Flat spot when accelerating (around 3-4k, under load)

[JedMeister]

Hi all,

 

I have had my Subi for a few years now and I love it! It's a 2lt '98 GX Liberty (Legacy in the rest of the world) Wagon.

 

But it has one minor issue, For as long as I've owned it, it has had a flat spot around 3-4000RPM when under load. I didn't notice it initially when I test drove it, but within a few weeks of buying it I noticed it.

 

It's most noticeable on the open road - under heavy acceleration, like up hill or overtaking. It accelerates up to the flat spot fine and once over about 4k (sometimes as much as 4.5k). Within the flat spot it just has no go.

 

Whilst it is generally between about 3.5-4k, sometimes it happens as low as 3k and doesn't resolve until 4.5k (sometimes even closer to 5).

 

A couple more things I have noticed about it:

- Seems much less noticeable when the car is cold (air temperature doesn't seem to affect it at all)

- It's always much better (for at least a few days-weeks) after it has new spark plugs.

 

The guy I bought it off said that he had replaced the coil pack recently (and I have no reason to doubt him, but no proof either).

 

I recently had the timing belt done and the mechanic siad that the old belt was one tooth out and I hoped that that may have been the issue but it wasn't...

 

The only other issue (which may or may not be related) is that the idle is a little unstable sometimes.

 

I had just got used to it but since I thought it may have been fixed it has been really bugging me...

 

I've had a bit of a google and this is the only post I found that sounds the same as what I have. Unfortunately there is no clear answer there...

 

Anyone got any ideas? Even ideas on how to properly diagnose it?

 

Thanks in advance.

[mikaleda]

sounds like tps sensor to me you can pull one out of a junk yard for like $10. i did this on my moms 90 legacy i just cleaned the corrosion off before i put it on and it worked great its still working actually.

[1 Lucky Texan]

TPS can def cause symptoms similar to yours.

 

Some other things to consider;

 

valve adjustment

 

plug wires (if aftermarkets were installed when the coil pack was put on, replace them with Subaru wires - many people have problems with 3rd party plug wires)

 

Engine Temp Sensor (since you mentioned a difference with temp)

 

 

are there any codes stored?

[Fairtax4me]

Knock sensor sounds likely. Or the MAF sensor needs to be cleaned.

Seafoam in the intake can do wonders for problem like this as well.

[JedMeister]

You guys rock! Thanks for the rapid responses.

 

After my googling I was thinking TPS or Knock sensor seemed like possibilities.

 

Just thinking though, wouldn't a dodgey TPS tend to cause the hesitation regardless of whether it was under load or not? (Going by the name, it reports throttle position back to computer...?)

 

@1 Lucky Texan - By valve adjustment, do you mean valve timing? If so, I would assume that that the timing belt replacement (and readjustment of the valve timing), or is there something else I'm missing..?

 

Actually the more I think about it, the more likely knock sensor seems to me. Or perhaps it actually is knocking? (Someother fault causing the timing to be to too advanced or something). I have never heard it knock, but the flat spot occurs under the same conditions in which it would be knocking if the timing was too advanced (and it didn't have a knock sensor).

 

@1 Lucky Texan - Excuse my newbness... but how do I find out 'are there any codes stored?' I'm assuming I'll need to connect to the computer somehow. After a bit of a google and I came across this on ebay. Is that what I need (connected via bluetooth to laptop or smartphone)? Or is is there a better way to go?

 

Can I just disconnect the knock sensor and see what happens, or will I need to get a replacement to test if that's the issue?

 

Guess it's probably worth checking the MAF sensor not sticking and run the multimeter over the TPS too

[1 Lucky Texan]

I just dunno if your car has hydraulic valve adjusters or not. If it has the 'bucket' /solid lifters, after a lot of miles ( 120K? 150K?) the valve seats can wear and the valves may not be closing alll the way. Not sure if you mentioned the mileage or other car details.

 

Here, a 98 would be OBDII and I was wondering if the check engine light was on, and codes retrieved, it may help with diagnosis. The link you have is one of many types of code readers.

 

Maybe someone knows if you can get a 'blink' code from your car.

[mikaleda]

you have to have a reader for obdII the one you linked to is a bluetooth reader i am not familiar with those it looks like it might be a two part reader but i'm not sure. you can usually go down to autozone and they will hook you up with one for your car for fairly cheap... usually.

 

or there are ways of checking your tps and knock sensor with a multi meter simple and easy but i wouldn't trust chiltons. i did this on my moms 90 legacy and the book was totally wrong.

Edited July 15, 2012 by mikaleda

[Fairtax4me]

You could use a code reader, but if there were anything serious enough to cause a problem right now, the CEL would be on. No CEL probably means no codes.

