All lights are pulsing?

Tiny
TBOAT13
  • MEMBER
  • 1998 DODGE STRATUS
  • 2.4L
  • 4 CYL
  • 2WD
  • AUTOMATIC
  • 130,000 MILES
I installed an external voltage regulator because PCM was not controlling voltage to battery. I noticed that, after I installed it, the dash lights and headlights were pulsing. I have checked the ground and tried different locations for sourcing power, but have not had any success correcting this issue. Any advice will be appreciated.
Tuesday, December 23rd, 2025 AT 12:01 PM

7 Replies

Tiny
CARADIODOC
  • MECHANIC
  • 34,415 POSTS
What kind of regulator did you use? Chrysler's 1970s unit is what is commonly used. Did you do any diagnostics on the system before modifying it?

Is the pulsing a regular surge up and down, an intermittent flicker, or something else? Measure the battery voltage with the engine running and tell me what you find. Is there anything you can do to cause the pulsing to occur, or to occur more often?
Was this
answer
helpful?
Yes
No
Tuesday, December 23rd, 2025 AT 12:52 PM
Tiny
TBOAT13
  • MEMBER
  • 12 POSTS
Original test found voltage in the 16 voltage range. Not seeing a voltage regulator under the hood, and not realizing that the voltage was regulated by PCM, I bought a new alternator. It came with specific instructions for testing, and that test showed alternator to be good. I installed the new one anyways. Since I was getting 16v I thought the regulator was the issue. Then I installed the external regulator. The pulse is constant and doesn't seem to change regardless of load. I have attached a couple of videos. Thanks.
Was this
answer
helpful?
Yes
No
Wednesday, December 24th, 2025 AT 8:55 AM
Tiny
CARADIODOC
  • MECHANIC
  • 34,415 POSTS
You do have the common regulator used for this modification. I did the same thing on my skid steer. While the setup will get the job done, there are some technical problems with it. This was the first electronic voltage regulator, first used by Chrysler on 1970 models. They have a really long list of "firsts" that truly benefitted car owners.

The first problem is this regulator has temperature compensation built in, but that's all. Charging the battery is a chemical process, and those slow down in colder temperatures. This regulator will bump up charging voltage a little in cold weather.

Other manufacturers like to put the regulator inside the generator, making it an easy-to-replace package deal. You have to buy both parts when one of them fails. With Chrysler's design, the alternator remains a stand-alone part, so that's all you replace when it fails. By putting the voltage regulator inside the Engie Computer, it can respond to everything the computer knows. Specifically, it knows battery temperature, engine coolant temperature, engine load, engine speed, and a number of "anticipation" controls. That means it knows for example, when the computer is about to cycle on the AC compressor or the rear window defogger. Both require extra current, so instead of that sudden current increase causing a drop in voltage, and dimming lights, it ramps up alternator output just before those higher loads begin. The external regulator has no way of knowing that. It can only respond after the loads occur and system voltage has dropped.

The second thing to be aware of, while likely not the problem here, is this regulator has a third, very important wire. That is the ground made through its metal housing. Relying on the bolt threads is not the best idea. Instead, consider scratching a clean surface through the paint to get the best possible connection. To check the connection, use a digital voltmeter on the lowest DC volts scale, then measure the voltage drop between a clean spot on the housing and the battery's negative post. Since mechanical connections are involved, there will be some slight voltage drop, but it should be in the order of millivolts. I'd expect to see less than perhaps 10 to 20 millivolts. If you find more than that experiment with additional ground wires between the engine and body sheet metal, and between the body sheet metal and the battery's negative cable. That will be the smaller cable of the two in the negative cable clamp.

The next concern has to do where you're getting the 12 volts to run the regulator. The Engine Computer usually has four 12-volt supplies. One is constant to maintain learned fuel trim numbers, sensor personalities, and any diagnostic fault code data. A second supply from the ignition switch turns the computer on. There may be a third supply that gets switched onto one of the outputs to turn on the automatic shutdown, (ASD), relay. That relay feeds 12 volts to the ignition coil(s), injectors, fuel pump or pump relay, oxygen sensor heaters, and it is the fourth 12-volt supply going back to the computer. That is the 12 volts the computer looks at to monitor system voltage. By monitoring the ASD line, it maintains steady voltage to everything that makes the engine run right. The lighting circuit comes along for the ride. It will be stable as long as the engine controls are stable.

