The famous stumble...fixed?

Attach the Advance table too, or if you want to be awesome combine them together, Timing Editor will make your life easier (link to latest version).

 

I'll pull my map too so that you can compare, but mods will change the optimal timing majorly (what turbo are you running, do you have intake or exhaust).

What's the name of the Advance table? I want to make sure I pull the right one.

For the timing editor, do I copy the values out of the table into a text file and load them up?

 

Mods:

VF-52 (JMP)

Stock airbox w/ HKS hard inlet pipe

Invidia Catted Downpipe, 3" mid pipe w/ HFC, and stock (or TurboXS) axleback

Stock injectors w/ AEM 50-1000 fuel pump

AVO TMIC

TurboSmart 50/50 BOV

Cusco Coils

1-step colder plugs

What's the name of the Advance table? I want to make sure I pull the right one.

For the timing editor, do I copy the values out of the table into a text file and load them up?

 

Mods:

VF-52 (JMP)

Stock airbox w/ HKS hard inlet pipe

Invidia Catted Downpipe, 3" mid pipe w/ HFC, and stock (or TurboXS) axleback

Stock injectors w/ AEM 50-1000 fuel pump

AVO TMIC

TurboSmart 50/50 BOV

Cusco Coils

1-step colder plugs

Table should be called Knock Correction Advance (KCA).

 

You should be able to just copy and paste into the timing editor if it is the one I'm thinking of. There is a base and an advance to paste into.

Here is my combined table, apologies for the delay, had to travel for a conference.

 

All good, I'm just getting around to replying to this thread too.

 

What's the name of the Advance table? I want to make sure I pull the right one.

For the timing editor, do I copy the values out of the table into a text file and load them up?

 

Mods:

VF-52 (JMP)

Stock airbox w/ HKS hard inlet pipe

Invidia Catted Downpipe, 3" mid pipe w/ HFC, and stock (or TurboXS) axleback

Stock injectors w/ AEM 50-1000 fuel pump

AVO TMIC

TurboSmart 50/50 BOV

Cusco Coils

1-step colder plugs

 

VF52 is similar enough in size to 16G that timing will be about the same, everything else is about the same or don't have an impact on timing. Main items that impact timing are Turbo, Intake, Header, Downpipe, Catback, maybe TMIC (but not likely in the mid/low ranges).

 

Anyway onto the map...

Here is my combined table, apologies for the delay, had to travel for a conference.

 

 

First notice notice how jaggity your 2d maps are, these cliffs are bad for engine smoothness. Don't worry factory Subaru maps are just as bad (probably why these cars are so unsmooth stock). You want the timing map to be as smooth as possible for consistent and smooth spark delivery, which leads to smooth torque delivery.

 

Ok onto the problem area, the Engine Load columns in question would be 1.35 through 1.56. You're map is asking for 22.19, 20.78 at 2.4k and 24.30, 22.89 at 2.8k. That's a far cry from what you are actually running, which usually means there is knock reducing your timing (what's your IAM at, I don't see IAM logged?). If IAM is at 1.0 and FKBC and FLKC are at 0's, then you might have a timing compensation table removing timing.

 

 

Here is my latest map for Stock Intake setup, which actually is asking for about the same amount of timing that your map calls for (I realized that my earlier numbers were for ELH, which wants a good bit more timing then stock UEL). Notice how much smoother both of the 2d maps are, NSFW's timing editor makes this pretty easy to accomplish, thought it is still time consuming.

 

Don't go copying and pasting my numbers though, even though it's well below knock threshold (~2-4* below), it is right at MBT for my setup. Which could knock badly on another motor. But for perspective I'm at 2.20g/rev by redline at 12psi.

Edited October 15, 2018 by covertrussian

Thanks for the detailed reply, I'll work on smoothing the lines.

IAM has always been 1.0 and I basically have no FLKC or FKC.

 

I'll see if I can find any tables pulling timing. If you have any suggestions, I'll take em.

