2003 GTX 155 Mid Range Surge & Knock Sensor Code

[bidema]

Hello, Thanks for the wealth of knowledge here! It has been invaluable!

So here is what I have going on. I recently purchased a 2003 GTX 155. It was said to have an electrical problem when I bought it so I took a chance on it. Turns out I gambled and lost. The engine was seized, so a "great" deal turned into an only so-so deal. I recently installed a 2005 155 engine w/50 hours. Got it in bolted up and lake tested Saturday. Engine runs okay. I am getting a mid range surge and a knock sensor code P0326. I am getting a 300 ish RPM bounce at approx 4200. I am also not getting to peak RPM as top RPM is 6500.

Worth noting, before putting it back together I did drain the fuel and replace with fresh along with having the injectors cleaned and calibrated. It seems to run/idle smooth and if I push past 5000 the surge goes away. Again though I am not seeing rpm's over 6500.

My feeling is that the knock sensor is a symptom not the cause, can my thinking be flawed here? Could the knock sensor be my culprit?

Of note also I have a code for P0113, which in my mind is unrelated and just in need of a new sensor.

I am inclined based on what I have read to either a) bring it to the dealer and get it diagnosed via BUDS, or b) try and MAPS sensor. I know thats throwing parts at it, but I hate the idea of spending money at the dealer.

Would the Candoopro help me in this situation? Not opposed to purchasing it, but again don't want to unnecessarily spend money either.

Any thoughts would be very much appreciated!! Thank you!

 

[Sportster-2001-951C-Stock]

PO113 - Map sensor full range - I'm sure the engine won't run well this way, probably reverting to a lookup table instead of the normal constant feedback.

MPEM will respond to high knock count by retarding ignition advance.

 

[bidema]

Okay, so update, even though my existing intake temp sensor tested with range on the ohm meter, I replaced it and my p0113 code is gone-at least on the trailer.

P0326 is still present, however I took another knock sensor off my old motor and plugged it in. The code goes away with the sensor loose and hanging outside the engine compartment(clearly false reading).

So next question, what are the chances a bearing going out in the jet pump would be picked up by the knock sensor as a motor knock? I can feel/hear something is not quite right and it feels like a pump bearing though I don't know exactly what that feels like, it seems as though it could be the culprit.

So I can "trick" the motor into running without codes and up to rpms, again on the trailer as high as I dare. I still think the surge at 4200 rpm may be present, but cant tell for sure out of the water.

At any rate, it seems logical that the knock sensor might be picking up this bad bearing?

Thoughts?

 

[Sportster-2001-951C-Stock]

Ah yes, po113 is intake manifold temp sensor, not MAP sensor. Sorry about that. MPEM will be happier if it thinks it knows engine temperature.

Yes, seems logical knock sensor could be picking up a pump bearing noise, might consider removing pump and drive shaft for engine testing but will have to plug the opened PTO spline or oil will shoot out of the PTO while running without the drive shaft installed.

But while the jet pump is off you can investigate if the impeller shaft bearings feel like they're binding or lose.

 

[mkov608]

I highly doubt the ECU will think a bad pump bearing is detonation inside the cylinder. The pump seal is a 100 hour replacement item, so if you don't know how many hours on on that part, I'd change it and inspect the bearing while you're at it. Be aware, jet pump parts (like all Sea-Doo parts) aren't cheap.

As for your lack of top end, in my case, I had an exhaust leak. Use some SeaFoam spray (follow the directions on the can), and if you have an exhaust leak, you'll find it quickly ... smokes like crazy! I changed my leaky J-pipe, and I instantly knew the problem was fixed. I was back to 7K+ RPM. Like you, my ski was hovering around 6.5K RPM.

 

[Sportster-2001-951C-Stock]

An exhaust leak will fill the bilge with carbon monoxide and displace oxygen, definitely a possibility for the surging and low power.

High knock count will retard ignition last I checked, this also will limit RPM's due to late ignition timing.

 

[bidema]

Great rabbit trail. That actually makes a lot of sense! When I pulled the seat off after a quick 10 minute ride I noticed smoke in the engine compartment. I chalked it up to the engine change. Since then I ran it my next two test rides with the seat off.

I questioned my j pipe when putting it back on but could not see anything visually. What went wrong with yours? Dd you replace it with stock? Were you throwing the knock sensor code too or maybe that is unrelated.

