Forgive me if this question has been asked before, but in a world of automation where things are replaced instead of repaired, I want to know why car batteries are still living in the 1950s. Let me explain. In the last two months, I have two vehicles die suddenly at very inopportune times, both due to a battery that had seemingly worked fine numerous times throughout the day, but then randomly died. Once upon a time, it seemed that cars would barely start and take a while to turn over to give you at least a little warning death was coming. Modern car batteries no longer seem to do that. They just go from working 100% to dead over the course of a few hours based on my personal experience.
I am aware some fancier vehicle have an integrated system that tells you when your battery is low, and certainly any repair shop and auto parts store has a battery tester…but what the hell? Propane tanks seemed to exist for decades without a gauge that told you how full they were, but all modern tanks seem to now have those gauges integrated into the nozzle because around 15 years ago, someone got a clue and said “what the hell are we doing?”. So why has this not happened with car batteries? My older Toyota and Honda do not have these fancy sensor systems, and while I could go to a repair shop/auto parts store, they are in the business of selling new batteries even if they are not needed, so there is the added anxiety of not knowing if they are telling the truth. So teaming millions, why is there no simple gauge built into a modern car battery that lets you see at a glance if you are about to be screwed or not? That way you don’t take a long road trip, dress up for a night out, of do a thousand other things only to have your trip ruined by a battery that left you stranded and needing a jump when that might not always be possible or convenient.
Because, there is no simple way to measure the “health” of a lead-acid battery.
The state-of-charge isn’t that important - that’s easy to control, with a decent charger. The problem is determining if a cell in the battery is about to fail, and that isn’t easy to do.
@Beowulff - I would buy that except for the fact that auto parts/repair shops have a device they use to determine if the battery is low and should be replaced (assuming they are being honest). Some vehicles have that sensor system too. Yes, you cannot detect every little flaw in a car battery that could cause it to fail, but if it is basically a chemical reaction that is creating the charge, isn’t there some way to say “hey, the chemical is running low” or “hey, this built in self test isn’t working right all the time, which means your battery is going to fail soon based on a historic baseline the battery is no longer meeting”. I refuse to believe that my car battery operates on witchcraft that makes it impossible to tell when it is about to die if not for some fancy testing equipment. Build that into my damn battery and just charge me $5 more for it.
In-store car battery testers check one (or both) of two things:
idle no-load voltage
heavy (fixed) load voltage
Neither condition is applicable if the car is operating.
And your understanding of what constitutes an unhealthy battery is naively simplistic. An end-of-life lead-acid battery has chemistry comparable to a fresh battery of similar charge level. The primary internal symptom of a ready-to-die battery is sulfation – sulfur from the sulfuric acid electrolyte sticks to the lead electrode plates, displacing lead stuck in solution and preventing clean discharge current or complete recharging. As far as I know, there’s no external “measurement” for that, other than the kinds of checks I mentioned earlier that simply cannot be done while the car is in operation.
I didn’t say it was impossible, I said it wasn’t simple.
How long does that test take? In my experience it, takes 15 or more minutes. The tester first charges the battery to a known state-of-charge, and then discharges it though a known load. A car isn’t a constant, static load - the amount of current the starter draws will change based on the temperature, the type of oil, and the wear on the engine. So, your battery health meter would have to take all that into consideration. You could probably get a approximate answer just by measuring the voltage droop as the starter was cranked. But, I would expect a lot of false negatives - batteries tend to die very suddenly (at least they do out here). On Monday they start the car just fine, and on Tuesday - “click.”
It’s not witchcraft, it’s physical principles. The testing equipment that does what you want goes for several hundred dollars, and is not feasibly integrated with the battery. Your notion that the problem can be addressed for $5 is, sadly, not in any way realistic.
If you had an economical way to monitor battery voltage and current, and you took the time and effort to build up a database of voltage/current time response signatures for batteries in various states of health you could use prognostic techniques to estimate a mean time to failure of the battery. You would likely need a database for every car/engine/battery combination you make.
Just not worth the effort for a failure mode that occurs 2-3 times in the lifetime of the car (unless you live out here in the desert, in which case it could be 5-6 times) and is generally not a safety issue when it does happen (unless your battery dies when you stall at that railroad crossing).
What they really need to do is get rid of that sensor which measures your anxiety level so that the battery fails at the most inopportune time - the same sensor that controls things like copier jams.
