ohm's law = Murphy's law * my IQ/3.1416

Randy

Trial and error have gotten me where I need to be but I'd like to know how to figure this with an equation. I am using nine LM34 temperature sensors with the ground terminal bias to 0.33 volts. This lets me measure negative temperatures ( to -33 degrees ) without negative values. The outputs go to an eleven channel·TLC1543 A/D converter·and then to a BS2pe where I subtract 33 for the corect temperature. The -ref is to ground and the +ref to 2.55 volts for scaleing. If you look at the attached drawing you will see that I am using a voltage divider consisting of two resistors (R1 and R2)·and a pot that is drawn as R3 and R4, all in series. The 2.55 volts at the wiper, or node between R3 and R4,·is applied to the +ref pin and the current draw is too low for my meter to measure. The 0.33 volts at the R1 / R2 node is used for the biased ground to all nine LM34· sensors. Combined they draw 0.567 ma. It seems to me that using the voltage and current I could figure the effective resistance of the nine sensors and using that number I should be able to prove all of the other values with ohm's law. Either I am screwing something up with the combination of series and parallell resistances or I am wrong about being able to deduce a resistance value for the sensors. I have seperated the dividers using the pot between ground and five volts and adjusting for 2.55 volts and using just R1 and R2 between ground and five volts but I cannot get to the 0.33 volts except by trial and error. How do you figure the values?

In the drawing I give the measured values of the pot from the wiper to the corresponding pin. It is a 10k pot that measures 10.5k and yes I know the numbers don't quite add up.

Randy

Tracy Allen

The LM34 does not have a resistance as such. It's power supply current is more like a constant current source of around 100 microamps and the current is more or less constant as the power supply voltage changes. You can find the typical curve in the LM34 data sheet. Typical, with sample variation. For a circuit, you would be better off with an op-amp to provide the virtual ground of 0.33 volt. Then it wouldn't depend on current at all.

The circuit will need a pulldown resistor from the LM34 output to ground. (~10kohm)

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

Randy

Thanks Tracy, that explains a lot. I didn't notice until now but the voltage varies quite a bit if I unplug some of the sensors. The current seems to stay unchanged regardless of how many sensors are pluged in. Does that seem correct? I have added the op-amp configured as a voltage follower using a pot for the + input and it is as stable as can be. I can now adjust the virtual ground to within 2 or 3 1000's of a volt and it stays rock solid. 0.30 volts works better for my application and it was easy to get there.

I do have the pulldowns, I just didn't bother with them in the drawing.

Thanks

Randy