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.
I have 100 LM358P Op Amps that are giving me different results with one or both outputs either flipping or not to the positive reference voltage. While the Op Amp will pass the basic test of negative switching to positive on both outputs and vice versa. The results are completely different when an HCF4030B two input Quad OR gate is added for LED control. It seems that the tolerances are different on these Op Amps. Out of 100 pieces, 11 work as expected changing a negative output to positive when the trim pot is adjusted. The rest will vary. Some, both channel 1&2 stay negative, some only channel 1, and some only channel 2. All 100 Op Amps were tested in the same board with a socket added.
The TI LM358 datasheet seems to cover mostly the A or B versions, not the P version. Should I consider a different model for more consistent outputs?
This electrical circuit has worked and been used for more than 10 years to display battery status in a 24vdc system. The hand drawn schematic only includes the portion being affected. Boards were once hand built in house, now they are outsourced to KimCo. New boards are seeing the same types of failures, and replacement Op Amps are a 15% chance of working.
Hello Nick,
I see you have pins 2 and 6 pulled up to supply voltage. LM358 is not an rail to rail input device. Therefore the output will be undependable; it might be high or low if both inputs are above the common mode range.
The solution is to reduce voltage on pins 2 and 6 to be at least 2V lower than supply voltage. .
B+ is from a 24vdc nominal source, it passes to pin 3 and 5 through a 10k pot and some other resistors. Pin 2 and 6 at the LM358P are supplied a fixed 15vdc from the L7815ACV voltage regulator. Noise is flitered with some electrolytic and ceramic capacitors at the regulator's input and output. Trim pot will then allow for the Op Amp to switch outputs to trigger a red, yellow, or green LED. Light is green when both Op Amp outputs are Vcc+(pin8), and red when both Op Amp outputs are Vcc-(pin4). Now, I'm seeing these Op Amps that won't work consistently with the OR gate installed on the outputs. This circuit has worked as is for over 10 years.
Nick
What does "with the OR gate installed" , is the OR gate being installed something new?
You can save some money by not installing R5 and R6 because they don't do anything.
I didn't design the circuit, and I don't plan on changing it, yet! I'm simply seeking answers as to the nature of why the TI LM358P won't give me consistent results. My original post was asking if I should try a different Op Amp LM358, LM358A, LM358B, etc for more consistent results. Out of 100 new LM358P Op Amps that I purchased recently, I've got 11 to work as needed in this board, the other 89 Op Amps won't turn one or both of the outputs positive when IN1+, and IN2+ approaches or exceeds the IN1-, or IN2- voltage.
The orginal post also mentioned that the very basic test on the Op Amp does show the output changing as you would expect. When the Op Amp is installed into the board using the HCF4030B quad OR gate, now the LM358P outputs fail to turn positive. The OR gate isn't a new addition to this circuit. As I said, this circuit has been working and used AS IS for over 10 years. The OR gate's purpose in the circuit is to limit the LED output to only one LED at any given time. As a simple battery level monitor, The OR gate and Op Amp must work together to display the green, yellow, or red LED. For this 24 volt circuit, any battery voltage at or above 21 volts is green. Any voltage between 20-21 volts is yellow. And any voltage below 20 volts turns on the red LED to suggest when to charge the battery.
Hi Nick,
section 9.3.3 of datasheet of LM358P answers your question:
As Ron already mentioned, it looks like the circuit is massively violating the common mode input voltage range specification of LM358P. And it's very hard to believe that this circuit has properly worked for over 10 years.
Kai
Nick,
Kai’s point is valid. When both input voltages exceed the common mode, the input stage will cutoff. That means no current flow, open circuit, as if the input transistors were ripped out.
With the input transistors gone, the rest of the op amp still has a lot of gain but no input signal. Most op amp designs are well balanced. That means the random mismatches in components will tip the balance one way or the other. Therefore, the output of any sample op amp could either be low or high. Usually not 50/50 split based on remaining imbalance. With your new devices that spilt is 10%,90% and not in your favor.
Ten years ago, the design of the LM358P was very imbalanced design from ripped out input to output. The results were consistent in your favor. The newer design of the LM358 has better balance when used correctly or incorrectly. Balance is not a friend of this application.
If there is a long term, change the input voltage to be less than 13V.
In the short term for PDIP, there is only LM158JG (ceramic DIP) which are $7 each. Otherwise try other manufactures that still use an old very imbalanced design. With these, they may all work or none will work.
Thank you for the information on changing the input voltage to the Op Amp for common mode range. It will be a simple change in the regulator. Unless I missed it, I think the Texas Instuments data sheet is missing electrical details specific to the LM358P.
Nick,
Page 12 of data sheet has data for LM358. P is not mentioned because that's the package code.
Suggested change for op amp circuitry