• Hello Guest, welcome to the initial stages of our new platform!
    You can find some additional information about where we are in the process of migrating the board and setting up our new software here

    Thank you for being a part of our community!

8v microsquirt rpm specific lean swing

Stiggy Pop

In the cool kids club
Joined
Feb 27, 2011
Location
Granville, MA
I have a persistent rpm specific lean swing from ~2,900-3,200 rpm on my car that I've been unable to work out, would love some other thoughts on what could be causing it.

The first year I ran the car on MS2, stock fuel system, stock intake, stock ignition and didn't have any problems. Last year I put in a Nathan intake, DW300 pump and -6 lines, and a microsquirt V3 driving LS2 coils paired in wasted spark with a stock VR crank sensor. The lean swing has been around to some degree since. This year I put in a freshly built bottom end, RSI st2 head and st3 cam and am still working on this exact same rpm range.

I have been trying to fuel it out, and can knock a point or two off the peak but I still end up with the same shape to the curve (and usually a rich condition coming out of it). I've had Kenny look at it as well, and thinking is that there must be an underlying mechanical cause.

Things I've tried:
Fuel pressure is verified good
Fuel pressure reg is brand new/functioning properly
Fuel flow is good
Injectors cleaned and balanced - good
System pressurized and air leaks addressed
Coils, wires and plugs all swapped - good
swapped to a BMW crank sensor

The car was wired with the micro harness with IGN1 and IGN2 split to drive the coils in wasted spark pairs. I get a good rpm signal from the VR crank sensor and am not seeing any sync loss during the condition.

We're starting to think it could be some kind of odd ignition problem, since we've ruled out fuel. The fact that it is rpm specific is interesting. We had thought about some kind of oddity surrounding the Nathan intake but the fact that it happens at different manifold pressures makes that seem unlikely. I might spend the time this week putting a stock intake manifold back on to rule this out.

Would love any ideas from people here on things I could look into. I'm kind of getting down to my harness or the micro, but not sure what I'd be looking for.

These should lead to download links, let me know if they aren't live:

MSQ
https://drive.google.com/file/d/0B_LBSR-F6k0yME9DNkczMHQyWHM/view?usp=sharing

LOG
https://drive.google.com/file/d/0B_LBSR-F6k0ySEZNbTNJcmZUeGc/view?usp=sharing

4lxrung.png
 
Does it do it if you back off the acceleration enrichment at the initial tip in? Looks like you get into it, it dumps a bunch of fuel and causes it to go way rich? I wonder if it is trying to correct for the initial rich part so it leans out at ~3k more then it should though your VE table doesn't have a weird spot in it there.
 
have not been adjusting AE, it will do it under steady throttle though.

Yeah I have this island of VE as I've been trying to add fuel into that 3k spot.
 
The two anomalous high VE entries @ 160kPa 2900-3200 rpm make it harder to see what's going on. I'd try logging some slow accelerations, with different load if you can, around the problem area and see what AutoTune suggests. Maybe you're seeing a tuned resonance in the intake/exhaust that's actually bumping up your VE by a lot around 3K rpm?
 
The two anomalous high VE entries @ 160kPa 2900-3200 rpm make it harder to see what's going on. I'd try logging some slow accelerations, with different load if you can, around the problem area and see what AutoTune suggests. Maybe you're seeing a tuned resonance in the intake/exhaust that's actually bumping up your VE by a lot around 3K rpm?

I have a pile of logs through this area, I'll see if there's a better log to upload. I have run autotune for probably 300 miles and dozens of pulls through this zone under varying conditions and it tends wants to add VE but has not effected a change to really smooth it out.

The last point is partially what I'm thinking, but it seems awfully stubborn to tune out. I wouldn't be surprised if I couldn't nail it, but I've had some more knowledgeable people helping out who are getting similar results. Part of the reason I might swap the intake just to see what difference it makes in that range.

One argument against the resonance point is that it's been consistent across two very different engines dynamically (one whooped out stock block, one zero mile fresh build with a head moving a LOT more air).
 
