[cparke]

This has been going on a long time, that's true, and the car is still starting and running (and I'm still standing) too!

The battery analyzer was connected under the hood at the charging posts. That's a junction point between the line to the battery and the line to the starter. While not as good as on the battery itself (which I will use next time), it's what I have so far.

I'm also skeptical of the starter/solenoid too. Especially since the previous starter seemed to start failing in this same way too, though it took a year for that to actually happen. In my imagination, it seems to me like the starrer in certain positions is shorted, but the short is small enough for now that it can still turn itself out of that position and crank. And with a fully charged new battery with maximum CA, it moves out of there very quickly, but at lower CA from a 2 year old battery, it takes longer to start and could burn out. The starter motor of the previous starter did burn out, solenoid was fine.

But this is all just my personal theory. That's why I wanted some input here if starters or batteries really can act like this, because I really don't know how they fail or if my theory is credible. If the starter can last longer on a new battery, that would be preferable, of course.

[GSB]

Originally Posted by cparke

The battery analyzer was connected under the hood at the charging posts. That's a junction point between the line to the battery and the line to the starter. While not as good as on the battery itself (which I will use next time), it's what I have so far.

I'm also skeptical of the starter/solenoid too. Especially since the previous starter seemed to start failing in this same way too...

Ah, yes, this explains a lot, and I'd suspect the starter first then, based on the symptoms AND where you measured the voltage with the analyzer. While cranking, the battery voltage will read lower and lower the closer you approach the starter terminals. Add a short to the starter motor, and that voltage drop gets worse.

The battery itself has internal resistance and the cables and terminals coming all the way from the trunk have resistance too, so the voltage drop across them will be very significant.

As an example, if the starter pulls 120A, it means the starter windings have a resistance of only 0.1 ohms!! The cables and terminals coming from the battery could easily measure 0.1 ohms too. Therefore, the voltage directly across the starter could measure HALF the battery voltage, only 6V during cranking.

Add a short to the starter, reducing its resistance even further, the voltage across it will drop even more.

[cparke]

Originally Posted by GSB

The battery itself has internal resistance and the cables and terminals coming all the way from the trunk have resistance too, so the voltage drop across them will be very significant.

Yes, of course. That's why most automobile manufacturers put the battery under the hood, not in the trunk. The low guage sized cable needed to bring the high current to the starter is very expensive too!

But, I think you're forgetting what are we testing? The point is on normal starts battery dropped to around 9.50 volt, which is normal. But on the slow crank start, it fell at this same measuring location to 7.99 volt. True, the voltage at the battery itself may have been a bit higher, but there's a huge difference in the voltage minimum between these two starts.

I did a few measurements at the battery in the trunk today. No slow cranks today, but even with the normal cranks, the voltage minimum at the battery during starting has varied from 9.30 volt to 10.45 volt. That still looks like a pretty big spread to me, but is it normal? The alternator output, on the other hand, is fairly consistent between 13.7 volt and 14.7 volt while the engine is running and when driving, including with headlights and all accessories on, so I don't see how the battery isn't getting enough charge from the car itself. The car is driven daily.

The fact that I don't get any slow cranks for a few days or a week or two after recharging the battery, but once they start happening, they can repeat more often, and I'm seeing Low Battery warnings on the dash occasionally, is a sign to me that the battery is involved here by not putting out enough CA for some starts at a certain point. The varying voltage drop amount when starting is still a mystery, and an intermittent short in the battery or starter is the best explanation that I've heard here.

[Tambohamilton]

Normal charging voltage is around 14.5v...im surprised yours varied so much, down to 13.7v...

[GSB]

Originally Posted by cparke

But, I think you're forgetting what are we testing? The point is on normal starts battery dropped to around 9.50 volt, which is normal. But on the slow crank start, it fell at this same measuring location to 7.99 volt.

I'm not forgetting this at all. That's why I added this...

Originally Posted by GSB

Add a short to the starter, reducing its resistance even further, the voltage across it will drop even more.

An intermittent short in the starter motor is very possible, which would overstress the battery, drop its voltage even further, and damage it.

Originally Posted by cparke

...even with the normal cranks, the voltage minimum at the battery during starting has varied from 9.30 volt to 10.45 volt. That still looks like a pretty big spread to me, but is it normal?

