Hello,

Our Protege started being difficult to start before finally requiring a jump. We replaced the alternator because the battery is only a couple years old. Had to jump again, and took it to O’Riley’s to test the battery. Their machine said bad battery, so we replaced it.

On my way back from the shop I noticed that at 40mph the battery light comes back on. The battery light turns off again when I get below 40mph. This happened with great reliability. Additionally, the car’s shifting is super off (automatic). It has no power to get out of 3rd gear, and there is a noticeable decrease in power to the AC when the battery light comes on.

We are kind of stumped on this one. Thoughts?

 

Lotsa variables here could be a bad alternator. This could have caused all of your problems from beginning all depends, on even a new alternator you could have issues. So did this just start to happen or was it over time? What kinda of work was done to the car prior to this happening. Always back track to what was done to car before this issue, new wiring ,mods, of any kind etc. I had a similar problem happen in my alternator after I replaced the bearings in mine. When I took apart the alternator to fix it everything is very tight with not much room. i replaced bearings and put it back together and all was ok until one day I had battery light come on then lights would dim then all was ok. Next time PW where slow and wipers where slow then ok. I pulled alternator out and found that when I put windings back in a few insulation spots got scratched hence when it was running it was grounding therefore causing problems. I took it apart and was able to reseal windings and all is ok. Ive heard that depending on non oem alternators you could get one that's not good. The test you can do is put a voltmeter across terminals of battery 1st you should have over 12v when you start there should be 14.2 v approximately. Then remove battery + cable and if car dies then your alt isn't putting out enough power and the car is only running off battery. For this reason is why I fixed my OEM alt bearings I didn't want to risk buying some china knock off that would last more than yr. Also check for any cables that are pinched you mentioned AC check wiring to that. Good Luck there's so many possibilities here that I couldn't possibly cover....

 

TLDR: there may be a bad batch of alternators out there (over voltage?), and troubleshooting is hard.
-----------------------

I'm having very similar symptoms to the OP. The alternator and charging system has been an endless headache, and there is also seemingly unrelated transmission madness, that might be related...

My 01 Protege 1.6L Auto needed an alternator (worn out brushes, wasn't charging) I got a rebuilt alternator (Ultima?) from Orielly's and had them test it at the store with their test machine. It spun up to 32 volts and failed test prior to leaving the store. I returned it right there and began looking for other options.

I decided to rebuild the existing alternator with a kit from ebay. Which worked for about 2 weeks. Not exactly sure why/how the rebuild failed at this point, but the thing got hot for sure.

I then ordered a proper AC Delco re-manufactured alternator (online, rock auto), which worked for 1 day, and began over charging at RPMs above 2500. Thinking the overcharge could be a bad battery issue, I replaced the battery, and alas it's still over charging.

Perhaps all 3 alternators suffer from being rebuilt with bad rebuild kits? Bad batch of MOSFETs out of China? or there is something else going on with the car... Combination of both?

At some point during all the charging system madness, the O/D off light started blinking (transmission code) and the car is now shifting in 'limp home mode' or whatever Ford/Mazda calls it. The automatic trans runs at full pressure and lurches on shift. Sadly there is no check engine code, just the OD off light, which I understand can't be read with a standard scan tool... Maybe over voltage could have blown up a transmission sensor?

At the moment, while the battery light is on (from over voltage), running on the freeway, the car doesn't shift into overdrive. Maybe the transmission controller is angry about getting a high voltage?

After some research, I understand that the PCM senses the voltage and turns the field windings on and off to the alternator. Perhaps the PCM is failing? Or a bad ground?

I want to check the grounds next, verify wiring/voltage from the alternator to the PCM. I will try to remember to follow up here if i find a fix... posts like this makes me think charging and transmission are related somehow: https://www.mazdaforum.com/forum/maz...y-light-41249/
Sadly no known fix yet.

 

Looks like the grounds was a good starting place:
TSB_8086.pdf

Gonna chase this wire down tomorrow.
Also might explain the testing problems with the rebuilt Orielly's one.

 

The grounds, wiring harness, and PCM are verified good on this one. I sanded and spray cleaned the digital grounds (odd white connector near battery cable on driver side). I ran a continuity test on the field winding PWM signal from the PCM to the alternator connector(tested good), as well as the feedback signal that picks out of the alternator's internal rectifier and sends to the PCM (tested good). I also found out the PCM turns the battery light on in the dash board. So that means the computer can "hear the battery voltage" and also "command the alternator charging" ... I attached an Oscilloscope to the PWM command wire to the alternator, which shows a normal PWM at idle, but then, if the RPMs increase, the PWM decreases. Eventually to zero.. Which is correct for the PCM to do, as the alternator field winding shouldn't be on... but somehow the alternator stays stuck charging after the computer requests that it stop.

