scooty

Ezone, I really appreciate you trying to help me find this problem. I am seriously considering buying an OBD link cable to hook my laptop up to the PCM and then using some OBDII software to interpret the info.

The TPS voltage is steady at .49 even when opened and allowed to snap shut several times. It increases steadily when the throttle is opened. I'm going to go pull the IAC valve again and check if it some how got stuck again.

ezone

Hey, I'm trying... best as I can do from where I am at.... Not sure what else, I'm fast running out of ideas here.
I might come up with something totally different if I could have the car in my bay at work.

ezone

Pull the battery cable off and short together for like 10 minutes to discharge capacitors and memory.... and retry the learn?

scooty

Ok it's back to the IAC valve again!

I pulled it off and noticed two things:

1) The valve was gummed up again. I remove the plastic portion and the rotating part feels sticky and gummy while trying to rotate it. I'm not sure how this happened because i cleaned it 2 days ago and it was rotating very freely by the time I was done. Anyhow, I have it sitting in a can of carb cleaner overnight. It is supposed to spin freely right? I can see the magnets having a rough time trying to rotate it fast enough if its gummy, this might explain the hunting idle.

2) I must have had a huge air bubble in the cooling system because I did not lose any coolant while pulling the lines of the IAC valve and the level in the radiator was low. I guess I didn't bleed it properly the last time I pulled the valve. I've heard this can mess up the idle.

ezone

Sweet!
https://www.civicforums.com/forums/p...31-post20.html
Keep cleaning.
YES it must move freely.

Already explained how an air bubble cannot directly affect the IAC control here: Air bubble is not the problem.
The IAC is not controlled by coolant temp directly.
Only the PCM controls it based on its temp sensor reading.
EDIT: There are more criteria the PCM uses to control the IAC, but its temperature input comes from the computers' coolant temperature sensor.

Coolant level would have to be low enough to mess with the temp sensor in the end of the cylinder head.

The IAC is heated by antifreeze in order to reduce sludge and prevent freezing shut of the air passage in cool and humid conditions (same as "carburetor icing"). It is not heated to control RPM.
There is no thermo-wax type temperature control like a FICV used to use.


Also, the IAC is a high spot in the cooling system. Air might collect there naturally if the system is low.

scooty

I tested the TPS sensor again as per the factory service manual, (see attached page from the manual), What do you think of these results:

[I opened and snapped the throttle shut several times while taking the readings and it did not make a difference.]

Resistance between pin 2 and pin 3 on the TPS sensor was 0.886 K Ohms. Manual says it should be between 0.5-0.9, so this seem ok

Resistance between pin 2 and pin 1 was 5.75 K Ohms. Manuals says it should be 4.5 K Ohms. This reading seems a little high. That's more than a K Ohm of resistance. Isn't that a significant amount when we're only dealing with 4.5 volts and only fractions of an amp in current?

Since the earlier voltage tests by back probing the TPS sensor appear to be within range (0.48 V - 4.69 V), the high resistance implies a higher than normal current load across that circuit, but I'm not sure if that would cause an erratic idle.

ezone

...............................Ummmm.......If you look at a wiring diagram (I randomly picked a 1995-97), the fixed resistance is between terminals 1 and 3. Terminal #2 is the sliding contact and measurement will change/vary with throttle position when read using either of the other contacts as a reference.

I can't tell you an answer about the resistance other than "maybe".
My opinion is that "I doubt it", but your results are not "by the book" so...do whatever you want. (Borrow one from a good car and try it!)

Here's why I think that:
The computer doesn't read/measure or use resistance of the sensor, it reads the output voltage. That's the only wire directly read inside the PCM, that carries all the information the PCM can use.
The other two wires attached to that TPS, (5VREF and sensor ground), are shared among other sensors.

My thinking is that as long as the voltage output is correct for any given throttle opening, it should be good.
But I also know it doesn't always happen that way, a skew of the output voltage readings at a partial throttle position is very difficult to determine even if you have another identical unit in your other hand to compare to. There isn't a special compass tool for charting opening angle vs. output voltage that I have ever seen. (At this point is becomes a diagnosis by process of proving everything else good.)

Something else here: When I am looking at stuff like this, I want to see what the PCM sees. That means a scanner and datalist.
Voltmeter readings by hand testing might not perfectly agree with values as seen by the PCM for various reasons (you are using a different ground point, sensor ground comes from within the PCM). When a value of .01v change can trigger some certain action, I want to be right in there to see it.
The PCM uses the 0.49v value to enable "idle program" conditions. It might turn those conditions off at 0.53v (random number!)....so if you aren't measuring exactly as the PCM measures, you might come to wrong conclusions.
If your TPS reads too low, you might be able to significantly open the throttle and cause the surge only because there is too much airflow and the idle programming is still in effect.
Make any sense?

The only thing I really ever consistently used resistance for when checking a TPS was a "sweep test", to make sure it didn't have dropouts or that "scratchy volume ****" effect when moving the sensor very slowly throughout its entire range of movement.
If the output voltages of a sensor are way out of whack, than I might pull out the ohmmeter to see why, but I gotta have a reason to go there.

