Random Data Plots - M42 Engine Parameters Stock vs 2.1L Stroker

[roguetoaster]

You da man bmwman!

We all want to see data plots and the MAF mod finally coming to fruition!

[bmwman91]

Working on it. I got the stuff installed back in the car today, but the RPM measurement is giving me some trouble. It seems like there is a lot of noise getting in to the signal, as well as some sensitivity issues. I have asked the guy that made the VR sensor conditioner board for some help with that. It would be even better if I could find the RPM pulse signal on the Motronic itself (the nice 5V square wave one that I am sure exists in there lol).

It is going to be hard to make an actual comparison of air flow between the two air meters (AFM & MAF). Based on the test logs I took, I think that the fact that they differ in age by 2 decades, and the fact that I have not made calibration curves for each specific sensor, is going to mean that no direct air flow comparisons can be made. According to the log, the MAF flows ~20CFM less than the AFM, which I imagine is really just due to the tolerance son the sensors (the data sheets seem to indicate that there can be up to +/- 5% accuracy variance between sensor units).

So, I will try to do a comparison where I measure the time it takes to go from 2000RPM to 7700RPM in 2nd gear at WOT with both sensors on the same day on the same section of road.

[varg]

It would be even better if I could find the RPM pulse signal on the Motronic itself (the nice 5V square wave one that I am sure exists in there lol).

Pin 74 is tach out, should be easy to use the signal from that. I'm guessing the noise that you're encountering is because of the wire running directly over the chip on the VR conditioner board. With a stock harness and an adapter running to my MS, I have had no issues at all with the crank or cam signals.

[bmwman91]

OK M42 fans, I finally have decent data for comparing the stock VAM and a MAF on the 2.1L MM engine. The VAM uses a stock M42 air box with a 1 year old Mahle paper filter, and the MAF uses a modified M30 air box with a fairly new Mahle paper filter (60% more surface area than the M42 filter).

Here is the quick and dirty summation of my findings: Using a MAF doesn't gain you anything on the M42, even with a strung-out 2.1L.

Surprising? Not really. Disappointing? A little. BMW really knows what they are doing (and yet we are constantly second-guessing their design choices!), and the stock M42 air box & VAM are just not restricting anything based on the data I am seeing. The filter element in the M42 has the same surface area as the M20 and S14, so obviously it is good for ~200bhp, at least. As for the VAM, I was expecting it to be at least a little restrictive at higher RPMs on my 2.1L beast, but so far that just does not seem to be the case.

We can all rest, assured that BMW made a very good air box & VAM sensor setup for the M42 when they designed this engine. So, that much is a big plus. The obvious downside is that it is a big nail in the coffin for the idea that there is more power to be gained from a MAF conversion. Sorry folks, it looks like the only areas where the M42 got handicapped are the stock software tune, and the conservative stock cams (and do note that hotter cams will not do anything good for your fuel economy, so even the stock ones have their purpose). The stock cams aren't really a handicap anyway since they traded power for fuel economy, so basically it's just the stock software that sucks. BMW really knew what they were doing in designing this engine (duh), except for the horrifically over-complicated and unreliable timing case setup lol. My MAF conversion adventure has been about a decade long, and I can now finally say that I have reached a conclusion, not necessarily the one I wanted, but at least I know what the deal is now.

So, here are air flow and AFR readings taken during 2nd gear WOT pulls.


There is an important fact to keep in mind here. The air flow values are from 2 different sensors. Air flow values are calculated as a function of the output voltage from the sensors, so the accuracy of the air flow values is subject to whatever tolerance deviations are present in these individual sensors. For example, an "ideal" M42 VAM sensor might have this as its transfer function: Air Flow = (5 x Voltage) + 30, so my logger reads the voltage and calculates air flow based on that. This actual sensor might have a transfer function like: Air Flow = (5.176 x Voltage) + 28.97, so obviously my calculated air flow value will not be entirely accurate. Neither the VAM or MAF that I am using have been put on a flow bench to measure their actual transfer functions...I am going based on Bosch data for "ideal" sensors.