One word commonly used to describe this type of pulsation is "hunting". The voltage regulator sees system voltage going too high. It responds by reducing alternator output, but there's a built-in delay. By the time it lowers output, that output voltage has already been rising for too long. By the time it cuts back, the drop in output overshoots and drops too far before the regulator sees it and catches up. This wasn't a problem in the 1970s when the biggest standard alternators were only good for 55 amps. At most you might see a small dimming of the lights when the AC compressor kicked on, but that was easy to overlook. Today with a computer hung onto every conceivable part of the car, various loads are being switched on and off all the time. The regulator circuit has to be able to respond almost instantly to maintain steady system voltage.

The next thing to look at is the fat, bolted-on output wire from the alternator to the battery's positive post. There will be some type of fuse device in that circuit. It used to be a fuse link wire spliced into that cable, usually very near the battery. Later model vehicles use a regular fuse bolted into the under-hood fuse box. Those bolts can work loose and are a good place to find arced terminals that need to be shined up. Like you did with the ground circuit, measure between the battery's positive post and the stud on the back of the alternator. These tests must be done with the engine running. If you find a significant voltage drop, or if the voltage is pulsing, we have to break the circuit down into a number of smaller sections, and check each one to find the excessive resistance. Expect this voltage to be in the order of possibly as much as 0.2 volts, and it will increase as more high loads, such as heater fans, head lights, and rear window defoggers, are turned on. The maximum allowable is generally 0.4 volts, but I don't like to see that much.

I want to add a comment here for the benefit of others researching this topic. It has to do with digital instrument clusters like most cars have today, and GM vehicles in particular. GM "AC generators" have their voltage regulator built in. They have a whole pile of design problems already, but system voltage sensing becomes an issue with digital dashes. Their voltage regulators monitor output voltage inside the generator, regardless of what's happening in the rest of the vehicle. The sensing wire in the four-pin plug on the side / back is empty. When the vehicle has a digital dash, that sensing wire IS in that plug, and system voltage is monitored remotely right at the instrument cluster. That ensures the dash is fed a stable supply voltage, with whatever happens on the rest of the car less important.

On Chrysler vehicles, there's no need for remote voltage sensing. It's part of the software in the Engine Computer. They call your instrument cluster the MIC", (mechanical instrument cluster). It is immune to voltage fluctuations, except, of course, for the light bulbs in it.

If nothing reduces the pulsing up to now, a possible solution might be to add a large capacitor to the two smaller terminals on the back of the alternator. I'll address that in a minute, but first let me explain how to diagnose this system. I mentioned all the places the ASD relay sends 12 volts to once the engine is running. One of those places is the alternator's field circuit. To generate a voltage and a current mechanically, you need three things; a magnet, a wire, and most importantly, movement between them. We use a coil of wire as it's more efficient. The rotor is an electromagnetic coil, and the movement is we spin that coil with a belt and pulley. There's a number of ways to adjust output, but only one is practical. We can't raise and lower engine speed as the electrical demand changes, but it's easy to adjust the strength of the electromagnetic field coil. At its highest possible output, that field current is only roughly three amps. (That's the three amps that also flows through the regulator's metal case, to ground). Under most typical driving conditions, that field current is only around one to one and a half amps, so it's not real significant. It IS enough, though, to cause a voltage drop across that spinning field coil. To begin testing this system, simply measure the two voltages on the two smaller terminals on the back of the alternator. Of course this has to be done with the engine running. During normal, or proper operation, battery voltage must be between 13.75 and 14.75 volts. I'll use 14 volts for my story. When you measure these two voltages, one of them will be 14 volts. If you find less than battery voltage, there's extremely little to check, basically a few inches of wire, because most of what could cause that would cause the engine to not run. You'd be diagnosing a failure to run, not a charging problem.

Assuming you do find 14 volts, the voltage on the second terminal must be less, but not 0.00 volts. If you find 0.00 volts on that second terminal, the brushes inside the alternator are worn. Those can be replaced, often without even removing the alternator from the engine. Cost of parts is around $12.00. Instead, most people just replace the entire alternator.

