Constant throttle stumble.

[VWJake]

“The stumble seems to not be so bad the first couple of minutes of driving, then gets worse when it gets warm.  The stumble is available at idle, just sitting there, as well as it noticeable  at most any constant throttle setting.  Transient response, even with that function turned off, is not bad at all- this is not a transient result- at least I can't see how it could be.”

I’m no end all be all tuning wizard, but a few thoughts rattled through my head this morning and again tonight as I read this new post. 
 

1)Air leak- engine warms and manifold or something twists/grows… raises o2 levels engine lean misses.

 

2) Engine is fine til warm. Fuel enrichment is covering a desire for the engine to run lean (maps). Acceleration enrichment resolves the issue as well. I don’t know haltech from Adam, but I’ll assume you can set base stoich and target AFRs. Perhaps fattening the maps will correct the issue. I know I know… but when you pull the plugs you hear angles sing… maybe they need to be a bit dirty… 

 

just a few thoughts from someone that doesn’t know enough to know he doesn’t know enough. Lol

 

-Jake
 

[ray_]

I suppose that thing is running open loop until it warms up a bit. That would be another reason why it's better behaved at first.

[irdave]

2 hours ago, VWJake said:

“The stumble seems to not be so bad the first couple of minutes of driving, then gets worse when it gets warm.  The stumble is available at idle, just sitting there, as well as it noticeable  at most any constant throttle setting.  Transient response, even with that function turned off, is not bad at all- this is not a transient result- at least I can't see how it could be.”

I’m no end all be all tuning wizard, but a few thoughts rattled through my head this morning and again tonight as I read this new post. 
 

1)Air leak- engine warms and manifold or something twists/grows… raises o2 levels engine lean misses.

 

2) Engine is fine til warm. Fuel enrichment is covering a desire for the engine to run lean (maps). Acceleration enrichment resolves the issue as well. I don’t know haltech from Adam, but I’ll assume you can set base stoich and target AFRs. Perhaps fattening the maps will correct the issue. I know I know… but when you pull the plugs you hear angles sing… maybe they need to be a bit dirty… 

 

just a few thoughts from someone that doesn’t know enough to know he doesn’t know enough. Lol

 

-Jake
 

 

There's a little correction based on coolant temp still- but I like the idea.  I'll turn it off in the morning so it'll just be the VE map and that's it- so I guess kinda' like @ray_ is thinking- at that point it would only be open loop, and if the actual AFR is different from the target AFR I'm changing the VE table to make it match by hand, instead of allowing the ECU to correct itself.  That little bit of coolant correction is the only correction still active, everything else is turned off, including transient (acceleration) functions and all auto-corrections.  It's just what it is.

 

I've thought that throwing a little fuel was helping, now I'm not so sure.  But yeah, if that's all it takes, I'm in.  Hell, as a joke, I think one time I set the entire table to 12.2:1.  All of it.  But at that time I still had massive vacuum leaks, who knows what was going on.  If I do get it sorted, it should run pretty good.  Just looked, at the 50-60*F that it's starting at, it's adding 10-15% until coolant temp = 86*F, then zero.  Probably at 86 pretty quick.  But I'll turn it off in the morning.  For sure much better test.

 

Oh, and for shits and giggles today- just to make sure I'm not chasing my tail- quick compression test showed 200 +/- 5, up here at altitude, so maybe that parts not too broken.

 

Thanks again to all y'all for thinking about this with me; I was getting pretty bummed out.

[jimk]

11 hours ago, irdave said:

he suggested simplifying the table (rpm only) and making the timing advance as the revs went up- made sense to me.  The table I had had more complexity and would retard the timing at about 4k- my guess at about torque peak.  It was also most advanced at lower rpms, then would retard a little as it went up.  So now I start at 375 at 0 rpm, 450 at 9k (whatever), with linear regression across. 

The timing table comes from the Motronic chip and replicates it.  Timing has been translated from AFM load to TPS load.  It does not replicate a centrifugal/vacuum advance distributor which is limited by it's design 

As rpm goes up at the same engine load (throttle position) the timing does advance.  As load goes up at constant rpm the timing advance reduces due to the increased cylinder pressure and concurrent increase in flame speed.  

