johnboat13

Okay, the p2646 code. It has persisted since getting the car, and just won't go away.

Here's what I've done; Checked the Vtec switch/solenoid assembly, by that I mean taken it off the block, cleaned it, took off the pressure sensor and used a compressor attachment and voltmeter to check for resistance changes, and it is working, took the solenoid itself off and triggered it with 12v, and that is working as well. Tried resetting the ECU, however the light came back on. Checked to make sure that the assembly was grounded, and it is. Changed the oil and filter, with the correct 5w20 oil. The light still comes back on after about 20 minutes of driving.

Now for the second part of this: The high beams are on in both the normal lights and high beam positions. There is a sound like a snapping/arcing/shock noise coming from underneath the hood every now and then.

I think the problems are related, maybe a loose connection or a grounding problem, but I could be wrong. Is there a connector that contains both of these things, or in the same portion of the harness? I checked everything that I could find this afternoon to make sure everything was plugged in, the only thing left is maybe the ECU, but I really didn't want to have to un-bury that.

Any help would be very much appreciated, and I can't get the CEL to stay off to save my life. I posted this in another forum but they didn't seem to have any insight. Thanks!

ezone

What year is the car? USA? Canadian?



You took the VTEC valve off of the head, was there any debris caught in the screen?





P2646 VTEC Oil Pressure Switch Circuit Low Voltage

This is what the PCM saw to set the code:
When the VTEC solenoid valve is ON, the VTEC oil pressure switch remains ON.

Sounds like the switch did not change states when the VTEC system was activated, or was stuck.

I have seen an occasional car with intermittent, random sticking of the pressure switch. Very difficult to catch in the act. I have replaced a couple just on a guess, for no other likely cause was found at the time. No comebacks, so I must assume it solved the problem.
I can't say this is your problem, just that I have seen it.

Couple days ago, I replaced a stuck pressure switch on a '12 Civic with only 400 miles. But it was constantly stuck in the ON state.



HTH

johnboat13

My bad, details: 2004 USA EX.

I checked the screen, and there wasn't any debris in there really, although I cleaned that too and put it back. I picked up a couple extra of those switches from the junkyard, and did the test with the ohmmeter and compressor, and those responded similarly as the original one in the car. It's possible that all three are bad, but it seems unlikely if they all respond the same.

And you're right, really hard to get it to react consistently. I'm going to try to pop that last switch I have in and see how it goes, and if it still throws the code. It's hard though to justify purchasing a new one for $90 when all three of these seem to act in the same fashion.

ezone

Did you happen to grab the entire VTEC valves, or just the switches?
If the solenoid or the valve stick and stop working, that would also cause problems.




The engine oil is full, right? Low oil level is probably the most common cause of VTEC codes. Can't keep sufficient oil pressure in the engine if the oil is low.

Did you do any oil pressure checks that are (should be) listed in the service manual? If the engine oil pressure is too low, the VTEC will be affected.

Service info says it needs at least 57 PSI @ 3000 RPM.




As far as the headlights, a good wiring diagram and some logical troubleshooting are going to be needed.


HTH

johnboat13

Yup, pulled the whole assembly, and yup oil is full, triple checked that. Thanks for the spec, that will come in handy.

Is it typical for ECU's to go bad? I assume that's the next logical step to go after the oil pressure is checked.

Matt_75

Did you physically actuate the VTEC valve?

An entire VTEC assy is $146.

A VTEC oil pressure switch is $44 and a new ECU is $405.

Why would you even contemplate the ECU at this point given the information you've been provided?

ezone

I can do this with the scantool (bi-directional controls) on some cars, I can do it manually on the rest of the cars with jumper wires.
I raise the RPM and activate the solenoid, I can hear the change in the sound of the engine as the valvetrain actuates, and at the same time I can watch the response of the pressure switch on the scanner or with a test light.

I can watch it all happen while driving, using the scanner.


Not in my area.

No, not logical, except to someone loading up the parts cannon.

For reference, here is a handy pic of a parts cannon:

Note the red shirt and the label on the battery....




Test, don't guess. HTH







.

johnboat13

Hah, parts cannon. I was just curious if the ECUs would exhibit symptoms like this if they were bad, or the likelihood that they go bad. I don't intend on buying one, if I can't justify a $90 switch, a $400 ECU is out of the question.

I don't have a scantool, so to manually activate it, I can stay in neutral, raise the RPM to 5k+, and then actuate the solenoid with some jumper wires, yeah? And monitor the signal of the pressure switch with a voltmeter?

ezone

It's one of my favorites.

You got a code reader?
Does it give freeze frame data?
What RPM, road speed (VSS), and coolant temp was the code set at?

Pretty much. 5k is a little high, try 3k or so. Apply power and ground to the solenoid, listen for the instant change in the engine sound (usually obvious unless you are deaf).
You could use an ohmmeter to check/monitor the pressure switch, it should change EVERY TIME the solenoid is activated. Actuate the solenoid about 100 times in a row, make sure the switch works each time. I think a light bulb on the pressure switch is much easier than a meter to see, that's why I said a test light.
I could even rig a couple of test lights up inside the car to watch the action while driving.

Since all I know about your car at this point is you have a code --with no supporting data-- I'm thinking the switch could be intermittent, meaning it didn't work one time out of several.
Heck, if the solenoid didn't work one time, that would give the same code too.
All the computer knows it that it did not see the switch change when it turned on the VTEC solenoid. YOU have to figure out why.
(Watching data on a scanner tells me if the computer saw the change too, a meter or a light can't tell you that info.)

Or maybe rodents chewed some wires? Big possibility here.


HTH?

johnboat13

No, no code reader, I only had it read by autozone/the honda dealer that looked at it.

And yeah, test light would be good. I'll give it a shot. If it turns out that it the switch doesn't register one of those times, where do you guys typically order parts from?

This will narrow it down anyways.

ezone

So what was the dealers recommendation?


I work for a dealer, I use OEM on most things.