My initial reaction was the same as bobxyz, since your fuel PWs are holding relatively steady over the RPM range, that there must be some resonant condition within the intake runners that is altering the volumetric efficiency over that short span of engine speeds. Since you are plotting MAP values, I am making the assumption that you are running the speed x density algorithm. In such case, manifold air leaks will generally not alter AFRs.

I also assume that you are not running EGO correction and as such this anomaly is not caused by EGO feedback (which should show up in the fuel PW).

If it is an intake runner resonance issue, the resonance will generally be RPM specific. Because the resonance is dependent on the speed of sound which is a function of air temperature, it may alter slightly with different operating conditions; however, in general I think that this resonant condition should always be present around this same engine speed. If it really is a resonant condition, then you should be able to alter the Ve map sufficiently to address the lean condition. If the resonance is very sharp and you have to make Ve in the target cell very high, you may have to alter the spacing of your RPM coordinates so that the interpolation algorithm doesn't result in really high Ve values on either side of the resonance. If you can't adjust the Ve value to address this condition, then something else other than resonance is at play.

Your argument against the resonance point may not be valid. If the lean condition follows the intake manifold, the resonance is determined by the physical characteristics of the intake manifold, not the engine. On different engines, the resonance will be altered slightly by the intake port dimensions and may be altered slightly by the up front plumbing.

As an observation, I have had no luck using the Tuner Studio Autotune algorithm and have given up on trying to use it. My preference is to adjust the Ve based upon log files obtained without EGO correction. I plot AFR, Target AFR, Ve and PW on one graph and RPM and MAP on another. On the log when the AFR deviates from the target AFR, I look at the Ve value that MS is using for the fuel PW calculation and then adjust the Ve value in the cell (or adjacent cells) as determined by the RPM and MAP coordinates. I only adjust the Ve values 1% at a time and then retest. When doing this type of measurement, I try to get the engine into a state where things are not changing very quickly (that is what I use the PW curve for). There is always a time delay between the calculated values of target AFR, PW and the measured AFR. This will depend on the response of your injectors, engine speed and operating conditions and the distance from the engine to the O2 sensor.

If this is a resonant condition and if the resonance is really sharp, chances are that when you adjust the Ve value to hit your target at the resonant peak you are going to screw up your AFR values adjacent to the resonant peak.

There are formulas that allow you to estimate the resonant frequency (engine RPM) for various runner lengths. They are grossly approximate and say nothing about how effective the resonance will be at altering the Ve of the engine. However, they could be instructive in determining whether your runner lengths are even in the right ball park to create this problem.

Finally, your O2 sensor is well away from the open end of the exhaust pipe? You wouldn't want the same kind of resonant condition on the exhaust system causing a hit of fresh air to the O2 sensor, although I think that would be a pretty low probability.
 
Last edited:
Correct, I had 15% EGO authority on before but have turned it off while I work this.

I noticed in a log from yesterday some obvious delay in wideband response. Kenny just confirmed it looks abnormal so I bought a new sensor. It's an NTK sensor that's at least 5 years old. Will put the new sensor in and see what changes I get, but perhaps that has been impacting my ability to tune this effectively.
 
looking at your EGO control:

its set to activate under 70% tps, you're at 50 at this time
under 120 kpa, you're at 117 kpa

log EGO parameters and see how much fuel it's pulling. doesn't seem like it should take out much, but disable it, try again, tune some more.

i set it down much lower (not needed if your VE table is dialed in)

tables needed quite a bit of love. they should look more like what is represented now.

give this a try and report back.

http://www.filedropper.com/kornely-new-brett1
 
brett brett brett.

looking at your EGO control:

its set to activate under 70% tps, you're at 50 at this time
under 120 kpa, you're at 117 kpa

log EGO parameters and see how much fuel it's pulling. doesn't seem like it should take out much, but disable it, try again, tune some more.

i set it down much lower (not needed if your VE table is dialed in)

tables needed quite a bit of love. they should look more like what is represented now.

give this a try and report back.

http://www.filedropper.com/kornely-new-brett1

If you looked at the log, you'd see that ego correction (and total fuel correction) at the time of the anomalous event is static, despite having a controller authority of 15%

the ve table being jacked up is a product of troubleshooting, again, going back to what it's doing and how it's behaving.

