Theory and History of Alternators and Regulators

[joe_db]

(The start of this is copied from another thread. This seems to be a repeating line of inquiry and I decided it might do better having its own thread)

Let us start with first principles:
There are only two ways to regulate an alternator - fixed set point and variable.
A fixed set point regulator increases field current when the voltage is less than the setpoint and decreases it when it is over the setpoint.

The second part of the equation is how storage batteries act. You can look up their data and find an acceptance curve. At a given state of charge and a given voltage, they only accept so many amps.

Thought experiment number 1. You have a huge alternator that can putout 1,000 amps at idle and has a fixed setpoint. You have a dead 100 AH wet cell. It will NOT be recharged in 6 minutes! What will happen is the battery charge will follow the acceptance curve, which might be somewhere between 1/4 and 1/2 C, so in this case someplace between 25 and 50 amps. As the battery charges the acceptance rate tapers off. Instead of being fully recharged in 2 to 4 hours, it might get half charged in two hours and take 4-6-8 more hours or even infinity hours to get all the way back to 100%*.
Here is the tricky part - the higher the voltage the faster it charges, but the higher the voltage the more likely the battery is to be damaged eventually if it just stays that high. Imagine filling a water balloon from a hose. If you set the valve to not have enough pressure to burst the balloon, it takes forever to fill it. If you turn it up full blast, the balloon will pop.
This is why so many sailors who bought big battery banks and big alternators back in the day found themselves with either slow charging and sulfated to death batteries or fast charging and boiled to death batteries.

More to follow.

* Easy way to think of this asymptote function - every X hours you fill up half the empty space in the battery. X hours is 50%, 2X hours is 75%, 3X hours is 87.5%, and so on. You never get to 100.

[joe_db]

Early attempts to solve this problem:
The first attempts were human controlled one way or another.
Some people made or bought regulators that could be adjusted manually. I had one of those for many years. You could turn it up for awhile and then when you judged the batteries near full you could turn it back down. This has obvious problems with human error, both in forgetting to turn the voltage up, forgetting to turn it down, and/or misjudging the state of the batteries.
The other manual method was called the "AutoMac".

What it did was for the duration of the timer (big twist knob) the regulator would be fully bypassed. There was no regulation at all - the alternator was what we call "full field", putting out as much as it possibly could. This actually could work well *if used and supervised* by someone who REALLY knew what they were doing. The potential downside was far worse and more dangerous than just leaving a regulator set to 14.8 instead of 13.8 volts, you could possibly see 15 - 18 - or even more volts and cause fires or battery explosions, not to mention really destroying the batteries and other gear

[joe_db]

This photo is similar to my regulator from 1994 to last year. It only had a single setpoint, but it was user adjustable. It was free, it came with a big alternator we ordered for a customer and he was not interested in a single voltage regulator and gave it back.
I generally had success with this, but three issues did crop up. The first odd one was until I stuck some ferrites on the leads, using a handheld radio near the engine would drive it to full output The second issue was if the connections got loose or corroded, it sensed less voltage than there was and drove the output too high. Some careful wiring work and dielectric grease solved that.
The third issue is one that I was never been able to fix: It burns out diodes. A stock alternator on an A4 with a 13.8 volt setpoint does not work very hard. Once you move away from stock alternator and the usual one car battery at a time setup, you are working your alternator much harder. Not only are you working it harder, you are doing it in a worst-case scenario. Best case is less field current and high RPMs, that runs cooler. Running more field current to get the same voltage and lower RPMs produces more heat, the cooling fan is going slower. My stock Motorola has lost diodes, the 10si I had lost diodes, and the shiny new Balmar lost diodes. Every few years is a visit to the alternator repair guy in Annapolis. The 10si didn't get repaired, it only cost $65 or so and the bearings were getting noisy too, it went right to the dumpster.

[joe_db]

Where we are so far:
If you can adjust the alternator voltage by hand, you can:
1. Set it for different battery types. AGMs, Gels, big flooded batteries, and car batteries all have different optimal charging voltages.
2. Charge much faster than the slow charge you get from 13.8 volts, which was the standard setting on the Motorolas.

