HT! I have had this problem for some time now, and still cannot figure it out.

If I was too messure any of my signal voltages and turned my lights on or hit my brake the coltage would got up!

In other words it seems like i dont have a voltage regulator anywhere on the car.
you can only imagine what thats doing to sensors like TPS.

for example... tps at an idle .45 then with headlights turned on .789ish and if i hit the brake pedal and lite up the tail lights .964ish. WTF?

I have replaced the ALT hoping it was the voltage regulator in it, no fix!
I have check and double checked Grounds, all of whihc check out fine.
I have replaced and tried many different computers... all of which, same thing.
I have replaced the TPS sensor as well.

H23 with A4 vtec head in a 97 civic ex. never had this issue untill I put the h23 crank in, had to swap to an obd1 dissy and do the little splice thing C1 & C4 and C11 & C14

Im not gettting any ELD check engine light, only getting a TPS light because voltage is too high and pissing off the computer. any advice?

 

I'd recommend going through the service manual's directions for checking the alternator wiring.

 

When the computer senses increased draw on the electrical system at idle, it usually opens the IACV a little to compensate. This raises the idle and therefore the alternator output to supply more power to the car. This is why cars usually idle higher when the lights and other electrical components are turned on. It's normal and there would be problem if the voltage did not increase with electrical draw.

 

Correct (for the most part). To elaborate:

The computer senses this extra power fluctuation via the ELD or electrical load detector within your USDM fuse box. If this ELD is malfunctioning, it may yield a charging issue. When there is excessive draw, the ecu switches something within the alternator to increase the magnetic flux produced within, thereby increasing the power output. To compensate for this extra load on the crankshaft (via the alternator's altered load), the ecu also raises the idle via the IACV.

If you are concerned about this, check your ELD voltage levels (especially if it is a overall governing electrical issue). Someone else may be able to help you out (I don't know much past what I have stated).

EDIT: If it is still within running specifications, the ELD may not trip a CEL. Don't know though...

 

no charging issues. just too much voltage at the tps, wiring was checked, it seems to be fine, why the extra voltage? which is throwing a voltage too high tps code.

 

I also noiced my wideband starts off at 4.4volts and when I give it any hard accel, it goes up to 20.0 and stays there, its acting all funny like static noise in the chassie, yet grounds and voltage drop test fine.

 

Make sure ecu is fully getting chassis grounded (its mounted and not floating around the floorboard). Check the grounds to the ecu. Make sure your TPS 0 volt and 5 volt lines are clean of noise and are appropriate voltages correspondingly.

This is odd. 4.4 volts seems high to start with. Shouldn't it be around 2 something? I know you checked but, engine ground. Run an extra one or something, and a hefty one at that (I have a 0 gauge ran to my block). I know this is a stretch, but if you have an oscilloscope, you could check your electrical noise levels by moving the probe around the questionable areas (but not many people have O-scopes handy).

That is all I got, sorry. Can't think of anything else at the moment...
/scratch head

 

A problem with the electrical load detector would likely cause too little voltage or none at all. It would be unusual for the ELD to be working "too well."

The TPS output is a percentage of the TPS input. For example, if you had 10v going into the TPS and turned it 10% of the way to fully open (lets assume the TPS is at 0.00v when closed and at max voltage with WOT), you'd get a 1v output. If you have 5v going into the TPS then 10% open would be .5v and the ECU would read it the same as the 1v/10v reading. Since the TPS reference voltage changes while the motor is running, the sensor must be tested with the motor not running and the key in the ON position.

You might want to test for drains on the battery. It could be that the ECU is overcompensating for a slightly excessive load. It's unlikely but an easy check. Disconnect the battery and check the resistance across the positive and negative connectors of the harness. Anything but infinite ohms is a drain on the battery. Small drains, like from the radio memory, are normal.

 

While I agree with most everything you said, I still suggest checking the ELD circuit for the following reasons...

The reason it could be working "too well" is simply because the current reading could be wrong for various reasons (the ecu thinks it sees a higher current flow than what actually exists). This can happen simply by raising the voltage level on the ELD load sense ground node. This can happen by a bad ground, bad termination, or bad ELD unit, etc.

Here is a so-so article on the ELD unit. What I suggest is just checking the ELD node going to the ECU and see what it reads. If it reads extremely high (close to 5 volts I believe) then it may be your issue. If it reads fairly low, then you can probably ignore it entirely, call me an idiot, and move on to other suggestions.

http://www.civicforums.com/forums/21...-detector.html