Some of the other things that have been tested (and may still be in the tune), besides disabling/enabling ego control: changing rpm breakpoints on the ve table, gross increases in ve values (in the 25-3200 range, not up top where you'd expect a better head and intake manifold to differ drastically in terms of fuel requirements from a near stock setup).
Dead time is likely incorrect for the injectors, battery voltage at the time of the event is reasonably consistent, as is coolant temp. The ecu is doing as told, however that doesn't seem to manifest in the o2 readings.

One thing that was noticed in another log is that there is considerable lag between the fuel event and the corresponding reading from the o2 sensor (i.e., you hit the gas in neutral, then almost 2 seconds later, you see a corresponding rich/lean swing in the wideband). While some lag is expected at lower rpms, and it is somewhat dependant on sensor location, in this case the sensor is within a couple of feet of the turbine outlet, and the lag continues up in the rpm band. Even in the datalog posted above you can see a delay in reading of a solid second at 2700rpms. What I'm used to seeing is a delay of maybe half a second at low low rpms to virtually no delay at higher rpms (when volume and speed greatly lessen transport time). Even a delay of half a second is a little bit excessive, but still workable.

Anyway, what I'm getting at is, I'm not sure you can trust the o2 reading any more on this setup, and as such, you can't really make a call regarding anything else, at least not until that has been eliminated from the pool of suspects.

Something else to check: did the AFR table get updated(the calibration one, not the target afr table under fuel) to match the wideband? If not, that could explain some of the errant readings as well...
 
Ken - Agreed on the above, the new o2 sensor should be here in a day or so and very interested to see what the data looks like afterwards.

There's a canned calibration for the NGK AFX that I've been using. I'll verify it, especially when I put the new sensor in and calibrate the controller itself.
 
If you think that the O2 sensor delay is your problem, make sure that problem is not originating in the wideband controller. Most wideband controllers have settings which average the O2 values over adjustable periods of time. As such, they do not introduce a fixed time delay; but, act like a low pass filter which slows the rate at which the controllers analog output follows the sensor signal. However, the maximum for this averaging period that I have seen is only 1/3 of a second.

The delay issue is theoretically easy to diagnose, but, may be more difficult from a practical perspective. Your lean spot is around 117 kPa / 3100 RPM. If you can arrange to operate the engine at that condition for a couple of seconds, it will become apparent whether the problem is due to transport / sensor / controller delays. If it is a delay problem, the lean condition should resolve itself as the sensor / controller report the correct AFR after the 'delay period'. If the lean AFR persists during the test, the problem is not some artifact created by a sensor delay.

Finally, if you think the rest of the engine operating conditions seem reasonable, it would be a really odd sensor problem that only creates strange readings at 117 kPa / 3100 RPM and gives reasonable readings at every other engine operating condition.
 
The delay issue is theoretically easy to diagnose, but, may be more difficult from a practical perspective. Your lean spot is around 117 kPa / 3100 RPM. If you can arrange to operate the engine at that condition for a couple of seconds, it will become apparent whether the problem is due to transport / sensor / controller delays. If it is a delay problem, the lean condition should resolve itself as the sensor / controller report the correct AFR after the 'delay period'. If the lean AFR persists during the test, the problem is not some artifact created by a sensor delay.

good point on holding it lean, I've done that on the highway and it will sit at 16:1 and run like crap with EGO pegged but unable to compensate.

Please note it's not specific to 117kPa, that rpm range sees a lean swing at all MAP. It is actually more pronounced at lower MAPs.