A certain amount of spread will be perfectly normal. But I don't know how MUCH is normal. If you measured the voltage and current draw with an oscilloscope instead, you'd see it continuously varying, with massive spikes as the engine compresses the air in each cylinder, and spikes from the starter motor as the brushes and commutators constantly switch windings. The average voltage and current will also vary, depending on the load, the temperature, the state of lubrication in the engine, the length of time that the engine is cranking, the health of the battery, and so on. Shorts in the windings or commutators of the starter motor will certainly make this worse too.

Originally Posted by cparke

The fact that I don't get any slow cranks for a few days or a week or two after recharging the battery, but once they start happening, they can repeat more often, and I'm seeing Low Battery warnings on the dash occasionally, is a sign to me that the battery is involved here by not putting out enough CA for some starts at a certain point. The varying voltage drop amount when starting is still a mystery, and an intermittent short in the battery or starter is the best explanation that I've heard here.

Yes, both the battery and the starter could be damaged. Ideally, I would replace both. You could start by replacing the battery, but I'd be nervous that a bad starter could damage the NEW battery too.

[GSB]

Originally Posted by Tambohamilton

Normal charging voltage is around 14.5v...im surprised yours varied so much, down to 13.7v...

Yes, this also seems a bit concerning. A bad battery could be overloading the alternator and regulator too.

[GSB]

Originally Posted by cparke

No slow cranks today, but even with the normal cranks, the voltage minimum at the battery during starting has varied from 9.30 volt to 10.45 volt. That still looks like a pretty big spread to me, but is it normal?

Another thing that suddenly occurred to me... The voltage readings you're seeing may not be that accurate. It only takes a second or two for the car to start, and that's not enough time for a digital multimeter to lock in an accurate reading, especially while trying to average a highly spikey, wildly varying waveform. A meter is designed to average a regular, predictable waveform like a sinewave. What's happening on a battery during starting is anything but regular or predictable, and is too short-lived.

Therefore, the longer the starter cranks, the lower the voltage reading will be, because the meter has more time to average AND because the battery is being depleted.

[cparke]

Originally Posted by GSB

If you measured the voltage and current draw with an oscilloscope instead, you'd see it continuously varying, with massive spikes as the engine compresses the air in each cylinder.

The battery analyzer tool that I am using isn't an oscilloscope, but it is reading and recording voltage around every 50 milliseconds or so (about 20 frames/second). In fact, I gave this device a test run first on a different car which has no issues, and it seems like you maybe just answered my question posted in a neighboring forum about why I was seeing high voltage spikes in the data from that alternator when the car is stopped and the RPM's are low: https://mbworld.org/forums/gle63s-gl...r-problem.html

Now, with the alternator on the BMW that we're talking about here, with no slow crank issue during the recording, I've attached a similar voltage graph for a 20 minute drive which included local and highway drive. The device was in the trunk connected directly to the battery. As with the other car, I think the high and low jumps are happening when the car is stopped at a light and idle (though it certainly is not as clean). The running voltage of the alternator is therefore mostly around 14.3 volts once it gets going, if we discount the spikes as spark plug activity. Does this graph seem good and reasonable for normal vehicle operation?

FYI - on the attached graph, the battery voltage dropped to 9.51 volts when starting, and the highest voltage spike reading is 14.99 volts. On subsequent starts during this same trip (all normal cranks), minimum voltage during starting (in order) was 9.10 volts, 10.28 volts, 9.66 volts, 10.07 volts, and 9.56 volts---that's the spread that I was talking about. With the slow start I captured recently, the low voltage was 7.99 volt, but I did not know (or think about) the wave recording feature at that time.

I will keep recording again until I finally capture a recurrence of the slow start situation that we're really interested in seeing. I think this device has good enough accuracy in its data, and an oscilloscope isn't necessary for these purposes and would just make the graph harder to understand.

[GSB]

Originally Posted by cparke

The battery analyzer tool that I am using isn't an oscilloscope, but it is reading and recording voltage around every 50 milliseconds or so (about 20 frames/second)...
Do this graph seem good and reasonable for normal vehicle operation?

I think this device has good enough accuracy in its data, and an oscilloscope isn't necessary for these purposes and would just make the graph harder to understand.

