Hi, Michael. First, your 3.3V power supply does change some things (see prior thread, "Interfacing A Pressure Sensor"
). You need a weak pulldown (say, 33K or so) resistor at the final output of the LM324 to get the output down to 0V (actually 20mV or so). With the pulldown, you're going to be somewhat limited as to maximum output voltage -- about 1.5V or so less than the
supply. That means you're talking about a maximum output from the LM324 of 1.8V. And that means you're throwing away 45% of your counts, which might be a problem -- it certainly isn't optimal. If you want to stick with the LM324, then set your pairs of gain resistors so that maximum output is about 1.8V (i.e. if your maximum sensor output is
200mV, make it a gain of 9 using 11K and 100K resistor pairs).
This inexpensive solution should get you better results than the initial effort of putting both inputs of the Wheatstone bridge directly at the PICAXE inputs. If you want to do better, you'll have to consider a rail-to-rail input and output quad op amp like the R-to-R op amp like the TLV2374ID, which is out of stock at RS.
As far as the voltage divider to read battery voltage, you'll have to provide more information. Basically, whether your 3.3V is coming off the battery voltage or not, you will have to be concerned with the Vin at the PICAXE input pin exceeding the voltage at the Vcc pin of the PICAXE. It's trivially easy to latch up a uC by exceeding that voltage, and you'll have to give that some serious thought.
If you're working off battery voltage to power your PICAXE (say, with an LM317), all you have to worry about is powerup/powerdown sequencing. It might be a good idea to use a smaller divider with diode protection, and then use the 4th op amp to boost it up, like this (view in fixed font or M$ Notepad): | | VCC | + | | | .-. | | |10K | | | | '-' | | ___ |\\ | .---o---|___|--|+\\ | | | 10K | >--o---->
The 100:1 divider cuts your 13.8V down to about .138V. That's not enough to cause serious conduction in the diode, but will protect the op amp input. Then the 11:1 ratio for the feedback will give a gain of
12, which will bring your .138V up to 1.65V, adequate to have a good handle on the battery voltage. Use a 1N4148 or 1N914 for the diode. This may look overly complex, but you're making sure that, under any circumstances, the LM324 input isn't seeing more than 0.6V or so, and also that the PICAXE input will never see more than Vcc. It's not elegant, but it should do, especially for a hobbyist with a limited budget.
Hi, Mike. You didn't describe the environment in which you're powering the PICAXE. It might be possible you're using this in a car, or something else.
The point is, you want to try to make things so you don't exceed the maximum input voltage at any pin. This is particularly important with microcontrollers, where the reverse current can cause mysterious glitches which might crash the uC program, even if it doesn't damage it.
If you're monitoring an external battery, and your regulator is getting power from that battery, you'll have a problem at turn-on, when the full 12V is present, but the regulated output voltage is ramping up. This may only be a millisecond or so, but it's enough to cause problems.
Let's assume a 12V battery is hooked up to your voltage divider with the PICAXE unpowered. As your PICAXE is powered up, your series resistance might prevent latchup (a destructive condition common to CMOS where reverse currents caused by a low impedance voltage at a pin above or below the power rails causes a destructive SCR effect).
However, it might cause unintended effects on the uC chip which prevent a normal reset or reliable operation. Better to be safe. Especially if you're cost is only an LM324, a couple of more resistors, and a protection diode.