Scooby Jonni

I think my MAF is on it's way out due to a rough idle. I am able to log the MAF's output with a datalogger and plot it on a graph or see it in real time. I'm just wondering how I would use this to diagnose a faulty MAF.

Looking at the graph I can see the MAF going up and down with the revs but I am not sure if it is also the cause of the fluctuation.

Any help would be greatly appreciated.

Jon

Splitpin

What type and model year of car do you have, and what sort of unit is your logger displaying - a raw 16/8 bit hex or decimal value, or have you converted it to a voltage, or a value in grams per sec?

Either way, on your graph the biggest giveaway is usually a "jagged" trace. At idle the voltage shouldn't vary more than +-0.06 or so, and, under acceleration you should see a nice smooth rise without any significant "teeth" - especially downward ones.

If you've been logging and saving data, it'd also be worth comparing a file from now with one from a month or three ago and trying to compare the sensor's behaviour at idle, and also a similar bit of driving under boost. You should spot any changes in response quite easily.

If you can tell us more about the type of car then one of us might be able to give you some guide engine speed/boost/MAF voltage guideline figures.

Scooby Jonni

Thanks for the reply

I should have said, my car is a V3 STI Type-RA (Late 96). It has an orange label hot wire MAF. The datalogger is an Innovate LM-2 which is measuring the voltage in decimal as seen by the ECU.

I can't seem to find my card reader anywhere so can't get any data off the unit! I'll try and get another tommrow and post some graphs up.

The voltage of the MAF was going up and down with the idle but it was hard to see if it was jagged as it was plotted on a tiny little LCD screen. I'll do some more logging tomorrow anyway at constant and sweeping RPMs for comparison. Hopefully the memory card will still have some old logs on it to compare with.



Jon

Splitpin

Does this system allow you to graph up the info on your PC (once you find the card reader!) either on some proprietary software or a spreadsheet? Will help you to see what's going on far more clearly than the unit's own little display.

Peanuts

iTrader: (15)

another way tocheck without the datalogging is to disconnect the maf at idle.
If the car stalls then usually the maf is ok.
If it carries on running lumpy then usually the maf is goosed.

Splitpin

No, that "test" proves next to nothing. Waste of time. at "usually".

Peanuts

iTrader: (15)

I'm trying to find a quick an easy way to give an indication of a goosed maf.
Another way would be to record the voltage (mV) at idle and compare to a good maf, this is how we used to determine the state of the maf with the powerFC commander in the good old days, sadly I've been running other ECUs and no maf for long enough to not remember the voltages.

Splitpin

I know you were. Trouble with that "test" is that while it might be quick, it doesn't assist in giving that indication, and can give dangerously misleading results.

It obviously has no relevance to the sensor's accuracy (or otherwise) at high mass airflow readings, it will pass good sensors as bad and, worse, pass damaged sensors as good. Hence better to not bother than go forward on the basis of incorrect info.

Again testing at idle doesn't examine the sensor's performance at high throughput, which is where you really need it to be accurate and where errors are easier to spot. Granted though comparing with a known good sensor - or with the same sensor some weeks or months previous, is a good diagnostic step.

Scooby Jonni

Thanks for all the comments.

I managed to find myself a card reader and have taken some logs but my car has decided to play ball and is idleing steaderly

Luckily there was a log of the car running a rough idle on the card. Here it is..

Black = RPM
Pinky red = Throttle
Green = MAP
Blue = MAF



From my experence, the MAP, MAF and RPM should all all have the same pattern at constant throttle. The RPM and MAP look good but the MAF is seems to be all over the place. Do you think that I may have found my problem?

Just for comparison, here are the two logs I took today..

Idle..



Sweeping RPM..



And just for good mesure, here is a WOT run I did a long time ago..

(the erratic pink line in this graph is the AFR)




As always, any help/comments would be greatly appreciated.

Splitpin

S*d, nothing like an intermittent fault to make you scratch your head.

That first "rough idle" graph. The first obvious feature is the almost totally regular intervals between the oscillations - 2.3 seconds or thereabouts. Also, I'm presuming the red key on the Y-axis corresponds to MAF voltage? If so, it seems to be bumping along just under 1.0 volts, +- quite a bit. It does appear to spike upwards during the moments when engine speed, and particularly MAP have risen.

That first graph certainly points to something being up, but, especially in the light of today's which generally look "normal" MAF-wise, I'm not sure it points a conclusive finger.

Questions: Does your software monitor the idle speed control valve position at all, and is there any chance that some sort of physical issue (for example a loose, split or partially disconnected pipe) might be in play?

The reason I ask is that the MAF voltage in that rough idle graph looks generally low. While that could be the sensor itself, the fact that today's looks normal initially suggests otherwise. Is it possible that air could have been getting into the inlet tract from another source? This would explain both the unusually low MAF reading, and also the ECUs difficulty in controlling the idle speed.

Scooby Jonni

Yes that's correct the red axis corresponds to all of the voltage values and the black is just for the RPM. It sounds as if the car is about to stall but then the idle control value kick in and prevents the stall but the car is unable to maintain the RPM so it drops back down. The MAF does seem to roughly match the other sensors but not as much as in the other graphs, as you say.

Unfortunately not, it is just a set of analog inputs that I have had to physically attach to the ECU wires. I believe there are two control wires going to the idle control valve which means I do not have enough inputs left to monitor it


Ah I see what your saying, thats interesting; the MAF voltage on the rough graph seems lower than on graphs where it was idling correctly. It is possible that there was a leak and I have unintentionally fixed it when I removed and reinstalled the MAF

I have just been outside and checked all of the hoses I can access but I can't find any leaks. I have also sprayed some soapy water on all of the silicone but can't see any bubbling. Again, the car idled perfectly all the way through

I guess I'll just have to wait and see if it happens again.

I was just looking at the logs again and I noticed these three abnormalities. In the first the MAF is going up while the RPM is going down and the TPS is constant. In the second, the RPM is going up while the MAF is going down. I've just noticed I've highlighted the wrong bit in the third I meant to point out that the upwards slope in the MAF just before doesn't correspond with any of the other sensors. Do you think these could this be due to a faulty MAF or is it just a the time delay caused by the air going through the FMIC?

Thanks

Splitpin

It is an explanation that would neatly explain the known facts. Doesn't mean it's correct, but in the absence of anything better...

Likely candidate would be the pipe that runs between the idle speed control valve and the turbo inlet pipe.

You wouldn't expect to. At idle the inlet tract is under vacuum so you wouldn't see bubbling on the outside of the pipework - the soap would be sucked into the pipe and any bubbles would be on the inside.

May be the best way forward given where you are at the moment.

While there probably are exceptions, failing MAF sensors tend to under-read or drop out completely. Thus the apparent lag between airflow and rpm/MAP would be more easily explained by the FMIC/pipework than it would be by a failing sensor. Keep a close eye on it though - at least you've got the right kit to monitor it.

Scooby Jonni

Ok, i'll double check that today

True, that thought did cross my mind. I read it somewhere and thought I'd give it a go but, as you say you'd need to put the pipe under pressure first.


I went for a long drive yesterday and afterwards the car was doing a fast idle (about 1,300rpm). I sat and watched it for a bit and then it dipped to about 800 and came straight back up again. I did this a few times at about 5 second intervals Unfortunately I didn't have the logger on me but i'll stick it in the car for next time.


Yeah having the ability to log the vital sensors is really coming in handy. That's good to know anyway. I'll put the logger and wideband lambda in today so I can keep an eye on things and try to see if I can see a pattern emerging. At least I know what i'm looking for.

Thanks again for taking the time to reply