Fascinating! Yes, this device is good enough, and provides an awesome visual reference.

The spread you describe still sounds perfectly normal to me, though your overall battery voltage (and therefore, capacity) may be a little lower than it should be, and may be why your DME is reporting low battery.

The spikes in the data while the engine is running do not surprise me at all. Hopefully, the device was in the trunk for this. Otherwise the device could be picking up all kinds of radiated noise/interference from the high-voltage coils and spark plugs discharging. Those radiate huge radio-frequency spikes that most equipment has a hard time filtering out. A simple oscilloscope would pick up even larger, sharper spikes. Smaller spikes (the noise or fuzziness) will come from the alternator and rectifiers too. All normal.

[cparke]

Yes, I said this was recorded at the battery itself in the trunk. But as the battery also functions as a capacitor to reduce the line noise that you talk about not sure why you think there would be less noise there than elsewhere in the wiring. Voltage drop alone does not have a dampening effect as far as I am aware. With the battery and sensitive electronics exposed to so much noise, really makes me wonder why they don't put a real capacitor also on the separate line to the fuses and computer modules.

[GSB]

Originally Posted by cparke

...as the battery also functions as a capacitor to reduce the line noise that you talk about not sure why you think there would be less noise there than elsewhere in the wiring.

There is definitely less noise across the capacitor itself because of the resistance in the long wires to the source of the spikes. The spikes become bigger and bigger, the further you are from the capacitor.
Furthermore, wherever there are sparks or inductive loops (coils, motors, alternators, loops of wiring), these all radiate noise through the air as well. The further you are away from them, the less the noise will be. And the better the metal shielding around your equipment (in the trunk, for example), the less the noise will be too.

Originally Posted by cparke

With the battery and sensitive electronics exposed to so much noise, really makes me wonder why they don't put a real capacitor also on the separate line to the fuses and computer modules.

They place capacitors and filters inside each of the computer modules to prevent outside interference. In fact, every single chip inside the module needs to have its own dedicated capacitor (right on the pins of the chip) for the same reason mentioned above for the battery capacitor.

[Konrad5288]

So at this point why not just go buy a new battery and see what happens. Or go to a junkyard and grab a $30 that tested good and try it out. As long as you register it correctly doesn’t matter if it’s lead acid or AGM.

Since you like running and analyzing data buy one of those OBD2 scanners that have a battery tester and test your battery a few times a day over the course of the week to see if your battery registers “good” every test. Pretty sure the tester sends the battery a load and analyzes the batteries response to the given load. And if you want to be cheap about it Return the tester afterwards, saying that it was faulty. the battery tester didn’t work correctly. When you took the battery to AutoZone it was bad even though that OBD2 device kept saying your battery was good

I’d never heard of the small car battery shorts before this thread. I didn’t realize a car battery could have intermittent shorts. I always thought the voltage would read lower with any type of fault in a car battery. Or higher. But typically lower. That might be some useful information in the future.

[GSB]

Originally Posted by Konrad5288

I’d never heard of the small car battery shorts before this thread. I didn’t realize a car battery could have intermittent shorts.

Yeah, intermittent shorts in a battery are rare these days, but they can happen due to impact or other factors, and it may only affect one or two cells. Settling of sediment is more of a soft short. Either way, the battery voltage always drops. Voltage readings over 12.7V usually indicate charging in progress or spikes from inductive components connected to the battery. A temporary over-voltage situation could exist if the battery has been over-charged, and that is damaging to the battery for sure.

[gbalthrop]

Originally Posted by Konrad5288

... I’d never heard of the small car battery shorts before this thread. I didn’t realize a car battery could have intermittent shorts. I always thought the voltage would read lower with any type of fault in a car battery. Or higher. [Now THAT would be "Remarkable".] But typically lower. That might be some useful information in the future.

NOT sure there is any "useful information" in this thread, Until/ Unless OP indicates that his Starter Motor has finally failed (current symptom: voltage drop during cranking > 2.5V), OR that he has found a loose/ corroded B+ Cable.

ANYONE ever seen Battery Voltage, battery NOT being charged, > 12.8V, EXCEPT for "Surface Charge"?