My main goal here is to answer the question of is this a tune issue or a mechanical issue. Kenny has done a hell of a job helping me with this car but I can't keep pushing it on him. What I'd really like to do is just bring the car somewhere and have it dialed in but I'm not going to make that investment until I can say confidently that I've done my due diligence and there are no underlying problems that will hold me back while I'm paying for the time. I'm a much better mechanic than I am a tuner...
 
Last edited:
so you have waste spark coils.. did you use sparkplugs designed for wasted spark?

Most spark plugs are designed for straight polarity since with a distributor or even coil on plug the current direction is the same for all spark plugs. with a wasted spark system half of the plugs fire in reverse polarity while there are special plugs made for wasted spark that have the same metal on both the center electrode and the ground strap.

firing a plug with reverse polarity that is not designed to be fired will move some of the ground material onto the center electrode fouling it out quickly and causing misfires.


https://www.youtube.com/watch?v=9TqB76mZzhM



Basically if you have the wrong plugs then you will get weird missing. if a cylinder or 2 miss the wideband o2 will read as a lean event, then you will throw fuel at it and it wont help anything. smooth your map over and get new plugs designed for a wasted spark system.








..how did you get ls2 coils to run wasted spark? have you got 4 of them in a coil per plug arrangement? if so ignore this post about wasted spark where engines that have 1 coil per 2 spark plugs..
 
Last edited:
yeah 4 coils they are just paired on the two stock micro ignition outputs. Couple of Volvo guys running this ignition setup I believe, plus tons of ls1 guys running them that way.

That said I've been thinking about some kind of odd misfire causing this but nothing evident as to where it'd be coming from.
 
so you have waste spark coils.. did you use sparkplugs designed for wasted spark?

Most spark plugs are designed for straight polarity since with a distributor or even coil on plug the current direction is the same for all spark plugs. with a wasted spark system half of the plugs fire in reverse polarity while there are special plugs made for wasted spark that have the same metal on both the center electrode and the ground strap.

Basically if you have the wrong plugs then you will get weird missing. if a cylinder or 2 miss the wideband o2 will read as a lean event, then you will throw fuel at it and it wont help anything. smooth your map over and get new plugs designed for a wasted spark system.

Perhaps off-topic since S. Pop isn't running double ended wasted spark; but, I respectfully need to disagree.

Every in-line air cooled four cylinder Honda motorcycle made from the 1969 Honda CB 750 up into the mid 1980s (I stopped paying attention after that) ran dual ended coils with wasted spark. I had three of them and they ran just fine with the garden variety NGK or ND plugs that Honda recommended for them, right up to 10,000 RPM and sometimes more (oops). I expect that Kawasaki, Yamaha and Suzuki fours of the same vintage use the same wasted spark arrangement with the same garden variety plugs.

If you are using fine wire plugs (platinum, iridium, ...), the breakdown voltage of the two different gaps may be slightly different because of the differing shapes of the positive and negative electrodes on the two plugs (conventional fat wire plugs will suffer less from this). However, that phenomena is not particularly significant and does not have any material effect in the real world.
 
Last edited:
I really like 142 guy's response, seems like great recommendation to me.

I've heard of "lazy" narrowband O2 sensors that need to be replaced but I don't know if the same issue occurs with wideband sensors?

If the transport delay is really long, then EGO correction may be unstable/bad. I'm haven't read up on the details of Auto Tune - if it's using straight MAP/RPM/AFR readings for correction, then it could also be doing bad updates. Even if it qualifies the samples with low MAPdot or RPMdot, a long transport delay could confuse it.

I'd try turning off EGO, using a smooth VE map (get rid of 2 spikes), and doing some very gentle driving with logging enabled. Post those logs with timestamps of weird sections for discussion.

One of 142 guy's earlier comments is interesting -- is your O2 sensor well away from the end of the exhaust, and have you made sure there aren't any upstream leaks?

Thinking about "lazy" sensor response, have you checked that your MAP hose isn't kinked? I'm not sure if it would slow down MAP response or if it could be related to your issues. Kinking your CAS/rpm wires won't matter -- they'll still run at the speed of light (or ~2/3rds if you're picky).
 
Back
Top