Ignition advance

[duxthe1]

I'm spit balling another idea out here. I'm wanting to advance the ignition timing on my 104.980 engine. Here's why.

Any one who has ever owned a 104.980 knows that its low end is soggy at best. Even after the mods to mine, its WOT off idle torque leaves a lot to be desired. If you tip into the throttle slowly it has plenty of torque to pull off in 2nd gear but if you floor it, it can't hardly move until the downshift to 1st.

I've been ruminating over this for a while and think I've figured out the root of it. The 104.980 is a very low vacuum engine. By watching the datalogs of my transmission controller I can tell that by 15 degrees throttle angle the intake manifold is at atmospheric pressure. The primary load signal into the EZL is engine vacuum so by 15 degrees throttle angle the engine is at full load in the EZL's eyes.

I postulate that is the primary culprit to the 104.980's soggy bottom When you just tip into the throttle, enough vacuum is maintained to deliver significantly more advance, resulting in abundant torque. When you floor it vacuum is dumped and along with it, the ignition advance. To test this theory I put a restrictor in line with the Ign module and added a small vacuum reservoir on the module side of the restrictor. This holds slightly more vacuum on the module after you go WOT for a short period. It definitely lessened the "soggy bottom" and I can feel the loss of power when the vacuum equalizes. Proves the theory, but isn't a real solution.

A real solution would be a fully programmable ignition with multiple load vectors, dyno tuned. Barring that, advancing the entire curve a few, up to several degrees. This is what I'm thinking. Here at ~5000 ft its not uncommon to advance the older distributor cars 5 degress for the altitude. I'd like to advance my 104.980 about 4 degrees. I feel that would give me enough extra to get some off idle torque without being too much timing on the high end. It is a knock sensor motor but I'd prefer to not be on the ragged edge as the combustion chamber is quite efficient in the 104 engines and don't require lots of timing to begin with.

So advancing the timing on these engines is no simple feat with the fixed trigger. My plan is to separate the flex plate and ring gear and reposition it in a slightly advanced position. I have a couple of 103 flex plates lying around but the parts catalog shows a different 104 part number for the 104.980 and I'm unsure of the difference. Speculating, I'd guess a different size and position of the timing vanes. I'll have the trans coming out of the 91 CE (hopefully soon) and will pull the flywheel with it. Meanwhile I'm going to mod a 103 flywheel for 4 degrees advance. I'll compare the two when I have both in hand and make the choice to use it or mod the 104 part.

[Frank Reiner]

d:

Consider a three tooth trigger wheel at the front of the crankshaft, with a VR in an adjustable bracket.

[duxthe1]

That's a great idea but space is at a premium down there. However even a makeshift setup can allow me to test differing amounts of advance. That way I could mod a flywheel to the best performing amount of advance instead of committing to an educated guess. I am planning to install a 722.6 into this 90CE so a flywheel swap in the relatively near future isn't going to be a major inconvenience.

[Frank Reiner]

d:

Once committed, the flywheel triggers could be positioned in excess advance and a delay device such as sold by MSD (that is adjustable) can be used to provide the signal at the desired time.

[97 SL320]

A fish tank air line adjuster would work. You could even use long hoses to the interior to allow for testing. It probably would be a good idea to use a check valve parallel to the restriction to allow the EZL to see vacuum more rapidly. In the 70'S / 80's, Ford used various inline restrictors with check valves but finding one today may prove difficult.

How about a switch over valve from the EGR / AIR pump circuit hooked to a timer / RPM switch?

[duxthe1]

I pulled the 722.6 out of the red coupe and the flywheel along with it today. I found the difference in the 103 and 104 flywheel. The 104 part has a double thick center section! As far as I can tell the vanes are in the same position and are the same size.

As great of an idea as the front mount setup was, there just isn't room without some significant engineering. I'm going to commit 4 degrees advance to this flywheel by cutting 9.2mm from the trailing edge of the vanes. A bit of testing indicated that the coil fires with the falling edge of the crank sensor pulse so a trim of the trailing edge should do the trick.

I would have installed the 722.6 and flywheel in the blue coupe today but Windows 10 is such garbage that I couldn't get the laptop to communicate with the TCU. Without being able to set up the TCU I stopped short of an install. The TCU is wired in just to the point of not interfering with any current systems.

I did do a bit of testing for a baseline since the trans didn't swap. Currently the engine idles hot at 7° advance with 12" of vacuum. I'm really curious if the engine will show any increase in idle vacuum when the timing is advanced. If it does, as I'd expect it should, it may cause a cumulative effect where the additional advance creates enough additional vacuum that the curve is kept slightly more advanced too. I wouldn't expect a night and day difference but in such an inherently low vacuum engine, every little bit should help.

[duxthe1]

Here's a pic of the flywheel in the process of being modded. The notched section is now completely removed and should provide 4 degrees of static advance to the entire ignition curve.

[whipplem104]

well I have to say I am curious on how this comes out. I have always wondered why the teeth were so big on these. I had assumed that is was poor sensor technology from when they were designed.
Great thinking outside the box.