You will notice that the AFR differs between the two sensors' WOT pulls. The fact that the MAF pull is ~5% leaner over most of the pull could indicate that the MAF is under-reporting flow, or that the VAM is over-reporting flow to the ECU. Or, it could be due to the "lumpier" MAF converter signal (the plot above has had some pretty heavy smoothing applied so it is easier to interpret visually) causing a little bit of confusion in the ECU, which I have seen cause lean conditions in the past. I find it unlikely that the MAF setup is actually flowing more air since the next plot of RPM vs time indicates NO performance difference between the two sensors. If you are also thinking, "damn, those are some rich AFR numbers for an NA engine" then I am right there with you. I need to find a local dyno shop that can work with the M1.7ECU to get things better dialed in. Either that, or MM did it on purpose so I could use 91 octane with their 11.5:1 compression build.

This next plot shows engine RPM vs time for four 2nd gear WOT pulls, two with each sensor. I unplugged the Motronic to reset it before doing the MAF pulls, and then unplugged it again to reset it after swapping the stock VAM setup in. This was to try to minimize any LTFT (long-term fuel trim) adaptations, although I don't know if those even effect the WOT maps. The runs were done on the same stretch of road within 40 minutes of each other. The <0.2 second variation between runs is probably just margin-of-error due to the driver (me) and slight road variations. VAM and MAF lines are all mixed in with no obvious difference.


Just for everyone's info, here is what a 2nd gear WOT pull looks like with a MAF. I think that it should be fairly self-explanatory why you cannot just hook a MAF up to the ECU. Throttle was 100% open, so don't worry about the fact that the log shows ~95%, it's just a calibration issue. The fuzzy yellow line is what comes out of the MAF sensor, and the blue line is my converter's output which mimics the VAM's output and goes to the ECU.


Here's a zoomed-in view of part of the plot. You can see that the MAF output is showing every intake valve opening once the throttle gets opened up a bit.


As I mentioned in a previous post, I was having issues with the RPM signal. It turns out that the physical installation of the RPM conditioner was just fine. The issues were in my logger's firmware, and I fixed them. Measuring RPM takes a lot more coding than measuring voltages! I cleaned-up the board installation anyway though since it wasn't really ideal to begin with.

[econti]

Do you think there'd be improvement with a MAP sensor?

[bmwman91]

Nope, I think that it is doubtful. The air metering part of the engine does not seem to pose any sort of meaningful restriction. A MAP (and a MAF) are both much faster sensors, so if you are running a more modern stand-alone ECU then you would see a big improvement in throttle response. With the stock Motronic though, it just isn't designed to take advantage of the faster sensors.

Also, I sort of forgot that in my original post's plots, the stock 1.8L engine air flow data was taken from the same MAF sensor when it was retrofitted onto a stock M42 air box. So, the big bump in air flow in the ~2500RPM area that I see on the 2.1L is ONLY there when using the stock AFM. It must either be some sort of resonance in the AFM's damper chamber, or the carbon track is messed up. It isn't the air box because I did not see the same bump in air flow when the MAF was hooked up to the stock air box on my old engine.

[gearheadE30]

Huh, very good info. Thanks for posting all of that.

You said that the M42 has a horribly unreliable timing case setup. I rebuilt mine awhile back, and while it is somewhat involved, there wasn't much that jumped out at me as being a terrible design aside from how the rails are retained (somewhat limited thread engagement in the chain guide retainer bolts) and how the complicated housing setup bolts to the front of the block. Granted, I've been a rather inactive member for a few years now, but what is so unreliable about it?

[bmwman91]

My beef with the M42's timing case is mainly with the deflector sprocket. The sprocket itself is prone to failure after 200k+ miles (and it is not something most people ever replace because it is stupid-expensive). When the bearing blows out, it introduces enough slack in the system for pistons and valves to make contact, maybe enough to bend valves, maybe not.

On top of that, the part of the timing case casting that it mounts to is also known to completely snap off. When that happens, there is a very good chance that you will need a complete head rebuild (bent valves) and maybe new pistons.

