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.

LM193: Doubt about Electrical Characteristics for LM193, LM293, and LM393

Part Number: LM193
Other Parts Discussed in Thread: LM393, LM293, LM139, LM339

Hi team 

Question 1:
Refer datasheet SLCS005Z –OCTOBER 1979–REVISED OCTOBER 2017.
Table 6.6. Electrical Characteristics for LM193, LM293, and LM393 (without A suffix)
at specified free-air temperature, VCC = 5 V (unless otherwise noted).

The first 3 rows of this table (Parameters Vio, Iio, Iib) all mention the test condition of Vo=1.4V. Since the supply voltage will be 5V at minimum, I am curious about this value of 1.4V, since is an intermediate voltage level (not 0V, and not Vcc), which seems to suggest that the output NPN transistor is being kept in its active (linear) state as part of the test fixture (or that it is fully ON and is conducting far more than 4mA - very unlikely). Is that the case? If so, what is the current flowing into the output pin in the test? Why is the value of 1.4V used here? Does changing either Io or the 1.4V value influence these parameters? For example, suppose Vcc=12V, and I wanted to know these parameters when Vo=6V(Vcc/2), do these parameters change significantly? Does changing the output current (collector current of the output transistor) also affect these parameters?

Question 2:
Let's assume that Vcc=12V, the output has a pull-up resistor of 10k to Vcc & there no connection of output to either input, and the positive input (pin 3) is held at 1.0V. The negative input (pin 2) is now ramped from 0V up to 5V. How does the current of pin 2 (neg input) change during that ramp?

Please kindly give your comment.

Best regards


  •   add screen shot for your reference.

  • Hello Lidong,

    The 1.4V is a legacy condition from the early 1970's when the devices were first introduced, and it has carried through to this day.

    The output *is* digital, but the "linear" tests (Vos, Ib) are tested in a test loop, similar to the op-amp test loop.

    The output is DIGITAL and NOT linear. The feedback loop forces the comparator output to digitally "PWM", and the output is then RC integrated to get an "equivalent" linear output voltage.

    For these tests, the output is pulled-up to 1.4V instead of the usual 5V or 30V.  The lower swing voltage was optimal for the test loop.

    There is no "real world" reason for 1.4V. It was a convenient value to use in the test system of the time, and it is still used today for backwards comparability.

    Changing the output conditions has very little effect in the input performance. All the gain is in the input stage and the output is a "dumb" NPN.

    As for question #2, since this is a bipolar PNP input, the bias current will slightly shift a few nA towards the more positive voltage.

    For example, at the 1V point, the positive (1V fixed) input current will drop from -25nA to -22nA, and the negative (swept) input current will rise from -22nA to -25nA. This is due to the "unbalancing" of the input stage base currents (note they are negative as the bias current flows OUT of the input pin towards ground).

    Please see the LM393 family appnote for more information:

     Application Design Guidelines for LM339, LM393, TL331 Family Comparators

  • Thanks for you quick reply.

    would like to ask a follow-up question to question 2, lets call this Question 3, under the same test conditions as per question 2.
    Question 3: Please refer to the appnote Paul recommended, figure 1. Lets assume that the neg input has ramped up to about 2V, 1V above pos input. The voltage at Q10 collector is set by V at pos input and the Vbe of Q1 & Q2. Q4 and Q3 have become cut-off.
    Consider what happens as the ramp continues up to about 10V (2V less than Vcc). Q3 B-E is now reverse biased at about 7V. At what voltage does this junction breakover? If breakover occurs, will current at neg input reverse direction ie: flow into that pin, rather than out as it normally does?

  • Hi Lidong, hi Paul,

    only wanted to say something about the "1.4V". At the time the LM139 was created TTL compatibility was a big issue in electronics industry. The output signal range of TTL was 0...0.4V and 2.4V...5V, simply spoken. 1.4V is just the midpoint between the two thresholds of 0.4V and 2.4V. The same for the input signal range of TTL which was 0...0.8V and 2.0V...5V. Again 1.4V is just the midpoint between the two thresholds of 0.8V and 2.0V Relaxed


  • Hello Lidong,

    I think I have an answer for you, but I want to confirm with Design first. I hope to have an answer early next week.

    You may also want to review the original LM339 application note, AN-74, written by the original designer in the 1970's:

     A Quad of Independently Func Comparators (Rev. A)

    Kai, I think the 2V logic levels are just a happy coincidence. The original customer for the LM339 was automotive (rhymes with "Melco"), and there was not much digital logic in cars in 1971. Slight smile

  • Thanks for the historical insight, Paul Relaxed


  • Li Dong

    Paul will be getting back today or tomorrow.


  • Hello Lidong,

    The PNP's in that process have large reverse BE breakdown voltages, over 36V, so they can easily handle the reverse biased junctions.

    I'm not a process expert, so you would have to talk to them about how/why that is possible...but it has been that way for 50+ years...:^)

  • Many thanks for your reply re: LM193 input voltage. I'm surprised that the input PNP BE junctions have such a high reverse breakdown voltage - I would have thought this is incompatible with high gain and low offset voltage. Perhaps this high breakdown V may simply be due to a diode in series with the emitter?

    My apologies re: having to communicate with you via Tech Support - TI changed their policy on which users may post on the E2E forum about October last year, and my email was not deemed sufficiently worthy of such a privilege.

  • Lidong

    As Paul mentioned, that is more of a process related question and we cannot provide any additional information on this.

    In regards to your comment about using Tech Support, can you elaborate?  There were changes to the forum last year but customers were not supposed to be impacted.  Regardless, we have received your question on the forum.  Maybe you just had to login using your work email address.


  • Hello Chuck, allow me to reply to your question:
    "In regards to your comment about using Tech Support, can you elaborate? :
    My name is Fabio Barone, I am the person who posted the original question, and the follow-up responses, Lidong has been kind enough to do this for me via the Tech Support system, since I am no longer able to post directly here on the E2E forum (please refer case number CS0479887). I was able to post directly on the forum up until about Oct last year when TI changed their policy regarding what email addresses are acceptable. Here is one example:

    So this has been a rather tiresome experience since I have to post on the CS ticket, wait for Lidong to copy it to the E2E forum, then I have to routinely check the forum to see if there have been any responses. Also tech support has char limit w

  • Fabio

    I am sorry for your experience.  You are correct that the policy change happened sometime in October.  It was not TI's intent to exclude customers from using the forum. To avoid problems such as you have described, TI emailed account holders that posts from email accounts such gmail would no longer be accepted.  The request from TI was to utilize work email addresses so our field can better support.  So if you can use your work email address, you should be able to directly post on E2E.


  • Hello again Chuck, the problem is this: the gmail account recorded in my TI account is indeed my work email address; it is the **correct** email for the company that I work for, Regenersys. I do not have any other company email addresses. The company has been operating with similar email addresses (for other staff members) since 2010.

    Interestingly, I can still upvote your E2E post, but I cannot reply, nor start any threads there.

  • Hi Fabio,

    why not using any other email address?


  • Thanks Fabio for sharing the information.

    I will close this post since we will support offline.