Tuning from scratch

[Jim Blackwood]

Now that I have the 340 Buick running under '411 control, it's time to make it run well. A bit more complicated than adapting a Camaro tune, redline is set to 7500rpm and MAP to 170. All the usual, fan controls and such are done. The car is fitted with a wideband and EGT gage but neither of those are connected to the controller. The injector flow is set to 4.4gm/sec for 35lb injectors. I have run my first datalog using OBDwiz but am not impressed with the available choices for PIDs. I have run datalogs before on MS/TunerStudio and am comfortable with that type of graph display. Can anyone suggest better datalogging software?

As it stands right now, after starting the engine leans out badly and wants to die (14-15:1) and then eventually goes very rich and settles down to a nice idle (11:1). My datalog may indicate it is still in open loop, or I may not have the right PID included. Is this determined by engine temp, O2 sensor temp, or a time constant?

For initial set-up I've made use of the squirt bottle technique to verify rich or lean conditions and adjusted prime pulse and cold enrichment (B3632) as well as wholesale adjustment of the main VE table and the OL table (B3605)

First off, it seems counterintuitive that if it runs rich in OL after all the enrichments have timed out, that the enrichments could have caused any sort of a lean condition. Something else must be going on here but I don't know what. This, I think is the most important issue for getting on with the tuning.

What I'm missing here is a thorough understanding of which enrichments do what and how to tune them. For instance, I can easily accept that I may be missing a critical table, or that I'm adjusting the wrong thing. The B3632 table for instance. If it decays based on B3650 then how can it suitably compensate for coolant temp? And if it doesn't then why does my mixture go lean soon after starting? I richened it up quite a bit.

If anyone would care to have a look at this, my Bin and datalog are on the laptop but I can attach them. I just have to post this first then move over there.

Jim

Got the bin attached but it said the log was invalid. Not sure what to do about that, it was only 15kb.

[brian617]

First off, it seems counterintuitive that if it runs rich in OL after all the enrichments have timed out, that the enrichments could have caused any sort of a lean condition. Something else must be going on here but I don't know what. This, I think is the most important issue for getting on with the tuning.

Time to work on the VE table. Your Idle speed and MAP are changing via cells in the VE Table. Don't change enrichments until the VE is correct.

[DavidBraley]

Jim,

Because you are using the 0411, maybe something like this might help. I could be wrong, but I think it's part of, or an off-shoot of PCM Hammer:

https://pcmhacking.net/forums/viewtopic.php?f=42&t=6575

[Jim Blackwood]

One thing that I think might be worth trying: Tuner Studio has a graphical log viewing screen which is really quite good and I seem to recall that it can make use of .csv files so there could just be some slight chance of importing log files from other logging software into it. No idea if the labels would be anything close to right but it seems like something worth the effort.

Jim

[kur4o]

You need to start with the basic. Spark, ve tables and maf. Than work on other stuff.

Also find a bin with l31 engine and use some of the tables as template for spark. I guess will be the closest match for the buick.

Is this engine ever been run by a pcm. Ripping some stock tables and interpolate them from there will also give good results.

[Jim Blackwood]

As much as I appreciate the basic tuning advice I'm a little beyond that. The biggest problem with starting by tuning the VE table is that you first have to start the engine and then run it long enough to warm up, hopefully without washing the cylinders down in the process. Then you need numerous restarts in order to make changes to the table. The engine is a well broken in fresh build but with all this rich idling I'm starting to see blue smoke. I have to do something about that over-rich idle during warm up.

Yes, I understand that changes to the VE table will affect literally everything. Yet first it is necessary to get there. Now my ignition tables are just about right, certainly close enough for a Small Block Buick. I could build one of those in my sleep. After fueling is done I can fine tune that map. I've studied the instructions for using STFT and LTFT. My first datalog does not show the engine getting warm enough for closed loop but it wasn't running very long. I was hoping to lean out the idle before doing that. Up to this point my efforts have been perhaps not the best and that's why I'm asking for advice.

