22 April 2018

Well, it’s better, I guess…

Took the alternator out and found it’s a CS130D model – I thought it was the previous generation, the CS130. The “D” version has more output, larger bearings, and better ventilation, though it is a bit larger.

It appeared to be in good condition so a cable connecting the alternator directly to the battery was fabricated, in addition to connecting the previously-unused “sense” wire to make sure the voltage is monitored directly at the battery.

At initial cold-start, battery voltage immediately went to 14.5, higher than before, but like before, after about 30 minutes of driving, it had decreased to 13.5-13.7V. The odd thing is that later during the same drive, sometimes it would go back up to 13.9-14.0V. If it slowly sank and leveled out, I’d think it was a temperature-sensitive voltage regulator and that would be the end of it, but this, huh.

A possible contributor might also be belt tension. Some sources claim that a serpentine belt is tight when you can just turn the longest span 90 degrees. Another source says that the belt should be tight enough that the engine can be rotated by turning the alternator pulley with a wrench. If it’s the former, I’m fine; if it’s the latter, I’m not. If the belt is slipping though, I’d expect it to get worse when a big load like the radiator cooling fan is switched on, yet when I did so when the alternator output was 14V, it didn’t drop much, implying that belt slip isn’t the issue.

The alternator might be defective, or maybe they all do this. The belt might be slipping, or it might not be. The alternator is mounted in the only place it can go due to the required reverse-rotation of the water pump. Because of the mount’s geometry, mechanical leverage increases as the belt is tightened. This would be a good thing if there was more space but there’s a chassis tube not far away, so there’s only a limited range to get it tight, and the next shorter belt won’t fit. Another possibility is to switch to the slightly smaller CS130 (though it’s only about 0.3″ smaller) but may still not allow getting the next shorter belt on.

Yet another variable is the pulley ratio. An aftermarket alternator manufacturer claims their alternators are good to 18,000 rpm, though I’m not sure whether  OEM units are the same. The ATI crank damper is 5.5″ diameter and the alternator pulley is 2.7″ (68.6mm) – an odd size regardless of units. Alternator speed is scaled by pulley diameter, so if the engine’s spinning at 8200 rpm, the alternator’s spinning at (5.5/2.7)*8200, or 16,700 rpm, so there’s a bit of wiggle room left. The math says that the alternator pulley could be reduced to 2.5″ to increase charging at lower speeds. The thing is, the belt routing is sort of like the letter “B”, such that it nearly touches itself as it passes over the water pump pulley. If the alternator pulley diameter is reduced, the belt surfaces may touch and greatly accelerate wear. Maybe a pulley could be added to increase spacing, but messing with pulley diameter is sort of besides the point since sometimes the voltage output is fine. Further muddying the waters is how sometimes, when I’d rev the engine up, battery voltage would increase, implying it’s a pulley ratio problem – yet other times it wouldn’t make any difference. Ugh.

Summed up, I’m not sure where to go with this, leave it alone and keep an eye on it, or chase it down until root cause is identified.  Lastly, this article is really helpful in understanding alternators.

In other news, I was wondering why my exhaust pipe was suddenly pointing off to the side. Turns out that the flex joint in the exhaust has completely cracked through. That’s nice, but it didn’t ruin my 100-mile drive through the back roads!

16 April 2018

Whether or not the alternator gets swapped out, it looks like its wiring needs revisiting. The alternator connects through the fuse box to the battery. The catch is, I added a battery cut-off switch upstream of the fuse box and bad things can happen if the switch is opened with the engine running. At best, the engine quits as it should, though I now think it won’t. Since the alternator feeds in downstream of the switch, it would likely self-power the system and the engine would keep running. At worst, not only would the engine keep running, but with no voltage reference, the alternator could generate voltage spikes high enough to damage the ECU. I’m too chicken to try it and find out.

The alternator feed-in point needs to be moved directly to the battery. As a related issue, the wire connecting the alternator to the battery is way too small (I blame the Painless Wiring kit but should have known better). It’s 10-gauge but needs to be much larger; the PowerMaster site claims that a 7-ft wire running 125 amps requires at least #6, if not #4. Yes, 125A is extreme, but having everything on in stop-and-go traffic with a flat battery could get there, albeit briefly. Wire size doesn’t seem like it could be the sole cause of the voltage drop, but it can’t be entirely ruled out either. Copper has a positive temperature coefficient, meaning the hotter it is, the more resistance it has, which causes voltage drops with current (this is why a hot starter won’t crank an engine). Still, air temperature rising only 20C wouldn’t be enough to account for the problem… only there’s more to it. The alternator wire runs down the center tunnel, sharing space with the coolant pipes. There’s a fair chance that it’s seeing around 60C or so, increasing resistance by around 16%. That still doesn’t fully explain the situation because the wire temperature in the center tunnel likely doesn’t vary much. Guess I just have to try it to  see; right now I still think there’s still a good chance the problem is an overly-sensitive voltage regulator in the alternator. Regardless, the alternator wire needs to be upgraded no matter what alternator is used, so this is a cheap experiment before spending money on other things.