Advice needed on initial startup of new engine

[trbugman]

Remember, when you make any changes to the spark table, you need to re-tune the VE table. Even small changes in timing can make changes in the VE table.
This is where the tinkering comes in, it's really hard to leave it alone. I guess it depends upon how much you actually drive the car. For my dd/work car, I tended to "play" with the timing in spurts and re-did the divisions on both x and y axis of the tables to really dial in the tune. I tended to favor the area around 3k rpm and about 60-75 Kpa for hwy cruise. I'll miss that car.. at the end it ran better than it ever did on CIS.

Once you figure out the tuning, you'll wonder why you didn't do this sooner.Have fun.

No doubt about that. I had a flat spot that no matter what I did with the timing I couldn't get rid of. Once I got the O2 sensor plumbed in I could finally tell what was going on with the carb. My idle jet was reamed out huge and was running 10:1 around town and drowning the engine in fuel. Put in the correct idle jet and it really woke the engine up. Now I have a lean spot between the idle and main(they probably need to both go up one size) but I don't want to put any money more money into the carb. I need to get the last 2 pieces done so I can start tuning fuel too.

[grelland]

That looks better, but your high advance cruse might need to go higher into your map than mine.
My engine idles in the mid 40's and under light cruse is in the high 30's around town and the 60's to 70's on the freeway. I also use the idle advance table to set my base advance and the rpm timing correction curve to try and hit my target of 900 rpm.

Here is my table.

You are right, I actually did that before I took it for a spin, and it seemed to be a bit better, and the logs shows a lot less fluctuations on the timing.

I did also re-tuned the fuel after changing the ignition, and the VEAL is still working on leaning out the 1500-3000 rpm area which makes sense, since I can feel a distinct flat spot in that range.

I have used a lamda delay table that someone (pildriver, was it?) posted here a while back, but I should really take the time to set up my own, as I can see that the VE analyzer and VEAL do come up with slightly different solutions.

Most of the time I am alone in the garage, and somehow I am not fast enough (or tall enough) to see whether the butterflies were fully open when I hit the "pedal to the metal", and it turned out I only had about 70% opening. That helped the top end a bit

For the first time, I have a CHT sensor under the #3 spark plug, and the highest I have seen so far is 145-150°C (305°F). I really dont know if that is "good" or "bad" as I have no experience with neither this particular engine nor the CHT temperature, but are there any good guidelines as to what to look for on the CHT? My intention has been to have a gauge so that I can monitor if things get hotter or colder as I tune, but I have no real experience on what the absolute value should be.

[Piledriver]

The VE analyzer in MLV is far inferior to the one in TS, simply because it lacks that delay table.
If you have registered TS, use VEAL, not the one in MLV, it will trash a perfectly good VE table with noise, as it only has one delay you can set, and it usually isn't enough at most loads/RPMs.(on MS3, none)

The big difference is the table allows VEAL to accurately map AFR samples to cells in the VE table even while accelerating or whatever. (works best if you accelerate/decelerate at a reasonable rate in third, but...)

The only way MLV is useful for fine VE table tuning is ~steady state, slow changes in the load/rpm, and manual tuning. Works, but MUCH slower and more error prone.

I can have a reasonable ~from scratch VE table with VEAL in ~20 minutes hitting a few freeway onramps and a hill or two for those hard to reach corners of the table..(had to do it the other day when I went back to MS2 CPU temporarily, and switched to "include AFR", which was off on the old MS2 tune)

The table I posted should be reasonable with most ACVW headers, (at least in the ballpark) but also depends on the WB you are using. Should be a lot saner than the default, anyway.

It was first set up with a $25 JAW1.03, which could clearly detect a single cylinder misfire at speed...
(found a bug in the old sequential code that way, it was handy for setting the delay time as well)

The 14point7.com OEM eval kit has about the same lag times, just has bigger output filter cap.

[grelland]

Great, How is the PLX compared to those? still in the ballpark?

[Piledriver]

Great, How is the PLX compared to those? still in the ballpark?