And yes, the engine was running under MS-II and running well. However I've not found MS to be particularly reliable. (over a period of nearly 2 decades)

So here's the sequence. Cranking may or may not produce immediate fire off. If it doesn't, a second turn of the key does. Then the engine settles down to a fast idle on the whatchacallit, PWM idle air thingamajiggit. (I know what it's called, give me a minute) Then it leans out and slows down to the point where it pops and crackles, begins surging and tries to die. If I nurse it past that using the squirt bottle the idle goes rich and it smooths out to a nice idle but it's puffing black smoke out the tailpipes, and I expect it'll run like that until it reaches operating temperature and maybe beyond. That's just not good for the engine.

What I want to do is lean that out. I'm GUESSING there are 2 ways to do that, with the main VE table and with B3605-commanded fuel in open loop. If that's correct, which one would you guys suggest? In MS I'd know what to do but this is quite a bit different.

I'm also not sure it is entering closed loop. I have a TPS error and a new one on order. Also a CKP error that I think I've corrected. But before testing again I want to lean out that rich idle. Yes I realize I will have to come back and re-tune it after the VE table is right. I'm less worried about that than I am about wiping out my rings.

Jim

[kur4o]

Fueling is derived by 2 tables maf and ve. You didn`t mentioned about the maf. Is it present at all. If it idles rich cut the ve tables in that region. If the startup adders are adding more than what is set, than air entering engine is not calculated correctly.

Bad tps can also throw a lot of troubles and if maf is not disabled correctly expect big headaches. If you have laying a maf plug it temporarily. You can dial it pretty fast compared to ve.

You must be sure that the inj data is spot on. If not switch back to stock inj for initial tuning.

Changing too many constants is not a good thing, you will never find the problem.

[Jim Blackwood]

Again, all good advice but...
First off there is no MAF, this is and can only be a speed/density tune. See photo. No place to put a MAF. My advance map is fine for this stage of the tune and I will fine tune it after fueling is set. It is very close, I'm quite familiar with what this engine needs and have been tuning small block Buicks for decades now, with EFI. I believe the changes for S/D are correct as I spent 3 weeks on the tune before ever turning the key the first time. The injectors flow rate is set to 4.4gm/min but again, I can double check anything y'all think I should look at. We've done this drill before. If a fault at the TPS will keep it out of closed loop I'd like to know. I suspect it could. At any rate all we are doing here is tuning fuel delivery or injector pulse width. The engine could care less how we get there, long as it gets what it wants, and it's a pretty simple device.

It would be nice to have the fueling algorithm to know what goes in and how. Mostly adders no doubt but for instance, when in open loop does the algorithm still use the VE table or does it substitute the open loop fueling table? And even is it uses both tables (which makes sense because the OL table has a temp axis) doesn't it make sense to adjust warmup in the OL table so the VE table stays flatter for initial tuning?

Based on my experience with tuning EFI systems I suggest the following initial process: Provided the advance map is reasonably close,
1) Enrich prime until the engine fires off
2) Enrich after-start for about a 1-2 second run
3) Add cold enrichment to get a smooth idle. With the '411 I would say this involves the open loop table and also the MAP based after-start enrichment. The question is, which one first? That may depend of whether or not the engine will continue to run. It might be advisable to zero out the after-start until later if the engine doesn't die. Then look at the AFRs and adjust accordingly.
4) After all of that, warm up the engine, adjusting the OL values incrementally, which will require several re-starts.
5) Finally, after the engine is warm and all the enrichments have dropped out, adjust the VE map for idle. Smooth the map into the surrounding areas using appropriate slope values. (Mostly a visual thing at this point)
6) Tune adjacent areas based on throttle response and AFR gage and blend map.
7) Either after cool down readjust fuel adders, or continue with STFT and LTFT and driveability tuning.