I had both happen to me in the span of about 1 year, and after the timing case casting broke I tossed the engine and had a Metric Mechanic 2.1L stroker built (was getting married, figured I should blow that $$$ while I still could lol). The "new" M42 has an M44 timing case, which uses a plastic deflector rail in place of the deflector sprocket. It also uses a slightly bigger set of oil pump rotors, which is a plus since it means more oil flow, but you need a custom bracket made to hold the crank position sensor. From what I understand, the 1994+ E36 M42 also uses an updated timing case without the deflector sprocket, and it is a direct bolt-in (same oil pump as the E30 M42, but also has a spot for the crank position sensor). I recommend swapping in that timing case to every E30 M42 owner.


The other gripe I have is with the driver's side guide rail. The top bolt that goes into the "adjustable" sleeve seems to have a tendency to strip the threads in the head, which will lead to premature wear-out of the guide rail and that top bolt rattling around inside the timing case (and in one member's case, wearing it's way THROUGH the upper cover lol).

[bmwman91]

Just to add more fun data in here, I had to SMOG test my car to renew the registration. This part of CA requires the whole "put your car up on a dyno and drive while sniffing the tailpipe" process. Anyway, I passed with flying colors, which is a relief since I was pretty sure that I would need to put a fresh cat on (and I usually need to every 4 years to pass SMOG). I had a new cat welded in 2 years ago, about 20 minutes before my last SMOG test, and I passed then with crazy-low numbers. Surprisingly, I passed this time with very low numbers as well. Apparently the Magnaflow cat that I have on there was recalled / lost its CA certification for some reason, but since I had it installed before that happened it is fine that it is on there. Maybe it is related to the fact that the cat rattles until it warms up.

Anyway, here are the test results. What a relief!

[stanz]

Thank you bmwman91, you are an m42 genius! Didn't you create a site dedicated to maintainence on the 318is?

[bmwman91]

Yeah, I have some articles at http://bmw.e30tuner.com but I have not really updated it in a number of years.

[stanz]

Yeah, I have some articles at http://bmw.e30tuner.com but I have not really updated it in a number of years.

So cool! I remember coming across this site years ago. I actually posted the link for someone asking how to repair their valve cover bolt holes. What a great resource. Thanks again for all your hard work and dedication to a wonderful car!

[bmwman91]

Hey no worries, I am just glad that people actually find my site to be useful!

[varg]

My beef with the M42's timing case is mainly with the deflector sprocket. The sprocket itself is prone to failure after 200k+ miles (and it is not something most people ever replace because it is stupid-expensive). When the bearing blows out, it introduces enough slack in the system for pistons and valves to make contact, maybe enough to bend valves, maybe not.

On top of that, the part of the timing case casting that it mounts to is also known to completely snap off. When that happens, there is a very good chance that you will need a complete head rebuild (bent valves) and maybe new pistons.

yep. This is only one of my beefs with the M42, and I'm afraid that I'm on borrowed time since my engine is probably around the 200k mark. When I had the front of the engine opened up I inspected the idler sprocket and its bearing, and all seemed fine. It spun smoothly, had little to no play, and had no visible signs of unusual wear. I could not afford to replace it at the time, and don't want to take the front of the engine apart again any time soon.

[bmwman91]

yep. This is only one of my beefs with the M42, and I'm afraid that I'm on borrowed time since my engine is probably around the 200k mark. When I had the front of the engine opened up I inspected the idler sprocket and its bearing, and all seemed fine. It spun smoothly, had little to no play, and had no visible signs of unusual wear. I could not afford to replace it at the time, and don't want to take the front of the engine apart again any time soon.

Yeah I know the feeling of not wanting to take it apart again. I had to fix crap in there once a year for 3 years in a row before the idler sprocket snapped right off of the timing case. That was the point at which I decided to look into the MM 2.1L stroker, an engine I had wanted since well before I was earning a paycheck! It was probably not the most responsible use of my money at the time, but since I was set to get married later that year I figured I would do it before I had to answer to anyone lol.

Anyway, should you ever want to do the timing case update, you should be able to get a used 1994+ M42 timing case for under $100 (I think it was 1994 when they changed to the plastic guide, I guess you would need to confirm that if you ever do the swap). It's just a lot of work, and while it is entirely possible to swap it without removing the head, it is a stressful hassle. You could always look into an M44 timing case too since it uses slightly larger oil pump rotors (more flow), but you would need a custom bracket to hold the crank position sensor and that would run you at least $100 from Metric Mechanic.