That chip can work down to about 0.007 volts without modification.
Need help with Op Amp Circuit
- Thread starter n4mwd
- Start date
Yes was wondering how it could oscillate without feedback. Thanks for clearing that up.
Regarding noise, 0.10V isn't much. I'm thinking radiation from house wiring, cell phone emissions, the MPU itself, etc., these could potentially be problematic. Anyhow, the MPU can take a few thousand samples and then average them, so that might be a solution since the input is ignored after initialization.
R1 can be changed to 10K with a marginal improvement (lowering of the 100 ohm voltage at point A).
Another theoretical solution is to change R2 to 1K (as in the original design) and connect it to ground through an NPN transistor rather than directly. Then the base of the transistor is fed by a cap that gets charged by putting the input pin on the MPU into output mode and driving it high (3.3V) for a short charging period. The cap keeps the transistor turned on long enough for the MPU to switch back to input mode and read the value. Then once the input value has been read, the MPU can drive logic 0 (gnd) onto the cap and force the transistor off. The reason for all this is because while using a 1K resistor for R2 will drop the voltage at point A by about a volt, it draws too much current through the off-board 100 ohm resistor. By switching the current on only for a few milliseconds, the external 100 ohm resistor should be able to tolerate the overload without damage.
It would be nice if there was a way to spread out the voltage more without overloading the off-board 100 ohm resistor (60mW).
Regarding noise, 0.10V isn't much. I'm thinking radiation from house wiring, cell phone emissions, the MPU itself, etc., these could potentially be problematic. Anyhow, the MPU can take a few thousand samples and then average them, so that might be a solution since the input is ignored after initialization.
R1 can be changed to 10K with a marginal improvement (lowering of the 100 ohm voltage at point A).
Another theoretical solution is to change R2 to 1K (as in the original design) and connect it to ground through an NPN transistor rather than directly. Then the base of the transistor is fed by a cap that gets charged by putting the input pin on the MPU into output mode and driving it high (3.3V) for a short charging period. The cap keeps the transistor turned on long enough for the MPU to switch back to input mode and read the value. Then once the input value has been read, the MPU can drive logic 0 (gnd) onto the cap and force the transistor off. The reason for all this is because while using a 1K resistor for R2 will drop the voltage at point A by about a volt, it draws too much current through the off-board 100 ohm resistor. By switching the current on only for a few milliseconds, the external 100 ohm resistor should be able to tolerate the overload without damage.
It would be nice if there was a way to spread out the voltage more without overloading the off-board 100 ohm resistor (60mW).
Thanks everyone for all your help. I finally put the order in for the project parts (~50) so I should be getting the chip in a few days. The only bad thing is that they don't seem to make a DIP version of the part so breadboarding will be impossible without some sort of breakout board. So I decided to go ahead and make it into the main board. I can tweak the resistors once its built.
That's why I sent you a link to 41 choices. Two of them are through-hole.
I wasn't able to find any DIP versions of theTLV1702 which is what I thought you said to use. I didn't want to get a different part because I might get one that was more like the 741 than the 1702.
When the board is done, it will have a lot of other stuff on it too, but I'll populate the TLV1702 and associated resistors and test that first and let everyone know how it worked.
Again, thanks for all the help.