Now I've been stuck at step 3 due to confusion about the cold enrichments but I think I'm starting to make sense of them now. For me at this point it is more important to protect the rings than to be concerned about having to go back and balance the tables several times. In the end we are only supplying fuel. The right amount, too much, or too little. I fully understand the VE table being the foundation of that fuel delivery, but in the end each engine condition has a single fuel requirement and the engine cares not how you get there so long as you satisfy it's appetite. The balance can be off a bit and still work acceptably well.

But, if I continue to wash out the rings I can easily convert this brand new engine into a hopeless oil burner. I must prevent that above all else.

Jim

[kur4o]

Good that you post the bin. I managed to look at it and was mildly shocked with the ve and ol table settings.

I attache a bin with stock engine cal segment so you can use for comparison.

The ve is increased roughly 3 times and the target afr is 1.3-1.8 across the board.

A note how to use the OL table to get afr use 14.7 / value of table.

so value of 8 will give you 14.7/8= 1.83 AFR

14.7/1.065= 13.8 AFR

First fix those 2 tables and than we can continue the discussion.

[Jim Blackwood]

Ah, thank you. The multiplier was what I was missing. That I can work with. So in essence for both those tables (at least) the entered value is a multiplier, and acts on the requested AFR? I don't remember the name of the AFR table but I remember seeing a reference. Don't recall it being a full table though and maybe just a value. Something else to look into. Anyway this information will demand a whole new look at fueling but I'll try to run a test sometime tomorrow. Very helpful. It'll be real interesting to see how it all shakes out.

So if I understand correctly, during OL the afr is multiplied by the VE and then multiplied by the OL table value? I can see that resulting in some seriously high numbers very quickly. No wonder it was rich.

Jim

[tayto]

As much as I appreciate the basic tuning advice I'm a little beyond that. The biggest problem with starting by tuning the VE table is that you first have to start the engine and then run it long enough to warm up, hopefully without washing the cylinders down in the process. Then you need numerous restarts in order to make changes to the table. The engine is a well broken in fresh build but with all this rich idling I'm starting to see blue smoke. I have to do something about that over-rich idle during warm up.

Yes, I understand that changes to the VE table will affect literally everything. Yet first it is necessary to get there. Now my ignition tables are just about right, certainly close enough for a Small Block Buick. I could build one of those in my sleep. After fueling is done I can fine tune that map. I've studied the instructions for using STFT and LTFT. My first datalog does not show the engine getting warm enough for closed loop but it wasn't running very long. I was hoping to lean out the idle before doing that. Up to this point my efforts have been perhaps not the best and that's why I'm asking for advice.

And yes, the engine was running under MS-II and running well. However I've not found MS to be particularly reliable. (over a period of nearly 2 decades)

So here's the sequence. Cranking may or may not produce immediate fire off. If it doesn't, a second turn of the key does. Then the engine settles down to a fast idle on the whatchacallit, PWM idle air thingamajiggit. (I know what it's called, give me a minute) Then it leans out and slows down to the point where it pops and crackles, begins surging and tries to die. If I nurse it past that using the squirt bottle the idle goes rich and it smooths out to a nice idle but it's puffing black smoke out the tailpipes, and I expect it'll run like that until it reaches operating temperature and maybe beyond. That's just not good for the engine.

What I want to do is lean that out. I'm GUESSING there are 2 ways to do that, with the main VE table and with B3605-commanded fuel in open loop. If that's correct, which one would you guys suggest? In MS I'd know what to do but this is quite a bit different.

I'm also not sure it is entering closed loop. I have a TPS error and a new one on order. Also a CKP error that I think I've corrected. But before testing again I want to lean out that rich idle. Yes I realize I will have to come back and re-tune it after the VE table is right. I'm less worried about that than I am about wiping out my rings.

Jim

you'd have to have an injector stuck open to wash the walls down. i took a tuning seminar many years ago, if anything you'd want rich from starting out. i'd rather change plugs vs bearings or pistons.

