paul_s0
paul_s0 New Reader
3/16/19 12:20 p.m.

Let's play diagnose the problem over the internet.  

Last week I went up into the highlands with my '08 Mazda 3 (2.0 manual). First time doing the trip in this car (have done it several times without issue in my previous '08 Mazda 3).  Just before starting the climb from sea level to over 13,000 fasl, we filled up with fuel.   After about an hour, probably around 8000 fasl, I started to feel a slight misfire, mostly below 3,000 rpm in 2nd (it's a steep climb and with the power loss and lots of hairpins from around 7,000 fasl up to around 12,000 fasl, you're in 2nd and 3rd most of the time).  It progressively got worse, and started to become a complete loss of power, until downshifting to 1st when it would take off again with normal power (for the altitude), change up to 2nd and it would be ok for a couple of minutes before it would repeat.   This started to worsen over the next 10/15 minutes, until I was down to only 1st.   At this point I pulled over whilst there was somewhere to do so safely, and as soon as I let the clutch in the engine died. 

Whilst there was no CEL, it felt like an ignition issue, so I jumped out and went over the ignition side,  but couldn't see anything amiss.  Just in case, thinking we'd just fueled up, I went and popped the fuel tank cap off, and was met with a big pressure differential (difficult to quantify, but seemed excessive).

I then went and started her up, and everything was back to normal.  We headed off, but 15/20 mins later the problem returned whilst queuing in some roadworks, I jumped out, popped the cap off again, and back to normal.  5 mins later, the same again, now at the peak of the climb at 13,000 fasl.  I popped the cap off again, banged it a few times (thinking it may be the vent in the cap), and set off again, with no repeat of the problem.  There was also no  repeat of the issue on the run back, so it's done around 800kms since the issue occurred.  I'd like to be certain that the cap was the issue, as in a few weeks I'll be doing the trip again.

Normally when I travel I take my OBD2 adapter just in case, but unfortunately it's gone for a walk, and I'm still waiting for it's replacement to arrive, so seeing live values is impossible at the moment.

Has anyone encountered a similar issue?  I did wonder if it was a fuel pump issue, but I wouldn't expect that to get better.

Cheers folks

 

Streetwiseguy
Streetwiseguy MegaDork
3/16/19 1:10 p.m.

In reply to paul_s0 :

The only thing a gas cap can do incorrectly is NOT seal.  Any tank ventilation is done through the evap canister, and if it is plugged, the car will be almost impossible to fill with gas, and you will eventually suck the tank flat as you drive.  I think your issue is elsewhere.

Nugi
Nugi Reader
3/16/19 1:29 p.m.

I live at altitude and it seems to make any ignition issues worse. Check which cyl is rough, do the unplug one at a time trick to isolate the problem when it happens again (after popping fuel cap to disambiguate!). Be sure your plugs, boots, coils look ok. Might be busy work, but if theyre easy to get to, it seldom hurts to check. 

Had the same gas cap issue on my 80s truck, but tuned out repops didnt have vents. More likely a vent line blockage on a newer car like yours if popping cap worked. 

 

 

paul_s0
paul_s0 New Reader
3/16/19 2:24 p.m.

Thanks for the suggestions folks.  There is some sort of vent into the cap, there's an external vent line up the filler neck, and a small hole covered with a grill in the inside of the cap.  Plugs I changed a couple of months ago.  I wondered if there could be something overheating in the ignition side.  Idle was smooth even with the issue, it just had no power.  When the problem was relatively mild, it would run ok on small throttle openings, but large throttle openings would actually slow it down. 

Years ago I had something similar with a 309 GTi, and it turned out to be a bad ignition amplifier, it took a long time to diagnose as the first replacement amp was bad too, and it would only happen after at least 30 minutes continual hard driving (competing).

I guess I just need to work through everything.

Streetwiseguy
Streetwiseguy MegaDork
3/16/19 3:25 p.m.

Any vent into the cap will be for extreme failure somewhere else, or it just looks like a vent.  Vented caps have been verboten since the 80's.

sergio
sergio Reader
3/16/19 9:28 p.m.

Take an extra coil and some spark plugs on the next trip. If it acts up see if the scanner shows a pending code to ID the cylinder. Change plug and coil see if it’s fixed 

nutherjrfan
nutherjrfan UltraDork
3/16/19 9:32 p.m.

Wait? Fasl? 309 Gti? Where are you located? cool I'm gonna guess south of France.

TurnerX19
TurnerX19 Reader
3/16/19 10:23 p.m.

I think the tank is failing to vent. Because your car was sold new in Peru it just might have a vented cap, but I doubt it, as most of the world forbids it. OBD2 scan might show why it is not venting, but when you released the vacuum in the tank it ran OK for a while after, so I think you have found the basic problem. Some tanks are strong enough to not collapse in this circumstance, some will show a lot of "shrink" until you release the cap. If possible look at it before and after releasing the cap.

paul_s0
paul_s0 New Reader
3/17/19 12:31 p.m.

As TurnerX19 says, I'm in Peru these days, I'm a Brtish expat, sadly not living in the South of France!  Apparently the Mazda 3 does have a reputation for issues with the evap control solenoids, maybe that should be my next port of call.  Thankfully it looks like my OBD2 adapter has arrived at the post office, I'll try and have a late lunch and get there tomorrow and battle the officialdom to get it released to me.