I appreciate your post I never would have thought of that but with my smoke deal it makes complete sense. I will try the seafoam. Where was your leak? Any other direction you have is very much appreciated as it fits my symptoms.

I did order the pump bearings and seals as a pm deal.

 

[Sportster-2001-951C-Stock]

Also an exhaust hose that's internally delaminated can cause an exhaust restriction.

 

[mkov608]

I'm not sure what went wrong with my j-pipe ... factory defect I guess. the ski only had 43 fresh water hours on it. My j-pipe was leaking at the 6 o'clock position between the flange and the stainless steel bellows. I replaced it with a like (used) item. Ski-doo hooked me up.

The Sea-Foam will reveal all your exhaust leaks for sure!

Good call on the pump PM. PMd my jet pumps at 80 hours since the skis were 12 years old. I have two GTXs; one had water in the pump, and I had to replace the bearing (pitting). The other pump was GTG. Ski-doo recommends a 5 year PM on these style pumps ... says they are bullet proof. I've seen videos where some will pull the cone off during witnerization just to make sure there isn't any water inside.

 

[mkov608]

Don't forget the tools.

Great rabbit trail. That actually makes a lot of sense! When I pulled the seat off after a quick 10 minute ride I noticed smoke in the engine compartment. I chalked it up to the engine change. Since then I ran it my next two test rides with the seat off.

I questioned my j pipe when putting it back on but could not see anything visually. What went wrong with yours? Dd you replace it with stock? Were you throwing the knock sensor code too or maybe that is unrelated.

I appreciate your post I never would have thought of that but with my smoke deal it makes complete sense. I will try the seafoam. Where was your leak? Any other direction you have is very much appreciated as it fits my symptoms.

I did order the pump bearings and seals as a pm deal.

 

[bidema]

Quick update. I did find a leak, got a fresh water one from Jess cairns-you ever need used parts he has a lot of them. He hooked me up with a fresh water pipe. It's on the way to me so as soon as I get it I will toss it on. Not sure it will clean up the knock sensor but one step at a time I suppose.

I also know what you mean about the hoses delaming. Had that on my challenger after I rebuilt the mercury in it, those are okay in this one.

As far as tools, did you find it necessary to buy the bearing and seal tool? I have an impeller removal tool but I held off on ordering the tools though I am concerned about setting the bear at the correct depth. Thoughts? Never done it before but after all the engine work I am not afraid to tackle it...

 

[mkov608]

I bought all the tools that were recommended in the service manual. As for what was necessary, I'd say the impeller tool, the bearing and seal tool, and a vice. The right tools always make the job easier, and since I didn't have to take it to the dealer, the tools paid for themselves quickly. I purchased all the tools to pull the drive shaft too since I replaced my carbon seal and bellows.

 

[bidema]

To come back and update so hopefully this helps someone else. I put a new intake temp sensor in like I said before, and that solved that, but then the knock code was still there. I followed the advice here and got a new "J" pipe. I installed that and thought before I do the pump bearings I would test run it on the trailer. Everything is good, knock code went away. If the weather allows this weekend I hope to get on the water and give it a good water test. If its fixed I may let the pump seals go until winter as we only have a few weeks of riding left.

Thanks so much for the great advice to look at that J-Pipe.

 

[bidema]

Water tested today. When I left the garage I had no codes present. I put it in the water and warmed it up. Immediately upon going over 5000 rpms the knock sensor code appears. Back to shore, plugged in a different knock sensor and just let it hang in the engine compartment, no change, as soon as I hit 5000 the knock code appears. Tried holding it in my hand no change. Putted around the lake a couple laps staying below 5000 and it ran well. Got back to the garage same thing on the hose. I am convinced it is not the sensor since trying 2 different knock sensors.

I just now(on the trailer unplugged plug "A" from the ecm. Popped the back of the plug open to visually inspect the wires, all appeared okay. Closed up the plug, plugged it in and started it on the hose and no code.

So two questions, likely this is in the plug or wires or ECM. Do I chase a wire? Is this a common thing? Second question, can the ecm store the code data and does it need to be cleared to get the code gone? I thought I had it....

Motor sure seems to run smooth right up to 5000 before the timing gets retarded by the ecm after the code trips. Then I get my surging RPM. I don't know how to hear a spark knock as this is my first 4 tec, but it sure does not seem to have a knock.

Anyone?

