Another MS conversion, Progress, at last. Options

[JamesM]

Alright, I am determined for you to get this sorted out!

After staring at your latest logs for a while, I can't explain exactly what is going on just based on the log files, but I am seeing a bit of a pattern related to the rapid RPM oscillations showing up occasionally around the 1800-2000 RPM range and specifically the events that proceed them.

Basically the conditions are just right that you are coming in to that really lean spot in your Ve table that I had previously mentioned right before applying part throttle every time. Opening the throttle like that while lean is enough to cause it to buck a bit. What I cannot explain is the RPM ripple that persists for a couple seconds after it. Not being able to feel for myself what the car is doing makes it that much harder to guess. That being said, as the proceeding events appear to be the same I think those should be sorted out before making any sort of a call on this one.

Long story short, sort out that lean spot in the Ve table in the lower 1800-2300 bins before doing anything else. If you notice those 2 columns have values down into the 20's while on either side of them you are up in the 30s and 40s. You may want to sort out that entire bottom row as well, I would suspect any value less than 40 anywhere on your table could be to lean.

Fix that and grab some more logs.

[JimN73]

OK, after James' post about everything below 40 being lean, I decided to adjust all below 40 cells to 40. Boy, did that run bad.

Sadly, I saved over it and can't post it. But I ran auto tune again and did a data log. They are the "b" files in the Dropbox box. Then I reset to 40 and did a data log - the "c" files.

Tomorrow, I think I'll try to blend the VEs bit and raise the bins in the 2000 - 3000 range and see if things smooth out.

https://www.dropbox.com/sh/p593bw6iyebgwjn/bANp3X25ic

[edwin]

Slightly odd suggestion
Why not disconnect the tps and disable it as an input in the ecu?
I chased a really odd surge with my setup. Turns out a small section in the tps had worn and showed 100%. Would be similar to interference
I would second the disable AE as well (no tps would do this anyway) as it will only mess with your base tune
Cheers
Edwin

[JamesM]

OK, after James' post about everything below 40 being lean, I decided to adjust all below 40 cells to 40. Boy, did that run bad.

Sadly, I saved over it and can't post it. But I ran auto tune again and did a data log. They are the "b" files in the Dropbox box. Then I reset to 40 and did a data log - the "c" files.

Tomorrow, I think I'll try to blend the VEs bit and raise the bins in the 2000 - 3000 range and see if things smooth out.

https://www.dropbox.com/sh/p593bw6iyebgwjn/bANp3X25ic

Yeah, 40 might have been to much of a jump there, especally in the overrun area, that was a guess based on your table. Best way to do it is to work them up slowly so you can see how your AFRs react. Ill take a look over your new logs tonight.

-James

[JimN73]

Slightly odd suggestion
Why not disconnect the tps and disable it as an input in the ecu?
I chased a really odd surge with my setup. Turns out a small section in the tps had worn and showed 100%. Would be similar to interference
I would second the disable AE as well (no tps would do this anyway) as it will only mess with your base tune
Cheers
Edwin

I know it's at least slightly odd, Edwin. The bucking issue has plagued this installation since day 1, I think I'm grasping at straws.

I know how to disconnect the tps, but not how to disable it. AE is disabled.

A new twist. Last week, James (I think) noticed that the tps was reading something like 30% at 5000 rpm. I recalibrated it and now it reads 10%. After reviewing old logs, here's what I found. All at steady rpm

October 2012. 5000 rpm, 52% -- 2700 rpm, 22%

May 13. 2700, 16% and 23% -- 2800 rpm. 20%

Today. 3200 rpm, -0.7% -- 5100 rpm, 44%

I know there will be some variation based on load, but it seems to be low and erratic.

I'll have to get in touch with Mark to get the part #, but I'm about ready to run out and get another.

[McMark]

It should be set up to basically ignore the TPS for now. Those numbers should have zero affect on fuel. I'm happy to supply a replacement if it needs it, but for now just ignore it. Unplugging it may work, but I'm not 100% sure.

[JimN73]

Yep. didn't change anything.

I'm rereading all of the info that I've received over the last couple weeks to make sure that I'm using all the advice that's been provide.

Another day.

[JimN73]

I was looking back over the advice and comments from the last week or two and realized that one of James' suggestions had not been done. We got sidetracked retuning and trying to reconstruct the ve table.

James suggested adjusting VE to get rid of the lean spots around 2000 rpm. That's what I worked on today. I started out with a tune that didn't run too badly and added to the VE at the lean spots. 5 iterations later, still bucks and still doesn't run really well, but the table is a lot smoother. The adjustments eventually spread out oer much of the table.

So, will you please take a look at these files and tell me if I'm on the right track or have I overshot? Since the bucking hasn't decreased I'm wondering if there is something else going on, but I'm not sure what to look for.

Many thanks.

I hope the names of the files/folders are self explanatory

[904svo]

After looking at your file I would say you still have a noise problem, Look at your battery voltage its all over the place also your TPS is going negative when you let off the gas.

[stugray]

After looking at your file I would say you still have a noise problem

I havent even looked at any files, but enough people have reported strange values in the logs enough to convince me you have a noise problem.
Try drawing out your wiring diagram in detail.
pay close attention to where all of the sensors are grounded.

Consider running any sensor grounds to a common ground point close to the MS ECU with a dedicated wire (twisted pairs in certain cases).
Ground everything at that point and that point only.

It almost sounds like your ignition is causing the problem because it seems to happen at specific RPMs. The noise might line up with some sensor process in the MCU only at certain RPMs.

