Help! '94 3.4L Datalogging

[WhiteCam94]

"mode 4" is what GM called the use of commands over the ALDL line to control certain functions. usually done by a tech1 for diagnostic, you can also force some interesting things to happen.

minimum IAC counts were slightly lower than i would have expected, which means a possible small vacuum leak, but it could also be how the throttle blades are set.

what all has been replaced recently?

Let's see..... I replaced the engine (long block), radiator, all the miscellaneous things that go along with that like spark plugs and wires, t-stat, etc. I also replaced the clutch, water pump, and hoses. This was about 12k miles and about 3 years ago. I just recently replaced the spark plugs and wires, trying to chase an under load misfire (which it fixed) and....I checked for vacuum leaks at the time and didn't find anything that spraying it with b12 exposed. But that damn rough idle has plagued this car for as long as I've had it.

So, mode 4 is the "command" feature in most scanners? Why is it called mode 4?

[RobertISaar]

called mode 4 due to the way the request is made. a normal request to the ECM to dump data is mode 1 due to the sent message being something like "F4 57 01 00 xx" with xx being the checksum and 01 being the mode, 00 being the message, so it's actually mode 1, message 0. mode 4 is more like "F4 56 04 xx xx xx xx xx xx xx xx xx", with all of those xx being the items that you're commanding. decoding them is always fun.

which calibration is in the PCM? there were a few that GM released specifcally for a rough idle caused by something TPS related. assuming the sticker is still there? if it isn't, have to pull it via the datastream.

and the 160* thermostat isn't helping matters either.

[WhiteCam94]

Wow I've got a lot to learn. I'm still trying to wrap my head around VE and things like that, let alone PCM "speech" protocols.....


Oh I'm glad you said that! after I put the engine in, it had a major issue with idle over heating. I think I put in a 165* t-stat. But yeah, once you were moving over about 20mph it was totally fine, but below that speed it would get up to about 250* on the gauge before the fan would kick on. I've since realized that there's no problem except the two different temperature sensors reading about 40* off from each other, i.e.; the computer's sensor reads 160* where the gauge sensor reads 200* (this has been backed up with laser temp readings), making it seem as if the car is overheating when it's actually not. I spent countless hours trying to figure that one out. Now I know that these cars come from the factory with the fan enable set to 225*...which is a little high I think. I changed the fan enable to turn on at 185* and shut down at 165* I think, should I bring that back up?

Tell me what you think, but there's a couple spots in the logs where it looks as if the TPS is a little whacky...Jumping from 25% down to 0 and then back up to about 17% in about a 1/4 second seems a little fast for my foot....

One of the o2 sensors seems like it's not switching through quite the same range as the other, like it's a little lazy or something. Like I said, I'm new so I'm still learning the basics.

Also, I was using a program called "Scan9495-V6" and there's a warning bubble for when the cam sensor signal is lost, and it seemed to just about correlate to the pattern of the misfires. Do you know what role the CPS plays in engine performance with these engines?


Pic of the label is attached.

[RobertISaar]

i would go back to a 195 thermostat and adjust fan temps to match it better. like 205 on, not the mid to high 220s that the factory cal is setup for. there are a re lot of things in the calibration that are dependant on coolant temp being around the 195 range. especially the stuff that wants a minimum temp of 176 before it acts normally/learns values.

i've never seen that program before, seems like he put quite some effort into it. i may play with it on the FWD applications too, can use not only the PCM functions, but ABS as well. i've never gotten the ABS mode 4 stuff to work.

[RobertISaar]

BKWX code.

i don't know if that is up to date or not.... calibration 16203291 was one update for the manuals, i can't cross-reference it though.

TPS jumping like that will cause issues.

O2 sensor could also do it, as would the cam sensor.



really, it seems like any of them could be causing it, or all 3.

[WhiteCam94]

I'll probably just end up replacing all three of them, I don't like having unreliable data...and as far as I know, all 3 of those sensors are factory. I'd rather tune once I know everything's good you know?


As far as the cal versions go, what are you using for a cross reference? I didn't even know they had revisions for these cars. It seems like GM kind of just threw the 94-95 3.4's together and sent em out the door!


I smoothed out the main spark table a little bit and that seems to have helped the idle a bit, although under WOT it brings timing down to about 19* for some reason. There's almost no knock either

[RobertISaar]

depending on RPM, 19 at WOT is correct.

keep in mind IAT also effects spark. at 100kPa and 90*F, 2* is being taken out just from that.

my crossreferences are data from GM through TSBs and the archive of BINs i have from these PCMs.

[WhiteCam94]

I've done some fiddling, things like messing with the desired idle vs ect and bpw multiplier for crank rpm...seems like it starts up a little faster now. I've also noticed that the stock spark table is extremely erratic, so I've smoothed that out and brought it up a couple degrees, then took it down at low rpm and high load because of a repeating knock at around 1200 rpm and between 90 and 100 kpa. She seems to be happier with the higher spark advance, and I'm not picking up any knock at all anywhere with 91 octane. The power increase just from an average of about 3 degrees timing advance is amazing! I realize that I need to be careful though, because no knock doesn't necessarily mean no damage, correct?

