2G sometimes it pulls, sometimes not

[Mike_GST]

Hello guys.

I looked around for threads like this, but none of them presents irregular working like mine.
I have now my '98 GST for one year and half and something is bothering me since ever.
My car sometimes pulls strong and sometimes not. No reason I can notice.
Same day, same temperature, same time, same conditions. I pull once, the acceleration presses my body to the sit. Then stop the car and pull again. The rpms goes up, the pressure gauge shows full boost, but the car has a very weak performance. In this same exactly pull, if I release and quick the accelerator, the car reacts full performance again! Pretty wird thing.

Everything in the car is stock. When I bought it, replace all the belts, spark plugs and fluids.
At first I was suspicious about the wastegate, that an erratic opening and closing could be doing that.
The car has VERY LOW MILEAGE, so being sitting could make some moving parts get sticky, I thought
Then I released that little rod that connects it to the actuador and moved with my hand right and left.
Very smooth movement! Is not the wastegate.

Well, I don't believe this problem is related to the air flow, neather from or to the engine. It couldn't be such a leackage on the lines, way to make the performance drop that much, without been noticeable in the gauge, which by the way, shows full boost.

Do you guys have any guess?
Thank you very much.

 

[Mike_GST]

At this point the thread title doens't make sense anymore! Just when I tryed to figure a small and rare issue, after a long period, it became a nightmare in just two months!
AKI 95 octanes gas didn't make any difference. And now I think the car is in the worst possible performance. Here goes the pattern I realized now. Even the stock boost gauge not been a reliable tool, a pattern can be observed on it. It has 3 marks splitting its scale in three thirds. Well, as long as I don't cross the first third of boost, the engine runs fine. If I cross this mark, car starts to shake, rpms don't go up, even though can hear the turbo spooling and the gauge showing boost increase. Releasing smoothly the throttle pedal until the boost goes back to the first mark or under, engine goes back to its normal working. Also, this day, I took off the back seat trying to hear something wierd coming from the gas pump. Nothing drew to my attention. To check the pump, took the rpms all the way to the red line more then once, without crossing that boost first mark and no failing episodes happened. After 6krpms I can floor the throttle without the engine showing that shaking performance, even though the pull is still weak. Long story short, the problem appears when from middle to high boost.
Boost leak went back to the most probably hypotesis. Still working on the leak test.

 

[ThunderChild]

This sounds like a boost leak, maybe around the intake manifold-to-cylinder head gasket. At low throttle/vacuum, it can handle a slight leak or cover it up. Then you go positive pressure, and it leaks out faster than the engine can consume it (or at least a higher rate than able to compensate for), then once you get to higher rpm and boost, you over-run the leak and the car gets timing and fuel "close".

That's just a guess, but I'd put a few bucks on it. Also, the stock boost gauge is pretty worthless, but it's at least indicating for you.

 

[90talontsi69]

Pull your timing or run a different intervooler

 

[Julio Mendoza]

Your turbo may be starting to fail.

 

[WisconsinMonSTAR]

NGK BPR7ES plugs @0.028"
New NGK blue wires

New bypass valve - I recommend the Evo IX stock valve, they're ~$150 brand new from extremepsi. That's the valve I got to replace my stocker, and I'm in the process of getting a flange so I can run it on my current setup pushing 20psi. It's a direct swap for the 2g.

Boost leak test

New knock sensor. You can tell for sure that it's bad if the back looks kind of gooey. It can still be going bad even if it doesnt look gooey. My car had very similar symptoms to your first post when my knock sensor was going bad. The times where it just feels like it's dogging out, even as RPMs increase and the turbo is producing boost, is most likely timing being pulled. It almost feels like the car gains 1000 lbs on the front wheels.

That's what I would start off with, because it will all cost less than $250, and your car will need all of those things anyways. Imo, fresh ngk plugs, wires, and a bov upgrade should be some of the first things anyone does when they buy a stock 2g.

Pulling the plugs will also help determine the issue to an extent. If they're all white and jacked up, you'll know you're running lean and should look at fueling. Super fouled out will tell you that youre running rich and probably have a boost leak. You need to swap those delcos out for NGKs amyways, and make sure they're gapped properly!

