# IC that can show the amount of battery charge without the need for a microcontroller

I have a lithium-ion battery pack that I want to show the charge level for example, by turning on 4 LEDs. Now I'm looking for guidance on how to do this. I'm told that I may not ask re specific IC's on this site - so whatever guidance is acceptable will do.

I don't want to use a microcontroller and I want it to be very simple and cheap ( using just an IC), just turning on LEDs is enough, but do this with high accuracy. (Well, the amount of remaining charge depends on the current, temperature, etc., and on the other hand, I don't want to know the remaining charge with an accuracy of one percent, because I only display it with 4 LEDs, but if it is in a way that the dependent parameters considered its great .). It is like power banks but just monitors the state of charge, not the charge and discharge process.

I have two methods in my mind

1. Using a suitable IC like:

1. If it is possible, for example, by measuring the voltage and current in one moment find out how much charge is in the battery. I don't know if it is possible or not.
• ("...do this with high precision"...: I assume accuracy, since I doubt high precision saying 25.000% when the battery was just fully charged would be appreciated. So you will need a thorough set of lab notes, with respect to the specific battery pack. And depending on just how high, you may need to repeat those every so often. And at a variety of operating temperatures. And then provide all of that here so someone willing to be sniped by the question can put in requisite time to design this analog system and somehow also achieve cheap, too. I will just enjoy the show, myself. Commented Dec 17, 2023 at 9:12
• Engineers need specifications and numbers. What numbers do you mean with words like "simple", "cheap" and "high precision" as they are generally mutually exclusive, and how many LEDs or what is even the pack voltage or capacity? Commented Dec 17, 2023 at 11:34
• @periblepsis Thank you for your attention. I clarified the question and added some information, but in general, I tried not to go into too much detail so that it would be easier to find the IC in question. Commented Dec 17, 2023 at 13:32
• @hosseinhamzekhani A VITAL question: Q1 How many of these do you wish to produce? 1? A few, 100's, 10,000s ? - it makes a major difference to the answer. || Also Q2 - what is your target cost for materials? Q3 Is labour important to you? (1 off - no. 10,000 - maybe) . Q4 Will this need to be certified by a regulatory body ? Commented Dec 19, 2023 at 12:14
• I have moved the Q&A philosophy Comments to chat; please do not continue the non technical discussion here. || I have left comments which are mainly related to the subject. || Please keep comments here related to technical issues. The chat version can be used appropriately if desired. (I mean to do the above yesterday but found the day only had 24 hours). Commented Dec 19, 2023 at 12:15

A microcontroller is liable to be among both the easiest and cheapest solutions. You need to specify an acceptable price.

You can buy "gas gauge" ICs that track current in an out of the battery and calculate residual capacity.

You can get a rough but useful indication of battery capacity from battery voltage, with some adjustment for temperature. Load may make a significant difference. A look at the discharge curves for the cells of your choice will show capacity against voltage at a range of load levels. 4 levels would be 'easy enough' with a resistor ladder and comparators. More than 4 starts to make a microcontroller very attractive.

4 LEDS may show as few as 5 capacity levels (if all off is used) or 16 if a binary code is used, or some intermediate number if some somewhat intuitive coding is used.
eg 0000 0001 0011 0010 0110 0100 1100 1000 <- 7 (or 8 levels with 0000) with 1/2 steps shown by two steps lit. Our current washing machine uses such a system - it's very understandable once the concept is realised. BUT a microcontroller will probably be lower cost.

Here is an example of an IC based solution using an LM324 - an LM339 could be used with some changes. This uses 4 x 100 Ohm resistors to create 4 voltage points and a 3V3 Zener reference. Very crude indeed and would need values adjusting. Apart from resistors and LEDs cost is 1 x LM324 IC = $US0.70/1 at Digikey. (10 cents from LCSC in China , probably). The 40+ year old LM3914 IC will do the basics of what you want . Superb Sparkfun tutorial here. Many examples of its application here About$US0.50 in 10's here plus a shipping delay.

This page Lithium cell charge level indicator describes monitoring a single cell. This can easily be extended to a multiple cell pack.

This combined LM3914 data sheet and application note provides many examples. The diagram on page 8 shows how it works in one mode and COULD be used to build your own - probably at a lower cost in some cases.

Note that the basic LED bar/dot drivers do not have temperature compensation, battery characteristic curve compensation, ... . This can be added with extra effort. A small microcontroller could do the same job with more frills, the ability to tailor the results in software and at a lower hardware cost. As TonyM pointed out in now deleted comments, the development costs of any design need to be taken account of, and performance verification may be harder with a microcontroller solution than with hardware.
That said, I consider that for this application a microcontroller is able to provide a very satisfactory solution, with more capability at less cost*, with due care taken. [*less cost: in small volume with no labour and design costs OR in very large production volume, all accounted for.]
I have produced hardware solutions made in 100,000+ volume that did not use a microcontroller for "good enough reasons" but would have been better and possible cheaper if they had used one. That is, I'm aware that a valid design choice may choose either method, depending on other factors.

• The half-step sequence (0000 0001 0011 0010 0110 0100 1100 1000) is an interesting way to increase the number of "levels". Especially if binary coding is not reasonable for the target audience. Commented Dec 19, 2023 at 10:48
• Downvoted, I'm afraid. To start with, it doesn't answer the question within the parameters stated, it contradicts them and pretends an MCU is OK. That's for comments only. Second, an MCU is not among the easiest and cheapest solutions. Unfortunately, and as happens so often on the site, this only considers the component cost and has no consideration for development or approval costs. As home project, it may be quick, may take long, depends if one can program. As a professional design: much longer and so much more expensive. Depends on the business you're in. (...contd) Commented Dec 19, 2023 at 11:14
• (contd...) A few will let you knock out source files, dump it on a server, get it on a board and walk off. Others (I'd say the majority) will need a decent docs for design, build and testing, plus testing that doesn't just show that it works (fairly easy) but that it never doesn't work (proper and harder). So that's to be amortised into the part cost. There's a wide range of industries and situations OP could be in and should be advised of in an answer. But again, the answer's trying to redefine the question. That's for initial comments, discussion and an OP-revised question first. Commented Dec 19, 2023 at 11:15
• This has been fixed so upvoted, thanks. But I think it's good that my comments here remain, as they detail the point for future readers. Commented Jan 26 at 9:52

Just a simple search such as https://www.google.com/search?q=battery+charge+level+ic might find some interesting results such as the TI BQ2050H

I suspect including battery chemistry in the search terms would help.

I would try lots of variations of the search terms and then use what is learned about useful terms in the parametric search facility in an electronic components distributors website.