Like many of you I took a risk and decided to buy a batch of very cheap eBay components. In this case I purchased 50 LM358 operational amplifiers for $2 USD.
I thought it would be interesting to try and determine if these op-amps match manufacturers specifications or if they are fakes.
- Gain a personal understanding of what the op-amp data sheet specifications mean
- Determine if the chips I bought match the official specs
I decided to test the op-amp by using a relaxation oscillator:
This oscillator allows us to measure several key attributes of the op-amp that I’m particularly interested in.
Using this circuit we can measure:
- Output Swing
- Low Level Output Voltage
- Slew Rate
This is the expected output for output swing for the LM358 as given by Texas Instruments:
To make things easier on myself I set my oscillator to operate on a supply from -7.5 to +7.5 volts.
I recorded the output of the oscillator and this is what we get:
We can see here that our output swing is 12.7V. Pretty close to our expected 12.5 volt.
I suspect the bottom of the square wave being slanted might be something to do with my ground not being perfect.
Low Level Output Voltage
The LM358 has an interesting feature in that it can get very close to the low voltage rail on its output. The LM358’s ability to do so depends on how much current it is sinking at the time.
During the low point in the cycle we are discharging around 4V peak from the capacitor via a 1k resistor; so roughly 4mA is being sunk by the LM358:
From the above graph when sinking 4mA we’d expect our output voltage to be somewhere around 0.4 or 0.5V above the low level input. So in our case we expect around -7.1V
Let’s see what we get in practice:
The V minimum here is -7.07V. This is quite close to our estimate of -7.1V
For a sanity check, lets put our probe on low level:
Our negative rail is setting at the expected -7.4V
The LM358 has a very slow slew rate of 0.3V/us so it should be fairly easy to test this.
To get better accuracy we increase our horizontal resolution and get in for a close look at the rising edge of our oscillator:
We can fairly easily see that it took 3 and a bit divisions of 10us each to cover our peak to peak voltage of 12.5V
This means our slew rate is simply 12.5V/33us = 0.37V/us
This is slightly faster than our expected 0.3V/us but I think it’s close enough that we can give this a pass.
In this case it seems like the ultra-cheap LM358s I bought from eBay for $2 for 50 seem to be legitimate. Obviously this testing isn’t comprehensive however it seems to confirm several key aspects match closely to the LM358 specifications. In this case I got lucky — there’s definitely a lot of people receiving fake components from ebay sellers!