Full load advance is close to but not exactly the WOT table in the Motronic.  It has been dyno tuned by another S14 user in Cali at Electromotive.

[jimk]

Have you dropped the OEM injectors in to see if the new ones are the problem?

[irdave]

4 minutes ago, jimk said:

Have you dropped the OEM injectors in to see if the new ones are the problem?

 

Not yet.  Suppose I should.

[AustrianVespaGuy]

Ok, so here's my thoughts after looking at the datalog (attaching screenshot):

 - I don't really see any 'stumble' in the RPMs, however I do see what look like a few 'spikes' or 'surges' such as the one right where the cursor is.  RPM seems to jump from 2300 to 2450 in half a second and then comes right back down.  Think this is what you're feeling?  I sorta doubt the engine is actually doing that, and think this is more likely an artifact of crank signal noise or reading of a 'phantom' tooth, but the ECU doesn't seem to be flagging any tooth errors either, thought the 'Trigger Tooth Count' varies wildy, though I don't really know what that parameter means. Does the S14 use a 60-2 VR trigger wheel?

 - Sensors all look good, so I'm pretty sure 5v ref/sensors/etc. are NOT your cause.

 - Acceleration enrichment is turned off, right?  While the TPS position itself looks pretty good, the derivative/TPS rate is a bit more wild, probably from a bit of noise.  Make sure it's grounded at the ECU, and when you go to put turn accel enrichment back on, make sure its activation threshold is ABOVE this level of noise.

 - AFR looks REALLY good, only going a bit lean right when you roll on the throttle, which is exactly what one would expect with no accel enrichment, so I also don't suspect any problems in you VE table/fueling in general.

 - That said these injectors might be a tad on the large side, only 1.6ms and 2.5% duty cycle is kinda borderline on the low side, but would need to see what they're at during idle, and again not something that would cause your stumbling issue.

 - Timing looks fine too, pretty stable in the 25-30 deg BTCD zone.  I know the plugs look good and you still had the issue with the stock distributor, but I don't want to rule out ignition just yet as the symptom still sounds like an ignition-related one.  What coils are you using and are they logic-driven or is there a high current driver for them somewhere in the system or ECU?

[Einspritz]

Have you checked that you have sufficient gauge wire for the fuel pump, that the pickup is clear, and most importantly the pump is sufficient for both pressure and volume?

 

The M3 has two pumps for the S14, and if either goes weird, or is on its way out, particularly in hot weather, then you (me) get surging. Btdt several times before.

[Dudeland]

Ok... so shooting from the hip.  AustrianVespaGuy I think is reading this correctly.  The screen shot from the data log looks ok. Nothing electronic seems to pop out as a misfire, considering the plugs are looking good and even, it would indicate that the problem is not isolated to just one cyl. 

 

So I am thinking of mechanical changes that you would have made when the swap took place.  Being not familiar with the S-14 and having to deal with my ECU challenges, you need to being to look at ways of air/fuel being introduced between the metering points(In your case it is throttle plate to O2 if I am not mistaken).   What is the EGR situation like?   Was part of that system altered?  What is going on with the crank case pressure,  where does it go?  I expect that perhaps with a more robust and precise spark it may exaggerate any small change in the mixture.   Having a robust spark and a great combustion chamber it would likely tolerate changes that would not reflect in a change in the combustion gasses, hence not picked up by the 02.

 

In the FWIW department I had a small hole, about 2mm in the collector of my header when doing my EFI conversion.  I thought that it was the silver bullet, so I popped for a bag of snakes from Ireland, and it made no measurable change in the 02 levels.  It needs to be a big leak.     

 

Nothing seems to be happening that the haltech detects and corrects for, so whatever is happening is doing it very quickly.    I respectfully suggest that you get it on an oscilloscope and see how clean the signal is going to the coils and injectors.  Ground loops and other anomalies can only be detected by back probing these devices.  As far as the haltech goes, it is sending a signal, but what is actually received by these devices can be something completely different. 

 

Picture of the engine bay may spark some new thoughts.  If you can kindly provide some snaps, and perhaps a short movie of the car at idle, perhaps ears that are much better trained then mine can pick something out. 

