Electrical Gurus-Check it out please!
 

I have a perplexing issue (at least to me) that has recently cropped up. Here's what's going on...
I had Palo Alto Speedo make me a tach. Shortly thereafter (and since verified unrelated) I noticed that sometimes the tach would be bouncy and almost at zero at idle. Sometimes it was perfect, but it was inconsistent.

For reference, my car has very little electical load. All LED lights, fuel pump and ignition are all the constant-use components.

I recently added an electric heater to the car. Direct battery power with a relay to switch it on.It has a variable fan, but pulls 40amps max.

If I have the car running, and I turn on the heater, the tach will go from flaky to dead nuts perfect. While watching the Megasquirt gauges, I also note that the ignition advance is flaky at idle, just like the tach, but perfect with the additional load of the heater.

In addition, and this is new, when I shut the key off, power to the gauges fades away over a few seconds, and the car keeps running for a few seconds, like a relay is slowing de-energizing. If the heater is on at the same time, it shuts right off like it should.

I have a USB socket with a digital voltmeter built into it on the dash. It charges at 13.8/13.9 at idle. Headlights drop it 13.7ish. Adding the heater will pull it down to 12.9/13.1, but it seems like its doing its job.

I have checked an cleaned all grounds.

Is this a sign that the alternator is going away? Voltage is maintaining, but maybe the amperage output is weak?
Thanks for reading!

Did you check the ground at gear box too?

Shut down is not right, frankly would scare the poop out of me if my car did that.

When you mentioned you have Megasquirt it made me wonder if you have electrical interference that is messing up the tach signal. I’m assuming you believe the car is running fine and the tach is just not showing the correct rpm.

In reading many posts here on EFI conversions, noise/interference seems to be a common issue.

Just a thought.

You need to observe the DC voltage with an O-scope, your DVM won't show the the ripple or transients.
Try an R/C device on megasquirt input voltage to snub the voltage and noise to input specs.
This all assumes good grounds all around

Thanks guys. Yes the indicated rpms are incorrect. The megasquirt rom is solid so the noise /issue is in the electrical system and the noise is causing the bounce. The load eliminates the. Lie somehow. Could a bad diode in the alternator do it? Hounds are good and clean.

Erik,
Greetings.....I've had two projects....one AEM and one M'squirt/MSD Streetfire that exhibited the "run-on" problem. I added a Diode to the Alt/blue wire as shown in this thread.....

http://forums.pelicanparts.com/porsche-911-technical-forum/982002-msd-diode-placement.html

and the problem went away.....

Have you spoken to Bob Ashlock about the tach issue?
http://ashlocktech.com/



regards,
al

Thanks Al. I will give it a try. It’s strange that it just started happening thousands of miles after the install. But we’ll see

OK, the diode did the trick for the run-on issue. Thanks Al!

But the tach is still flaky unless there is an electrical load placed on the system. I suppose I could install a resistor in the power feed to the tach? I wonder if a light bulb would do it? trying that now...

glad the diode did the trick. Just fired up a 3.0, this morning...diode not required...some need it, some don't.

Call Bob Ashlock to discuss the tach issue. He is the expert.

regards,
al

also, to clarify, the tach input is shared with the input to the megasquirt. The megasquirt reads steady and correctly, so something is amiss with the tach power/ground/signal interpretation.

I jumped the ground direct to battery with no change.
I jumped power with no change.

I put a test light between the Tach positive terminal and the power to the tach and that works like a charm. So the tach wants more load on the circuit for some reason. Best way to address this?

ok i installed an instrument light inline and its all good. light barely even lights up, so that is odd also. measured voltage at the power lead is correct.