Somehow the transistor that runs the field winding gets stuck on if the connector is plugged in even at 0v. If I unplug it, the alternator stops charging. The problem BJT is shown on the schematic, but is crammed into the rebuild electronics that are attached to the connector. (in the pic attached)

I tried grounding the alternator case to the digital ground location. No luck. I need to buy an alternator that doesn't charge when the computer tells it to not charge, and or find something with internal regulation that doesn't rely on the PCM. It would need the same geometry for the mounting bracket, offset to the pulley. Maybe some one-wire hotrod stuff? Sigh. This car has like 120k on it. The reman-aftermarket shouldn't force it to the cube shape yet.

 

So reading ur posts I see that at idle it seems to be ok?. When u get rpms up over 2500 then it starts acting up corrrect? Which Alt is in there now one rebuilt or from store.The voltage regulator maintains voltage at all speeds thats why its called that. I would be checking anything that will come on eg temp sensors,anything that comes on. So that would mean that something is grounding bad there causing the alternator to go squirelly trying to keep up to charge battery. Does everything dim or operate slow power windows wipers fan etc, or you get check engine lights? Also at that rpm try unlpugging a sensor and watch voltage.What kinda volts are you getting can u test it? Also if you start car what voltage are you getting to battery should be ard 14volts then remove cable and at 2500 rpm see what happens the voltage regulator in alt should keep constant, if its going up then alt is bad. If its too high then it will boil the battery dry. The alt should run car fine,all the battery does is for starting nothing else. Further more you checked signal going to pcm is good but thats from only alt connector ? You have to check that connector and wire going to pcm . Check volts at wire at PCM and check that you dont have a pinched wire in harness from alt to pcm. U mentioned cleaning grd wire or wires? You need to check all grds in engine bay theres a multi connector by drivers side strut it has about 6 grds in it make sure they are clean and properly tightened down. Check all motor grds to fenders etc and see what you get good luck.....

 

Yes, ~14ish volts at idle. With the field winding seeing a 20%ish PWM from the computer.

Yes, as the car revs up, the PWM from the computer to the field winding decreases, as the computer can sense the voltage going up.
The battery light turns on around 17 volts. The battery light is controlled by the computer.

There is a remanufactured AC Delco alternator in there right now. It has been on the car for 24 miles of driving. Most of which with the battery light on in an over voltage situation.

Yes. The computer is trying to get the alternator to turn off here. But the alternator is not turning off, it's charging when there is no need to charge.

I'm not sure what the meaning is on this one.

I expect a bad ground could involve this behavior. I intend to explore the grounds more. Somehow the field winding transistor is getting stuck on. How a bad ground does this, I am not sure. I cleaned up the grounding connector on driver side prior to my post and images.

It's not that things are slow, it's that they are getting over voltage (fast motors, hot headlights). Battery light, and O/D off light are on. No check engine. Other posts I have seen described the headlights blowing out, I don't intend to test headlights during over voltage.

I have tried starting the car with the alternator-control-connector unplugged and there is no charging. This verifies that the high current side of the excitation transistor hasn't burned and failed as a short circuit inside the alternator. If the transistor fails short, the current would always flow (i drew an arrow on the image), so the unplug test verifies this is not the case. With the connector unplugged, the battery starts at 12ish volts and diminishes. I have not unplugged the connector while the car is running partly for safety reasons. I may build extension wires to do this..

This alternator design requires the computer command to charge/not-charge. The PWM indicates the computer is at least attempting to tell the alternator what to do. I agree that this is probably a bad alternator, or there is a bizarre grounding issue. I have already ordered a 5th alternator that will be here tomorrow. The Ultima brand one was over voltaging the oriellys test machine right out of the box, it hadn't left the store yet. The AC Delco one i got from rock auto seems to be doing the same thing in my car, I did not have it tested, but could. Maybe I can find some special aftermarket one that doesn't have this problem?

Yes, this has happened to 1 battery already.

I checked continuity for the feedback to the computer. I chose not to get scope trace, as the computer PWM goes to zero with rev. So whatever the computer needs to hear from alternator, it's probably hearing it.

