8v microsquirt rpm specific lean swing

[Stiggy Pop]

I doubt that it will make any difference but try setting the GeneralSettings-CLT/MAT/Battery Lag factor back to the default of 50. I've noticed that the MS2 so called "smoothing" function tends to result in lumpy stair-step changes instead of the desired smoothing.

For the ridge, can you feel a boost in power? In other words, do you believe the VE, and hence power, really do pop up over the narrow RPM range?

I'll give that a shot.

Re: the power - it's kind of hard to say since it's such a narrow range, but no if I was going to draw a curve I'd put a dip there. What I feel in the seat is revs starting to come up with MAP and a slight hesitation as I cross the 3k range.

 

[wennstroma]

I got the voltage supply fixed on mine and started driving it more and leaning out the ve table under higher load to get rid of my misfires. So I've got some datalogs that coincidentally show the rpm range behavior that is similar to yours.

The ve table is still pretty rough after some changes I made earlier this year. The 115 values are from the past when I had closed loop boost control turned on. I'm just using a flat, open loop table for boost control right now, so I haven't been hitting those cells. The high value at 100kpa and idle speed is from when the clutch is let out too quickly and the engine almost stalls.

Though my table is generally flatter than yours (from includeafr?), there is still a noticeable ridge at 3000rpm in and out of boost. The datalogs show it going lean a little bit even with the higher values in the ve table.

If I drive it some more this week, I will try to take some logs specifically in that rpm range under steadier conditions and different loads.

 

[Stiggy Pop]

Though my table is generally flatter than yours (from includeafr?), there is still a noticeable ridge at 3000rpm in and out of boost. The datalogs show it going lean a little bit even with the higher values in the ve table.

This is interesting, our VE tables are very reminiscent of one another. I'd be interested to look at some logs.

No logs handy until tonight but I did end up fixing a voltage drop issue which had no effect on the way the car runs, but was another good thing to fix/eliminate. I also spent a day swapping the stock intake manifold and my old IC pipe back on the car. It also had no effect on the way the car ran in that rpm range. I did some tuning for it but the VE table really didn't change much at all at low rpm, and the little swing at 3k was persistent. So that has eliminated the Nathan intake as the source of some kind of peculiar airflow thing. As a side note the back-to-back difference of the intake manifold on the built engine/high flow head is pretty impressive! The car was much less willing to rev and felt noticeably "choked off" with the stock intake on.

The car is running really well everywhere else so I'm trying to not let this ruin my party. I can cold start it with no foot, idles, cruises well, and I can keep my foot in it until 6800rpm with steady AFR. So it's really just this ~500rpm window that gives me headaches.

Will keep picking away at it. Have thought about trying to borrow/rent another microsquirt just to keep knocking things off the list of possibilities. It would be nice to hold the car at 3k rpm where it leans out and really see what the coils are putting out - but we're not sure how we can accomplish that yet? I'll throw up some logs later of healthy voltage and the old intake manifold just for posterity. Weather looks good this week so hoping to spend some time playing with the tune, going to take a shot at turning on AFR Target, maybe fiddling with the Req Fuel value and rescaling the VE map.

 

[wennstroma]

This is the only one I got today with a steady throttle input, from record 550-650. Holding it steady at that rpm for longer would have to be done at higher speeds on the freeway on a steep hill or something.

msq: https://drive.google.com/open?id=0B_ELhGS5O-W4bktuV1g0UURJWGc

log: https://drive.google.com/open?id=0B_ELhGS5O-W4M3lld0ZrcXh0ckk

The attached msq is from after the logs were taken, so the ve table is not the same one used in the log. I also cleaned up the low load, high rpm areas and removed the low rpm stall spike. I hit the largest spike in the middle a few times today by lugging it some from lower rpms and the autotune bumped one cell from 115 up to 119. I'll see if I can get some better logs later this week.

Our engine set ups are somewhat similar. I have a NIW intake, tube header, 530 head with some porting and 38mm exhaust valves, KG2T cam. The bottom end is newish with ~8.0-8.5 compression ratio. I am using an MS2 on a v3 board. Injectors are 79 lb/hr, ignition is dsm cas with dsm ignitor and wasted spark coil packs.

I found some old msq files when I had a stock intake and exhaust manifold, T cam, different injectors, and different ignition system and the ridge is still there at the same narrow rpm range.

 

[Stiggy Pop]

Veddy interesting. Will take a look tonight. I'm not really sure what I can get from the info but it's definitely interesting to see another example of similar behavior. I have no real idea what is causing it, and even if it is a tuning issue why would the car demand so much fuel/VE at that particular rpm? It's puzzling and more than a bit frustrating. I've done a lot of mechanical due diligence, all for the greater good, but with nothing turned up I guess I have to dig back into the software side of the world.

Here's what it looked like with the stock intake back on. 3k misbehavior examples can be found around the marks in the log.