If you do find less than 14 volts on the second terminal, the input side of the charging system, including the voltage regulator circuitry, is working. My next story will cover the output side of the charging system, when necessary, but that involves little else besides the alternator itself. Typically, you'll find that second voltage to be between 4 - 11 volts. The lower that voltage is, the greater is the difference between them, meaning the stronger magnetic field is being produced. That means more output current is being developed.

If your regulator circuit in the computer is open, meaning a break in the circuit, you will find exactly the same voltage on both smaller terminals. There's no current flow through the field, coil, so no magnetic field is being developed, and with no current flow, no voltage is dropped across it, so you end up on the second terminal the same as you started with on the first one. This is the condition you're trying to solve.

In this diagram, you'll see the feed wire to the field coil is dark green / orange. The "control" wire is dark green. In this story, we've determined the field circuit is open, and it appears you correctly found the defect is inside the Engine Computer. Before going through the work of the modification, (or spending our customer's money), we had better verify the alternator itself is okay and doesn't have an additional problem. A real fast way to do that is with a "full-field" test. All that involves is grounding the dark green wire while the engine is running. Be aware this is definitely no time to raise engine speed. Doing so has the ability to send system voltage well over 30 volts, enough to destroy multiple computers and burn out any bulbs that are turned on. Alternators and generators are very inefficient at low speeds, but if working properly, they will develop plenty of increased voltage to be seen. Monitor battery voltage, or watch the brightness of the head lights when you ground that wire. If voltage goes up, the alternator can be said to have "not failed". It still may not be working to full capacity, but that's a story for another day.

Next, I get to discuss some problems with this dark green wire. This same circuit was used on Dodge Shadows. We had three of them, donated by Chrysler, to my community college, that I used for training. On those models, there was an inline connector halfway between the computer and the alternator. I never did, and never allowed anyone to pierce wire insulation to take measurements. Instead, those connectors made dandy test points. That is a perfect place to touch a jumper wire, and to ground, to full-field the system. If you don't have an inline connector, or know where to find it, you're presented with a new problem. You can do this test right at the back of the alternator, however, both wires go through a black plastic block and emerge as two thin tabs bolted to the two small terminal bolts. When the system is working properly, it's easy to tell which tab goes with which wire by the voltages. It's when you have your problem of an open circuit, both voltages will be exactly the same. Which tab do you ground to perform the full-field test? If you try the correct one first, charging voltage will go up. If you grab the wrong one first, you'll blow the fuse feeding the ASD circuit. No harm done, but it is embarrassing. I do have a relatively simple trick to figure out which is the correct terminal to ground, but I can't give out all me secrets at once.

Okay, as promised, back to my thoughts on adding a capacitor. We're looking for the cause of alternator output going up and down. That is a result of unwanted current flow changes between the two smaller field terminals, and that is caused by unwanted voltage fluctuations between them. A capacitor, for anyone not familiar with them, can be thought of as nothing more than a very small battery. The large ones basically want to oppose a change in voltage across them. Their use in this type of circuit is to absorb the voltage increases, and provide a little current when the voltage drops. It smooths out the voltage. This is not a part I would buy just to experiment. As a former tv / vcr repairman who has scrapped out hundreds of units, I have boxes full of these "prizes". Find any friend who is into electronics, or visit any tv repair shop, if you can still find one, and they will have all you want, probably for free. They're rated in voltage and "microfarads", which is their electrical storage capacity. I'd look for something in the area of 100 to 1000 mfd. The voltage just has to be higher than what it will "see" in the car. Common voltages are 50 and 100 volts. It refers to the voltage it can withstand without arcing and shorting internally. These larger capacitors are referred to as "electrolytics", and as such, they do have a polarity. Get that polarity wrong with so a low voltage, and it's extremely likely nothing exciting will happen. For testing, it will work fine. If possible to figure out, the positive terminal should go to the 14-volt feed terminal, and the negative to the dark green wire / terminal.

If the fluctuations stop with the capacitor installed, that may not necessarily be the solution, but it would be a clue.