Us golf cart enthusiasts live and die by the battery. I have a NOCO Genius Mini 3 charger that analyzes and charges a battery in 8 stages:
Analysis. Can detect any kind of lead acid battery, regular, deep cycle, AGM, gel, etc
Diagnostics
Recovery. When detected, it applies a high-voltage low-current pulse charge causing battery sulfation to convert back to the active material, regaining lost battery capacity.
Initialize
Bulk charge
Absorption charge
Optimization and finish charge
Maintenance
My vintage Melex golf cart came with an extraordinary set of batteries, (6) 6 volt Trojan Gel Cells that cost $315 each. When I bought it, they were in dire shape. The previous owner used the old fashioned buzz box charger, not suitable for gel cells. Cables were poorly connected to the point that one battery ***melted *** around the terminal. Another battery only read 4 1/2 volts, a sure sign of death. I charged it, after sitting a week it was back to almost nothing.
I bought the Mini 3 Genius. After only a few charge cycles (which originally took 2 days each!!) all batteries are back in action.
I am convinced that a $5-$10 device (in small quantities) to provide the health status of a lead-acid car battery is feasible and relatively easy to make.
Such a device needs to measure the instantaneous current and voltage of the battery together with the time acquired and the battery temperature, save them into some non-volatile memory and analyze them once in a while.
The device needs to be initialized with the battery AH and type (99.9% of car batteries are of perhaps 5 types).
There are a few parameters that indicate the health of a battery:
Obviously idle voltage shows state of charge.
Change of voltage vs. current shows internal resistance.
Change of voltage with time when a large current is drawn shows sulfation amount.
Ability of battery to hold charge (indicative of sulfation, electrolyte quality and level) is also shown by coulomb measuring - integrate charge & discharge over time and compare with observed charge level.
All this is achievable with a good Hall sensor for the current measuring and reasonably good sensors for voltage and temperature. And of course a small processor with on-chip flash. An ESP32 module will provide all processing power and A/D needed PLUS Wi-Fi and Bluetooth connectivity to be controlled via a smartphone for $3-4 in single units.
The thing is that even $5 is a lot for standard car equipment. Average industry EBDITA are around 8%, so such an addon will need to add about $60 to the average car. Features of cars are major decisions, and most people experience battery failure as a common, easily figured out event, happening once every 4-5 years. That is, if your car has trouble starting and the battery is over 4 years old, just change it.
Now, an aftermarket manufacturer could sell this for, say, $50-$100 after investing maybe $200K in development and make some money. I would be skeptical about this being a big hit.
Yes, but those are deep-discharge batteries that are charged over many hours. The charging circuit on a car has very different priorities. The battery outputs hundreds of amps for a few seconds when you start the car. Then the charging circuit needs to charge the battery back to full charge fairly soon. There’s no time to do any analysis and conditioning of the batteries, because the driver may stop the car after a few minutes and expect to be able to start the engine again.
A cheap digital thermometer costs more than $5. A cheap voltmeter costs more than $5. Ammeters cost more yet. Add memory and an analysis function and I think you’re dreaming about the feasibility and the price range.
No, it does not. It shows alternator output. Rest voltage is indicative of the battery’s state of charge.
I think you guys are making it more complicated than it needs to be. The load test is still the definitive measure of an automotive start battery, this could be done a couple different ways, but manufacturers aren’t going to include that in the vehicle, they just aren’t.
An accurate voltmeter would probably work for this during starting if the operator knew what to look for. AC conductance testers are also fairly good at spotting defective batteries as well, this might work. Better ones calculate the CCA, over time a battery will still be serviceable but the capacity is reduced. It’s on its way out. I’ve always thought that an integral AC battery charger would be an excellent accessory as well. It pays to spend a little time every few months at least checking out the electrical system and making sure everything is swapping electrons around as they should.
I’m talking about components not complete devices. An Allegra ACS758ECB-200B 200A Hall sensor is a bit over $3 in quantity of 5. An NTC sensor is in the single cents and volts are measured directly with the microcomputer. A good enough micro like STM32F103 is around one dollar and has more than enough power and memory and can communicate with the CAN bus. For wireless add another dollar. The $4 ESP32 is an overkill, but includes a 2-core MIPS-like processor, 4GB of flash and over 500MB of RAM plus 2 radios and an antenna.
PCB, housing and cables extra, but no more than $5 even in small quantities.
BTW, if you would order 1000 units from a China manufacturer you will pay around $5/unit.
I do these type of stuff and have the experience. Thing is, there are much more interesting things to do than a battery health device.
You are right, and I meant the same, but used the wrong verbiage. It was “idle” from the POW of the battery, that is neither charging nor discharging.