Also, I forgot an aftermarket boost gauge in that list. A nice simple mechanical one will do fine.

 

[miliman13]

Mike, man I have given up on my thread, after several months of walls, based on the same issue.
I did every thing you did and more.. But members naturally assume your just missing something supper obvious, and thus you must be lectured relentlessly on simple concepts that are frustrating to hear... I know the struggles man... I wish you well and hope to see a solution.

 

[Mike_GST]

0-30psi gauge already in my hands. I'm still searching for the connections and stuff for the leak test. I'm gonna use a tire and a needle valve as an air compressor, hope it works
Since I ran the engine until dry the tank out of that low octane gas, can hear some very smooths "POW-POWs" comming out of the exhaust on idle. That made me remember something. When I took the car to the mechanic, to replace the ordinary (filters, fluids and belts) when I just bought the car, he told me he put a flushing product into the gas tank. Well, as the issue became noticeable in a progressive way, a hipothesis could be dirt getting loose and making the filter get saturated in a very short period, restricting gas feeding to the engine. It has only 3k miles on it, but depending in how bad was the tank interior... Got a Mitsubishi original filter (two actually) and I'm in the middle of the replacement. That damn 14mm nut is welded shut. Gave up on the open wrench, will buy a flare wrench next saturday and pray to God to brake it loose with no damage.
Since I took all that stuff out to reach the filter, now i can (hardly) see the knock sensor. Will try to stick my cell phone closer and take a picture of its face to check how it looks.

 

[WisconsinMonSTAR]

I just replaced my fuel filter a few weeks ago, and using a line wrench is a must! I rounded mine off even with a line wrench, and as a last resort had to use vice grips while unscrewing the filter from the top.

I'm not trying to condescend you at all. I've run into many issues where I was overlooking basic things and looking way further than I needed to. To me, this sounds like a boost leak issue (95% guaranteed if using the stock 2g bov) or a dying knock sensor. I use a $30 portable air tank from walmart to perform my boost leak tests. You can even fill those tanks with a bike pump.

I highly recommend swapping out the delco plugs simply because the NGK plugs are cheaper and 10x better for these cars.

Be patient and don't lose hope. You'll feel great once you figure it out and get it running well!

 

[Travis92talon]

It sounds like its fuel starved at higher boost. Fuel filter swap is a great idea as a first step now that you mentioned the fuel tank flush. My tank was full of rust when i opened it up to change the fuel pump and i ended up getting the tank cleaned and relined by a professional company since i couldnt see putting a brand new pump in a rusty tank.

 

[Lou98GSX]

if you have an air compressor handy you can use an impact gun to loosen the nut while holding the filter with a open end wrench . i have taken many filters off that way .

 

[Mike_GST]

Will get tomorrow the 14mm line wrench for the fuel filter, look for NGK plugs and parts for the leak test. A friend of mine had a great idea: throw some ice around the 14mm nut, so it's going to contract a little and maybe come off easily.

Here goes pictures of the looking of my actual Delco plugs (3kmiles) and the knock sensor. Please let me know if you notice anything wierd, or if they are normal.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

 

[Travis92talon]

That plug looks like it's running slightly lean. Normal plug should be more tan than grey. Lack of fuel under boost could cause this. Car will retard timing to try and compensate which kills power. I can't see gap on the plug from that angle but make sure they all gapped correctly. Fuel filter would be my first step and making sure all plugs are gapped right. I wouldn't mess with knock sensor until you retest the car with new fuel filter.

 

[Mike_GST]

Line wrench, WD40 and ice cubes around the nut/line for 30 minutes didn't make it easier. Actually, I ended up rounding that f***ing nut. Son-of-a-*****!!. Had to take a rest to pull myself togheter again! Will figure something out. Starting with the vice grips and ripping the filter from the top, right WisconsinMonSTAR? What's next?

Good news are I got all the parts for the leak test and a friend of mine will borrow her 12V tire air compressor.
Anyways, no filter no play.

Plug gaps are 0.8.

 

[Mike_GST]

Took the fuel filter off holding the bottom nut with vice grips and turning the filter from the top. Still had to apply a lot of force. Nut head is wasted!

One idea is to cut the line end, remove the nut and then use a connector like this:

You must be logged in to view this image or video.