There's one more rather important problem with doing this modification. That is the Check Engine light is going to be on all the time. The fault code will be "Field circuit not switching properly". Along with all the things the internal regulator knows by being inside the computer, it's circuitry is also monitored for proper operation. The dark green control wire was disconnected to run to the new regulator. No current will flow through the computer's regulator circuit and that will be detected. There are roughly 2,000 different diagnostic fault codes that can be set. Roughly half of them refer to things that could adversely affect emissions. Those are the codes that turn on the Check Engine light. With unknown or questionable charging system operation, low voltage to the ignition coils and injectors can result in poor engine performance and increased emissions. That's why this code turns the Check Engine light on. A lot of people drive every day with that light on. The problem is to set any fault code, there is always a long list of conditions that must be met. One of those conditions is always that certain other codes can't already be set. A number of the self-tests the computer constantly runs will be temporarily suspended, so a new problem can develop that you have no way of knowing about. Many of those problems can be very minor, but turn expensive if ignored. Mostly that involves overheated catalytic converters. Also, when the original problem is finally repaired properly, all those suspended tests will resume, and that's when the Check Engine light turns right back on and we have to start the diagnosis all over. That makes for frustrated car owners, and frustrated mechanics who had no way of knowing there were other, new problems.

Now that I've shared all this, it occurs to me you weren't looking for a no-charge condition. You found charging voltage to be a fuzz too high. Was there some other symptom you were chasing? 16 volts is a little high, but if the regulator circuitry was shorted, the system voltage would be considerably higher and the battery would be spitting acid and its sides would likely be bulging.

There's two more things to consider. The first is when you added the external regulator, the one to the computer has to be disconnected. The one with the lower "target" voltage will overpower the other one, then the two will fight each other. The most likely symptom of that would be those pulsing lights. A second issue, somewhat hard to explain, is controlling the charging system is not actually done by holding field coil voltage constant. Instead, it is switched from full-on, (maximum output), to fully-off, (no output), about 400 times per second. The percentage of "on-time" to "off-time" is adjusted to change the average output voltage. The resulting pulsing is smoothed out by the battery. (That's another sore point with GM charging systems and repeat generator failures). The external regular switches the same way, called "pulse-width modulation", or "pulse-width dimming", and is why such a tiny, lightweight transistor can control such a high current. Switching current flow off through a coil of wire sets up a huge reverse voltage spike. That's highly desirable in ignition coils, but aged batteries in GM vehicles can't dampen and absorb those spikes. Those spikes are what destroys a lot of their generators and causes elusive engine running problems. On your car, if the original regulator is still in the system, one regulator might be switching on while the other one is switching off. With two different target voltages, the alternator is responding to two different commands. Just like when you would hear a whistle when tuning a really old AM radio, if one regulator is switching at 400 cycles per second, and the other is switching at 399 cycles per second, they'll go in and out of sync and you'd see that as a pulsation of the difference, or one cycle per second.

That's getting way more technical than necessary to solve this. First, be sure the internal regulator is not connected at the same time. If you aren't sure, just unplug one of the wires on the external regulator. If the charging system keeps working, there may not actually be anything wrong with the computer. For my better comment of value, given the age of the car, I would seriously consider finding a used Engine Computer from a salvage yard. There will be plenty in the pick-your-own-parts yards. Do a search for "Pull-A-Part" to see if there's one in your area. I've been to 16 of them. All are very clean and well-organized. The last computer I bought for one of my vehicles cost $25.00.

Let me know if that helps or if you have more questions.
Was this
answer
helpful?
Yes
No
Wednesday, December 24th, 2025 AT 6:20 PM
Tiny
TBOAT13
  • MEMBER
  • 12 POSTS
I think I will go back to square one. I am going to reconnect ecm and re-test. However, I have a couple of questions. When I connected external regulator, I cut the two small wires that go to plug on back of alternator. I then tied external regulator to that plug. This takes regulator in ecm out of the loop, correct?
The other questions I have are in regards to buying a used ecm.
The options I see online ask if I need:
Low emissions vs federal emissions.
Firm suspension vs standard suspension.
Also, would it need to be programmed or can I just match part numbers?
Was this
answer
helpful?
Yes
No
Thursday, December 25th, 2025 AT 9:11 AM
Tiny
CARADIODOC
  • MECHANIC
  • 34,415 POSTS
First question. Yes, if you cut the wires, the internal regulator is removed. The two on the external regulator go to the two on the alternator. No polarity, meaning either wire can go to either terminal. Then the entire modified charging system needs a 12 volt power source. I put this drawing together to show it better. On '70s and '80s models, for all cars and most trucks, every dark blue wire under the hood got 12 volts switched onto it when the ignition switch was in "run". By the mid '90s that 12-volt wire was commonly dark green / orange. On the voltage regulator, that 12-volt wire is the one at the top peak of the triangular plug, like in my sad drawing. (For some '80s trucks, that 12-volt wire was often red).