What remains to be done:
1. Sensing the alternator temperature and not burning diodes or bearings out.
2. Sensing battery temperature and not boiling or ruining batteries in the summer.
3. Sensing battery temperature and not undercharging batteries in the winter.
4. Switching automatically to a lower float voltage and not forcing a human to remember to do it or have to pick a lower charging voltage to avoid ruining batteries by overcharging them.
5. Provide expensive complicated stuff for marine electrical shops to sell and service

[Dave Neptune]

Joe, I used a switch to take my lil 35 to full field. I did this to get a bit more out of the alt for a SHORT time. I would start and once she'd idle I'd hit the switch and the A-4 would bog down. I used RPM to regulate the output voltage and would hold it to about 17v. I had a big battery bank and would only do this if they were low. I would run and monitor until the alt would start to get hot then switch back to the regulator. I had to be very careful and also be sure no one bumped the switch. Was lucky as I never cooked the windings or a battery. I also did not use it as much as I thought I would. I tried a potentiometer but that was to confusing so I went back to OFF-ON.

Dave Neptune

[joe_db]

Joe, I used a switch to take my lil 35 to full field. I did this to get a bit more out of the alt for a SHORT time. I would start and once she'd idle I'd hit the switch and the A-4 would bog down. I used RPM to regulate the output voltage and would hold it to about 17v. I had a big battery bank and would only do this if they were low. I would run and monitor until the alt would start to get hot then switch back to the regulator. I had to be very careful and also be sure no one bumped the switch. Was lucky as I never cooked the windings or a battery. I also did not use it as much as I thought I would. I tried a potentiometer but that was to confusing so I went back to OFF-ON.

Dave Neptune

If you added a timer it would have been an AutoMac

[joe_db]

Now a quick look at "3 stage charging" aka "smart charging" aka "3 step" charging and maybe they'll up the number to 4 if there is an equalize mode.
You will see these terms a lot and it actually is not as complex as it sounds.
A modern multi-stage regulator really only has 2 voltages, bulk/absorption and float plus maybe an equalize setting. Here is how they get the stages:
1. Starting off with a low battery, the alternator is full-fielded. This is the bulk stage. The voltage will not be at the bulk/absorption level because despite trying as hard as it can, the alternator just can't get there.
2. At some point, maybe 2 seconds later or maybe 4 hours later, the voltage hits the bulk/absorption setpoint. Now the regulator starts backing off the field current to keep the voltage from rising higher. This goes on for a period of time determined several different ways. You cannot see this cutover with a voltmeter, but you can with an ammeter on the alternator output. Details to follow on exactly how the regulator decides how long to do this.
3. At some point the regulator cuts over the float setting. You will see this happen on a voltmeter.
4. If there is an equalize function, which is not a given, there will be a manual way to set it and it will raise the voltage to maybe somewhere around 15 to 18 volts. This can be quite destructive, details on that to follow. Do NOT do this unless you really know what you are doing. This is NOT a way to charge batteries, it is a maintenance task.
Here is a graph of what goes on. The colors are different battery chemistries, I think red is wet, purple is AGM, and green is gel. Don't take these for hard numbers, this is to show the relationship between time and voltage.

[joe_db]

Here is the big question - HOW do you know when a battery is full and cut over to float?
You would think this would be a solved problem by now, but it very much isn't. The following methods have been tried:
1. Using a human. A person is responsible for changing a setting or flipping a switch. About as reliable as the human operating it and as accurate as whatever instruments that human has available. Huge opportunities for error.
2. Using a timer. This was what an AutoMac was, a spring loaded timer switch. Later on smart regulators would run in bulk/absorption mode for a certain amount of time. This time setting did not necessarily have anything to do with the time needed.
Below is a Heart Interface Alpha. These have three settings, bulk-absorbtion(acceptance), float, and time. You get bulk/acceptance charging for however long the timer is set for and then it switches to float. Note that that may have no relation to the amount of charging needed. Also note their suggestions for setting the charger are IMHO not correct.
Note - if you find one of these today, remember they were not especially reliable when new and that was a long time ago.

[ndutton]

While I have been very impressed with the battery charging presentation offered here, particularly that the early description included a great plain English analogy (filling a balloon), I hope at some point the discussion mentions what is to be expected from our battery and charging system and should those expectations vary with battery chemistry?

