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.

Output sink current of LM339-N

Other Parts Discussed in Thread: LM339-N, LM339, LMC7215, TLV3691, LPV7215, TLV3701, TLV3401, LP339, TLC339, TLC3704, REF3325, REF1112

The Output sink current of LM339-N is specified as 6mA min.

Since this part has an open collector output, does it mean that with the Vcc value being used, the pull up resistor must be sized to allow at least 6mA of current into the collector of the output stage transistor when the output is low, for the part to function properly?

For instance, if Vcc = 5V, then should the pull up resistor be at most 5V/6mA = 833 ohm?  If not, how should I be interpreting that spec.?

Regards,

Nitish Agrawal

  • Hello Nitish,

    Yes. That would be the lowest value pull-up resistor you should use. But it is very rare that a value that low would be used. You do not want the pull-up resistor to load the output too much.

    Sinking means current going INTO the output pin. So it is the maximum current going to ground when the transistor "switch" is closed (6mA). That "resistor" can also be a load like a LED, relay or light bulb.

    The size of the pull-up resistor depends on the load. The value needs to be high enough to pull the output load up quickly, but not to overload the output when it is low (it also wastes system power - and that adds up as you add outputs).

    You can use a 1Megohm pull-up if you wish - but the tradeoff is rise times since it is purely the R-C time constant of the resistor and the load capacitance when going high.

    The typical pull-up resistor is 4.7k to 100k when driving other CMOS or TTL gates. Lower resistor values (<10k) are used when higher speeds are needed (less influence from output capacitance). The LM339 is specified with a 5.1k pull-up.

    So you do not *have* to draw 6ma for it to function. The minimum pull-up current needs to be low enough to overcome the output leakage current (100pA), but no more than 6mA. 10k is a good value, or 100k if you want to save power and speed is not critical.

    Regards,

  • Hi Paul,

    Thanks for your response and explaining how to choose the pull up resistor value.
    I have a lingering question:
    The max value of output sink current per datasheet is specified as 16mA (min value 6mA)....that is what was making me think that I ought to have at least 6mA into the comparator's output transistor's collector when transistor is ON.
    Why wouldn't they say that the max value of sink current is 6mA?

    Regards,
    Nitish
  • Hello Nitish,

    Don't confuse Min and Max and Typ..

    The 16mA is a "typical" value, as it is in the 'TYP" column. Typicals are not guaranteed.

    Because the 6mA is in the "MIN" column, it means 6mA is the MINIMUM amount of current we guarantee the LM339 can sink under those conditions.

    Typically it can sink more (16mA), but we only guarantee 6mA. So there is a 10mA "cushion" between the typical and minimum. When designing, you should not expect to get more than 6mA across all devices. You can sink more current, but it is at your own risk because if the device fails, the warantee is void since it exceeds the specified 6mA.

    If the 6ma was in the "Max" column, then that would mean we only guarantee that it would only sink up to 6mA, and that any currents above 6mA means that the device is defective ...which is obviously backwards from above and not practical.

    Regards,

  • Hi Paul,

    Thanks for pointing out that 16mA is the typ. value. What you say makes sense.

    My application is really power critical and output rise time is not of concern. I was thinking of using a 47Meg ohm pull up resistor with a 4V supply. That would be high enough to overcome the 100pA of leakage current. What do you think?

    I also note that supply current is 2.5mA (max). That seems a lot for my application.
    Does TI have a low cost nano power comparator with Vcc ~4V? Also, looking for a 2.5V nano power voltage reference.

    I would greatly appreciate your comments and recommendations.

    Regards,
    Nitish
  • Hello Nitish,

    47Meg is way too large. 10Meg would be as high as I would go. But the supply current will still dominate.

    We have many micro and nanopower comparators.

    The lowest supply current comparator is the TLV3691 at 150nA, and many in the TLV3401 /2 /4, TLV3701 /2 /4, LPV7215, and LMC7215 / 25 series.

    Note that these micro and nanopower comparators have "push-pull" outputs and do not need a pull-up resistor, since a a pull-up resistor wastes power at micropower levels.

    You can view the low power comparators here:

    www.ti.com/.../comparators-products.page

    Pay close attention to the single and quad pinouts, as they can differ. The lowest power quad comparator (<20uA) that would drop in place of the LM339 (same pinout and open collector output) would be the TLC339, TLC3704 and LP339.

    The lowest power reference we have is the REF1112, which can be run as low as 900nA. You will need to place two in series to get 2.5V. Or the buffered REF3325 at 3uA.

    Regards,
  • Hi Paul,
    Thanks for the max pull up resistor value and the part recommendations. I will look them up.
    Much appreciated!
    Regards,
    Nitish
  • Hi Paul,

    I am looking at the part TLV3401 (nanopower open drain output comparator).
    What is the max value of pull up resistor we may use at the output? There is no pull up resistor recommendation in the datasheet, except for a figure on the first page of the datasheet where I notice a 1Mohm resistor pull up.

    I have a question about the convention of representing inverting and non-inverting inputs (I want to make sure I am interpreting the different representations correctly).

    The TLV3691 datasheet shows the comparator inverting input with a (-) and the non-inverting input with (+) symbol or sign. I am used to seeing this.
    The TLV3401 and TLV3701 datasheets show a bubble on one of the inputs and no sign on the other. I am assuming that this is just another way of representing the inverting and non-inverting inputs (bubble represents inverting input while the other is non-inverting input....is there anything more to it?).

    Your clarifications would be greatly appreciated.

    Kind regards,
    Nitish Agrawal
  • Hi Nitish,

    The TLV3401 is a CMOS device, so the leakages are very low. Again, you can use any large value resistor you wish, but you still have to deal with lower leakages and the time constant to charge the output capacitance.

    Because this device is a micro-power device, targeted at low power applications, the 1M resistor is used to minimize power dissipation.

    Remember, when the output is "low", the resistor is dissipating power. So with a 5V supply, when the output is "low", the circuit is dissipating 5V/1Meg = 5uA - just for the pull-up! This is why micro power circuits avoid using pull-up resistors and use "push-pull" outputs.

    The TLV3701 is the Push-Pull (or "totem pole" output) version of the TLV3401. The output can go high (source) or low (sink) by itself - so it does not need a pull-up resistor. Same pinout, same basic specs, but you do not need a pull-up resistor.

    Yes. The "bubble" means inversion, or negative input. So the "bubble" is equivalent to the "-" input. This is a standard logic gate symbol.

    Regards,
  • Hi Paul,

    Thanks again, for the timely response. I will avoid using the open drain output comparator to reduce power dissipation.

    Thanks also for clarifying the bubble convention was also used in opamps.

    Regards,
    Nitish