When battery is being charged, whether by Alternator with engine running, or by external Battery Charger, the Voltage Value measured at battery posts is ~ 14V, or more. If you disconnect charger or turn off engine and immediately measure Battery Voltage at battery posts, you will see value ~ 13.0+V, which is "Surface Charge" which dissipates to a reading closer to 12.7V after 15 to 30 minutes.

But don't "Believe ME", do your own TESTS, and REPORT your findings.
George

[GSB]

Originally Posted by gbalthrop

ANYONE ever seen Battery Voltage, battery NOT being charged, > 12.8V, EXCEPT for "Surface Charge"?

Nope. Except in a case of abusive charging at a much higher voltage, which also resulted in catastrophic damage.

An intermittent slow-crank, non-start condition that miraculously cures itself on the very next crank is fascinating. Yes, it can only be the battery, starter or connections.

You're right that every thread has very little "useful information" until a solution is posted. But everybody's experience can help narrow down the solution. That's what a forum is for. Any other "useful" insights are welcomed.

[PhaseP]

Originally Posted by Tambohamilton

Normal charging voltage is around 14.5v...im surprised yours varied so much, down to 13.7v...

ECU controls the alternator voltage output in these cars, by a wire that goes to the alternator's voltage regulator. And ECU can decide that battery had enough charge and doesn't need too much charging and tell voltage regulator to reduce its voltage. Or it can decide that ambient temperature is high and tells it to reduce the charging voltage also. 13.7V is not an indicator of problem by itself.

The voltage drop spread OP seeing between different startup is also expected, as long as they each had different speed of cranking. The starter is a dc motor, as it spins generates its own counter voltage. (Counter EMF https://en.wikipedia.org/wiki/Counte...romotive_force)

The current flowing through starter is equal to (V battery - Counter EMF it generates)/resistance of its windings. Slow spinning starter will generate low counter EMF which will result in high current draw. Fast spinning starter will generate high counter EMF which will result in less current draw.

High current draw from battery will cause voltage drop from it because of its internal resistance.

I would say the starter is likely the issue, but seeing battery voltage warning on the dash is from my experience about the alternator not charging the battery.

[GSB]

Originally Posted by PhaseP

I would say the starter is likely the issue,

Likely in this case.

Originally Posted by PhaseP

...but seeing battery voltage warning on the dash is from my experience about the alternator not charging the battery.

Not likely in this case. The posted monitoring graphs show proper charging voltage.

I had an 11-year-old battery in my 335i operating at only 10.5V after a full charge with full cranking capacity. Nothing wrong with the alternator, just one bad cell. Charging on a fancy charger made no difference.

[cparke]

After missing getting a recording of a slow start incident this past Wednesday (although I did get a Low Battery indicator using radio briefly with engine off later that same day that was recorded), I finally captured the episode this past Sunday! Even better, it includes a continuous recording from the prior start and drive as well!. My only regret is I didn't keep the recording going and do an intentional third start right after, as these incidents make me nervous that car won't start again after I just had it running.

So we started off Sunday evening with the car having been parked 2 days straight, and the battery's resting voltage having dropped to 11.89 volt. This is certainly low for no apparent reason, but car still started just fine! (despite battery voltage dropping to 8.02 volt). While parked for about an hour after that initial 15-20 minute drive, battery resting voltage was mostly stable at 12.67 volt. Seems like car charged the battery back up pretty good in such a short time. On the following start, voltage dropped only to 9.44 volts when starter was first activated.

The next 2 starts (same evening, see below) is where it gets interesting. First start battery, dropped from 12.43 volts to 9.01 volts while starting. Next start, it dropped 12.32 volts to 7.45 volts!

Overall Graph:




First start (normal)




Second start (slow, and very low voltage!)




The two close-ups of the starts are scaled to the same timeframe so they are comparable, and I notice even the apparently normal start was really just "stronger" but did seem to be cranking almost as long as the slow start. As mentioned and show, the battery voltage before the start was the same before both of them, and certainly not even abnormally low.

So I don't know what to make of all of this. As mentioned previously, the slow start has never been the first start for the vehicle of the day, and my experience is even if I shut the engine down immediately and re-start, the following start will be a normal start, not another slow start. The following day after the recording incident above, resting battery voltage before starting was 12.28 volts, and the low voltage during three normal starts were 9.14 volts, 9.06 volts, and 8.85 volts. While a simultaneous ammeter recording of the starting sequence showing how many amps are being pulled by the starter with these voltage drops would be nice, I don't have equipment that can do that presently, and is it really going to be useful?