Joe, twice in this thread you have referred to early attempts to solve a "problem" without describing exactly what this problem is. Is it the time it takes to charge a battery? Is it the battery's ability to hold its charge? Is it battery life? Is it a combination?

Would a battery system that is able to supply refrigeration for 3 days away from the dock, that lasts 10 years before it's ability to do so drops off have a problem?

[joe_db]

While I have been very impressed with the battery charging presentation offered here, particularly that the early description included a great plain English analogy (filling a balloon), I hope at some point the discussion mentions what is to be expected from our battery and charging system and should those expectations vary with battery chemistry?

Joe, twice in this thread you have referred to early attempts to solve a "problem" without describing exactly what this problem is. Is it the time it takes to charge a battery? Is it the battery's ability to hold its charge? Is it battery life? Is it a combination?

Would a battery system that is able to supply refrigeration for 3 days away from the dock, that lasts 10 years before it's ability to do so drops off have a problem?

More to come - this is not my only task today
The problem is how to reliably tell when batteries are full. It is not as easy as it sounds.
Battery chemistry will get in here at some point.

[joe_db]

More ways to tell how full a battery is and a mid 1990s high point in regulator design - also a low point.
This is what came after the Heart Interface timer regulator and it's predecessors. The Link System did things not attempted previously and not done afterwards until very recently with VERY expensive systems that probably cost half of what my boat does.
Here is the Link 2000:

[joe_db]

The Link system controlled one alternator, one inverter-charger, and had 3 shunts. It knew the current in and out of two battery banks and it sensed alternator output. It kept track of amp-hours used and remaining in the banks.
The system did not try and use a timer to change over to float, it literally looked at the amp-hours used and would fill the battery back up and then switch over to float. Since it knew the current going into the battery and the current going out of the alternator, it knew that 100 amps out of the alternator but only 10 going into the battery meant someone had turned on a 90 amp load, not that the battery suddenly went dead and needed massive charging current after being full for 12 hours.
It had soft-start to gradually ramp up the alternator so as not to destroy the belts. It also had alternator output limiting by amps, which was very useful with a big alternator on a small engine. If 80 amps was all the engine and/or belts could support, you could dial in 80 amps. It would not restrict the charging voltages as long as you were at 80 amps or less, which is vastly better for fast charging then just turning the voltage down to try and limit the load. You even could use this with small stock alternators by turning max output way down.
It also controlled a big inverter with a 100 amp three stage charger of its own. When you fired up the genset or connected to shore power then it switched over from inverting to charging.
This seems just about perfect, right? No crude timers, no guessing, no humans forgetting a switch, no trying to imply from third hand info what was happening, the system KNEW what electrons were going where and how many to send to the battery. The only modern thing it did not do was IIRC it had no temp sensors. It was just about perfect *when it wasn't screwed up*
More to follow.

[ndutton]

The problem is how to reliably tell when batteries are full. It is not as easy as it sounds.

Do we have the same problem in our cars?

[joe_db]

Do we have the same problem in our cars?

Not like a sailboat. For decades cars have had a single set point regulator and tried to pick a voltage that charged fast enough yet didn't boil the batteries to death. Cars don't routinely use most of their battery capacity and then try and charge up as quickly as they can before repeating the process. If you ever actually do cycle a car battery that way, it will be ruined very quickly.
Some high end cars that do have a ton of electrical stuff DO have sophisticated charging systems that sense amps in and out of the batteries. They are a PITA because battery replacement becomes a dealer-only operation unless you have the very expensive tools needed to synch the car computer with the new battery.

[joe_db]

The demise of the Link system:
You might wonder why we can't do now what we could do back then. The answer is simple, the damned things broke all the time!
The first issue is amp-hour meters are not very accurate. Back then 100% for sure they would drift off of the true value and screw the system up. If that wasn't happening, the system would just become defective in one or more ways. It was apparently a bridge too far for the tech of the day. We always installed a second simple regulator and a switch when we sold one of these systems because we knew sooner or later it would be needed.

Next up - what we have today that replaced all of that.