Any 3 wire sensor generally uses a ground, a voltage (5v in this case), and a signal return (output voltage). A general rule of thumb for the vast majority of all manufacturers was .5-4.5 working range, give or take a smidgen. The general ranges from 0.0-0.5 and 4.5-5.0 were used to tell if the sensor or wiring was causing the voltage to be out of acceptable range (internal self diagnostics).
However, those are not hard and fast rules. The outer extreme ranges can be minimal as long as the PCM can recognize an open circuit and a short circuit as determined by the signal return voltage...they could use 0.2-4.8 and still be able to tell open or short if they wanted to.

99% of Hondas TPS are set to give 0.49V at closed throttle, and WOT is whatever it is.

(GMs used to spec 0.44v at closed throttle on some cars.)

Holy crap that was rambling..... I haven't even had coffee yet.

ezone

Big question: Did recleaning the IAC do any good?

scooty

Problem Solved!!!

It was the IAC valve all along. Looks like a dip in a can of carb cleaner overnight did the trick! The car idles like it's brand new with no more surging and hunting.

If the idle starts acting up again I'm just going to replace the valve with a new one. Thanks Ezone for walking me through the diagnostic steps. I was very close to buying a throttle body, TPS and IAC valve. So happy I didn't.

I hope the MPG picks up somewhat now.

ezone

Holy poop! 40 posts to reach this point LOL. Glad you persisted and found it.


I expect gas mileage will probably suck until warm weather and summer blend fuel hits the pumps.

Air up your tires.

My in-town mileage dropped a full 5MPG (sure seemed like it was all of a sudden) back in late October or early November, and that was before the temps dropped.... It's not even really cold yet.
My daily drive was getting 33-34 MPG most of the summer, suddenly it became 29 and still dropping.
Plus I stuck some 18" rimz on it earlier this month. That's killin it even more.

scooty

Crappy ethanol winter blend fuels. Try tuning a 600CFM Holley 4-barrel to run on it without getting worse than 12 MPG!!

I wish Honda would bring their turbo diesel Civics to the State already.

FredS

Hello scooty and ezone,
Have followed this thread with interest. My car has the minor surging you mention ie increasing to 1000 -1200 rpm while coasting in netural until stopped ect. Have cleaned IAC in position no change. Am told this is usual with 7th gen civic? This slight surging has always been there when I bought the car at 19000 miles 2 years ago

scooty

No it is not normal. I can take a video of my idle at neutral to show you what is normal car if you would like.

It took me several times to clean the IAC valve before I finally got it clean enough. Did you remove it to clean it?

FredS

Thank's for the promt reply. The slight surging is almost the same as your first video. It is nothing like that shown in your second video. It occurs when coasting to a halt in neutral. On going into neutral and taking my foot off the throttle the revs drop to approx. 1000 then surge to approx. 1200 then idles at 750 when stopped.
I have tried cleaning the IAC valve twice by spraying cleaner with the engine running and once by dismantling and cleaning. I have not left it in carb cleaner overnight.
I understand this is a "fly by wire" throttle control and the first car of this ttpe I have owned
The car runs very well except for this and clutch judder but that is another matter.
Thank you for your help.

ezone

"Fly by wire" means electrically actuated throttle. That started with 8th gen (06 and newer) in North America.

05 and older use a cable operated throttle.

FredS

My car is a British 2004 5 door Civic SE model with the 1.6 litre vtec engine.
ezone is correct it does have a cable operated throttle.
I was trying to describe the occassional feeling of vagueness between the throttle pedal and the engine response.
Apolgies for any misunderstanding

ezone

Ok, it's a different engine from what we have over here so components and locations may differ..... but the principles and strategies of operation are still the same.

As probably already mentioned elsewhere in this thread, look first for vacuum leaks, then unmetered air leakage.....Did you check out the rest of this thread?

Then an idle relearn after you figure out the original problem and fix it.

scooty

I would start by reading this thread beginning to end, you'll find a lot of valuable info. I would also start by looking for vacuum leaks, get a can of break cleaner and start spraying around every single vacuum hose. Dont forget the pcv valve, not sure where it is on your 1.6 but on US 1.7s it easiest to access through the left wheel well with the wheel removed.

When did this start happening? Was the iac valve ever messed around with before?

scooty

I think I might have a blown head gasket now

I've been trying to bleed the cooling system since I lost some coolant while removing the throttle body.

The only symptom I have is that the coolant overflow fills up almost to the top but never over flows out of the top, yet.

I have a little hand crank pump that I use to transfer coolant back from the over flow to the rad after it starts filling up. I've got a sharpie mark on the overflow and as far as I can tell I'm not losing any coolant at all. Because each time i transfer the coolant back to the rad, the level in the reservoir settles to the same point.

I did notice a little steam escaping from the rad cap, so I replaced it with a Honda cap from the dealer ship and have not seen the steam since.

