TD_D mod journey. From stock to the 'bastardmobile'

[Td_d]

Thanks for that comprehensive answer - that's why I though I'd continue on the thread, I'm sure others are interested.

Well, checked out the logs this morning - a little too rich now, but not significantly, needed to lean out the bottom end a little. The knock is a very predictable pattern I've noticed - which means I should be able to tune it out, I just need a little help how

It's always at around 2500-2600 RPMs; Just entered open loop (10); load is between 1.3 and 1.45; Throttle delta is low (2-3%) so I don't think it's tip in; IAM (DAM) is still 1; FBKC generally get two knocks - i.e. pulls -2.8 and then rapidly drops (within 6-8 cells) so I guess under 1.5 sec; have not been logging timing advance, so will need to add that too. FLKC on the whole is zero - some time it has added -1.4 to that table cell, but drops it very quickly. It seems to coincide with the remaining 'murmer' from the compensation, so I wouldn't be surprised if it's leaning out a little, and then correcting. It's also not at WOT, generally low partial throttle.

Have a look at this table section - although this time I got -4.2, which rapidly declined. What particularly struck my attention is the AFRs lean out as load and RPMs are increasing at that point - I found some funnies in the primary open fueling table he used at exactly that point, which makes me believe he was trying to tune out the stumble with the fueling table, as opposed to compensation. I made the two cells slightly richer, let's see, otherwise I'll drop ignition timing a little - but timing is as high in the adjacent higher load / RPM cell in timing, with no knock - but AFRs are richer.

View attachment 907

Here's the full log: http://dl.dropbox.com/u/3099895/log0058.csv

On the fires and pops, yeah, I'm a kid at heart too, especially with this car! I've also got a BOV on since the aim is to take the boost well beyond what the stock BPV can handle (this is a built engine afterall )

As soon as the LC1 get in, definately must check the correlation between the tables and WOT AFRs...

 

[Td_d]

Wow, combined corrections well under 6% variance - mostly closer to 3%. Should be even better after the tweaking of the low end I've just done. I'm a little impressed with myself :lol:

View attachment 908

 

[HolyCrapItsFast]

I can't see the attachments. There is something wrong with the links.

Looking at the log it looks to me that you are having issues between 2400-3800rpm and between load cells 1.10 and 1.75.

You could take several approaches to fix it. The first one is to simply remove about 4* of timing from that region of the Base Timing Primary Cruise and Non Cruise tables.


The other way (which I believe to be the right way in this case) is to richen up the fuel in those areas. There are two ways to do that and you can do one or both.

First since this is happening in closed loop only as it seems, you can go to the closed loop targets table and zero out the region of the map that is causing the issue. This will simply make your closed loop targets the same as open loop targets.


The other is to richen up the open loop targets in the effected areas and that will apply a change to closed loop by default since closed loop references open loop targets.


Now I did notice you are using his fuel tables which is okay with me if they work but one thing I must insist you do is richen up the higher load cells as I have done in this example. I feel that 11.8:1 at 6800rpm and 3.20 load is way to lean for a Subaru even with meth.

Also you mention you have a BOV. That kinda changes some things. I thought you were still using the BPV. That said you will never get rid of the backfires so with a BOV you have to live with them. Also the large variations in A/F learned corrections and your stumbling issues can easily be attributed to this. It is nearly impossible to get perfect trims and smooth operation with a BOV on a MAF car unless you know the strategies for tuning it.

What you can do to minimize the effect is to tighten up the valve a little bit till you JUUUUST start to hear surge and then back off a tiny bit. You will notice a vast improvement in both trims and shifting. Trust me I have this whole BOV in a MAF car thing down pat. The best BOV that I have found for our application is the Synapse Synchronic BOV w/the anti stall valve. Well worth the investment and is the most tunable BOV on the market. Plus I now exactly how to set it up to work properly.

Unless you are using Speed Density or a Blow Through MAF you probably don't want a BOV unless you know me! :lol:

 

[Td_d]

Hey Holy - thanks for the help - I also figured that it's fuel, especially after keeping a close eyeball on the AFRs and how they react around those areas - way too lean. I've implemented the changes above - including richening up the top end, rather safe that sorry!

The BOV is a pretty decent one - Turbo Smart dual port - it's actually a recirculating BOV, so I think closer to BPV than straight out ricer :lol:

On the trims - I don't know what' going on with the image attachments, I'm having problems using the site's upload facility - it loads up, shows, and then doesn't seem to save, arg. However - after the tweaks - drumroll - learnt trims, all the way up to open loop, are between 0.5 and -1%! Even the combined correction is in the range of 3% to -3%.