 

Shooting from the hip again, injector timing may be off, perhaps you are getting random drops of fuel on the back of the intake valves which randomly pop into the cyl. 

 

 

 

 

 

[irdave]

More later, but injectors and injector dead time / offset.  I put the stock injectors in this morning and made up some number for offset.  My car has never run smoother.

 

I have to feed the wife NOW.

 

Thank you!!!!

[irdave]

12 hours ago, jimk said:

The timing table comes from the Motronic chip and replicates it.  Timing has been translated from AFM load to TPS load.  It does not replicate a centrifugal/vacuum advance distributor which is limited by it's design 

As rpm goes up at the same engine load (throttle position) the timing does advance.  As load goes up at constant rpm the timing advance reduces due to the increased cylinder pressure and concurrent increase in flame speed.  

Full load advance is close to but not exactly the WOT table in the Motronic.  It has been dyno tuned by another S14 user in Cali at Electromotive.

 

This was regarding the timing of the squirt of the fuel in the fuel injection cycle, not ignition timing...

[irdave]

6 hours ago, AustrianVespaGuy said:

Ok, so here's my thoughts after looking at the datalog (attaching screenshot):

 - I don't really see any 'stumble' in the RPMs, however I do see what look like a few 'spikes' or 'surges' such as the one right where the cursor is.  RPM seems to jump from 2300 to 2450 in half a second and then comes right back down.  Think this is what you're feeling?  I sorta doubt the engine is actually doing that, and think this is more likely an artifact of crank signal noise or reading of a 'phantom' tooth, but the ECU doesn't seem to be flagging any tooth errors either, thought the 'Trigger Tooth Count' varies wildy, though I don't really know what that parameter means. Does the S14 use a 60-2 VR trigger wheel?

 - Sensors all look good, so I'm pretty sure 5v ref/sensors/etc. are NOT your cause.

 - Acceleration enrichment is turned off, right?  While the TPS position itself looks pretty good, the derivative/TPS rate is a bit more wild, probably from a bit of noise.  Make sure it's grounded at the ECU, and when you go to put turn accel enrichment back on, make sure its activation threshold is ABOVE this level of noise.

 - AFR looks REALLY good, only going a bit lean right when you roll on the throttle, which is exactly what one would expect with no accel enrichment, so I also don't suspect any problems in you VE table/fueling in general.

 - That said these injectors might be a tad on the large side, only 1.6ms and 2.5% duty cycle is kinda borderline on the low side, but would need to see what they're at during idle, and again not something that would cause your stumbling issue.

 - Timing looks fine too, pretty stable in the 25-30 deg BTCD zone.  I know the plugs look good and you still had the issue with the stock distributor, but I don't want to rule out ignition just yet as the symptom still sounds like an ignition-related one.  What coils are you using and are they logic-driven or is there a high current driver for them somewhere in the system or ECU?

 

The s14 uses no trigger wheel.  I'm using a 60-2.  The trigger tooth count I believe is just reporting the tooth count based on whatever the sampling rate is- at the time sampled.  I would imaging that if I turned the sampling rate up high enough we would be able to see it cycle through smoothly.

 

I've heard that the 5v sensor fault could be software related and that an update might resolve the issue.

 

All everything is turned off, so correct, no acceleration enrichment.  I, too, thought the derivative TPS was out of whack.  I'll make sure it's above this rate.

 

VW 1.8T coils, full logic on coil, no external driver.

 

Thank you for your help.  Seriously.

[irdave]

5 hours ago, Einspritz said:

Have you checked that you have sufficient gauge wire for the fuel pump, that the pickup is clear, and most importantly the pump is sufficient for both pressure and volume?

 

The M3 has two pumps for the S14, and if either goes weird, or is on its way out, particularly in hot weather, then you (me) get surging. Btdt several times before.

 

I've seen you bring this up before and it stays in the back of my head.  I haven't replaced the wiring running to the pump yet, although the car didn't originally have an electric fuel pump, did it?  A '76?  But it has been something I've been wanting to do- run a bigger wire.  I will.