I chase these types of problems in advanced avionics and this screams noise issue coinciding with some rate based process in the Software. It is also the hardest to track down. If you had a friend with a Digital Storage oscilloscope, you could look at each individual signal while operating looking for the noise.

Stu

[904svo]

After looking at it again I'm willing to bet you have a wide band 02 on it, the heater
circuit on the 02 sensor must be wired thru a relay then to the battery also the ground for the heater must be run separate to ground on the engine.

PS the battery voltage controls the pulse width to the injectors as the voltage goes
up and down your fuel flow changes

[stugray]

Oh, and another thing that can drive someone crazy troubleshooting these things: See if the datalog rate can be increased from the rate you are currently using.
Sometimes problems cannot be seen at certain sample rates of the system (aliasing).

Stu

[McMark]

All sensors, including a custom CHT, ground through the MS. I'll look at my diagrams about the WBO2 heater ground. The wiring harness has a bunch of noise shielding (foil tape, shielded wire for crank sensor, shielded wire for MAP sensor). RWilner had LESS shielding on the exact same harness design and didn't have any bucking issues.

Jim, can you try to build a log with consistent rpm and TP? From what I'm seeing the MAP 80-88 with RPM 2300-2700 bins look rich. But what would be more informative to me would be long (10s+) sustained RPM and throttle position. Keep the values stable long enough to get a good long run at the data.

[904svo]

One goofy problem I had with MS was the case was grounded to the body. I mounted
it on a piece of plywood, noise problems went away.

[JimN73]

I'll do it in the morning.

The VE1 files are the ones that I believe James thought were lean in the 2000 range. I did several tunes and logs up to #5. I will not be surprised if it is rich. But, the car runs as shittily at #5 as it does at #1.

I changed the 94-101 MAP at 500 to 1000 cells, too. There was never a change in rpm, no matter what I did.

The new ECU made a great deal of difference, but there still seems to be something going on that overpowers everything at some rpms.

[JamesM]

I’m a day late here but I finally got around to checking out your latest logs. Going to try and keep this short as I just typed out a long response and my browser took a (IMG:style_emoticons/default/stromberg.gif) and lost it.


Logs are looking pretty good, Ve map will need fine tuning eventually but everything looks to be pretty much in the range now where you will not be having a problem from lack of fuel, Mark is right though you are a bit on the rich side in those areas he mentioned, doesn’t look to be so much that it would cause an issue though.

I wish I could see/hear/feel the bucking myself as bucking can feel a lot of different ways so its hard to guess at what it is. I am seeing that RPM fluctuation again around 2000 RPM, I am guessing this is where you are still feeling the bucking, is this right? how is it driving everywhere else? Even in that RPM/load range though the AFR looks to holding pretty steady at a good place, so congrats on that, I think we might be able to rule out the fueling as the problem.

So, if the problem is seemingly not with fuel that leaves us with spark to look at. Now I may be wrong on this one, but from the log files it does not look like it is misfiring during the problem area (this is why it would help to see it in person though or get a better description of the bucking) so I would assume you are still getting spark to burn the gas, perhaps just at the wrong time which is why it is running like (IMG:style_emoticons/default/stromberg.gif) Now this could be just that your ignition map does not work well with your engine and the timing around that range needs to be adjusted but as we are seeing some data that looks like it could be noise, I think we should investigate that possibility first. You could try adjusting the timing a little bit around that area and see what happens, but i am not holding my breath as your current ignition table is pretty close to what everyone else runs on type 4s.

I think the next step is to log your ignition trigger data and see what we can see.

Plan B would be to swap back to a regular distributor and run megasquirt for fuel only as a test to see if anything changes. (its the ignition side of megasquirt that tends to be sensitive to noise, running fuel only it is a lot more tolerant as the events do not need to occur at such an exact time.) this could at least help confirm it is a spark problem.

[904svo]

Make me a data log a 1000 rpms, 2000 rpms, 3000 rpms with the car standing still
for a lease 15 seconds for each rpm range. Also measure your battery voltage at the
battery while doing this test and report the values.

[JimN73]

All sensors, including a custom CHT, ground through the MS. I'll look at my diagrams about the WBO2 heater ground. The wiring harness has a bunch of noise shielding (foil tape, shielded wire for crank sensor, shielded wire for MAP sensor). RWilner had LESS shielding on the exact same harness design and didn't have any bucking issues.

Jim, can you try to build a log with consistent rpm and TP? From what I'm seeing the MAP 80-88 with RPM 2300-2700 bins look rich. But what would be more informative to me would be long (10s+) sustained RPM and throttle position. Keep the values stable long enough to get a good long run at the data.

Here they are

[stugray]

McMark,

RWilner had LESS shielding on the exact same harness design

Shielding helps mitigate radiated emissions. The Noise I suspect is conducted emissions (CE).
CE problems almost always are a result of poor grounding, ground loops, or inadequate noise suppression.
One thing that I suspect is leakage from the ignition into the chassis. (recall the other thread about poor running where the OP and others stated that they saw arcing from their wires to the shrouds at night?)
If this happens, the noise spike will be noticed by the sensors, but can also drive a shift in the ground. If all sensors are not grounded at one point, they will see varying levels of shift based on where they are grounded.
The battery voltage data jumping around is an indication of ground "bump".
The battery voltage should be the one thing that is rock solid in this system.

And just because you think everything is grounded at one place does not mean that it IS.
We do ground isolation checks before connecting the ground to the final connection point.

Stu

[904svo]

Jim having trouble downloading file. Email it to me at 904svo@comcast.net

Thanks Larry