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.

selecting comparator

Other Parts Discussed in Thread: LM393, LM193, LM393-N

I am new to comparators and i am having trouble finding the one I need. The power rails are +-15vdc, prefer two comparator per package. Surface mount, something like an SOIC8. The output needs to be CMOS rail to rail, so the output can go close to the positive or negative rail depending on input. I may be able to do this with an open drain or collector  but I am not sure. The input will be comparing two DC voltages to check for polarity. Input voltage offset may be an issue, as the DC voltage is in milivolts.  Anyone know a comparator that can do this?

  • Hi David,

    There are no 30V CMOS comparators. That is why you are having trouble!

    Do you need low bias current on the inputs?

    There are plenty of bipolar 30V devices, but none with "R-R" outputs. All are open collector outputs. Depending on load - the output is going to saturate within 0.1 to 0.2V of the negative rail. Positive swing will be within millivolts of the rail due to the pull-up (depending on load currents).

    The lowest 30V device offset available is the LM319, but it has a level shifting output that cannot swing to -15V, only to "Ground", so that leaves you with the LM393 family with it's higer offsets..

    Is there any way you can run off just the +15V supply and level shift the output? You will have many more choices, including CMOS devices, if you can drop the total supply under 15V.

    What are the input levels you are comparing? Do they go below ground?

    We'll need a little more details about your application..

    Regards,

  • The input signal maximuim is +-50mV when compaired to input common, so they do go below ground. The input voltage comes from a shunt resister, so there is not really any bias current available. The output of the comparator drives mosfet gates, so output current would be tiny.

    The actual voltage supply rails for the comparators are divided from +-15 down to +-9volts for the mosfet safety.  

  • Hi David,

    Most of the TI comparator line-up has an open-drain/collector output and that appears be the case for those that will operate with +/-15 V supplies. It should only be a matter of adding a pull-up resistor from the comparator output pin to the V+ supply line. The output will pull up to essentially the V+ level providing the stage following the comparator doesn't draw much current through the pull-up resistor. The low-level output characteristics of the output device are such that they allow the low level to pull close to the V- supply line. Again providing that the current through the pull-up resistor is kept low.

    The comparator input offset may be an issue if you are attempting to compare two voltages that are only millivolts apart. Most comparators are not especially precision and have voltage offsets in the millivolts. If the voltages being compared require finer resolution than the voltage offset of the comparator their difference may not exceed the offset.

    A common general purpose, dual comparator is the LM193/293/393. Actually it is a family of comparators having moderately different levels of electrical performance. They are bipolar comparators, not CMOS, but that shouldn't be an issue for most applications. They are available in a wide variety of package types including the 8-pin SOIC.

     

    Regards, Thomas

    PA - Linear Applications Engineering 

  • David,

    I was slow to observe that Paul had posted and already mentioned the LM193 while I was preparing my response.

    Regards, Thomas

    PA - Linear Applications Engineering

     

  • So the data sheet of the LM393 says the input can not go below -0.3volts. So what is the point of +-15v supplies?

    In my application the comparator could see 0v on one input and -5vdc on the other. Based on what the data sheet says , some protection diode would clamp and -0.4v and the device will overheat.

    Am I missing something?

  • Hi David,

    Yes...you are...:^)

    Remember that an op-amp or comparator does not care if the supplies are "Split" or "Single". It's little world lives between the V+ and V- pins. "Ground" is irrelevant to the opamp (since the vast majority have not had "ground" pins for 40 years).

    So the "-0.3V" is referred to the V- pin. So on a "split" supply of ±15V, the lower CMVR limit would be -15.3V. Your 0V and -5V is well within the CMR range.

    Yes. It is confusing. Some datasheets are written as "single" supply, where VCM is referred to ground (V-). Some datasheets, such as the high-speed devices, were written for "split" supplies, where "zero" common mode is actually mid supply ("Ground"). Even worse, a few datasheets have a mix of both.

    See Bruce Trumps blogs:

      http://e2e.ti.com/blogs_/b/thesignal/archive/2012/05/08/op-amp-voltage-ranges-input-and-output-clearing-some-confusion.aspx

      http://e2e.ti.com/blogs_/b/thesignal/archive/2013/04/16/rail-to-rail-inputs-what-you-should-know.aspx

    Regards,

  • on the LM193/393 comparators, what happens when one of the inputs exceeds the supply, but the other input is with in supply?

    example, supply is +-5vdc, but input is 0v and 10v

  • Hi  David,

    The older LM1/2/393 series does not have ESD protection diodes, so there is no clamp to V+ on the upper limit. The LMx93 also has a robust PNP input stage that allows the input to go up to 30V above the GND pin (note that is "GND", and NOT V+), regardless of supply voltage.

    See notes 4 and 5 on page 5 of the LM393-N datasheet. This should answer both your questions.

    As long as one of the inputs is within the "legal" common mode range, the output state should be correct. But it may also affect respone and recovery time.

    But DO NOT exceed the negagative limits. Because there are no ESD clamps, going negative starts turning on the substrate junctions - which can cause currents to flow in places they should not and can cause "bad things" to happen (including latchup).

    So on your +/-5V application, applying 0 to +10V to the input should be okay and should  not cause damage, but I would add series resistance to limit currents. A 10K series resistor should be enough. However, be aware that you are violating CMVR specs and voiding our warantee.... Can you divide the input  voltage down?

    Regards,

  • Thank you for the insight, i reduced the inputs down inside the voltage rails.