Oh forgot to mention, do keep the AVCS table in mind too. The more AVCS advance the less timing will be tolerated before it knocks, on my setup for every 5* of AVCS increase, I have to decrease my ignition timing by 2*. I'm running 5* of AVCS in those sections, stock is 10*. I didn't see actual AVCS values for yours, you should log AVCS advance degrees instead of OCV Duty cycles in this case.

 

Start logging IAM, since it might temporarily go down and be back up by the time you pull a learning view. As for compensations, check out the "Ignition Timing - Compensation" section, that will have most of them.

Edited October 10, 2018 by covertrussian

Just an observation... this discussion is SO FAR off now from the original discussion and issue.

 

Now being discussed is heavily modified, stage3 cars that are being self tuned and exhibit non-smooth/linear delivery of power. Not a light stumble exhibited in stock form and seen in Cobb staged maps.

 

You cant have a "scientific inquirey" without a control. Far too many varibles are being introduced.

A stumble is a stumble, no matter if it's fuel induced or timing induced. Everything I've discussed here can apply to stock cars too.

 

2005 LGT Stock Timing maps is as smooth as the Appalachians:

Edited October 11, 2018 by covertrussian

A stumble is a stumble, no matter if it's fuel induced or timing induced. Everything I've discussed here can apply to stock cars too.

 

2005 LGT Stock Timing maps is as smooth as the Appalachians:

 

 

I read all your posts related to this and other topics always in awe. Then i fire up various applications and find the tables discussed like "oh ahhh, I see"

Slow progress (too much time at work).

 

I started smoothing my timing lines (thanks to NSFW for his timing tool, it works really well), and I can feel a difference when driving. My last revision I got into a DAM lowering event and it took me a minute to realize that my AVCS table was too advanced from other testing. So I dialed that back. Here is where I am at for the moment. Going to take the 20 min drive to Costco this afternoon and will see how it goes.

 

 

 

 

Here is my combined table, apologies for the delay, had to travel for a conference.

 

[ATTACH]269636[/ATTACH]

That's looking much better, since you are increasing timing in a lot of cells do keep a careful eye on knock.

 

I've also been testing my resmoothed map this week. This one is similar to the one I posted before, but has 2* less timing due to cold air intake.

 

With this map I'm no longer getting constant throttle stumble (like when going up a hill in 4th gear at 40mph), but I do still feel a stumble of sorts if increase the throttle angle too quickly. Which is probably related to a different table.

Making good progress. On my 7th revision (working on minor tweaks to #8) and I've managed to get most of the stumble out and the transition from idle coasting to going is so smooth. Here is where I'm at:

 

 

This is the delta from where I started...

 

 

 

Tuned out the small FKC at moderate load that I would encounter. Running clean for FKC/FLKC so far. LV is very clean at 0-2 so far.

 

For the fuel line input, I tried a standard length (remember I have an STi FPR now), 2ft, and 3ft. Honestly, the 2ft and 3ft lines didn't help at all for the stumble. I went back to the standard length because I didn't like having all that hose in the engine bay where it could get snagged.

 

Last short run I did, the acceleration was noticeably better, to the point that I need to now figure out how to keep the nose from lifting so much. But, honestly just having it drive smoothly between transitions is so nice.

You are starting to push a lot of timing above 2.00/grev. Keep a very close eye on knock in those areas.

 

Instead of guessing, I would first find the full boost MBT (minimum best timing), since it's fairly easy with dyno/virtual dyno, then I would work in reverse to populate the 0psi column:

 

For example: if I'm running 12psi and my MBT is 21*, first I need to find how much timing to increase: 12psi x 1ign = 12ign to increase by

Then I increase my MBT timing by that number: 21 + 12 = 33ign at 0psi

 

I know this is a backwards of the typical start at 0psi and then reduce 1-1.25*/per psi, but since we have a full boost timing already tuned, we can work in reverse to calculate the in partial boost and 0psi columns.

 

For the fuel line input, I tried a standard length (remember I have an STi FPR now), 2ft, and 3ft. Honestly, the 2ft and 3ft lines didn't help at all for the stumble. I went back to the standard length because I didn't like having all that hose in the engine bay where it could get snagged.