 

[mkov608]

First I suggest you read this: http://www.contactmagazine.com/Issue54/EngineBasics.html

Have you considered that your sensor and wiring are fine, and that you really do have a detonation problem? In the worst case scenario (you do have a detonation problem), if you don't fix it, it will destroy the engine.

 

[Sportster-2001-951C-Stock]

Yes, be concerned about detonation. Especially for an engine equipped with a puffer of some sort.

If reseating the "A" connector seems to have eliminated the knock count code than perhaps one of the terminals aws making poor contact (for whatever reason)

I think the computer will store all the various codes for retreival using the factory diagnostic communication port until engine is operated for XXX successful restarts with no occurance or they are cleared using the diagnostic equipment.

Poor or low fuel quality can be enough to cause detonation, leading to high knock count and eventually the computer responds accordingly by attempting to reduce knock count via retarding ignition timing (and perhaps even adding fuel, depending on algorithm and severity).

So if you are experiencing high knock count then approach as you have been, your troubleshooting process seems to be going down the right path of eliminating possible causes.

I think the knock sensor will detect knock long before you can hear it, some algos actually use the knock sensor to obtain the highest possible amount of ignition advance and thus power/octane available from the fuel and conditions. That is, the computer advances timing until knock count reaches some predetermined (reasonably low, such as 5~20 events) then calls that maximum advance.

If you were to somehow disable or circumvent the knock sensor hardware, you might not achieve full performance but 1st you'll have to fool the software. I'm pretty sure the factory guys did anticipate someone would leave the sensor disconnected and thus have some sort of test embedded.

Once the computer is satisfied by the values being recorded by the various sensors and no concerning parameters are exceeded (such as temperature/knock count/RPM/etc.) the computer will leave limp mode and operate normally, engine will perform as if nothing had happened. Codes will be stored for XXX restarts. Some codes may be retained as preliminary pending codes, not current fault codes, assuming a particular fault has occurred only once or rarely.

I suggest for further study, referencing the entire list and description of codes this system is capable of generating, to help understand it's potential for self-diagnostics.

 

[bidema]

I appreciate the responses, you guys have already helped me more then you know.

My reasoning that I have a wiring problem it this. If I plug in a completely different sensor and hold it in my hand, wouldn't the ecm think the knock sensor is in place correctly, but held in my hand there is no way it could be picking up a knock is there?

You are right though I am very concerned about there being an actual knock. The engine that came out of the ski had piston #2 completely destroyed, needless to say I don't want that, thus only short water tests until I get this figured out. I bought the ski non running so I have no idea the behavior before the previous motor blew.

Again my reasoning that its in the wiring is to have the second sensor showing a knock when held in my hand. Feel free to correct me if my thought process is flawed here.

Side note, I have a full tank of fresh premium fuel(I drained and replaced before even starting the new engine), also had the fuel injectors cleaned and calibrated before starting it as well.

Would I hear a knock similar to a car engine?

I have read several other reports where people say they found a wiring gremlin. Just not sure if I would be ill advised to splice into the wire harness to run two new wires for the knock sensor.

 

[mkov608]

Some interesting facts from that article as relates to Rotax engines. Rotax does a great job of squeezing a lot of power out of a very small engine. That makes detonation even more serious. From the article "An engine that is making 0.5 HP/in3 or less can sustain moderate levels of detonation without any damage; but an engine that is making 1.5 HP/in3, if it detonates, it will probably be damaged fairly quickly, here I mean within minutes."

So how does this align with a 1503 cc engine that produces 155 HP? Well if my math is correct, 1503 CCs = 91.71 cubic inches. If you divide 155 by the cubic inches you come up with 1.69 HP/in3. So that means your engine falls into the author's later example ... "damaged fairly quickly." Those supercharged engines are even more prone to damage when detonation occurs.

I recently read a post from a guy overseas who had a problem with one of his fuel injectors firing anytime he put the DESS key on the post. He had a bad ECU. So in your case (previously destroyed #2 piston), maybe it's not a mechanical problem. Perhaps the ECU is sending fuel to the cylinder at the wrong time, sending the wrong amount of fuel, or some other combination of problems. I don't know if a dealer can put these engines on a scope, but I'd look into that before I operated the engine anymore.

 

[Sportster-2001-951C-Stock]

Alternatively, tossing out thoughts here... How about the fuel injectors, have they been recycled from the original engine or swapped?

A plug that's in a cylinder experiencing detonation can take on an appearance as shown in this photo.