The emissions is basically one system for vehicles sold originally in California. The federal system is for the other 49 states.

I've never heard of a different Engine Computer based on the type of suspension. Unless you're aware of some type of optional sport package, assume you have standard suspension.

As for programming, I'm going to tell you more than I know. For '95 and older vehicles, computers were a simple plug-in-and-go affair. If you find one with the same part number, it will work, but we normally go by application, not part number. When a computer is sent in to be repaired or rebuilt, it gets the latest updates, then, instead of a part number that ends with "AA", the part number gets increased to "AB". That last letter steps up one time for each update or advancement in the software. When there is a really significant change or modification, it gets an entirely new part number that ends again with "AA". You can have three, four, or five computers, all with very different part numbers, that all are for the same application. They will, of course, be different between four, six, and eight-cylinder engines, but within each group, some Engine Computers are compatible with different engine sizes.

Right around '98, give or take a few years, they began requiring a four-digit PIN be entered with a scanner. There is a way to read that from the old computer, with Chrysler's DRB3 scanner, otherwise you need to call the dealer. I used to work for a very nice Chrysler dealer, and the people there like me, so they give me these numbers over the phone. I found out they do that for my friend too who they do not know. You can find the DRB3 scanners on eBay. I don't know if aftermarket scanners can do this, but if any can, I'd expect Autel products to be able to do it.

This next topic is harder to explain than it really is, but it could be extremely important. If your car has the optional factory-installed anti-theft system, you can plug in any correct Engine Computer, and / or any Body Computer and it will work. The anti-theft system lives in those two computers. If the replacement computer doesn't have anti-theft programming, it will learn that from the other computer the instant you turn on the ignition switch. Once learned and that anti-theft programming is activated, it cannot be undone.

Here's where the potential problem comes in. All of the Engine Computers and all of the Body Computers have the anti-theft programming built in. It is simply not activated if the car doesn't have that option. If you plug in a used replacement computer from a donor car that has the anti-theft option, that computer will immediately teach that to the other existing computer as soon as the ignition switch is turned on. Now you'll have a crank / no-start condition because both computers are waiting to see the "disarm" signal, but that's never coming since the system isn't on that car. At that point, the only solution is to replace both computers at the same time, with two that have not had anti-theft activated. That's usually impossible to know when buying them from a salvage yard, because the people there usually don't know whether a certain car had that.

I've seen or been involved with this problem three times. The first one was a learning experience for a coworker who later explained this to me. It was on a '92 Dynasty without anti-theft. For the second one, a friend with a body shop where he specializes in rebuilding one and two-year-old smashed Chrysler products, repairs were done except for replacing the crushed Engine Computer on a 2003 Neon. He borrowed a computer from another friend's car and verified everything was working properly. What he didn't know is the repaired car had anti-theft, and the Body Computer activated it in the borrowed Engine Computer. Engine ran fine. Next, he put the friend's Engine Computer back in his car, and now that one was a crank / no-start. That's because the anti-theft had been activated and it taught that to the Body Computer. Both computers in the friend's car had to be replaced.

If you buy a rebuilt Engine Computer from the dealer or from a legitimate computer repair shop, it will come with the anti-theft programming not activated. It will be activated automatically if your car has that system. Once that occurs, you don't want to put that new computer in another car without anti-theft or this same problem will occur.

The third vehicle that this happened to was on a '95 Intrepid donated by Chrysler to my community college after it had been used for training. That car did not have anti-theft. A student borrowed our Engine Computer to see if it would solve a problem on his car. He did have anti-theft, so his Body Computer taught that to our computer. When he put our computer back in our car, that Engine Computer activated the anti-theft in the Body Computer. Now we had a crank / no-start. To add to the frustration, my tool room attendant kept ordering used computers, one at a time, thinking he was going to get the engine to run. It was very difficult to convince him both computers had to be replaced at the same time before the ignition switch was turned on.

To boil this down, there is nothing to worry about if your car has factory anti-theft. Any replacement computer will self-activate that programming if necessary, but remember, that programming cannot be deactivated, so it can't be used later in a car without factory anti-theft.

