LM393B: Operating voltage for LM393B

Ranjeet Malhotra

Prodigy 50 points

Part Number: LM393B
Other Parts Discussed in Thread: LM393LV, TLV4021, TLV7021, LM393

Hi,

We are looking to use LM393BIDDR in one of our applications. The device is used as follows:

- IN2+ is connected to 2.5V

- IN2- is connected to external logic and can be either 0V or 3.3V

- OUT2 is PULLED UP to 3.3V using 100K pull up.

We are looking to power the chip using 3.3V. When doing some bench testing - we observe the following behaviour:

(a). When IN2- is 0V -> OUT2 -> 3.3V (this is expected)

(b) In certain instances when IN2- is 0V -> OUT2 -> 0 (this is not expected behaviour)

Looking at the data sheet there is a footnote (1) on Table 6.7 (Page 6/52)- "The voltage at either input should not be allowed to go negative by more than 0.3V otherwise output may be incorrect and excessive current can flow. The upper end of the common-mode voltage range is limited by Vcc-2V. However only one input needs to be in the valid common mode range, the other input can go upto maximum Vcc level and the comparator provides a proper output state. Either or both inputs can go to maximum Vcc level without damage".

A couple of questions:

1. Can the device in our application be operated at 3.3V? The datasheet seem to suggest operating voltage between 2V-36V but due to footnote (1) above .. we think that our application will be encountering conditions where both inputs could see voltages exceeding Vcc-2V and hence the output will function correctly?

2. Will the issue be resolved if the device is powered using 5V?

thanks

regards, Rsm

over 2 years ago

0 kai klaas69 over 2 years ago

Guru 140200 points

Hi Ranjeet,

why not scaling down the input voltages and by this making them falling into the common mode input voltage range by the help of some simple resistive voltage dividers?

Kai

0 Ranjeet Malhotra over 2 years ago in reply to kai klaas69

Prodigy 50 points

We have looked into this - but we are very short on PCB space. If we have to scale down voltages - we will need to add many resistors which we are trying to avoid. We have a local 5V rail available which we should easily be able to connect. Hence the suggestion of using 5V (assuming it is a fix!!).

Thanks

0 kai klaas69 over 2 years ago in reply to Ranjeet Malhotra

Guru 140200 points

"In2-" is still violating the common mode input voltage range which can result in a degradation of speed performance of LM393B.

A 0402 resistor has a length of 1.0mm. Really no space for 0402 resistor(s) ?

Kai

0 Clemens Ladisch over 2 years ago in reply to Ranjeet Malhotra

Guru 258350 points

The headroom to the positive rail is temperature dependent. 5 V can fix it for some temperatures. See section 2 of the Application Design Guidelines for LM339, LM393, TL331 Family Comparators:

I would just use a low-voltage device like the LM393LV or TLV7021. If you want to save space, consider a device with integrated reference like the TLV4021.

+1 Paul Grohe over 2 years ago

TI__Guru 51446 points

Hello Ranjeet,

#1. Yes, but as Kai explained, the input range is narrow (0 to 1.3V on a 3.3V supply), and your inputs are violating the spec on both pins (output tends to go high on the B version).

Please see the LM393 family application note, sections 2.4 and 2.5:

Application Design Guidelines for LM339, LM393, TL331 Family Comparators

You may wish to look at the new LM393LV, which has Rail to Rail inputs and is designed to be a low voltage (<5V) drop-in improved replacement for the LM393 family, and the inputs can go up to 6V without damage (even with a 3.3V supply).

LM393LV data sheet, product information and support

#2. For the LM393, we would recommend that you use the 5V supply, to give a 0 to 3V valid input range. The input range shrinks at cold temperatures, so the 3.3V input could be problematic at cold. Both inputs out of range is technically an invalid condition.

For the LM393LV, it could be run on 3.3V with a 3.3V input (and up to 6V).

Amplifiers

Amplifiers forum

LM393B: Operating voltage for LM393B