Cheers folks

paul_s0
paul_s0 New Reader
4/25/19 10:47 a.m.

So, the saga continues.  Bear with me please folks.

Having received the new OBD2 adapter a week or so back, I was busy logging like crazy every drive.  Whilst I had no repeat of the symptoms, and no codes, I have seen quite a lot of variations in LTFT and STFT, sometimes totalling over 30% adjustment, which to me is excessive, however it is very variable, sometimes it'll only be 5% or so total.  Average MPG on this car is around 5% less than the (mechanically identical) red hatch I'd had previously, over the same routes, to have a comparison.

This week saw me heading up North again.  As I got up around 8000 - 9000 fasl, I was a bit down on power but no real misfires, and I wasn't sure if it was just the altitude.  I stopped to use nature's bathroom and stretch my legs, I also opened the fuel cap just to see, and whilst there was a slight pressure difference, it was very little, and as I jumped back in and fired her up, I got my high coolant temp alarm, dash gauges showed fine, but Torque showed 105 deg C (in Lima it's between 85 and 90 C normally).  I got moving and it dropped to 101 deg C, and downstream O2 sensor was showing a lean condition, as was the AFR measurement (although only ever so slightly).  I stopped again for a minute, had a look over everything, found nothing, and carried on, rather nervously, but wondering if it was all altitude related.   Temps finally started dropping to around 95 deg C, with downstream O2 showing 0.4-0.5V (I can't access the upstream O2 data).  During this period, sum of LTFT and STFT varied from -28% to +46% (!)

Things seemed to stabilise until an hour or 2 later, getting close to my destination, a good few thousand feet lower in altitude, I had to pull up for an hour waiting for the road to reopen (roadworks where they close the road for hours at a time...), as we set off I didn't notice anything, as it was only 20-30 mph in a queue.  As it opened up, I went to pass a truck, and had to abort - over 3000rpm / 50% throttle it was like hitting a brick wall.  As soon as I could I pulled over and popped the fuel cap again, but there was no noticeable pressure differential at all, so it doesn't appear to be a fuel vacuum issue.  I turned it all off and on again, and set off.  Power seemed normal (for the altitude) below 3000 rpm and less than 50% throttle, but above that it was struggling.  I limped it the last few km, and parked it outside the accommodation for an hour or so.  I again had a very cursory look but I had to get down to the facility ASAP, so headed off again (only 2 miles, mostly downhill).  Power seemed a bit down but it would just about rev past 3000 rpm and take more than 50% throttle, however still with a slight miss.  The next morning before heading to work, I popped plugs on 1 & 3 out, they looked ok, possibly slightly lean, gap looked a bit big.  The plugs are 5 months/4000 miles old.

I went to get my feeler gauge and...realised it was in my other toolbox in Lima.  So they went back in as is.  I also checked the coil pack resistances, all within factory specs.  I then did a hard reset of the electrickery, and went to work.  On the way it all felt mostly fine, just a slight miss passing 3000rpm.  I also found something in Torque I hadn't found before - the misfire counters.  I found it had registered 3 misfires on cylinder 3.  On the way back from work, 2 on cyl 1, 2 on cyl 2, and 1 on cyl 4.  So whatever the misfire is, it's across all cylinders.

Reading up that evening, on MS3s, there is something called the spring-stretch mod.  These are coil on plug, with a spring for the contact.   Several MS3 owners had random misfires around 3000 rpm, and it was basically that the spring wasn't making good enough contact with the plug, especially with Denso plugs (which I have), so they stretch the spring slightly to improve the contact.   So I gave it a try.   The entire drive back I didn't have any noticeable issues, apart from at one moment - there is a stretch of the road which never got resurfaced after the landslides of 2016, it's a few hundred feet of gravel.  I can never resist having a play, and after that I had a slight hesitation for 30 seconds or so, which went away again, which was rather peculiar.

The rest of the trip back was ok (well ignoring the terrible driving standards and awful traffic), but still logging 1 or 2 (or 4 in one case) misfires across various cylinders every drive cycle, but hardly noticeable, if at all, whilst driving.  Today, back in Lima, doing the normal run to school and work, the sum of LTFT and STFT are slight less, but still far too big a range : -16% - +32%.  Also logging a couple of misfire (cyl 1 and cyl 4 this time).  There doesn't seem to be any consistent correlation between fuel trims, vacuum, load, rpm, etc.  About the only generalisation would be that the LTFT is more often than not at the lowest negative at idle, and the highest positive at higher rpm, but not always - say 70% of the time.

The first thing I'm going to do is gap the plugs, possibly followed by a plug change - I'm running Denso ITV16, which are supposedly equivalent to the factory NGKs, however some MS3s have reported issues with Densos (different plug, but still).   I've now found a place with the factory NGKs in stock, at a reasonable price, so I may swing by there.  The lean condition at altitude had me worried though, although the entire drive back was normal.  I was wondering if the fuel pump was getting warm after 8 hours on the road and starting to fail, but it would do so on the drive back too, and it didn't.  With regard to general stuff, no coolant consumption (nor oil), no evidence of HG issues (I will try and do a comp test at some point though), and as it's intermittent I think it's electrical/fuel related anyway.

I'm all ears to any suggestions or similar experiences...

You'll need to log in to post.

Our Preferred Partners
6tELHQY6pEF8jKZLtQSsEAdQVv3JGamqKzeZzZtY37dUuf8LPj5GooWBNfQ9Z080