Thanks, that is almost exactly what I have done on the breadboard circuit. Instead of trying to use pins 7 and 8 of the LM3918 to provide a Low and High reference, I use a resistor divider fed from the stabilised voltage supply. I too have shorted the high order LED pins so that the last green LED (the fourth) will stay lit (you have used an extra LED).

Using the 7815 is less parts than using the op-amp maybe I should have thought of that and it was mentioned I think in this thread ? It uses 4-5mA at least compared to 2-2.5mA for the op-amp version. We can also replace the op-amp with a single pass transistor, a bit less accuracy but we'd be down to 1.5mA.

On the breadboard version I am trying to get it to be accurate enough, will take a picture in a sec.

 

OK I have an issue - LED 9 always lights up regardless. And when it needs to really light up, it just becomes brighter. As if there is some internal leakage on pin 11. I have tried two LM3914 chips, both the same behaviour. Weird.

Edit: no problem. I had grounded pin 9 to stop some other problem. Removed the link and now it works properly. As you can see I have used 3 LEDs and arranged them so that they come up at the correct voltages (you can see the links between them). I also removed the op-amp because it was pulling 4.5mA at best, and used a simple transistor (the square one with a hole at the middle right).

 

I attached my schematic using the LM3914 to monitor the battery voltage.

I use a LM7815 to get 15V from the 26V battery instead of using op-amp.

I will write a description on how my circuit works tomorrow as I have some other works to attend to. Read through my schematics and we shall discuss it later.

cheers

Allen

OK, here is a brief description of my circuit.

Based on your calculations, you want your Battery voltage to be halved before sending to the circuit for level measurement. The voltages of your interest are:
24.3, 23.0, 22.3 & 20.5 and after divided by 2 they become 12.15, 11.50, 11.15 & 10.25

Since my LM3914 is operated at a supply voltage of 15V, the voltage to be monitored falls in the range 10V to 13V. But I made it to monitor from 9V to 13V so there is some allowance on the lower range. Using a voltage divider, I would drop 2V between 15V Vcc to RHi, 4V between Rhi and RLo, and 9V between RLo and GND. As you've noticed from the datasheet, there are 10x 1K resistors between RLo and RHi. So I am using this 10K resistance to drop 4V across RHi and RLo.

That's how R4 = 5K drops 2V Internal 10K drops 4V and R1 = 22.5K drops 9V. There are 10 comparators in the LM3914. Since we are dropping 4V in the window of comparators, so 4V / 10 = 0.4V. Each output would be corresponding to a potential difference of 0.4V on the input window. This is like zooming the range of voltage of interest (9v-13v) and display them on the LED display. And that's how the output LEDs are not in running order as the voltage change between RED and ORANGE LED is 11.15V - 10.25V=0.9V which is 2 times greater than 0.4V.

Whether it is advisable to common the outputs together to one LED, that I am not too sure. I couldnt find any reference to this. But judging from the internal block diagram, I think it should be OK as the comparators are usually having open collector outputs. The datasheet also mentioned that it is a good practice to have a 2.2uF cap between the V+ of LM3914 and GND so I'll leave that up to you.

That very much wrapped up the description of how the circuit works. I am still studying how to get a reference voltage of 13V from the VRO and ADJ block.

cheers,

Allen

 

OK I have an issue - LED 9 always lights up regardless. And when it needs to really light up, it just becomes brighter. As if there is some internal leakage on pin 11. I have tried two LM3914 chips, both the same behaviour. Weird.

Edit: no problem. I had grounded pin 9 to stop some other problem. Removed the link and now it works properly. As you can see I have used 3 LEDs and arranged them so that they come up at the correct voltages (you can see the links between them). I also removed the op-amp because it was pulling 4.5mA at best, and used a simple transistor (the square one with a hole at the middle right).

That was awesome. Looks like you have solved most of your problems. If you read the datasheet carefully, outputs 9 & 10 are a bit different from the others. The LM3914 allows 10 pieces of themselves to be cascaded to form a chain of 100 LED display so I think that's the reason why.

cheers!

Allen

 

The current consumption is as follows:

4.5mA for the LM3914 standby
5mA for each/one LED
4.5mA for the 7815

Total is 14mA at all times.