d123
Advanced Member level 5
Hi,
If you don't need that much precision, and you are aware that NI may find those margins a challenge, maybe one "good" poor guy's solution, and none are "good" for 20mV differences, is to use bog standard 15V > 7812 Voltage Regulator > 7809 Voltage Regulator > resistor divider, to reduce ripple as much as possible, and don't hold your breath, and use paralleled resistors as was suggested by another member, but also choosing one resistor to have a specific positive temperature coefficient and the other to have an equal negative temperature coefficient, or you could look at a floating gate voltage reference (supposedly 10ppm/ºC for grade B, C and D have higher ppms), Intersil make one or two, called x60003, 4.096V and 5V, probably cost rather more than top grade TL431, anyway NASA use(d) them for some reason.
I'd seen that zeners have a tempco of 2ppm/ºC, can't promise that is so, maybe worth checking, as could be incorporated into making an "ultra" stable power source...
It's also unlikely to work, but perhaps worth trying to extend/amplify that 20mV window somehow before feeding it into your comparators, as that's an excrutiatingly minute value to manage, one ripple on a regulator blows the resistor ladder way off, not even thinking of 25ºC increment in temp in a little plastic/metal box, just check regulator and reference min and max output values. Or use some precision super low Vos op amps for four LEDs, a bit pricier than uA741s but good for demanding circuits...
Best of luck, spaceships, femtoAmmeters and medical equipment weren't built by mockery and a can't do attitude, best of luck!
If you don't need that much precision, and you are aware that NI may find those margins a challenge, maybe one "good" poor guy's solution, and none are "good" for 20mV differences, is to use bog standard 15V > 7812 Voltage Regulator > 7809 Voltage Regulator > resistor divider, to reduce ripple as much as possible, and don't hold your breath, and use paralleled resistors as was suggested by another member, but also choosing one resistor to have a specific positive temperature coefficient and the other to have an equal negative temperature coefficient, or you could look at a floating gate voltage reference (supposedly 10ppm/ºC for grade B, C and D have higher ppms), Intersil make one or two, called x60003, 4.096V and 5V, probably cost rather more than top grade TL431, anyway NASA use(d) them for some reason.
I'd seen that zeners have a tempco of 2ppm/ºC, can't promise that is so, maybe worth checking, as could be incorporated into making an "ultra" stable power source...
It's also unlikely to work, but perhaps worth trying to extend/amplify that 20mV window somehow before feeding it into your comparators, as that's an excrutiatingly minute value to manage, one ripple on a regulator blows the resistor ladder way off, not even thinking of 25ºC increment in temp in a little plastic/metal box, just check regulator and reference min and max output values. Or use some precision super low Vos op amps for four LEDs, a bit pricier than uA741s but good for demanding circuits...
Best of luck, spaceships, femtoAmmeters and medical equipment weren't built by mockery and a can't do attitude, best of luck!