 

Knock sensor has to be connected. If you unplug it the ECU will default to safe mode where you get almost no spark timing advance, and no power. Often the knock sensor will fail without setting a code. Knock sensors are cheap on Ebay, cheap enough that even if it doesn't fix the problem you won't feel like you've wasted anything.

 

Your thinking that a TPS problem will show only with a change in throttle position is correct. The TPS typically has nothing to do with a problem that shows up only at a certain engine speed. If the TPS were a problem you would notice changes in engine power by varying throttle position.

[JedMeister]

Thanks again guys, I really appreciate the input.

 

@1 Lucky Texan - I neglected to mention mileage it's done nearly 220,000km (~138,000mi) so I'll try & find out if it has hydrologic or solid lifters.

 

@mikaleda - Yeah it's a bluetooth one so you can connect to it via a laptop or smartphone (according to my googling there's a few different apps that'll do it).

 

@Fairtax4me - Thanks for clarifying. No CEL so there'll be no codes...

 

So I think I'll get on eBay and get that Bluetooth connector & I'll also get myself a new knock sensor & see how I go.

 

Thanks again guys.

[JedMeister]

Hmm...

 

Been having a look on eBay for a knock sensor and it seems that there are 3 different ones. The AA031 (which I'm pretty sure isn't the right one - it's for older Legacy/Liberty models), the AA061 and the AA070. I've read in places that the only difference is the plug. Now from the (poor eBay) photos that I've seen the plug on the 061 and 070 are the same and the 031 is different. The main difference between the 061 and 070 is that the 061 has a non-detachable lead that plugs into the wiring harness (like this) and the 070 has a short lead that can disconnect from the knock sensor itself (like this).

 

I've had a look and mine looks just like the 070 (has the detachable lead) and the cheapest one I can find (inc postage to Aust) is this one but it says it's for "2.5L 1999-2005" which is obviously not what mine is. Do any of you guys have any idea about this? Otherwise I might just shell out the ~AU$30 and try my luck.

[bluedotsnow]

check out http://opposedforces.com/parts/ great parts reference and they delineate between USA and Europe.

[Fairtax4me]

They all work the same AFAIK. There were a few different connectors, (some were one port, some were two port, but both have only one wire to the sensor) but you can pop the white harness connector off of one and put it on the other if you need to. (Ive done it) Very easy with a small screwdriver.

[mikaleda]

or you could do like i said earlier and test it. the chiltons is right on vaules on the ohm meter you just have to figure out where their telling you to connect your leads. the knock sensor is easy just connect the positive lead to the sensor and the negative lead to a good ground it should be 560K ohms i think i don't remember.

[Fairtax4me]

Even if it checks out fine for resistance values, it doesn't necessarily mean the sensor is working properly. In this case the sensor is likely sending an erratic signal that says knock is actually occuring when in fact it may not be. Also, if the resistance value were widely out of spec, the ECU would pick up on that and set a code for it, since that's part of it's check sequence. A mere visual inspection of the sensor housing will tell you more about the health of the sensor than a resistance check.

[JedMeister]

Thanks heaps again guys. I'm blown away with how helpful you all are! Wish I'd found this forum a few years ago and then perhaps I would've fixed this issue ages ago... I guess it isn't fixed yet but I got a good feeling...

 

@bluedotsnow - Thanks for that link, that is super excellent! It doesn't specify Australian models, but from what I can gather the Aussie models are the same as the European ones. And it turns out (assuming that is how it is) the knock sensor on mine is the AA070 (like I thought) which is the same part no as the US '99-'03 2.5lt (as advertised on the eBay listing I was thinking of buying).

 

@Fairtax4me - Thanks for your reassurance!

 

@mikaleda - Thanks for your suggestion, but reading on other threads/forums regarding knock sensors, it seems that testing knock sensor resistance isn't really enough, especially for a problem like mine (like Fairtax4me said). Apparently the only reliable way to test a knock sensor (or any other part) is to test it while the problem is showing (ie for me - while my car is experiencing the flat spot). After talking to a mate of mine (bit of an electronics whizz) he reckoned the only way to reliably test it would be using an oscilloscope on the knock sensor with the car on a dyno. That way the flat spot could be recreated while the signal the knock sensor is sending back to the computer could be monitored.

 

Long story short, it's cheaper and easier for me to just buy the knock sensor and hope for the best, so that's what I'll do.

 

Thanks again everyone for all your input. I've just bought that US part and will put it in as soon as it arrives. I'll try to remember to post back to let you all know how it goes.

[bluedotsnow]

do you have an after market air filter? like a k&n that allowed more air flow? I have a K&N stock box air filter that allows more air flow than a tru stock air filter. I had noticed a slight improvement initially but then it faded and recently my car felt sluggish even after a nice oil change with lubromoly. then I filled up with premium and WOW it goes like its in FWD restored the head into headrest pull I'm used to. at your next fill up try using a higher octane fuel, please let us know what kind of fuel and air filter you a running I'm curious since this fixed my flat spot problem.