They intentionally undersize batteries in cars these days, partly to save on cost, and partly to save on weight since weight affects mileage. I personally always put in an oversized battery (in terms of cranking amps). Not only does this make the vehicle more likely to start on a cold winter day, it also makes the battery last a lot longer since it isn’t getting discharged as much compared to its overall capacity.
I don’t replace batteries every 4 or 5 years. In fact, I usually don’t replace batteries at all unless there is something wrong with the vehicle. I recently had to replace the battery in my old pickup truck, but then I also had to replace its alternator and starter and I had to fix some of its wiring.
CCA is important when exposed to extreme cold, like in Pennsylvania, but the way they get those high cold cranking amps, the batteries are made with more plates to increase the surface area of the sponge lead, consequently they are thinner and more prone to physical damage. If you don’t actually need 950 CCAs, it’s probably better in terms of longetivity to use a battery that has a higher reserve capacity that also meets the minimum CCAs for your location. A battery charger will pay for itself, hot weather is tough on batteries because they tend to self discharge quickly, short trips tend towards a chronically undercharged battery. They don’t last well under those conditions. You might look funny charging a battery in July, but this is when the damage is done, they only fail when it gets cold because they no longer have the current capacity.
Sounds plausible. But how much of an advance warning would this give? How long between a fault is detected this way (measuring current & voltage each time the car’s engine is started) and when the battery is too weak to start the car?
By the way, I don’t think you included the battery temperature sensor in your cost estimate… Sensor itself is cheap but now we’re talking about a box with 3 analog inputs, with one sensor to be stuck on the battery, one sensor to be clamped around the positive lead of the battery, a power connector and a data bus connector. That’s a lot of extra steps in assembly.
Let’s imagine the manufacturers do the work to develop the instrumentation and algorithms to measure battery health and predict future battery health.
Most drivers are real good about putting gas in the car within 50 miles of the “low range remaining” light coming on. If not, the car absolutely positively *will *strand them by the roadside within a hundred more miles. That’s an incredibly useful teachable moment. Even the most fact-resistant airhead will learn to respect that light after a couple strandings.
Many cars now have lights or messages that come on to say “oil change due”. How many people schedule that work the very same week? My bet is <1%. After all where’s the immediate feedback that harm is being done?
Given the vagaries of battery measurement as so well explained by others above, any battery health meter is either going to fail to alert to a deteriorating battery in harsh conditions, or frequently cry wolf in benign conditions. Consider that car batteries last 1/3 to 1/2 the life of any given car. Any given driver will only experience 1 or 2 teachable battery moments before they get a different car.
As a result of the above, the driving public at large will not learn to value or respect the new warning device. Even if some tiny number of gearheads and/or worrywarts might.
It happens that my wife’s car needed a new battery just yesterday. It doesn’t get driven much or for long, the battery is ~5 years old, and in the last several months has cranked very weakly a few times then recovered after a long drive. Then it was stone dead a couple days ago. As a former motorhead I wasn’t surprised at all that it happened, just looking at the age & history. I was surprised it was that particular day. It had cranked strongly and started fine that very morning.
So it’s off to the local national chain car parts store. $150 later I have a new mid-grade store-brand 600 CCA group 34 battery. IOW, I got out of there about as cheaply and generically as possible. Batteries for my other car cost over $1000.
I don’t know too many drivers who’d choose to go drop $150 (much less $1000 :eek:) on the strength of their car’s battery health-o-meter changing from “fine” to “likely to fail sometime within the next 6 months”. Which is exactly what my experience and intuition told me was the state of that battery three days ago.
The only way that might work is if the manufacturers will be able to piggyback on the car computer, UI, and sensors for an added cost of a couple of dollars.
BTW, in most cases, a dying battery will announce itself during cranking with a tired cranking noise which most people recognize.
Also, I would recommend to keep such a device in the car, charged, it is useful not only for jump-starting but for many other purposes.
Agree. And perhaps this could be a way for the car’s computer to prognosticate the failure of the battery (or a bad electrical connection at the battery). Each time the car is started, the computer would measure engine RPM when the starter is engaged. The computer would throw an error code when the starting RPM is below a certain threshold value. The algorithm would also compensate for temperature. (The starting RPM will be a little bit lower in a winter climate due to thick oil.)
There seems to be a general belief that modern car batteries die suddenly, whereas in the past you got some warning from weak cranking. I suggest this may be because modern engines almost invariably start at the first attempt.
Thus the number of crankings the battery can perform will drift downwards with no-one the wiser until it falls from one to zero, with no warning.