Can this really be just a failing battery? My battery analyzer measures the internal resistance at 7 milliiohms and gives a SOH around 21%. I also went into INPA and saw that the IBS is a bit more generous and rates the battery's SOH at 33%. Ovbiously the battery is 31 months old and not new. The new replacement battery (identical model) I got as a warranty exchange (yet still haven't installed it yet) reads an internal resistance of 3 milliiohms. Battery does seem to be self-discharging rather quickly, or IBS isn't charging it enough. If I recharge the battery overnight using my 2-amp smart maintenance charger, I didn't get even one of these incidents for 10 days! On the flip side, how could the battery alone vary in voltage drop so dramatically between individual starts? I have not had a single incident yet when the car failed to actually start during one of these slow start episodes.

Finally, people on this thread have been pretty confidant that a voltage drop below 8 volts during starting must be an intermittent short somewhere. I am concerned these starting episodes are shortening the life of my starter. A bad starter could damage the new battery, while continuing to use a bad battery could damage the starter. If this is actually the battery, am I safe squeezing more life out of it by charging it weekly, or should I just install my battery exchange instead and hopefully not worry about this for another 2 years or so?

[GSB]

I would not be one bit concerned with those results. The battery drop you're worried about is an instant in time, when the starter motor has to overcome the inertia of an engine at varying rotational angles, and therefore, at differing stages of compression. The starter motor itself will come to rest at various rotational angles that will draw different amounts of current as it overcomes its own inertia, added to that of the engine. The spiky waveform at that point could be the brushes and commutators switching extremely low-resistance, high inductance windings inside, and perhaps cylinder compression loads.

The first revolution of the starter motor will draw very different instantaneous currents (with different waveforms) each time. A high-current scenario will pull a known low-capacity battery to lower a voltage, but only for a moment.

I don't see any evidence of intermittent shorts in there, except perhaps, worn and tired brushes and commutators. Nothing about that starter (in its present condition) will kill a new battery, and there is no danger of an old, worn battery damaging the starter. EXCEPT that one day, the battery may not have enough capacity left to complete a start.

Leaving your tired battery on the maintenance charger is a great way to extend its life. My 11-year-old battery was still cranking healthily with a resting voltage of only 10.5V. Keeping it on a GOOD maintenance charger extended its life by another 9 months, but eventually, it did leave me stranded.

[cparke]

GSB - in my OP, I asked, "Does starter normally vary in how many cranking amps it needs from one start to the next? That could explain some of this."

Seems your response just above confirms this is a very plausible explanation for what I'm experiencing. The problem, then, is indeed the battery itself, not the starter or wiring or anything else. At times when maximum cranking amps are needed because of random positioning of the motors, the battery voltage drops to the extremely low levels below 8 volts and the starter doesn't get enough power initially and struggles to get turning.

So we're saying that the root cause is a failing battery losing its charge and cranking amps over time? I'm concerned these slow starts are taxing on the starter motor's windings, and could lead to burning the starter motor out from extra heat generated when it can't turn initially due to insufficient power, as well as eventually getting stranded and needing a jump start.

So my options at this point are either keep recharging the battery regularly, or change the battery? Still, why does it seem the battery can't be fully recharged by the alternator as well as by the maintenance charger? I know the drives aren't 12 hours straight, and the charging voltage (and amperage) erratic, but this setup is supposed to be the normal way the car maintains the battery. Which is why I wonder if IBS is not charging the battery aggressively enough, or the alternator not putting out enough amps? Recall that it's odd that I've never seen this happen on the first (cold) start, only after drive time and several consecutive starts, at times when the battery should have just been recharged!

[GSB]

Originally Posted by cparke

GSB - in my OP, I asked, "Does starter normally vary in how many cranking amps it needs from one start to the next? That could explain some of this."

Yeah, the initial spike at first turnover will always vary in size and shape, but once the starter motor is in full swing, the current will be fairly consistent from one start to the next. Your plots have clarified what we're really looking at.

