Hi
I have noticed that there is a latch up on the opamp chip everytime I power it up. For example, I have tried to use TL8048N and OP07 which automatically latch up to postive rail. I have been told that the voltage is out of the range especially in input common mode voltage. So I went to look at their datasheet and found out that TL084's input common mode voltage within +/- 15V and OP07 only accept the voltage within =/-14V.
At the moment, I use a single power rail instead of dual. Someone said that the op-amp such as TL084 could not detect the voltage within 3V and I could not find that information in the TL084.
By the way, I am using the op-amp for current sensing which will later connected to ADC. Again I do not understand when I have been told that op=amp must have input common mode votlage to 0V and Positive supply. What this does mean?
Please help me to identity this issue?
MM
Hi there,
It sounds like you may need to look up "Input common mode range" and find out what
that means. I know it's a little difficult to understand at first, but it's not really that bad.
With the older style op amps like the TL084 that have FET inputs the input stage needs
a certain bias voltage, and that bias voltage has to come from somewhere, so it ends
up coming in the form of reduced input common mode voltage range.
What this means is that even if we power the op amp from +15 and -15 we may not be
able to allow an input voltage of more than +12v or less than -12v, or very strange and
sometimes unpredictable things may happen to the output. The reason is that the
op amp is specified this way and so they are not responsible for what happens outside
the input common voltage range. If the manufacture said, "Do not use this op amp
with inputs below +3v when the minus power supply is 0v", then we would immediately
know not to do this, but they dont do that...instead they specify an input common mode
voltage range and then we have to calculate the allowed limits ourselves.
The TL084 is no exception. If you connect the minus power supply terminal to ground
which is 0v dc, you will be lucky if the input works below 3v dc, but surprisingly it may
be even as high as 4v dc. That means you shouldnt let the input get lower than the
minus supply rail plus 4v. With a -10v supply, that means the input can not be allowed
to go below -6v, and with a supply of 0v (single supply operation) the input can not
be allowed to go below 4v.
Not all op amps are like that though, and it just so happens that you happened to
choose an op amp that is definitely not the best choice for this application where you
have only one power supply.
For example look at the LM358, which is a very common op amp used for many dc
applications. The input common mode range includes ground with this op amp, so
you can actually allow the input to go as low as 0v dc and it will work just fine.
The upper limit though is roughly Vcc minus 1.5v, which for a +5v power supply
comes out to 3.5v. That means we should not allow the input to go higher than
3.5v when working with a +5v power supply.
Many of the older FET input op amps have a really nasty input common mode range
so you should probably avoid them and use the newer type op amps or at least one
that is specified for operation with 0v input. The LM358 works pretty well this way
so maybe you should take a look. There are numerous newer design op amps out
there too that actually have rail to rail input range and also rail to rail output range,
if you happen to need that too.
In addition to the input common mode range issue there is another important issue
that comes up when trying to measure current with an op amp. That is, the input
offset voltage.
The input offset voltage is a theoretical input voltage that exists even with no input
voltage. The op amp thinks it has a small voltage like 5mv even when the input
is connected to ground. What happens is this seeminly small voltage gets amplified
by whatever gain you have the circuit set up for, and that introduces a significant
dc error.
In your drawing i can not tell what the value of your feedback resistor is, but say you
have it set up for a gain of 100. With a 5mv input offset, multiplied by the gain of
100, the output would go to 0.5v. Yes, that's a whole half a volt for no input at all.
That means with the input sense resistor of 0.010 ohms it makes it look as if we
already have 0.5 amps flowing through the sense resistor. Now if we were trying
to measure 100 amps that might not matter too much, and even if we were trying
to measure 10 amps it might not matter much, but if we were say trying to measure
1 amp that would be an error of 50 percent! That's not good at all so what we could
do in that case is try to find an op amp that has a much lower input offset voltage
so that we can reduce the error. If we could find one with only 500uv offset for
example that would reduce the error to only 5 percent for measuring a 1 amp current.
If we need better than that, then we need to look at better op amps with very
low input offset spec's like 50uv, which would reduce the error to only one half of
one percent which is usually good enough.