MSQ: https://drive.google.com/file/d/0B_LBSR-F6k0yZC1sMlZadFlxSWs/view?usp=drivesdk

MSL: https://drive.google.com/file/d/0B_LBSR-F6k0yTWlvQWs3cnE2VkE/view?usp=drivesdk

 

[wennstroma]

Meant to post this last week. As you said, its interesting to compare if nothing else. Mine has never been dyno tuned or looked at by someone with more experience who really knows what they're doing. That would be the best way to see if it is just a tuning issue.

I went back and forth over a hill where I could hold a load in third for a few seconds at a time. But holding it steady in that range is a little difficult, because that is where the turbo starts to respond quickly. That is part of what always made me think that some kind of "missing" MAPdot AE was being compensated for, but the MAP isn't actually changing fast enough to trigger one.

MSQ is the same as my last post.
MSL log: https://drive.google.com/open?id=0B_ELhGS5O-W4UFVlTWlNU0IyNWc

1465-1570 is 100-120kpa and looks pretty good. The area leading in and out is rich. If I reduce the ve table for the rich areas, it will make the step at the 2800rpm bins even more drastic.

4910-5220 75kpa tapering down to 70kpa. A little rich leading in, but otherwise smooth. The 3k rpm ridge in the ve table below 80kpa is almost non-existent though.

6250-6295 is 85-100kpa. Just a short pull, but the large ve values kept the afr consistent.

7040-7120 is 100-150kpa. Another short pull with larger throttle opening. The ve1 value flattens off at 108 right as the afr starts swinging lean, and makes it look like that 120kpa/2800 rpm cell could take even more than the 109 value its already at with steeper drops on either side. The first record range I listed is like this too.

On the drive back, I payed more attention to how the car runs in that range. While accelerating the way I normally would in second and third gear, I notice that as the turbo spools up and the engine starts to really pull, it flattens off right in that rpm range, just for an instant, and then it immediately gets past it and pulls smoothly on up from there.

The lean to rich swings in the logs seem to be from 2850-3000rpm and the bins I have set are 2400,2800,3200,3600. I could change the rpm bins so one centers over that range with a huge ve spike, and the adjacent bins within 100-200 rpms back at normal values and see how it reacts. Its been really hot, so I haven't driven it since last week.

 

[FreeEMSFred]

Could it be an efficiency gain from the wastegate opening? Are you using little toy stock turbos? Or something with hair on its chest like a big Holset? The effect of wastegate opening was visible in logs from my ute with the big Holset, but I imagine it'd be more obvious on a small turbo.

In any case, you've got a peak efficiency region, great, give it the axes zones it wants, and the fuel it needs, and let it perform. If with a big VE hump and consistent lambda output you still have weird torque behaviour, then wonder what's going on?

Disclaimer: This post was powered by French bubbly.

 

[Stiggy Pop]

Could it be an efficiency gain from the wastegate opening? Are you using little toy stock turbos? Or something with hair on its chest like a big Holset? The effect of wastegate opening was visible in logs from my ute with the big Holset, but I imagine it'd be more obvious on a small turbo.

In any case, you've got a peak efficiency region, great, give it the axes zones it wants, and the fuel it needs, and let it perform. If with a big VE hump and consistent lambda output you still have weird torque behaviour, then wonder what's going on?

Disclaimer: This post was powered by French bubbly.

It's a Holset HX35 on a tube header with a Tial 38mm gate with a 13psi spring (and I'm currently running off the gate, keeping boost down). Built block, RSI head, definitely not stock airflow.

Bold: Yeah, I feel like I've been trying to pump VE/Fuel into it (think Kenny has too) but not getting the desired results. And yes, I'm wondering like hell! I'm going to look hard at the ignition system and make sure I'm not somehow having a slight rpm specific misfire that's driving the lean AFR spike.

 

[linuxman51]

it's been asked before I'm sure, but did you guys degree the cam when you installed it?

 

[Stiggy Pop]

it's been asked before I'm sure, but did you guys degree the cam when you installed it?

no not scientifically since we didn't have the tools handy. We set it straight up and rocked out. I get where your heads at though, and it's something I've got on a mental list of things worth doin'.

Will mention though we fought this a bit with the old engine/head with the milder cam in it too (also just ran straight up).

 

[wennstroma]

Did the rsi cams come with specs to degree them?

I was also chasing ignition problems for an unrelated problem (high load blowout). I swapped in new plugs and made some new ignition wires at the same time. Didn't seem to change this ve table behavior.

it's been asked before I'm sure, but did you guys degree the cam when you installed it?

I set mine according to the kg specs. I had to advance it more than one full tooth to get it where they wanted though. Here's a post i made about it in my project thread. http://forums.turbobricks.com/showpost.php?p=3488139&postcount=34

I had run it straight up before I did my re-ring with the similar ve table. Also had a stock head with a t cam on it for a while when a valve seat fell out of the ported head, and it acted similar then too.

 

[FreeEMSFred]

I guess my point was, throw big numbers at it until it's a dip to the rich side, or follows what you want, then wonder what's going on.

Don't wonder until you've given the engine what it's asking for. If the hump is smooth and without spikes, it's unlikely to be misfire caused IMO.

Box the area in with appropriate axis bins, and push the one in the middle up and up and up until it causes you some issue, or it runs right.