And on the filters bottom use something like this:

You must be logged in to view this image or video.

And between them a high pressure (60psi) gas resistant hose, fixed by clamps.

Another idea is to keep the wasted nut, use the vice grips again, but mount on it an adapter like this:

You must be logged in to view this image or video.

The same thing for the filter side as the first idea.

Do you guys had any experience on that?

 

[Mike_GST]

Yesterday I mounted the leak test and run it just for curiosity, since it was already dark and hardly believed I'd figure something out on that condition. At first I tested the device on its own, let it reach 30psi (end of the pressure gauge scale), closed the needle valve and wait a minute. No leaks detected.

You must be logged in to view this image or video.

Then connected to the compressor side inlet. The pressure hardly reach 6 psi. After needle valve closed, pressure dropped fast to 3psi and then slower to 0.

Sounded a very noticeable leak coming from the BOV recirculation hose.
Simply covering it with my hand, reached 10psi and pressure didn't drop that fast after needle valve closed.

Today I tested for leaks with the recirculation hose covered with a PVC cap and a clamp. Big difference!
Still could hear something like a dripping coming from inside the engine. Then I took out that tinny vacuum hose from the top of the BOV to check how much air was leaking through the TBI. Started to loose pressure faster!

Than I decided to remove the intake hose where it’s connected to the TBI and cover it also with a PVC cap and a clamp, that way I wouldn’t be fooled by the TBI and internal engine leaks.

Here goes the test results, considering there’s no leaks through the BOV, neather TBI. As I’m using a tire as an air tank (MacGyver style), we can only evaluate how fast the pressure leaks, not how much pressure it can handle. Wouldn't be that purpouse anyway. I think it gives a good idea on how bad or how well my system is doing.

22psi drops to 3psi in 65 seconds
22psi drops to 5psi in 48 seconds
15psi drops to 5psi in 35 seconds
13psi drops to 5psi in 31 seconds
10psi drops to 5psi in 23 seconds

Even avoiding losses through the TBI, I still could hear that dripping noise from inside the engine. Maybe it’s something related to the turbo oil line connected to the sump. To be very sure, I'll try to mount the leak test straight to the compressor side outlet hose.

What do you guys think about the leak test so far?

 

[Mike_GST]

Since I definitely need a new BOV, do you guys know if it could be replaced by a VW / Audi 4cyl turbo ones? They cost half of a chinese HKS or Greddy, brand new easy to find, genuine made, recirculating and also flange-free (clamped to the intake hose) just like 2g DSMs. Sounds like a perfect plug&play, just gotta check if the hose diameters are the same. Bad thing is they are plastic made too, but replacing it every 10-15 years of use is something I wouldn't mind at all.

http://produto.mercadolivre.com.br/...o-prioridade-golf-audi-18-turbo-058145710-_JM

 

[chrysler kid]

Those plugs look very hot, as mentioned system is running too lean. You need to replace them soon and re check your gap. I'm not aware of what technique you are using to measure the gap but the plugs you pulled out look like they have been running lean

The fuel filter is a great start, I am sure that Brazilian gas is not as clean as it should be and your filter has plenty of sediment stuck in it.

Your drop in boost pressure test doesn't seem too bad, it never hurts to replace the blow off valve as I'm sure the condition gets worse as the plastic valve gets warm with the engine. You should verify the diameter of the VW part is the same as your blow off valve, perhaps see if the local dealer or junk yard has one you can measure.

Sorry for not checking back with you I didn't realize the condition had gotten worse. The plugs being so burned white really indicates to me they are getting too hot and not getting enough fuel.

Ethanol can be good for tuning and turbo cars, but in a factory fuel system the injectors must send more fuel to accommodate for the octane difference. The next free maintenance modification I would suggest is to hot wire or rewire the fuel pump to increase fuel pressure. The modification can be done very cheaply and can increase your fuel pressure.

From your pictures your engine bay is extremely clean.

So to summarize I would do the plugs and wires and fuel filter and if you notice the car is better after those I would plan a day to do the fuel pump re wire. I don't think your knock sensor is bad but I really think it is pulling timing because your car isn't getting enough fuel. If you had a bad boost leak it would cause the car to run rich, which is not the problem you appear to be having.