[duxthe1]

Thanks. When I saw you replied to this thread I figured you were going to hit me over the head and tell me that the 104 doesn't like a lot of timing.

I just really feel that this motor wants a few more degrees. If it ends up being too much, at least it isn't a big deal to pull the 722.6 back out.

Going back to the soggy bottom end torque, I've got yet another lead. I've set up my transmission controller this week and started to use the airflow potentiometer signal for a load input. It was quite illuminating to see a datalog of this signal. From a standstill when you punch the throttle you can see the airflow signal spikes pretty big. The ecu sees this spike and signals enrichment for the increase in airflow. What follows this spike is a very quick negative spike (as much as half the positive spike) as airflow and control pressure find its equilibrium. I believe that this negative spike is causing the ecu to pull fuel when it needs its most. The fix for this could get quite involved. I am going to try to play with the system and control pressures to see if I can tune this behavior out of the system. My big advantage now is being able to see it in the datalogs. Of course I'm not going to do anything until I get the modded flywheel in.

[Frank Reiner]

Originally Posted by duxthe1

Thanks. When I saw you replied to this thread I figured you were going to hit me over the head and tell me that the 104 doesn't like a lot of timing.

I just really feel that this motor wants a few more degrees. If it ends up being too much, at least it isn't a big deal to pull the 722.6 back out.

Going back to the soggy bottom end torque, I've got yet another lead. I've set up my transmission controller this week and started to use the airflow potentiometer signal for a load input. It was quite illuminating to see a datalog of this signal. From a standstill when you punch the throttle you can see the airflow signal spikes pretty big. The ecu sees this spike and signals enrichment for the increase in airflow. What follows this spike is a very quick negative spike (as much as half the positive spike) as airflow and control pressure find its equilibrium. I believe that this negative spike is causing the ecu to pull fuel when it needs its most. The fix for this could get quite involved. I am going to try to play with the system and control pressures to see if I can tune this behavior out of the system. My big advantage now is being able to see it in the datalogs. Of course I'm not going to do anything until I get the modded flywheel in.

Re: KE potentiometer signal

That pot is ignored if water temp is above 175F; in any event it only signals for a a rather brief enrichment. If the WOT switch on the throttle shaft is closed, enrichment is controlled by ECU open loop mapping via the EHA. In a KE control pressure is constant, as opposed to the variable control pressure of the two preceding versions and their external control pressure regulators (WUR).

Consider using manifold absolute pressure (MAP) as the primary load metric.

[duxthe1]

I'd previously used MAP and it left a lot to be desired. Between 15 and 20% throttle this engine reads atmospheric pressure on the manifold. That leaves 80% of the load range unaccounted for. The MAF spike wont affect the shift mapping and MAF is a much better indicator of load. I'm planning on keeping the MAP reading to use as a trim.

[97 SL320]

The FW notch length might control ignition dwell.

Where is your MAP sensor hooked to? If near the throttle body I'd expect very low vacuum.

[whipplem104]

His problem with vacuum is he is at 6000' in altitude. There is no air.
And I did not know I was so pessimistic about things that you would automatically assume I was going to disagree with you.
But just FYI they do not like a lot of timing? But a bit more than Mercedes ran them at. And at your altitude I am sure a few degrees more will help a lot. We used to do that all the time for the older cars that you could adjust them on when I lived there.

[duxthe1]

I wouldn't say pessimistic I just know you've spent significant time with a 104 on the dyno.

I was also thinking that notch length may control dwell. I doubt I've whacked off enough to be a problem.

[duxthe1]

I got the 722.6 transmission in tonight and along with it the modified flywheel. Confirmed with the timing light that I'm at 11 degrees at idle now so I did get the 4 degree advance I was looking for. Didn't even measure vacuum because I was pretty beat down by the time I had it running. I did notice the idle was higher than normal for a bit but corrected itself after only a bit of running.

Only drove 4 miles home so not much time on it yet. Good torque off idle, I'm anxious to see how it pulls off with the a/c on because stock these cars can barely even move with the load of the A/C compressor in a second gear start. I did notice when getting on it hard a power loss at 4K RPM which might be the knock sensors pulling timing. I'll admit I have the lowest octane available here in the tank and I'm not too surprised. I'll burn this tank down and fill up with the good stuff in a couple days. I say good stuff but up here our premium is 91 octane (85 in the tank now). What a rip, we pay premium price for what would be mid grade anywhere else.

I also have another cool bit of kit in the works. Expecting that I may have knock issues with the extra advance I have been planning on a water injection setup. I have a spare AdBlue dosing valve which cycles with a pulsewidth. It just so happens that I have one PWM output left in the PCS transmission controller and I have tested it and it will pulse the valve. I have an 80 PSI diaphragm pump ordered in also. I just need to plumb it up, which means making a few brass fittings and mounting everything. With everything plumbed I can set it up to inject based on the engine load signal from the airflow potentiometer. Should be pretty cool.