If you're listening for detonation by ear, expect it presents itself a bit differently in this engine turning 5,000 RPM compared to a station wagon engine turning 2,500, frequency will seem much higher otherwise essentially similar. The higher frequency can mislead you into not recognizing it as detonation.

At least I can say the 1st marine engine I ever heard detonating, I didn't recognize right away due to the higher frequency. Come to think of it, many people have never experienced an engine without an automatic spark-knock timing retard, have they?

Anyway, I don't expect you will obtain detonation while on a trailer unless somehow ignition timing is way out in left field and then I expect it wouldn't run well at all in that case.

 

[bidema]

Fuel Injectors have been cleaned and recalibrated by Brucato FIS. They have done several sets for me. Anytime I get a new watercraft I automatically send them there as a cheap PM deal. More important on the two strokes then the 4, but still worth it to me. My plugs all look good at this point and time. I only have maybe two hours on the motor since I put it in.

Sportster, would I be able to definitively hear the knock with the seat off in the water or on the trailer? I just don't hear it. It runs perfect and smooth right up to 5000, then the code trips and the timing retards.

Again, there is NO WAY the ecm should be getting a knock sensor code from a sensor plugged in but held in my hand should it? I just don't see how it could. I am really not thrilled with the idea of hacking into a wire harness, but everything sure points to a wiring issue and not a detonation. I have read on the forum of others who had wiring issues. Clearly I don't want to be running it any more then necessary to track this down, but if I was having a knock wouldn't I hear this?

 

[Sportster-2001-951C-Stock]

I'm not a 4-tec technician so I cannot speak with absolute authority. (FYI)

I have some thoughts though......

I expect it's unlikely this engine will detonate often enough to set codes while sitting on the trailer, under most any conditions. ie: I believe etonation can't occur without a load in gear underway. One case I had the sensor mounting bolt was lose and the sensor was rattling around on the bolt, knock count was out of range in neutral no load, falsely.

Thus if you're getting knock faults while on the trailer I'd say it's false, or there's some noise you can't hear. I've seen false knock before on other engines (not 4-tec), that was coming from a lose motor mount and another from lose exhaust components banging and clunking. It's a common "rough test" to take a small hammer and smack repeatedly against the block near where the sensor is mounted to confirm the ECU is detecting knock.

Newer systems are considerably smarter than older ones were though, they only check for knock signal for a few milliseconds following each spark event. Older not-so-smart ones will count about anything rattling if loud enough in the right harmonic, such as a rod bearing knock, even.

If the knock sensor is in your hand, it's not electrically grounded to the engine block. Assuming a ground is required then it's not picking up knocks but with no ground the ECU might consider it defective (as if disconnected)?

It should be a fairly easy task to run some resistance checks of the harness and perhaps sensor from the ECU end of the harness, no?

This knock fault may not even relate to the sensor circuit directly, it's possible the harness may have been pinched/cut at some point or there's a ground wire somewhere that was overlooked, causing a ground loop.

 

[bidema]

Okay, so wanted to report back now that I finally solved the problem. It was indeed a bad wire(or maybe resistor) in the wire Harness. I am not a whiz with tracking down electrical problems thus the duration of my problem here. When I finally started testing, I went first from the pin on the plug at the ECM to the plug going into the Knock sensor. I used a simple continuity tester for ease and what I found was no current in one of the wires. It appears that the two wires coming off the ECM plug go into some sort of resistor? I can't say for sure if the short was in there or further down the line, but after gathering the courage to tap into the main wire harness, I ended up splicing into the harness a new wire on the back side of the plug going into the ecm and on the back side of the plug going into the knock sensor. I did it this way so I could still use the plugs, though someone with better soldering skills could have likely soldered straight into the plug. I knew for sure testing it the wire was not carrying current so I was pretty sure it was the issue.

Put it in the lake this beautiful holiday weekend and it ran perfectly! Right up to almost peak rpm, and on the gauge I touched 60 with a full tank of fuel. Ran 3/4 of a tank of fuel through it and about 50 miles and never missed a beat. I am thinking I got it solved.

Thanks everyone for the help!!

 

[Sportster-2001-951C-Stock]

Good job man, very nice!

 

[Hawk388]

I would still pull the plugs and inspect them for discoloration. The best way to do this is take it out, ride the snot out of it, pull it off the water or onto the beach and check all plugs right after a full throttle run. If they're all paper-sack brown to a slightly darker color, you're good to go. If any of them are white, something isn't correct in that cylinders combustion process.