Frustrated Modbus

[gadelhas]

Hello,
How is everyone? I got a new question, it is related to this project but not to modbus. If no one minds I am just going to keep asking questions on this post. I am working with a MCP4209 DA Converter. I want to output 0 to 10V. So I want to run the output of the 4209 into a non inverting 741 op amp. My problem is, the lowest value out of the op amp I can get is 3.97V. Is that all the closer a 741 can get to the lower rail. I have 12V going to the positive rail and ground hooked to lower rail. I get my 10V out when my 4209 is putting out 5V.
I have a 10K resistor hooked from ground to pin2 of the 741. And a 10K resistor from pin2 of the 741 to pin6 of the 741. Input from the 4209 is running to pin3 of the 741.
If I run the 741 at + and - 12V all works good but I do not want to add a negative voltage supply to my project. Is there anyway around this?
Thanks
Shawn

Hi try to use a LM358 Insted of the 741. Power the LM at +12V and ground and post the results.

[mtripoli]

I'm guessing you mean "MCP4902"? A couple of basic questions... you have compensation in your feedback loop? If you don't have a cap there you're asking for problems, especially with a 741. If you really only need single-ended output go with one of the devices from Microchip designed for rail-to-rail single-ended use (or whoever, the MC stuff is pretty inexpensive). Don't forget the offset voltage from the opamp itself; it also gets amplified (by two here).

Are you decoupling as you should? Caps on the rails as described in the doc's for the DA?

[shawn]

Hey
thanks for the replies on the op amp. I tore a lm358 out of a board and tried it. It worked good. I have some mcp601 but they are only good for 5.5V which is not enough. Where does the cap go the example circuits I looked at don't show a cap.
thanks
shawn

[gadelhas]

Hey
...I tore a lm358 out of a board and tried it. It worked good...

Great!!!

[mtripoli]

Look at the attached schematic. Whenever you design a circuit around an opamp you should use a compensation cap in the feedback loop as well as good power supply decoupling. In most documents this information is "left out" because it is a "given" (the schematic gets very crowded showing every decoupling cap).

The formula for calculating the "Ccomp" cap is shown in the schematic with a very simple explanation (it really is beyond this forum to explain -3db break points etc.). The point is, you can design yourself a very nice oscillator if you ignore Ccomp and bypassing (another term for decoupling), especially when using breadboards and flying wire construction.

In the example, figure out what the highest frequency is that you will be passing though the circuit and calculate the value needed and then use the closest common value available. If this is varying DC I wouldn't get too worried about it...

There are often overlooked parameters using opamps (and other devices, even by "pro's") such as CMRR and PSRR (a little homework for you) that do not work properly without a bypass cap. A 0.01-0.1uF cap is mandatatory on every IC in a design. The additonal cap such as the 1uF Tantalum (notice one cap is ceramic, the other is tantalum, or could be electrolytic - another homework assignment ) shown is optional but you as the designer of the circuit are responsible for deciding that (ever looked at a circuit board and seen caps soldered across pins on parts; they "decided" incorrectly).

Consider these things as "law" and you will find that your circuits behave correctly not only on the bench but in actual use, over time and temperature changes. Ignore them and one day your circuit may "flake" on you; by the time you get it back to the bench it'll work "correctly" again... until the next time...