 

always glad to help!

[JedMeister]

Just have a standard air filter. I have read that a high flow air cleaner is the cheapest, easiest way to get a bit more go out of the Subi.

 

I have run it on Premium ages ago but didn't notice any difference at all.

 

Still waiting on my new Knock Sensor to arrive, shouldn't be too far away.

[Setright]

The same-age Subaru's I have owned would behave like this when I tried to run high octane "fancy pansy" fuels in them.

 

No problems with regular, additive free 95 octane. That's RON 95, the European octane scale...which has higher numbers because we rank fuels in the best theoretical quality.

 

"Premiun fuel" would be RON 98.

 

 

Have you tried filling up from a different brand/station?

[bluedotsnow]

MPG calculation does not lie I was getting like 20-25 city/highway just yesterday on the way to a hiking trip to tahoe I did 304 miles on on 10.9 gallons of premium. averaging 70mph with cruse control, 20 miles of off roading, and about 5 stops along the way. 27.88 mpg! next time I fill up I'm going to just use mid grade to see if I get same results then back to regular for a comparison.... if I see a change I'm going to put a new stock subaru air filter in run regular and compare. will post results.

[bluedotsnow]

The same-age Subaru's I have owned would behave like this when I tried to run high octane "fancy pansy" fuels in them.

 

No problems with regular, additive free 95 octane. That's RON 95, the European octane scale...which has higher numbers because we rank fuels in the best theoretical quality.

 

"Premiun fuel" would be RON 98.

 

 

Have you tried filling up from a different brand/station?

 

 

fancy pansy!?! lol its just a more stable fuel that ignite less prematurely. if your pistons fuel "wad" gets "busted" before TDC Top Dead Center of the stroke your waisting fuel and may be causing damage to your engine. the ECU adjusts engine timing to compensate but......

 

since switching to premium I have noticed I am seeing and clearing a p0420 code much less frequently than when running regular octane.

 

as I understand it premium leaves more of a residue so it has the potential to clog injectors faster.

[Setright]

Aye...Octane rating is a measure of how easily a fuel will self-ignite under heat/pressure.

 

Fancy pansy....a collective noun for fuels with names like "BP Ultimate" and "Shell V-Power"

[JedMeister]

Here in Australia 'Regular' is 91, 'Premium' about 97-98 (depending on where you go) and I have seen some others at around 100(+/-). I use Premium+ in my motorbike and you can notice the difference in that, definitely goes better/smoother/more power (I guess it comes down to power-to-weight). But never noticed any difference in my Subi.

 

On a sad note I don't think my Subi is OBDII compliant Like I think I said above, I got one of those cheap eBay bluetooth OBDII connectors and no joy (won't connect to ECU). I guess it could be a faulty one though... My car definitely has the plug, but when I pulled it out, it's only got a few wires in the back of it. I guess I need to test the eBay bluetooth OBDII connector on another car.

 

Although I don't know if it's conincidence or not, but ever since I had it plugged in and spent some time trying to get it to work, the flat spot has been really bad when it's cold and right down around 2,500 RPM (if you recall originally the flat spot was much less noticeable when the car was cold, now it heaps worse - once the car warms up it's like it was before).

 

Do you think it's possible that just plugging in the cheapo OBDII connector has caused this? Or is it just coincidence do you reckon??

 

Still waiting on my knock sensor to arrive...

[mikaleda]

have you done a visual check on your knock sensor? it wouldn't hurt to try, and have you tried a different octane since your problem started. that might tell you some thing there. the main job of the knock sensor is to adjust timing to compensate for knocking due to cheap gas which is why it didn't matter what fuel you put in your subie before, you might notice a difference now if your knock sensor is bad. just a thought might be worth a try.

 

as for the reader not working your car is obdII it is probably just the cheap reader, you get what you pay for.

Edited July 24, 2012 by mikaleda

[JedMeister]

Yeah I've done a visual check and it looks ok but I figure with something like that a visual check is never enough.

 

As for the flat spot, it's had it as long as I've owned the car (~4-5 yrs) but I just got used to it. Different fuel has never made any noticeable difference in any way, shape or form.

 

Although I haven't tried any other fuel in since it seems to have got worse (I only hooked up the cheapo OBDII thingo night before last...) Once I put the new knock sensor on I'll try again.

 

BTW, lots of people say that the knock sensor needs to be tightened to the correct torque, is that important?

[mikaleda]

yes if it is too loose or too tight it will make a difference easy thing to do since it is only one bolt. another thing to remember is to put the wire the correct way also, it has to be pointed in the right direction. and yes visual checks are not a good test.

Edited July 24, 2012 by mikaleda