The temp gauge is always where it has always been, the heat works perfect. Also, both upper and lower rad hoses are hot after driving for a bit, so I know coolant is circulating past the thermostat.

I just got back from a 200 mile trip and it never over heated. The coolant was at the sharpie mark before I left (half way between min and max). Immediately after I got back I checked the coolant and it was about 3 inches from the top of the overflow.

I let the engine cool for 4 hours or so and the coolant was still at the same place, maybe a tiny bit lower. I removed the rad cap and heard some bubbling sound and the level in the over flow reservoir immediately dropped about half an inch.

This has been going on for a week now. If it's the HG I'm not sure why the coolant in the over flow never reaches the point where it actually overflows. You would think during that 200 mile trip I would have lost some coolant out of the reservoir but I didn't.

Anyway, I'll do a compression test tomorrow and post the numbers here.

P.S. The idle behavior is still perfect

ezone

SOME change in the liquid level of the reservoir is normal as the temperature changes. Heat causes liquid to expand, cooling causes it to contract (shrink).
The radiator and cap control the pressure, pressure relief, and reverse flow that causes fluid to be sucked out of the reservoir when the radiator is under a vacuum.

It is all designed to keep the radiator completely full of liquid at all times under all conditions.

Any leaks can disrupt the process, akin to a pinhole in a drinking straw.
I would expect this was pressure escaping, not vacuum, if the HG is leaking. Squeezing the radiator hose prior to removing the rad cap should give a clue as to whether the system is pressurized or not.


This sounds odd to me, in that once you remove the radiator cap there is absolutely no way for coolant to flow out of the reservoir back into the radiator. You broke the seal, lost the vacuum necessary to "pull" the liquid into the radiator........
Unless there is a problem with the cap, the surfaces in the radiator neck that the rubber cap gaskets must seal against, a problem with the tube between the neck and the reservoir like loose fit, leakage, or blockage.

It might only need a certain amount of room in the system for the "air" bubbles to pass out of the radiator, so only a certain amount of coolant is displaced to the reservoir. OR it might be leaking bubbles slowly enough that it just hasn't belched a lot of coolant yet. I can't tell.
If the reservoir starts out much fuller than whatever you have it at, it may overflow.
Like if you start out just above the higher mark it might belch.....if you start out at the lower mark, that leaves more room.

Compression test will not show you this problem.

I would stick a funnel-fill funnel on the radiator

and let it run, watch for an endless slow stream of bubbles coming up.
If the engine seems to never finish burping its air out, then my next step is to put shop air pressure on each cylinder to see and prove which one is leaking.
I'm darn lazy and I don't like pulling a head without definite proof, and that test is proof enough for me.


You have read threads where I talk more about this method, right?

scooty

I couldn't find that funnel thing anywhere locally around here so I'll have to buy it online. Meanwhile I did a compression test any way as I already have the gauge. The results are as follows:

220, 217, 217, 225

These were on a fully warmed up motor, dry test, with the throttle wide open. They appear to be completely normal and fairly good for a 120K mile motor. Oh well, at least I know there isn't any major internal engine problems.

I also removed the overflow reservoir and removed its hose. I tested the overflow hose for leaks by blowing into one end and blocking the other. It was fine.

Ezone, are there any more tests I could do before sinking money into the funnel and coolant color changing exhaust gas test kit?

Also is the reason you were not recommending the compression test because the potential head gasket leak could be so minor it won't affect a compression test?

ezone

Good numbers. Well under 10% variation too.

Stick the reservoir hose in a container of water set up where you can see it, watch for bubbles coming up while you let it run?
All you need is a way to see the coolant level change or air bubbles coming up through it. I mentioned the funnel because that is what I use, most convenient for me.

Some of them don't want to leak much until the engine has a load on it, and pressure in the radiator reduces how much gets past the head gasket.
It's like $32, and damn handy if you do any regular work on cars. http://www.amazon.com/Lisle-24610-Sp.../dp/B001A4EAV0
Too many will pass this test, yet still have a head gasket problem. Not reliable enough for me.
Correct.

scooty

So blown head gasket it is then

I attached some rubber tubing to where the overflow reservoir hose attaches to the radiator filler neck and immersed the other end in an open container full of coolant. Please see attached video to the see the bubbles:


There are no bubbles at idle but you can clearly see a stream of bubbles when I rev it up.

Is this enough confirmation of a blown HG to warrant removing the head and replacing the gasket?

ezone

Was thinking....No funnel...

Fill the radiator to the just a hair below the top of the neck, then apply air pressure to each cylinder in my usual manner.
If the level raises at all, that's a leaker.

ezone

Posted at the same time! Ignore the previous post now.....

As long as the cooling system was completely burped, then you have enough evidence to continue IMO. If possible I would go ahead and do the air pressure tests on each cylinder so you know where to look closely after it is apart.

scooty

Don't have access to air pressure. I'm going to start tearing into the motor.