I may end up doing more on this, but the continuity testing verified the wires aren't pinched together, and the computer's output signal is getting all the way to the alternator. I can also verify the computer can see the battery voltage (not sure where) but because the battery light turns on from computer, it must know what the voltage is.

yes I found that connector as stated above. will look more at grounds.

Thank you

 

I should have also mentioned in the previous post, I tried touching a ground wire from the star ground location on the driver side fender, over to the body of the alternator. This didn't seem to make a difference.

I also tried adding a 10k pull down and a 1k pull down resistor from the E connection to the chassis ground. The goal here was to promote turn off during the low portion of the control PMW. No luck.

 

So have you tried to start car then remove battery cable and put a volt meter on it and run at 2500rpm and see if volts start to climb or stay constant at 14v Im not sure if it with be @ that since theres not load on Alt. Watch it and see what happens if it goes up then its the voltage regulator circuit in Alt although u said that u put in a good alt? As per sensors when car is running at 2500, start one by one to unplug a ,sensor ,evap sol , temp sensor, egr valve etc. Just dont unplug cam sensor and see if the volts still climbs or stays constant. This obviously is a difficult problem basically u need to isolate a sensor from alt. A alt needs load attached ,to see if a load is pulling too much, and causing voltage spikes. But no matter, the alt has to provide contsant 14V to the battery and keep car running, so if u have a skinned grd cable from motor to body or wherever it can cause headaches. Try using a jumper cable and disconnect the car grd cable to battery, and go to the motor or another good ground to battery grd and see if that works.

 

I think I have solved the issue with this particular protege, but I'm not sure which of the few tweaks actually solved the problem.

Problem 1: This car has bad grounds. There are 4 multi point gnd bus' ... they all have paint underneath them where they connect to the chassis. This means the bolt threads are passing most of the current. There are two connectors under the headlights (requires removing the bumper...) one near the battery, and one in the cab under the driver side fuse box. I removed all of them, scraped the paint, cleaned everything, reaffixed some dielectric grease. Nothing looked super corroded, but I did it anyway, with the paint being a red flag for typical grounding practices. The headlights looked pretty gross, but the black was not really a burn, just road grime that had affixed. very weird. I also cleaned the ground on the transmission which is under the battery box. It was shiny and un-corroded underneath the filth, but did it anyway.








Problem 2: Yucky stuff in the fuse box... maybe some dielectric grease became conductive with some yucky stuff? over heated? all of the terminals were ok after i sprayed them. one seemed like it'd got hot at one point. i did not bother tracking down what all of the wires were. cleaned and moved on. The large fuse box connectors require sliding the green bar out. odd connector, but seems to work. Everything here got wiped down and sprayed with contact cleaner.


Problem 3: Maybe a bad transmission sensor?
As people mention the weird shifting being related to the alternators blowing up, I replaced a sorta ugly looking top transmission sensor. I had enough of the car apart so it was easy to replace, and not super cheap at ~102 dollars in 2023 money. I had considered replacing the rear transmission sensor as well (under car) but it oozed when undid the bolt, so i just tightened it back up.


I've put a few miles on the car and no alternator failure. OD off blinking stopped, and no check engine light... Golly, what a pain.

Of the 6 alternators i've handled in this journey.
1) had a lot of miles on it, failed a single brush. quit charging. almost like the current in the car was heating that brush more than the other? no clue.
2) Had chunks in it and wouldn't spin direct out of the box. returned
3) Failed test with over voltage instantly before leaving the store. returned.
4) used a rebuild kit from ebay on alternator 1... ran for 2 weeks or so before it quit. simply stopped charging.
5) B+ stud fell out during install. attempted to reset the stud. alternator ran for 15 seconds and turned into a giant short circuit. The test machine couldn't even turn it.
6) random Amazon purchase. claimed 'new' but was clearly a rebuild. Installed after doing all the above repairs, and seems ok for now.

The 10v pwm from the pcm was sort of a surprise. I thought perhaps it was high and blowing up the guts of the alternator. I guess 10v is normal based on some tips I got from someplace. I'm glad I did not replace the PCM as it would have been a waste of money in this case. This is a CA car, so the corrosion isn't really all that bad. I expect east coast already had these crushed.

A combination of poor aftermarket parts, and unrelated systems makes this one a doozie. I put this thread together as it would have saved me a dozen or two hours if it'd already existed on the internet.



EDIT: there was an AC Delco alternator between 4 and 5 that would over charge the car. So 7 alternators... that was the state of the car when I started the thread.