[Jim Blackwood]

Been losing posts lately but oh well. Just went to check on something and now I get to practice my recall. So anyway, thanks for those two maps, they worked and look like a reasonable place to start over. I'm now on the other side of Stoich which is a better place to be. For now. As a point of interest, this controller uses a different algorithm because it takes the commanded AFR into account. Others use AFR numbers only in the optional autotune function. It means one is an adder where the other is a multiplier and that can result in one or two orders of magnitude difference by the end of the formula. Can't use the same numbers or the same range in other words. Worth remembering. Here you stay close to 1 which is unity. There you might be at 16 or higher with no decimal point.

So, back on track. Next, it would make a lot of sense to connect the output of the wideband to the PCM so it can be datalogged. Is there a preferred way you guys have found to do that?

Also, what is your favorite datalogger and why? The one I've used is purely a random choice.

And finally, what I went to check on and something I think can benefit you guys. The TunerStudio log viewing app is called MegaLogViewer and it is a pretty darned good graphical interface from EFI Analytics. You might have to register but it's free. It will open CSV files as well as a bunch of other formats and IIRC lets you select up to 16 color coded traces on up to 4 screens, and you can scroll through the datalog, resize, etc. so it makes it easy to compare blips and dips. Pretty feature rich, I'm still not familiar enough with it to know what all it will do. There ya' go, my effort to give back to the community.

Jim
(don't forget my 2 questions above)

[kur4o]

Here is a map of the pcm algorithm, so you get a better understanding how to tune.

The value from ve table is converted with charge temp, cyl volume and rpm to a mass of air entering engine in g/s. Now this your base scalar[mass air in g/s] for all fuel calculations.

The mass air is divided with the AFR target to get fuel mass needed for each cylinder. Than fuel mass is converted to inj PW by some calc involving rpm and fuel flow of inj.

So you need to set the Open loop afr table once with the preferred target and forget about it.

The best tuning is done in open loop by comparing the commanded AFR and actual AFR by adjusting ve table. How do you plan to tune boost with stock pcm, do you have 2 bar map or some external enrichment will be done.

Do you have a copy of the xdf.

For logging, tunerpro can be used but the built in obd plugin have some bugs. I still need to test the latest version. If all is good you can log any pid you need that the pcm support. Some custom patch can be done to support more than you need, like wideband on egr or ac input.

[Jim Blackwood]

Made some progress, it seems like the issue I had with the OBDLinkLX was very likely a Bluetooth interference problem although I can't be certain because in the process I upgraded a bunch of drivers and did a few other things that could have affected it. Still, the breakthrough came after I had turned off or deleted all the bluetooth in the area and then reconfigured just the one. So the bottom line is that my tuning laptop with PCMLogger now works. I have connected my WB-O2 output to C1-pin 55 and turned on the flag for B1301-EGR System Enable so I THINK the last thing to do is to include the EGR in the Log Profile for PCML. Anyone know how that should be done? I've heard it is a fairly common method for logging the WB. Once I get that I will run a log. I can post my xdf but not from this computer. I'll get that done though.

Jim

[Jim Blackwood]

kur4o I wanted to run something past you before I get too deep into the VE table. I'm working on the area between 30 and 40 kPa and 400-2400rpm trying to get a stable idle for warmup. I took the approach of richening the VE table and leaving the open loop table alone. At this point the mixture is still just a little too lean to support a steady idle but I have increased the values in those cells to 2.8000 which your comments should lead me to believe would be way too rich. This points to there being another problem somewhere, the question is, where?

I have checked and double checked the injector flow rate. I flow tested the injectors myself on my own flow bench and they range in a tight cluster around 35 lbs at 40psi. I used an online calculator to convert that to 4.4 gm/sec and entered that value.

Can you think of any other thing that could possibly cause this discrepancy? I'd sure appreciate any leads you can give me.

Jim