 

I don't really see the 2-3ft of lines helping personally. More rubber does work like a damper, but you start running into harmonic resonance issues eventually too, at least with vacuum ines. With vacuum lines there is a maximum length, after which resonances goes way up, thus you want to increase the inner diameter of the line as you start stretching it far.

 

2-3FT might not seem like much, but you have to consider how many feet the line traveled from the gas tank too. Thus I think it's better and cleaner to just install an inline pressure damper instead.

 

Last short run I did, the acceleration was noticeably better, to the point that I need to now figure out how to keep the nose from lifting so much. But, honestly just having it drive smoothly between transitions is so nice.

 

Stiffer rear springs should help with that, I have to do that on my high (for front wheel drive) power car. But other tricks are having negative rake, (most subaru's tend to have positive rake). But this is getting off topic

Edited October 22, 2018 by covertrussian

I've also been testing my resmoothed map this week. This one is similar to the one I posted before, but has 2* less timing due to cold air intake.

 

With this map I'm no longer getting constant throttle stumble (like when going up a hill in 4th gear at 40mph), but I do still feel a stumble of sorts if increase the throttle angle too quickly. Which is probably related to a different table.

 

Update to this map, I went through and got rid of most of the negative timing, and further smoothed out partial throttle acceleration zones. This just about eliminated the increasing throttle stumble while going up hills in 4th @2krpm:

Next update.

I'm on rev 10 now of my self tuning.

I had to replace the valve cover breather hoses, mine were ceramic and cracked in several places.

 

This is my current total timing table:

 

The key for me was realizing I was going the wrong way with AVCS.

I was going by a thread that was talking about fixing stumble by jacking up AVCS to 25 or 30.

That wasn't helping my case.

Going the other way to 10 or 15 was much better.

 

Right now the idle is dead smooth, almost can't feel it running. The acceleration is really smooth and consistent. The stumble is basically gone. Only time I felt it in the last week was at 50mph in 5th and <.30 load, I have to hunt to find it now. The shudder when you transition from load to coast abruptly is gone. This is the most fun I've had yet driving this car. No FKC/FLKC at this point. LV is really good. I'm a happy camper right now.

If you think about what more AVCS is actually doing, it starts making sense why it causes the stumbles.

 

Increasing intake AVCS increases overlap, which leads to these items:

1. As backpressure increases, hot exhaust gasses re-enter the cylinder, heating up the cylinder and the cool dense charge. This leads to detonation and powerless, this lowers knock threshold, requiring less timing to avoid detonation.
   
2. These inert exhaust gas dilutes the fresh mixture which causes the EGR effect. This greatly reduces the flame front speed, requiring you to increase timing to regain torque and get you back to MBT.
   
3. Last effect that most don't think about, as you increase AVCS, the intake valve closes sooner. This matters because the piston starts rising up while the intake valves are still open, the longer they are open the more air/fuel is pushed back OUT of the cylinder and into the intake manifold. When you close the valves sooner, the piston can compressor more air, which increase your dynamic compression ratio, by quite a bit actually (someone smarter then me can do the math to figure out by how much, would love to know myself). As you increase compression, burn efficiency increases which means you need LESS time to get a full burn, aka less ignition timing to get to MBT.
   
   

 

Where does that leave us? Two of the items make the motor want less timing, while one wants the motor to have more timing, in the end they might cancel each other out, but in my experience restrictive turbo & manifold setups absolutely hate overlap. I've turboed an NA variable valve timing motor, where it lost power as soon as the high cam kicked on (had more lift and a ton more overlap, while low cam had 0 overlap). Swapping in a factory turbo cam, which has much less overlap, made the car be insane. Smaller turbo setup (read anything short of GT30+ 1:1 setups) simply don't like a lot of overlap and will knock & stumble.

 

 

Perhaps this is why Subaru went with a dedicated EGR system on FA20DIT's, to be able to increase EGR, without increasing overlap.

Edited November 20, 2018 by covertrussian

Been trying to diagnose this irritating stumble for years now and I feel like I'm getting there.

 

My stumbles are becoming less frequent and less severe.

 

Recent things I've done that has had a big impact is cleaning my PCV valve and replacing my torn turbo inlet.

