simple Op amp circuit not behaving.

[JWoodrell]

ok so i made a simple op amp circuit to scale battery voltage down so my PLC could read it.
My input voltage (battery pack) initial target range is 20V - 30V, and my output voltage is 0-10V for the analog input card. to keep it simple

once i get it where i like it i will probably change the scale and offset to have 23.5V-26V as 0-10V output

so in my mind that should be a perfect use for a differential op amp setup.

It mostly works, but the value is slightly off for each of the 4 op amps on the chip  by a few tenths of a volt.
the opamp is a LM148 quad chip.

my question is should I be worried that they are different from each other even on the same chip or is there something I can do with the feedback loop to help eliminate that variance from chip to chip.

[Andy Watson]

Only a +10V supply to the op-amp?

It looks like you are taking the inputs to the op-amp above the voltage of the positive supply to the op-amp  - they won't work in this configuration. For example - consider the voltage at the positive op-amp input - I make it to be 15V.

Edit: Also, expecting to get 10.0V out when the supply voltage is only 10V is a little too optimistic - even for a rail-to-rail op-amp (which I don't believe the LM148 is).

[dannyf]

The input can exceed the rail supply voltage.

Unless this is one of those beyond-the-rail opamps, it is not going to work.

[rob77]

The input can exceed the rail supply voltage.

Unless this is one of those beyond-the-rail opamps, it is not going to work.

could you name those beyond-the-rails opamps ? the other day i was searching for one with a common mode voltage allowed to go to the top rail and max supply >=30V(+-15) , and found nothing except LM358... but that one got no trimmable offset.. so i put the project aside

[c4757p]

Off the top of my head, there's LT1494/5/6. No V_OS trim on that one either, but fairly low offset to begin with.

[rob77]

Off the top of my head, there's LT1494/5/6. No V_OS trim on that one either, but fairly low offset to begin with.

thanks certainly does the common mode part  ... but seems to be the slowest OPamp ever 1V/ms sleeve rate (volt per millisecond, not microsecond!!!) , gain bandwidth product 2.7KHz ! it's so slow, you could safely do other things till it swings from rail to rail

[c4757p]

1V/ms sleeve rate (volt per millisecond, not microsecond!!!) , gain bandwidth product 2.7KHz ! it's so slow, you could safely do other things till it swings from rail to rail

I like to use it as a tea timer.

[IconicPCB]

Single rail supply both inputs posiive... use a Norton opamp.

By the by... if only measuring battery voltage, bandwidth is not an issue. Battery voltage is not likely to oscilate at some megahertz.

[rob77]

Single rail supply both inputs posiive... use a Norton opamp.

By the by... if only measuring battery voltage, bandwidth is not an issue. Battery voltage is not likely to oscilate at some megahertz.

thanks for reminding me of the Norton amp. existence you're right  those have no issues with common mode input at the top rail. apparently it's time to dig up those LM3900 chips i sourced some time ago (sourced them just because i found them interesting)

that common mode at top rail and slow OPamp is not related to OP (sorry for the little hijack - but it's still the same topic )... btw.. for battery voltage 1V/ms sleeve is ok, but for current sensing it's slow.

[Zero999]

Off the top of my head, there's LT1494/5/6. No V_OS trim on that one either, but fairly low offset to begin with.

thanks certainly does the common mode part  ... but seems to be the slowest OPamp ever 1V/ms sleeve rate (volt per millisecond, not microsecond!!!) , gain bandwidth product 2.7KHz ! it's so slow, you could safely do other things till it swings from rail to rail

The open loop gain is also low.

I suppose to achieve such a wide common mode range, compromises need to be made.

[JWoodrell]

thank you for your input guys.  its a project I'm working on with my father, he is across the country so i have to wait until he tells me how the changes help (im having him take the supply+ up to the 20V, which should give it enough room to do its Op amp stuff. 

the speed is not important it will probably take several hours for the battery voltage to be drawn down.  its for an automated robot cart thing that needs to know to go home and charge when its batteries get low.

I'll just have an external clipping circuit to keep the final voltage below 10V.  which should never trigger, but just in case so as to not blow the analog input card on the PLC

[dannyf]

You may want to think what exactly you are trying to do with this circuit.

In case you haven't noticed, it can oscillate.

[Rupunzell]

The output will be stuck to the V+ rail.

Input voltages are too high for the divider ratios and the Op-amp is not powered-biased properly.

Use a +/- power supply bypass and ground proper, then reduce the input voltages to 0.2 & 0.3 volts. Begin here.

[Leadfootin]

Try 100K to 1M input with 10k feedback resistors. Be careful to match the input pairs and feedback pairs to get accuracy.