No idea, have never used one. Let us know

The newer 14point7.com units have a neat 2-step voltage feature at power on to allow analog calibrating the "known" AFR voltage outputs to the ECU. Only digital comms are ~perfect.

I just calibrate 20:1 at fuel cut and richen it up past 10:1 at idle to rail the high and low ends and use the "custom WB" setup and tweak the upper/lower voltages to match 10/20:1. Short of cal gasses (or the cal voltage output) it's the only way to be ~sure.

[andy198712]

Where's you find a wideband controller for 25bucks!! Lucky!

[grelland]

Nice view from my office window today

ofiice.png

It works decently now, but I need to sync the TBs as there is a stutter just as I touch the accelerator, but it disappears as I give it a bit more, so I suspect the one side is opening before the other. Also the idle is a bit too high at 1800 rpm when warm, which makes driving in the morning rush a bit of a challenge

[trbugman]

How's the tuning going?

Some good information from the MegaMeet this year was posted.
It tells you how to use scatter plots to help determine what tuning algorithm(sd, maf, itb or alpha-n) is best for your engine.


http://www.nbs-stl.com/tuning/

[grelland]

Thanks for asking.

It is going quite well. I got it up and running, but I did see some challenges in that I have a somewhat weak MAP-signal (it idles arouund 70 kPA), so I am currently using ITB mode.

I did spend some time reading up on how to determine the crossover point from SD to AN, and what I ended up doing was to take a data log file from a good long run, and dumped the data into excel, and filtered out all MAP data except the points that were between 89 and 91 kPA. The idea was to find a point (or rather a line) where the MAP reaches 90% do that the AN mode can take over at that point.

The filtered data looked like this:

tps - rpm at 90% MAP.png

I put that line into the ITB load Switchpoint table, and went from there.

The most annoying issue I have at the moment is that it jitters a bit just at tip-in. Or if I coast and just as I touch the throttle it jitters a bit. I got it a lot better by adding fuel in that region, but still not optimal, but I have no problems running the car in normal traffic, and I currently use it daily

I might have a bit too large butterflies, as I am using 52mm ones, type SP1 from VW Speedshop. I am loooking into making an eccentric pulley for the throttle cable, so that it gets less sensitive at low throttlen, and then increases the opening towards WOT.

These TBs have one butterfly on each side, so that I might also be experiencing an issue related to the fact that the firing order of the VW engine makes the time between vacuum pulse form syl 1 and syl 2 (seeing the same "plenum under the almost closed butterfly) different than the time from syl 2 to syl 1. Therefore I suspect that the volume below the butterfly does not have time to fill up with air at near idle on half of the firing events on each side.

I am not sure that made sense, but the point being that due to the firing order the firing on the 1-2 side of the engine is like 1-wait-wait-2-1-wait-wait-2 etc.

Other than this, the car goes like stink I will see if I can get a short video of it with the wonderful noise of the Python exhaust

I will try to get it on the dyno next week to get some figures

Roy

[trbugman]

Sounds like your getting a handle on tuning.


I'm hoping that this will help me with my ITB's when I get around to installing them.
I wonder if something like this could help you get a better map signal.

Did you ever get a chance to do a map log? That could give you a better idea what the cylinder pairs are seeing.

One of MegaSquirts greatest features is how powerful its logging features and software are, the hardest part is making sense of them.

[grelland]

That looks really interesting, another option i Was thinking of was to improve my current MAP sensor system. I have TBs with one butterfly on each side, so there is a "plenum" under each butterfly as the runners come together before the butterfly. So I have AN-6 hoses from this plenum on each side to a small tube, about 12mm diameter, and 120 mm long, centrally mounted, and then a 4mm hose about 20 cm long, to the MAP sensor. I have played with the thought of adding a larger "plenum" connecting the two sides. not just to improve the MAP signal, but also maybe for providong a larger volume for the engine to suck air from; or as you suggest, add one MAP senor either side, and then add/mix the signals in the MS.

I will see if I find a log file that is representable for the current status of the tune.

Thanks for the input