 

The pump is a 5 series, 6 cylinder pump, so here's to hoping it's big enough!  I'm not as cool as Jim to have a fuel pressure tap for the ECU to read... 

[irdave]

5 hours ago, Dudeland said:

Ok... so shooting from the hip.  AustrianVespaGuy I think is reading this correctly.  The screen shot from the data log looks ok. Nothing electronic seems to pop out as a misfire, considering the plugs are looking good and even, it would indicate that the problem is not isolated to just one cyl. 

 

So I am thinking of mechanical changes that you would have made when the swap took place.  Being not familiar with the S-14 and having to deal with my ECU challenges, you need to being to look at ways of air/fuel being introduced between the metering points(In your case it is throttle plate to O2 if I am not mistaken).   What is the EGR situation like?   Was part of that system altered?  What is going on with the crank case pressure,  where does it go?  I expect that perhaps with a more robust and precise spark it may exaggerate any small change in the mixture.   Having a robust spark and a great combustion chamber it would likely tolerate changes that would not reflect in a change in the combustion gasses, hence not picked up by the 02.

 

In the FWIW department I had a small hole, about 2mm in the collector of my header when doing my EFI conversion.  I thought that it was the silver bullet, so I popped for a bag of snakes from Ireland, and it made no measurable change in the 02 levels.  It needs to be a big leak.     

 

Nothing seems to be happening that the haltech detects and corrects for, so whatever is happening is doing it very quickly.    I respectfully suggest that you get it on an oscilloscope and see how clean the signal is going to the coils and injectors.  Ground loops and other anomalies can only be detected by back probing these devices.  As far as the haltech goes, it is sending a signal, but what is actually received by these devices can be something completely different. 

 

Picture of the engine bay may spark some new thoughts.  If you can kindly provide some snaps, and perhaps a short movie of the car at idle, perhaps ears that are much better trained then mine can pick something out. 

 

Shooting from the hip again, injector timing may be off, perhaps you are getting random drops of fuel on the back of the intake valves which randomly pop into the cyl. 

 

 

 

 

 

 

Picture of the engine bay!  Ha!  Need some new porn?    I'll get one for you.

 

Injector timing may be off.  This is one of the things one of the guys from Haltech brought up and is something I actually changed- good thinking.

[irdave]

Alright, here's where I'm at.

 

The dizzy made no difference.  I'm thinking that the COP took some slop out of the system, as well as cleaning up everything else- like the vacuum leaks, that the problem became more prevalent.  So I put the COP stuff back on- it's nicer to look at anyway, I think.

 

As Jim mentioned above, we'd been talking about reinstalling the stock injectors.  So I did that this morning.  Old injectors and original offsets / dead times that we came up with to get my car started.  Hit the key and it immediately idled smoother at 800 rpm than it ever did at 1100 with the FIC injectors.  Gob smacked is appropriate.

 

Took it for a drive and the VE table was demanding lots of attention, which is kind of wrong for tuning by VE- that change shouldn't have required almost any change at all.  So that led me to thinking about that offset / dead time.  I did some digging on the webs and found some numbers for a Bosch injector that was close.  Loaded those up and while the table is still off a little bit, I was able to drive the car for 50 miles without bothering to tune it at all.  This is the smoothest my car has run in the 2 years I've had it.  Thank you all for the help and support.

 

From here, I need to try and find the exact numbers for the stock s14 injectors to get those in the software.  I'm interested to see how it affects the VE table.

 

What I think the problem was based on what fixed it- I think the FIC injectors were operating beyond their capability, albeit on the low side.  They have an advertised stable pulse width of 0.6ms, but it appears that something was causing that to not be the case.  The dead time / offsets for those injectors were included in the box- they could be suspect as well.  Maybe their testing device and my use device aren't calibrated the same. (Haltech is made in Australia, so who know?  )  But it appears to be something between the injectors and their specs.

 

Gods, what an education I wasn't looking for.

 

I'll be riding my mountain bike tomorrow.  After that, I'll look into tidying up that TPS ground, getting the VE table dialed, and for enquiring minds, I'll pull some more data.

 

Again, thank you all.  Really.

Edited October 7, 2021 by irdave