Now, to get the wideband. THere's a production problem with LC1's - so I'm either going to have to go with an AEM or get the newer Innovate MTX-l, which I wouldn't mind - but will likely have to order directly from the States. Parts availability problems, what else is new...

 

[Td_d]

Oh - and by the way - I like the backfires

Our traffic cops aren't anywhere near as enthusiastic as yours, thankfully!

 

[Td_d]

Riched up the fuel maps, knock's basically come right down - only two instances at 2000-3000 RPM (1.1 load) of -1.4 and 1500 at 1.3 load (also -1.4) - zero knock at high load and/or RPMs now, and the car's driving like an absolute beast - did a couple of WOT pulls up to 6600 or so. Definately slightly rich (pop, pop, poppity pop at gear change - well, more like pow!), so can't wait to scale the open loop now, probably get a little more power out leaning it out.

 

[HolyCrapItsFast]

Before you do that, now since you changed your target fuel you need to recalibrate your MAF! :lol:

 

[Td_d]

Yeah, I know... jeesh, this damn MAF! I can feel that the compensations are out a little as well, so need to rescale and then do a load comp run, again!

I meant to ask you - on the previous logs, but more curious than anything - if you see the dynamic timing advance ramp up (positive to +8), and straight thereafter a small knock (as which points the ECU backs off) - at the 3000 RPM area or so - I would assume that this is nothing to worry about, or even try and tune out - just the ECU doing its job?

 

[HolyCrapItsFast]

I believe that to be correct. But it really depend what the load is also. If your load is more than 2.00 then maybe I would back down the dynamic advance table in that region just a little bit.

 

[Td_d]

Load is very low - 1-1.2 - so I don't think it's a biggie.

 

[HolyCrapItsFast]

Then it is fine. As you suggest the ECU is doing its job!

 

[Td_d]

Another quick question - in calculating the load comp tables - how do you differentiate for the spreadsheet between cruise and non-cruise? CL/OL 8 and 10?

 

[HolyCrapItsFast]

I make two files from the original log and run each one separately in its own instance of the tool. For each one just copy either the cruise table or the non cruise table respectively from RomeRaider and paste it into the ROM tab of the tool.

The one I use for cruise, I filter out the all the CL/OL switches with a value of 10.

For the non cruise both OL and CL are relevant because you could still be in CL in non cruise situations, and most of the time you are if you are just accelerating normally. So that file I will leave unfiltered.

 

[Td_d]

Okeydokes - that's what I thought it would - just wanted to double check! BTW - congrats on passing the 200 mile mark - that first WOT pull is sheer heaven!

The only annoyance is that my tables are in bar absolute, so always have to check that the headers are correct. I assume that to convert for the tool it's absolute bar, less 1 atmosphere (and then converted to PSI)?

 

[HolyCrapItsFast]

I am all smiles!

 

[Td_d]

Yeah, it's the most awesome feeling... I've been tearing around like an absolute bastard since the completing the maps, I forgot how awesome these cars are at full tilt!

 

[HolyCrapItsFast]

The thing is I was expecting to see a difference but not like this. It is really night and day and not much has changed other thane the turbo. The nice thing is it doesn't spool any later than what was there originally.

 

[Td_d]

Tuning the 3000 RPM stumble is proving quite stubborn. I have another theory - its probably worthwhile to scale the CL MAF with the compensation table zeroed first, and then do the compensation - I get the feeling that the two are giving each other 'feedback' which is skewing the results each time (if learning are off, compensations results in a certain correction).

Will try that...

 

[HolyCrapItsFast]

Exactly. This is why I started with comp table set to zero. I forgot how long ago we did that but we did it at some point.

Sent from my Desire HD using Tapatalk

 

[HolyCrapItsFast]

Say, I was wondering if you could give me a log of when the stumbling issue occurs and with the following parameters and if you could give me the uber CAN log that would be good!

A/F Correction
A/F Learned
A/F Sensor Ratio
AVCS
Engine Load
Engine Speed
Atmospheric Pressure
Manifold Relative Pressure
Closed/Open switch
(If Available) Commanded fuel primary or Target fuel
Mass Air Flow
Mass Air Flow Voltage
Roughness Cyl 1,2,3,4
(If you still have them)TGV voltage, Left and Right