The MFI Diaries, Part II: Triumph (NOT the LBC)

[304065]

I have triumphed over ignorance, fear and apathy! I am the owner of a correctly tuned MFI 2,2 E!

Many of you may recall the sad tale of how I attempted to set throttle correlation for the MFI stacks by using a feeler gauge, to no success. The MFI Diaries: How NOT to set Correlation

The resultant 3000 rpm idle left me convinced that the throttle plates were just too far open. So, heeding the advice of Grady and others here, I backed the adjusting screws out all the way, closing the throttle plates.

Side note to Bosch. The throttle plate adjusting screws on the port-side stack are IMPOSSIBLE to adjust using conventional tools. I tried a 9mm crows-foot wrench to no avail. Finally I resorted to unbolting the stack (six 10mm nuts, don't let the washers fall down in the engine!) and using a deep socket to undo the lock nut. With the nuts loose, a long standard screwdriver could be used to adjust the plates to full closed.

Now, here's where the pundits will chime in and say that you don't want the plates full closed, because they will stick when the engine heats up. To which I say, I haven't experienced that yet, and given the wear ridge in the throttle body, I don't think there's any chance of an intereference fit. Maybe if the stacks were new. Perhaps there are other air leaks (I started with the airbleed screws full CLOSED) that make it impossible for this engine to idle correctly with the plates anything other than full closed.

So, with the port-side stack reassembled (I swear, having done this a couple times it only took me about eight minutes to R&R) I started the engine with all the throttle rods disconnected. It was a little tricky, insofar as the vibration tends to make the plates open a little, so that when you set correlation, you have to make sure that they're all closed, or the idle will jump around, throwing off the readings. Anyway, at around 1500 RPM (the thermostat rod was still cold, and the idle mixture still very rich) I was able to get all the stacks pulling about a "7" on the synchrometer, with only very minor tweaks to the airbleed screws.

It was then very simple to adjust all the throttle rods to the precise length necessary to avoid disturbing the idle. The throttle links needed to be made longer than 145mm, but I made sure they were the same length by comparing them before I put them on. The MFI rod was exactly 114.0mm, and marked with orange paint.

The result? Smooth, even idle, no pops, no stumbling. And the best part? The MFI thermostat even works correctly-- once the engine was showing an oil temp of about 165F, I would rev the engine and the idle would jump up to about 1500 for a few seconds, then drop back down to 900 where it belongs. Pure magic.

Then it was time to check the AFR. With the thermostat cold, idle mixture was around 12.8, which corresponds to a CO of about 4.6%, not bad. When the thermostat heated up and the rod kicked in, the mixture leaned WAY out, to about 16, which is lean of stochiometric, throwing NOx instead of CO I think. That will probably require an idle mixture adjustment.

With the engine revved to 5000 (NOT under load, I know, it's not good for it) the AFR was around 11, or about 9.4% CO.

The old rule of thumb is 5-6%, 7-8% at 3000, but this is a good starting point. What matters to me is the AFR under LOAD. . . and I've got the RPM module for the LM-1 to help me figure that out.

All that remains is to tighten down the throttle rods and mark them, and also tighten up the locknuts on the throttle stops.

[Jandrews]

Great work, John!

So, what would you say was the critical findings? Rod length? Throttle plate adjustment? MFI pump calibration?

Thanks for the write-up.


JA

[304065]

The key to making it idle correctly was throttle correlation. The throttle plates were set too far open, and weren't adjusted properly relative to one another, for progress to be made. The difficulty of adjusting the port-side stack encouraged prior owners to be lazy about setting correlation properly.

The rod length is from the book, not much skill involved in setting it right. The MFI pump still may need a couple of clicks, but with the LM-1, adjusting it is a matter of science, not a guessing game. That will make the biggest difference I believe.