LM3914N Resistor Values?

Why don't you get those values empirically? Just breadboard the circuit using the variable resistors as drawn, adjust them for the responses that you want, then take the variables out, measure them and then substitute precision resistors in their places. You'll have to remember that the IC might have tolerances internally that might make it necessary to use very tight tolerance resistors or you might have to resort to variable resistors anyway.

Cheers!!

portable, simple read-out meter for testing solar battery

Firstly, you can build the circuit as shown, adjust it to suit your needs, then measure all the values you need. The deeper issue, though, is that the measurement is useful ONLY because it has a mathematical connection to the physical situation. All measurement is mathematical; those numeric values printed next to the LEDs are the only meaning you can get from the measurement.

LED output is power-hungry, an LCD output will take less energy and is visible in bright sunlight. I'd use an LCD voltmeter, and run it on batteries. The 12V case is similar to auto electric monitors, have you looked at what's commercially available?

Your right of course, just did not have pots in the bits and bobs box with values near those stated, so will have to go and get some from the nearest town (I live way out in the country). The readings at the end of the day are not life threatingly critical, so a bit of variance is OK, was looking more for repeatability if we decide to build more of the same.

Sorry I have no idea what you mean "the measurement is useful ONLY because it has a mathematical connection to the physical situation"

As far as LED current consumption is concerned, the circuit if you look again, only turns on the LED's briefly every so often. As far as LCD meters are concerned, yes you can buy cheap LCD volt meters off ebay, but they tend to need a seperate power supply from the one they are measuring. Also LCD meters rely upon the user knowing what the numbers mean, where as coloured LED's give a better Go, No-Go read-out, for the application I have in mind. Thanks for the input anyway.

What I meant was, the updating of this circuit for 48V has side effects on the accuracy you can achieve, and the addition of the bridge rectifier changes the calibration significantly. A measurement is easy with a meter ($10 or less); why would one prefer the colored lights?

I'm particularly disturbed by the bridge rectifier; someone will need to know the polarity someday, and that HIDES the information. I lost my car for a few days, when a battery polarity was misidentified...