If your car doesn't have anti-theft, you must be careful to use only replacement computers that don't have it activated. In your case where the engine does run but you're trying a different computer as a test for some other problem, if you suddenly have a no-start condition, that used computer has anti-theft activated and it just did that to your Body Computer. If you reinstall the original Engine Computer, it is going to self-activate due to learning from the Body Computer. Now you must replace both computers. If those computers are otherwise okay, they can be used in other cars that do have anti-theft. If you recognize this right away when the donor Engine Computer results in a no-start, you only have to replace the Body Computer due to it just being activated, but then the original Engine Computer can be reinstalled as it hasn't learned the anti-theft activation yet.

I know that can sound complicated, so sometimes I describe it as someone brushed paint on your shirt. Now that you have that paint, which can't be removed, you can wipe some on your hand and wipe it on someone else's shirt. The only way now to take the girls out tonight is for both of you to change your shirts at the same time. If only one of you does, the other still has paint he can wipe on your new shirt.

Now that I shared all that, I don't know if that applies to the Stratus and Cirrus models, but I suspect it does. I think I'd feel better if you reconnected the internal regulator, then take the voltage readings for me to analyze. If you do that, tell me what you find, and remember on which tab you find which voltage on that little black box on the back of the alternator.
Was this
answer
helpful?
Yes
No
Thursday, December 25th, 2025 AT 4:14 PM
Tiny
TBOAT13
  • MEMBER
  • 12 POSTS
Well, I reconnected the internal regulator and now I am getting a steady 14.2v and no pulsing lights. I am glad you led me back to this point.
I don't know what changed. The only thing that comes to mind is, I replaced the battery after I hooked up the external regulator. I figured the old one got cooked from the 16v or so I was getting. I never thought to try reconnecting the internal regulator. Thanks so much for all your help and the wealth of information you provided.
Was this
answer
helpful?
Yes
No
Saturday, December 27th, 2025 AT 9:02 AM
Tiny
CARADIODOC
  • MECHANIC
  • 34,415 POSTS
Happy to hear you solved this. The battery is indeed a key component in the charging system. I could go into a really long story about multiple repeat generator failures on '87 and newer GM vehicles, but it all boils down to aged batteries that still seem to be working just fine. These generators develop huge, harmful voltage spikes. The battery dampens and absorbs those spikes, but by the time they get to be two or three years old, enough lead has flaked off the plates that they lose their ability to do that. Ford and Chrysler charging systems work similarly but they don't have this problem.

It's good that you recognize that 16 volts is too high, but I wouldn't get excited as it's just slightly above an acceptable 14.75 volts. Two of my vehicles commonly run over 15 volts right after startup, then come down after a few minutes. At 16 volts, the regulator is working, even though it is a little high. If the regulator circuitry is shorted, it will run the alternator wide open, then expect to see system voltage exceed 18 volts at idle.

Due to all the additional internal circuitry, it is pretty much impossible for a voltage regulator to short completely. When looking at the two voltages on the two smaller terminals on the back of the alternator, one will have full system voltage. The key is the other one. If you find 0.0 volts on it, it's much more likely you'll find the dark green wire is grounded somewhere between that terminal and the computer. Even if the regulator did short, you'd still find a good two or three volts on that second terminal.

Others researching this topic may read this. For their benefit, I'd like to add one more comment. Along with a really long list of "firsts" for Chrysler, they were the first to develop the "AC generator", and they copyrighted the term, "alternator". That was first used on 1960 models. Mechanics didn't know yet how to diagnose these very simple systems, so they came up with a trick to remove a battery cable, then if the engine remained running, they assumed the alternator was working. There's a whole pile of problems with that "test". In fact, today, a mechanic caught doing that might be lucky and get one warning, but more likely they will be fired.

The first problem is the battery is needed to assist the voltage regulator in doing its job. All AC generators put out three-phase output, similar to what's found in large factories. After it passes through the internal "diodes", one-way valves for electrical current flow, it looks like the top waveform in the drawing below. The "ripple" voltage shown here as 0.5 volts, (the difference between 14.0 and 14.5 volts) is very easy for the battery to smooth out. In fact, the battery is preventing it from going much above 14.5 volts.

