasheet. If your transistor can be expected to have an I/I gain of 150 to 250, you almost never go near designing for a finished product with a gain of 150.

    Negative feedback is a good solution, too, and can be implemented on as few as one (1) transistor. When using multiple transistors, the uncertainty margin is very high and negative feedback is basically mandatory. A perfect example is a transistor driven by an operational amplifier.
     
  11. johnnymoha

    Thread Starter New Member

    Apr 2, 2013
    4
    0
    I see. Thank you. That's actually what I'm doing ATM. I've been working on a simple battery low voltage protection circuit where a comparator is driving the gate voltage of the Pmos. These questions of mine arose from needing to know what gate voltage I should be driving from the comparator's output to make sure that I can still supply all the current I need.
     
  12. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    The short answer is: as high as the mosfet allows. Yes there could be problems if say the circuit was running at 3.3V, but still you can never go wrong by driving the transistor to the highest allowable voltage as it will lower the Rdson and thus the dissipation. For example your transistor states Vgsmax ±8V and the curves go up to 5V, so I would simply drive it with 5V and be done with it. If you need to go lower with the voltage because some other reasons, then you need to look at it more carefully and find out if it still can hold the current you want.
     
  13. crutschow

    Expert

    Mar 14, 2008
    13,000
    3,229
    Fully ON is in the linear region where it has the minimum ON resistance. For proper switching you always want to apply as much gate voltage as practical (within the Absolute Maximum Limit). You don't want to operate the MOSFET as a switch with less than the minimum voltage as used to measure the ON resistance stated in the data sheet.

    The maximum drain current is listed in the data sheet in the "Absolute Maximum Ratings" table.

    For operation in the saturation region as an AC amplifier, the gain is determined by the Forward Transconductance value given in the "Electrical Characteristics" table.
     
Loading...