 

Just thought I'd check my two cents in.. it may help

 

Sent from my F8332 using Tapatalk

This issue with running smoother with 10-15 degrees of AVCS makes me think that Subaru must have changed the cam specs between the 05 and 09 years - presumably at the same time as the facelift.

 

My wife's 09 is super smooth and it uses 40 degrees of cam advance. According to the specs I found for my 05 (here), that would mean 26 degrees of overlap at .050 lift, which seems like an unreasonably large amount. However, 10-15 degrees of advance would leave -4 to +1 degree of overlap at .050, which is totally reasonable.

 

For comparison, my car (with BC272s and dual AVCS) is running -4 of overlap (maybe "4 degrees of separation" would be a better way to put it) and it's pretty smooth. Not as smooth as my wife's 09, but better than my 05 when it was stock.

Thanks! This was really helpful!

One of my ongoing theories is, the 07+ can get away with much more cam advance because their TGV's are closed during those valve events.

I can see how that would help, but 26 degrees of overlap at .050 is still hard for me to believe. Regardless of how nicely atomized the fuel is, random nontrivial amounts of it would be getting blown out the exhaust port at low RPM.

 

Apparently with LS1s (not apples to apples but the cylinders are only slightly larger) you can't pass tailpipe emissions tests with any overlap at .050. Mine has 9 degrees, and it's prone to surging at low RPM (like 1500 RPM in parking lots). With 3x the overlap I'd expect the surging to continue into higher RPM.

I've made a lot of progress with my 05 but it still isn't quite as smooth as my wife's 09, especially after a cold start. So it finally occurred to me that I should compare the "Timing Compensation (ECT)" tables in both...

 

They are both zero from 140-194F, but at lower temperatures the 09 ramps up gradually whereas the 05 stays zero until 50F and doesn't ramp up quite as much.

 

So I re-did the 50-120F cells in my 05 based on the 09 table. I have about 1.5 degrees additional advance after a cold start at 50F (which is about what the temperature is out here), and as the engine warms up the advance tapers slowly rather than just going to zero at 68F.

 

I've only done two cold starts so far (both today) but I think the extra advance is a win.

My wife's 09 is super smooth and it uses 40 degrees of cam advance. According to the specs I found for my 05 (here), that would mean 26 degrees of overlap at .050 lift, which seems like an unreasonably large amount. However, 10-15 degrees of advance would leave -4 to +1 degree of overlap at .050, which is totally reasonable.

 

For comparison, my car (with BC272s and dual AVCS) is running -4 of overlap (maybe "4 degrees of separation" would be a better way to put it) and it's pretty smooth. Not as smooth as my wife's 09, but better than my 05 when it was stock.

 

Can you elaborate a bit on this.. I'm afraid my lack of understanding how AVCS really works is getting the better of me. I'm not sure what you mean by .050 and .006 lift. That's inches or? I'm not sure if you are using the 07-09 timing chart that was posted by covertrussian or if you have a timing chart specific for 05-06.

Cam duration and "valve events" (opening and closing of the intake and exhaust valves) are always measured in reference some particular amount of valve lift.

 

In the Subaru world, 0.006 inches is usually the reference, for example the stock intake cam duration is 240 degrees, and BC 272 cams have 272 degrees at 0.006 inches of lift.

 

In the GM world, 0.050 lift is more commonly used, and those same cams would be referred to as 204 degrees and 222 degrees.

 

The difference between the .006 and .050 durations depends on the cam's ramp rate. Two cams might have the same .006 duration, but if one has a more aggressive ramp rate it will have a larger .050 duration than the other.

 

For example... I just noticed that the stock cam in my C5 was 204 degrees at .050, which is the same as the stock 05 LGT cam, which is a funny coincidence. But the duration at .006 degrees is 270 degrees for the Corvette, vs the 05 LGT's 240 degrees. So the Corvette has a much slower ramp rate. Perhaps because it has one huge heavy intake valve rather than two small light ones? I dunno.

 

I got information for my 05's stock cams from a page out of a factory service manual, and it's exactly the same numbers as in covertrussian's 07-09 diagram. So maybe the cams are indeed the same, and lots of overlap just works better than I thought was possible.