Originally Posted by cparke

Seems your response just above confirms this is a very plausible explanation for what I'm experiencing. The problem, then, is indeed the battery itself, not the starter or wiring or anything else. At times when maximum cranking amps are needed because of random positioning of the motors, the battery voltage drops to the extremely low levels below 8 volts and the starter doesn't get enough power initially and struggles to get turning.

Exactly. Though I wouldn't rule out that the starter motor is becoming old and tired too.

Originally Posted by cparke

So we're saying that the root cause is a failing battery losing its charge and cranking amps over time? I'm concerned these slow starts are taxing on the starter motor's windings, and could lead to burning the starter motor out from extra heat generated when it can't turn initially due to insufficient power, as well as eventually getting stranded and needing a jump start.

Yes, extra heating in the windings would be a concern with prolonged slow starts, but that's not really happening here. Furthermore, the lower battery voltage at start means less power and heat in the windings. You'd definitely need to worry if the starter motor stalled, unable to turn over while still drawing power, but I'd imagine that the ECU would shut it down if there's no crank.

Originally Posted by cparke

So my options at this point are either keep recharging the battery regularly, or change the battery? Still, why does it seem the battery can't be fully recharged by the alternator as well as by the maintenance charger?

Correct. The alternator isn't charging long enough, and nor is it a true smart charger. Additionally, the car is drawing power from the battery even when shut off, so an external charger can compensate for that in an old, tired battery that no longer has capacity or endurance. Your battery has reached an elevated level of self-discharge too.

Originally Posted by cparke

I know the drives aren't 12 hours straight, and the charging voltage (and amperage) erratic, but this setup is supposed to be the normal way the car maintains the battery. Which is why I wonder if IBS is not charging the battery aggressively enough, or the alternator not putting out enough amps?

The IBS is not a smart charger. It commands rudimentary steps from the alternator to compensate for the age of the battery, protecting it from damaging overcharge more than undercharge.

Originally Posted by cparke

Recall that it's odd that I've never seen this happen on the first (cold) start, only after drive time and several consecutive starts, at times when the battery should have just been recharged!

The massive current draw on the first start, and on several consecutive restarts, will deplete the battery faster than it can recharge from the alternator at this point.

[PhaseP]

OP, your symptom of long difficult start when car has been driven with warm engine matches how my starter was behaving before it completely stopped working all of a sudden one day. The first time I had noticed starter was having issues with especially warm startups, I had bought a new starter to be on hand ready for replacement. From that time it had taken a year for the starter to finally not work at all and I had to replaced it. I had taken a part the solenoid and the starter motor of the old original starter to see what went wrong. One of the two brushes of the starter motor which is close to the engine when assembled was stuck in its track with baked and caked carbon dust from brushes and grime etc. The spring behind it was not able to push it freely down on to the commutator of the starter rotor to make contact. The solenoid contact plate, which is a sizeable copper plate had bowed out a little also, but the brush was definitely the culprit for its final death. It was stuck there.

My conclusion from this was that when starter got hot inside warm engine bay, it was somehow making it more difficult to make full contact either on the brush or the solenoid contacts or both. Something was expanding a bit to cause the difference.

Now I never had battery warning on my dash, but I was going through batteries every two years with that original starter last years before it died. Even when it was working fine at cold starts, it was cranking slower than it should. This actually wears out the battery. The starter draws a lot of current from the battery. That is why you see the voltage drops when starter is cranking in your measurements. That is expected. But a slow and long running starter will both draw more current at a given time and do it more. The slower it turns the more current it will draw (less counter EMF generated by it) This reduces the lifetime of the battery each time it happens slow by slow.

After I replaced my starter, which I was amused by how fast the new one was cranking the engine, I didn't have to replace my battery again. Still on same battery, don't remember how many years it had been, at least 5.

Finally, people on this thread have been pretty confidant that a voltage drop below 8 volts during starting must be an intermittent short somewhere.

As I wrote above the starter itself is like a short, it draws a lot of current. The starter rotor from my failed starter which I kept because it was such a manufacturing marvel to me, has single loop windings. Those are basically short circuits to the voltage source, except the counter EMF generated on the starter windings counters that to some level.

I am concerned these starting episodes are shortening the life of my starter. A bad starter could damage the new battery, while continuing to use a bad battery could damage the starter.

My guess coming from my experience starter is going bad, taking out the battery with it.