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.

LM293A: Slow Respond Issue Of LM293A

Part Number: LM293A
Other Parts Discussed in Thread: LM393B, LM339

Hi Teams

Customer selected LM293A in their new project for overcurrent protection but they found the respond time is slow than datasheet spec.

The schematic have shown below. We measure the output in PIN1 with 5V VCC and we already removed all cap cap and resistor connected to PIN1 (Just like output circuit). And the waveform I capture shown in diagram2, the respond time from output high to 0V is about 6us. (The green one is comparator output, the Blue one is reference voltage and the yellow one is current sensor output)

My questions:

  1. Datasheet listed the typical respond time is about 1.3us, I wondering to know are this spec work for both positive transient and negative transient(Output from Hight to Low)
  2. The typical respond time for negative transient is  measure from high voltage to 0V or other specific points?
  3. Are LM393B could achieve better respond time than LM293A?

  • Hello Gabriel,

    Are they *sure* they removed all the capacitors on the output?

    It's a little hard to see the scope divisions - but is the output going all the way up to 5V?

    Does the rising edge of the output look similar?

    What else is the output driving? Long trace? Cable?

    Are all the voltages at the inputs below 3V? (It looks like they are <1V, but I want to make sure).

    Does replacing it with a new device show similar behavior? Is it getting warm?

    Do any of the inputs go BELOW ground at any point?

    The actual output starts dropping just slightly after the input (near the trigger point), so the actual response time  is fairly quick.

    Because of the open-collector output, the output falling edge should be the fastest. But it looks like the output sinking a large current, such as a large value capacitor, causing it to stretch out as it goes into current limiting.

  • Paul

    Thanks for your reply.

    I am sure the output have removed all cap that connected to PIN1.

    The output voltage going from 5V pull-up to 0V. The PIN1 is open circuit and it indeed connected to a long trace.

    I didn't found abnormal thermal in LM293A and I already replace it with new one couple time. But the result still the same.

    The input signal level is within recommended common mode range (5v-2v=3v).

  • Hello Gabriel,

    Does the rising edge (low to high) of the output look similar? The rising edge should reflect the time constant of the output capacitance and the 5.1k pull-up resistor.

    A long trace on a tightly packed, multi-layer board can get up to several hundred pF.

    What is at the other end of the trace? GPIO pin? MOSFET gate? What else is on the output line?

  • I didn't measured the rising edge previously. I need to check with customer tomorrow.

    The other end of the output is open circuit, originally it connect to MCU GPIO with a RC filter. But I already removed the RC filter.

  • Paul

    The issue have been solved. It's lead by the output RC filter. Customer remove the wrong RC in their PCB board, after remove the correct one the issue fixed.

    By the way, the input signal of LM293A will as high as 3.1V with 5V power supply. It seems we have chance to violate input common mode range when the input

    reach 3.1V in some condition. I want to know what will happen if we reach 3.1V? 

  • Application Design Guidelines for LM339, LM393, TL331 Family Comparators says:

    Between 0 V and 3 V (VCC - 2V), the device is fully operational and will function per data sheet specifications over the full specified temperature range. The VCC - 2V limit is the recommended upper input range limit to be utilized for all designs.

    The input range from 0 to 3.5 V (VCC - 1.5V) is valid at 25°C and above. The range between 3 V and 3.5 V will vary over temperature due to the VBE's of the transistors changing at -2.1mV/°C. This results in the input voltage range changing at -4.2mV/°C over temperature (note the negative sign!), necessitating the VCC - 2 V over temperature specification. Using the VCC - 1.5 V limit is the main cause of It worked fine on the bench, but it fails at cold complaints. Do not make this mistake!

    The range between 3.5 V and 4 V is the gray zone, where the device appears to still function at 25°C and above, but critical specifications are deteriorating, such as offset voltage, bias current and particularly propagation delay as the input stage is gradually cut off. These effects may not be immediately apparent. Operating at low temperatures will cause failures. Operation in this zone must be avoided.

  • Hello Gabriel,

    Great! I suspected there was still a capacitor on the output - at least a few nF.

    Clements beat me to the LM339 family application note!

    In short...You will probably be okay at 3.1V, especially if this is intermittent and not part of normal operation. If there were to be issues, you will see them at the lower end of the cold temperature range. The LM339 family "feature" (as explained in the appnote section 2.1 and 2.5) will keep the output in the correct state, but repose time and bias current may be slightly affected.

    If you do see failures at cold, and the input is 3.1V, then all warranties are void and not eligible for a warranty claim - so that risk is up to the customer. Our advice is to keep everything below 3V.