I'm driving the LM317 with a LM2937 5v regulator. The voltage drop across the test resistance
(I'm actually using a 220 ohm, 20 turn trim pot) is what I would expect at the different test resisance values...2.0vdc = 200ohms, 1.0vdc = 100 ohms.
However...Meter accuracy goes "crappy" at about 16 ohms and below. I've measured and compared the voltage drops against various test resistance values and they are correct. The circuit is doing a good job. But the meter seems to lose its accuacy at low ohms (<=16).
Just to make sure you understand where the inaccuracy is coming from, use another voltmeter to measure the voltage that is being feed to the panel meter. Is it the same for all values of test resistance? If not, then the problem could be elsewhere.
Just to make sure you understand where the inaccuracy is coming from, use another voltmeter to measure the voltage that is being feed to the panel meter. Is it the same for all values of test resistance? If not, then the problem could be elsewhere.
I couldn't find a spec sheet for the meter. Does is have accuracy specs?
I notice they sell for about $10, so I would think they must be pretty cheap probably made in China junk. Possible the low end accuracy is just not very good. The thing runs off a single supply: that makes it hard to process low level signals accurately.
the current source isn't behaving linearly then. often op amp current sources require closely matched resistors (better than 0.1%) to operate properly (something most articles about current sources fail to mention). better details and a better understanding of how these circuits work can be found here: https://www.electro-tech-online.com/custompdfs/2012/05/snoa474-1.pdf
the other thing that may be happening is that you are using a single supply op amp, and with those low resistances you are up against the -rail limit of the op amp's output stage.
I couldn't find a spec sheet for the meter. Does is have accuracy specs?
I notice they sell for about $10, so I would think they must be pretty cheap probably made in China junk. Possible the low end accuracy is just not very good. The thing runs off a single supply: that makes it hard to process low level signals accurately.
If not...I could...but I realy want to keep the final design using a single power supply and not have to worry about a battery...
My project currenty has two supply voltages...12vdc and 5vdc.
It sounds like the meter has a linearity problem with measurements near gnd. If so, then offsetting the measurement range away from gnd might solve the problem.
Try adding a fixed resistance of 20 ohms or greater between gnd and Rx, then using the offset feature of the meter to make "zero" in a new place.
Of course, to top of the range will move up by the same amount, so make sure you still have enough headroom in your current source.
It sounds like the meter has a linearity problem with measurements near gnd. If so, then offsetting the measurement range away from gnd might solve the problem.
Try adding a fixed resistance of 20 ohms or greater between gnd and Rx, then using the offset feature of the meter to make "zero" in a new place.
Of course, to top of the range will move up by the same amount, so make sure you still have enough headroom in your current source.
I'm a little confused on this one...
The panel meter input shares a common ground with its supply....so how can install the 20 ohm resistor "between" Rx and ground?
what circuit are you using? if it's the LM317, the minimum output voltage is 1.25, which means if the current source output gets down to 1.25, it will get very nonlinear.
what circuit are you using? if it's the LM317, the minimum output voltage is 1.25, which means if the current source output gets down to 1.25, it will get very nonlinear.
I'm a little confused on this one...
The panel meter input shares a common ground with its supply....so how can install the 20 ohm resistor "between" Rx and ground?
The measurement ground still goes to power ground. But measurement positive will always be 200mV higher than what it normally would be. Use the meters offset calibration to subtract the 200mV error.
(these numbers assume a 10mA current source)
Unclejed, the LM317 is being used as a floating current source, so it's minimum Vreg applies to the current setting resistor, not to the actual output voltage.
try this circuit. the emitter resistor (R1) sets the current, and as show is very close to 10mA. R3 is the resistor being measured. this also has the meter ground referenced.