Seeing as you are a tuning guru, are there any pointers you can give me as far as messing with fuel tables goes? Tunercat doesn't seem to have much in the definition file for modifying fuel, which means I need to keep learning to use tunerpro....sigh.....

I really do appreciate everything you've helped me with btw man.

[RobertISaar]

yes, you can damage without running into knock.... ideally, you would measure power on a dyno to know where, but there is an optimal point of advance for every RPM/MAP combo. passing it will place extra strain on bearings, pistons, rods and headgaskets at a minimum. this is due to the pressure in the cylinder rising too soon and causing the piston to be driven back down before it reaches TDC.

because of that, i would assume this would be easier to see on engines with smaller cylinder counts since they'll have less cylinders providing positive power.

anyways, one you hit that point, more advance won't increase power or will even decrease it. if you're certain it is making more power, i'd say if anything fails, it will be due to just exceeding their limits, not due to "bad tuning".... gives you an excuse to find/make something stronger too.



which fuel table(s) are you referring to? the main and idle VE tables are 16 bit, which are an oddity in the GM OBD1 world. i also can't get any conversion to make sense, which is why i just left it in raw values.

[EagleMark]

which fuel table(s) are you referring to? the main and idle VE tables are 16 bit, which are an oddity in the GM OBD1 world. i also can't get any conversion to make sense, which is why i just left it in raw values.

Wonder if this was GM early attempt at Grams*Kelvin/kPa ?

The air mass per cylinder can be determined from the VE table using the following formula:
g/cyl = VE*MAP/charge temperature
Ve is in g*K/kPa,
MAP is in kPa,
charge temperature is in degrees Kelvin.

This is default in EFI Live but you can also use VE%, which actually does not work properly because even stock comes out to over 100%...

[RobertISaar]

possibility. i don't know enough about the concept to be certain.
here is what they look like raw:
3.4

3100


now, the 3100 got significantly better heads, yet is shown with smaller values. they both use the same injectors even. there is no conventional BPC value in the calibration either(or injector flowrate or cylinder size or......), but there is one that seems to be hidden in the algorithm, i believe it is also a 16 bit value, even though it's the same between a 3.4 and 3.1.

these, if you were to use the normal X*(100/255) VE% conversion come out to 400+% range...

i just look at the logs and change each cell by a number that i feel is safe to get them in line.

[WhiteCam94]

As far as damaging the engine goes, I'm going to be pulling it and rebuilding it into a 3400/3500 combo motor at some point in the next year or so, so shortening the life of the engine isn't really that big of a deal to me. It seems as though it does have a lot of extra power now though. I wish I had friends with a dyno! Man would that make life easier.

On those VE tables, what is the vertical value representing? I need to learn VE math, bad. It doesn't seem to me that spark is where the big power is lurking as far as tuning goes anyways. That just answered a question I had regarding tuning though. I needed to know if I'd have to switch to the donor cars PCM file for the 3500 heads, but it doesnt look like I do. I'll just tune, tune, tune.


Also, is there any way to "re-write" a speed density PCM to be able to handle MAF tuning? I've read tuning a MAF engine is easier.



You guys are geniuses. I feel like an ape dragging his fists, just with a wrench in one hand....

[RobertISaar]

the vertical portion would be the scale for the values in the table. i really can't assign a value to them, but i can tell you larger value = more fuel.

being a P66 unit, there probably is an input that could be used for high frequency MAF, but i don't have one that i can pull apart and dig around in.

and reading that and actually believing it isn't something you should take without experiencing first. i prefer speed-density, but i'm also a lot more used to it. when you understand and anticipate for the inaccuracies a MAF sensor can have, it probably is more or less equal.

also, everyone starts there. no one gets into the game without feeling like the small fish in the big pond. and once you feel like a big fish, you realize how much larger the pond is and keep cycling through that feeling.

[WhiteCam94]

Okay, I had figured that much. It's just the product of a calculation, and not any real-world value like pulsewidth? It seems like most of this stuff is actually pretty easy once you get into it. I'm having a roaring good time as it is, even without understanding everything 100%.

So, then tell me this: if I go boosted with this PCM, do i just rewrite all the tables to be able to interpret the new voltage output on say, a 3-bar MAPS?

I know it's always tough in the beginning, but it's nice to know there's helpful fellows out there who can answer my stupid questions with intelligent answers. You guys rock.

[RobertISaar]

not as far as i can tell.... it could be like what Mark suggested and be some strange multiple value calculation rolled up into a single value. or it could just simply be a 16 bit value that works well enough to create a desired pulsewidth.

if you go boosted, i would recommend patching the code to emulate a 1BAR MAP sensor like what i did with nAst1. then for anything related to pressure above 100kPa absolute, then add in tables for those. i will say this though, there is NO room for anything on the T-side, but plenty on the E-side.