Are there any local tuning shops who work on import cars in your area? I would want to call and ask them if they noticed that vehicles they tuned needed more fuel to run on the Brazilian grade fuel. If there aren't any I would do the fuel pump rewire regardless because it requires only wire a relay and some time and it will give more fuel

Best of luck bud
-Ben

 

[Mike_GST]

Made three more tests.

FIRST: covered the intake hose end that connects to the TBI and mounted the leak test in the hose that connects to the compressor side outlet and goes to the intercooler. Throw easy 15psi into it and let it rest. In this condition didn't hear that dripping/bubbling noise. After 2 hours pressure was still the same. Incredible! Hoses and intercooler are fine! (1st line on the diagram)

SECOND: mounted the intake hose back to the TBI and let everything else same the first test. Still no bubbling noise. Results below (2nd line).

THIRD: mounted the leak test again in the compressor side inlet and the intercooler hose back to the compressor side outlet. Except the BOV recirculation hose closed (no doubt it's bad), this one tests the same condition as the car works. Bubbling noise back. Results below (4th line).

The 3rd line represents the test I made the day before.

You must be logged in to view this image or video.

As you can see there is a small leak through the TBI and injectors, but a larger one through the turbocompressor. In this second case, probably the pressure goes away through the turbine axel housing (IT ISN'T A BEARINGS SYSTEM RIGHT?). I imagine the housing has little holes where oil is injected to lubricate the axel. I'd speculate that during its working, spinning 200 thousand RPMS, the oil pressure in such a small gap, would perfectly seal it and this leak would simply disappear.

Anyways, maybe those numbers are far better than the acceptable and I'm just making too much big deal from it.

The ones that already made that test, please, how does it look?

About the BOV to replace, the stock one for VW / Audi 4cyl turbo has 25mm of diameter for intake and recirculation and 5mm for vacuum; Eclipses is 34mm and 6mm. I could build an adaptor do make it fit, that wouldn’t be a problem, but the point is, wouldn’t this downsizing change the overall working of the engine?

Chrysler kid, I measured the gap using a caliper rule. About the adding of ethanol in our gas always made classic imported car owners mistrustfull about it. Anyway, I never heared about tunning those cars just because of the ethanol mix. Authorities say they do a lot of tests before approve this changes to the fuel, trust what they say is a much different talk. Brazilian cars right now are almost 100% build "flex", we can use 100% gas (25% ethanol) or 100% ethanol or mix them in any proportion, but we still import a lot of cars that suppose to run just on gas. The thing is they work and never hear or read about problems or lack of power. Will keep you tip about the fuel pump rewire in mind. Thank you.

 

[chrysler kid]

Well good luck then sir, a boost leak would cause your car to run rich. Your plugs don't tell that story

 

[Mike_GST]

Well good luck then sir, a boost leak would cause your car to run rich. Your plugs don't tell that story

Maybe the fuel filter is in a much worse condition than the bov, causing plugs whitening. And maybe, there something else going bad and I still didn't figure it out. Last saturday I bought all the parts to fix the gas line end and mount the new filter. Will work on that this week. As soon as I do this, will test the car still with the bad bov. If the car stops to choke during medium and high boost, well then the filter was the worse problem.

 

[Mike_GST]

Ok guys, fuel filter new mounting system done and filter replaced. I was afraid of leaks, but I checked for it by accelerating the engine with the battery connected by wires from outside the engine bay. Worked like a charm! No leaks detected.
The connections were made of stainless steel. If you go through this, please do not use copper/brass alloys, it oxidizes gas in a few days, in worst case scenario contaminating the hole tank. The only special thing was made to the female connection: there was used the filter nut (cuted off from the old filter) and welded to the bar before machining, because of the sealing system. To the male part I used an o-ring sealing system.
From now on replacing fuel filter will be piece of cake.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

After all, that choking behavior is still happening when turbo boosts. I also notice the radiator fan turning on more frequently. I checked the cooling system when I bought the car and I remember I had to wait forever to see the fan running.
The new VW/Audi BOV is arriving today, will get it in the post office this afternoon. Still have to buy a rubber hose to adapt the valve to the car, since VW/Audi BOV is 25mm and the Eclipses BOV is 34mm.
I’m also ordering NGK BPR6ES plugs.