Just staring at a hump in the AFR and saying "the VE won't be smooth if I make the AFR smooth" is a bit crazy

Being shy with your tuning is not a way to make progress. Make changes, bold ones, until it's close, only then settle down.

Am I making any sense here? If not, blame the ozzy wine. Or the 24x19 / 27x17 / 22x21 fully adjustable tables. Hard to say at this time of night. Where's my glass, ahhh, that's better.

 

[linuxman51]

I think we tried that

 

[Stiggy Pop]

Did the rsi cams come with specs to degree them?

No, would be great if someone has some info. I actually bought a spare Stage 3 from Mitch brand new a few weeks ago and it just shows up in a box, no card or specs provided (not particularly surprised)

I guess my point was, throw big numbers at it until it's a dip to the rich side, or follows what you want, then wonder what's going on.

Don't wonder until you've given the engine what it's asking for. If the hump is smooth and without spikes, it's unlikely to be misfire caused IMO.

Box the area in with appropriate axis bins, and push the one in the middle up and up and up until it causes you some issue, or it runs right.

Yeah I follow you. I feel I've been trying that but like you said maybe I'm being too conservative with it. Most of my "throw a ton of fuel at it" results have just been moving the curve up and down the Y axis. So I end up with 9:1 dog **** on either side but with a similar shaped hump. It's like I need a whole other 12x12 table just for 2500-3500rpm.

I think we tried that

I think you've thrown everything but the kitchen sink at this thing!

I'll figure it out one way or another, do appreciate you guys thinking it out with me it's been helpful.

 

[bobxyz]

Out of curiosity, are you running un-timed injection and are perhaps seeing an artifact of injection crossing over valve opening time? I don't know how much this varies in MS2. Can you try semi-sequential just to make sure injection timing is always starting at a fixed rotational position?

 

[Stiggy Pop]

Out of curiosity, are you running un-timed injection and are perhaps seeing an artifact of injection crossing over valve opening time? I don't know how much this varies in MS2. Can you try semi-sequential just to make sure injection timing is always starting at a fixed rotational position?

You set me off reading for the morning, thank you. I do have the injectors on two outputs paired the same way I've paired the coils for wasted spark, so I believe I should be able to try semi-sequential.

 

[Stiggy Pop]

In any case, you've got a peak efficiency region, great, give it the axes zones it wants, and the fuel it needs, and let it perform.

Is there any precedent for such a narrow and fairly drastic peak in VE if things were truly functioning properly though? We can watch it go up and over the hump at consistent MAP and PW, I have a hard time buying that it could just suddenly need half again more fuel because the stars align right at 3k rpm. Also I've mentioned it's dragging on it's face over the hump, not giving a little bump of power.

just thinking out loud..

 

[FreeEMSFred]

50% more fuel required for a 500 RPM wide band? Not that I've seen, no. That sounds more like a software bug than a real world thing? Or an exaggeration?

Dragging on its face could be a result of rich or lean or misfire or whatever.

I saw one case of obscene VE required at a certain RPM/load and that was a case of misconfigured output angles such that the injections were merging together and the overlap was reducing fueling relative to what was intended. I have no idea if you have that level of control with your setup, though, and your symptoms are not the same, and no where near as extreme.

Looking back at your screen shots I don't see a 50% hump in AFR at all - it's much more modest. Exaggeration, or I'm reading them wrongly.

As for the required points to test this theory, 5 should be sufficient, and you could probably get away with 3. One on the knee on each side to contain it, one in the middle to produce the peak, and optionally another in between on each side to help better approximate the curve.

If you don't have fully configurable tables... then you could just borrow some columns from the top end to experiment with, instead. IE wipe out your top two or three bins by pushing the whole lot to the right (inc axis bins), and insert some new ones in between your trouble zone with the space made available. Try stuff. Don't be shy.

 

[FreeEMSFred]

On the other hand, depending on how filtered and/or slow your lambda readings are, maybe it is a nasty misfire that's just showing up in the traces as a smooth hump despite being wild swings? If it were an LC1 or 14point7 device connected to FreeEMS you'd see some pretty sharp swings for genuine rich misfires. Having 80Hz logging helps, too...

 

[FreeEMSFred]

I opened that log of yours up (after some clean up of the guts of it) and the first thing that struck me is seeing quantisation in the VE trace! I've never seen that before. On the other hand, I've not opened a MS log since 2007 and you certainly don't get quantisation in VE in FreeEMS!

• Injection timing is fixed at zero
• Timing error spikes to +/- 12.7 seemingly randomly
• You can't keep your right foot still
• There's no visible evidence of the engine "dragging its face" in that shot, the RPM curve would flatten and then recover if so
• No "accel enrich" in the area
• Stable dwell

And, interestingly, the VE/PW spike that you added is off centre from the AFR trace. I'd blame it on the wideband lag, but it's 0.7 of a second or so, which is MILES more than any good wideband does. In any case, the off centre VE should produce a waver in the AFR trace, as in, it should go rich, then lean, and it doesn't. As if they're disconnected, or your wideband is ultra slow.

That log is at about 10Hz. Almost unusable. :-/ I'm spoiled.