When a battery cable is disconnected while the engine is running, that ripple voltage increases and the peaks could easily go to 16 to 18 volts. That voltage goes right back around to run the field coil. With a higher voltage, you get higher current and a stronger electromagnet. That causes a higher output voltage to be induced resulting in higher output current. That cycle continues until it's very easy to reach well over 30 volts. For demonstration purposes this has to be done on a generator test bench because on a vehicle, 30 volts will destroy most of the computers, any bulbs that are turned on, the diodes in the alternator, and on GM generators, the built-in voltage regulator. The only thing that can save it is all generators are very inefficient at low speeds. Raise the motor speed on the test bench and that 30 volts is reached and exceeded in a heartbeat. Once the smoke comes out, you can't put it back in.

Without the battery in the circuit, the voltages in my top waveform are more likely to be 12.6 volts and perhaps 16.0 volts. Some voltage regulator designs operate by looking at the 12.6 volts. They increase generator output in an attempt to get that up higher. As they do, the 16 volts goes to 18 or 20 volts, but the regulator keeps seeing and responding to that 12.6 volts. Stuff gets burned out.

Other regulator designs look at the 16.0 volts and cut back on generator output in an attempt to get that down. Without the load of the battery to help with that, it can't cut back far enough. It keeps dropping and dropping field current until there's insufficient output current to keep the engine running. Injectors and electric fuel pump alone take over ten amps. The engine stalls, giving the false illusion the generator is not working.

Those are instances where you get invalid results when removing a battery cable. Another problem occurs when one of the generator's six internal diodes has failed. You lose one phase of the output. That looks like the lower waveform. Ripple voltage is very high, 5.2 volts in this case. Under the "full-load output current" test, the most you'll be able to get is exactly one third of the generator's maximum rated capacity. Today it's very common to find generators capable of developing 100 amps. With one bad diode, roughly 33 amps is all you'll be able to get. That's not enough to run the entire electrical system under all conditions. The battery has to make up the difference as it slowly runs down over days or weeks. Removing a battery cable with this condition will cause the same results as before, but maybe more pronounced. Even if the engine does stay running, you haven't identified this defect.

The only proper way to test the charging system is with a professional load tester. It takes longer, about a minute, to connect it to the car, than it does to make the tests. You can start the testing yourself with an inexpensive digital voltmeter. For those who don't know how to use one, this article is a dandy starting point:

https://www.2carpros.com/articles/how-to-use-a-voltmeter

They're using an "auto-ranging" meter which is an expensive option you don't need. Any of us can help in setting up a standard meter. You can find a perfectly fine meter for $7.00 at Harbor Freight Tools, or similar models at Walmart or any hardware store.

Measure the battery voltage with the engine off. It will be very close to 12.6 volts if it's good and fully-charged. At 12.2 volts, it's good but fully discharged. If it's around 11 volts or less, it has a shorted cell and must be replaced.

Next, measure the battery voltage again with the engine running. Now it must be between 13.75 and 14.75 volts, give or take a little depending on which textbook you read. If it is, that just means it's okay to continue with the professional tests. You'll need to visit a shop for that. Those testers will measure that charging voltage again, plus the full-load output current and the ripple voltage. A few testers can make paper printouts. Those will display ripple voltage as an actual voltage. Most testers just show ripple voltage as "low", (good), or "high", (bad), with a series of flashing lights.

The last problem I haven't mentioned yet has to do with worn brushes inside the generator. That becomes common by around 150,000 to perhaps 200,000 miles. On some designs, like Ford's from the early to late '90s, and any Nippendenso units on '90s and newer Chrysler products, those brushes can be rather easy to replace. Don't bother trying to do this on GM's generators. You'll need an entire rebuild kit because there's no way to disassemble them without destroying parts. The clue to this problem is the no-charge condition always starts out as an intermittent problem. At first it acts up for a minute or two. Those periods get progressively longer and longer, usually over a period of weeks or months, so you do get quite a bit of warning. The secret to finding this defect is any testing has to wait until the problem is occurring. A mechanic may advise you to keep driving for now, and come back when it acts up more often.

More information can be found in this article:

https://www.2carpros.com/articles/how-to-check-a-car-alternator

Some test results can be misleading, but they're a good place to start. Here's where you can find a list of other wondrous articles:

https://www.2carpros.com/articles

Bring on any other questions related to this topic, otherwise, please come back to see us again with your next problem.
Was this
answer
helpful?
Yes
No
Saturday, December 27th, 2025 AT 10:40 AM

Please login or register to post a reply.