 

[chrysler kid]

If it is relevant I was messing around with my eBay type s blow off valve trying to get it to sound like a real type S without any success. I changed the diaphragm type to a type s ngr imitation and notice I can get the bov To flutter with the large spring, but with the soft spring the car will only hold 5 psi of boost and build boost very slowly because the bov wasn't sealing

Anyways a boost leak from the bov makes the car very sluggish, I could still feel the boost kicking in but it wasn't until about 3k rpms I felt anything

With the bov operating correctly the t25 hits full boost at 2500rpms with a good torque spike with it.

The other question with the vw blow off valve is to see what pressure does a factory Volkswagen run, you may be over powering the valve at the 15psi spike that the t25 will give.

Also be sure you re tighten all of your clamps on the boost piping they do tend to get a little loose.

Also the evo 8? I think blow off valve is a direct upgrade and made out of metal. You could try to call a local performance shop to see if they know of anyone local who may have upgraded and could sell you a stock part, or see if a Mitsubishi dealer could source one

For all of the people who despise the fake eBay type s blow off valves they legitimately do hold pressure and work very well. Now if I could just get it to make any noise at all when it releases boost I would be much happier

 

[Mike_GST]

The new BOV holded only 5psi with compressor on, 2psi with compressor off. Those are the same exact numbers I was getting with the stock (maybe bad) BOV.
This made me suspicious about the way I'm doing the leak test.
Lets call the vacuum line as intake line, just to avoid misunderstandings in what follows.
The intake line provides vacuum to pull the piston up the way to release boost.
But when WOT, the intake line, instead of vacuum, provides boost.
And here comes the question:
Is this boost also needed in the intake line, to push down the piston the way to help the spring sealing the valve?
If the answer is YES, than my leak test is faulty and I need to do some changes on it.

The other question with the vw blow off valve is to see what pressure does a factory Volkswagen run, you may be over powering the valve at the 15psi spike that the t25 will give.

A3 / GTI 150hp 9 psi
A3 / GTI 180hp 12 psi
TT 225hp 16 psi

Those models all use the same BOV.

 

[laster969]

Hello guys.

I looked around for threads like this, but none of them presents irregular working like mine.
I have now my '98 GST for one year and half and something is bothering me since ever.
My car sometimes pulls strong and sometimes not. No reason I can notice.
Same day, same temperature, same time, same conditions. I pull once, the acceleration presses my body to the sit. Then stop the car and pull again. The rpms goes up, the pressure gauge shows full boost, but the car has a very weak performance. In this same exactly pull, if I release and quick the accelerator, the car reacts full performance again! Pretty wird thing.

Everything in the car is stock. When I bought it, replace all the belts, spark plugs and fluids.
At first I was suspicious about the wastegate, that an erratic opening and closing could be doing that.
The car has VERY LOW MILEAGE, so being sitting could make some moving parts get sticky, I thought
Then I released that little rod that connects it to the actuador and moved with my hand right and left.
Very smooth movement! Is not the wastegate.

Well, I don't believe this problem is related to the air flow, neather from or to the engine. It couldn't be such a leackage on the lines, way to make the performance drop that much, without been noticeable in the gauge, which by the way, shows full boost.

Do you guys have any guess?
Thank you very much.

sounds like a wastgate or pressure clutch

 

[laster969]

That's exactly what I'm saying. Don't floor it at low RPMs. Go half throttle until 3500RPM then roll onto the throttle. See if that helps with consistency. Also the stock gauge especially in 2G cars doesn't mean squat really. It's more of a decoration. Like @BlueArcher said, its just calculated from what the MAF sees, not actual manifold pressure. You could have all your air spewing out of boost leaks and the boost gauge would still be pinned. Go do what's called a boost leak test. Here is how you do one:

You must be logged in to view this image or video.

You must be logged in to view this image or video.

Now while doing the test, fix the leaks. Fix all of them. Even a small one can throw numbers off for the ECU. If you eliminate the boost leaks, then the problem isn't from that and we can move on to another theory knowing that a common issue has been fixed. If it fixes it, then you are done and the issue is resolved.

Like the video just made my own boost leak tester.