The MFI Diaries: How NOT to set Correlation

[304065]

As part of the Check Measure Adjust routine one of the steps is "Correlation." This means making sure all the throttle bodies are drawing the same amount of air.

In order to set correlation, you use a device called a synchrometer, which is basically an airflow meter.

My correlation was off significantly between stacks, and the #2 and #3 stacks are hard to adjust because the MFI pump is in the way. So I decided to set ALL the throttle plates identically.

How to do this? I looked at a web site for another throttle body vendor, which recommended using a feeler gauge. So after ordering replacement gaskets for the stacks from Pelican, I removed each stack and used an .005 feeler gauge to set the clearance between the edge of the throttle plate and the side of the stack.

After meticulously setting the clearance, I bolted everything back up, removed all the throttle rods, and fired her up.

Result? 3000 RPM idle! DOH! Correlation WAY off between stacks. That's with the airbleed screws all the way closed, too. The engine is just pulling too much air through with the throttles at the stops.

I guess .005 clearance is too much. Also, the throttle body is worn from the plate slapping shut, so it's probably impossible to set these with a physical tolerance.

Anybody have ideas for setting up a homemade flow bench? I could use my shop vac on one side of the TB and the synchrometer on the other.

Back to the drawing board. . .

[BK911]

I set all my throttle valves closed at idle, not the .005" you used. Then I adjust the idle screws to syncronize the airflow into each intake using a syncrometer.

[Grady Clay]

John,

Yes, a shop vac will work just fine. You need the kind where the motor doesn’t use the vacuum air flow to keep it cool. You should devise a plenum where you can measure and control the low-pressure side. The flow is easy to measure with a commercial syncrometer (Pelican) or devise an adjustable restriction and manometer like the Porsche Factory tool.

These work equally well setting up MFI throttle bodies, carbs, and EFI throttle bodies.

You might consider putting the shop vac very remote as they are noisy and obnoxious to work around.

I had two nice quiet “vacuum pumps” designed to keep the tension loop in high speed ¾” tape data recorders at 50,000 feet (USAF surplus.) I had an absolute reading Mercury barometer to measure the low pressure side and a simple air bypass to adjust it for daily atmospheric conditions. I have the Factory P-235 that is now 35 years old.

Best,
Grady

[Zeke]

Why do you need a flow bench? If you get the right tool, you can adjust the butterfly on the left bank forward with everything in place excpet the air filter housing.

I agree .005 is too much. The possible ridge in the TB's you mentioned is one possibility and worn shafts are another. In CM&A, I think it says that if you have to use more than 8 half turns on the air correction screws, to go back to the butterflys. I did the butterflys untill all were pulling about 7 on the Synometer and then dialed in the rest with the screws. I started with the screws all screwed out a little and the same so I could go either way when making final adjustments.

But, the various linkages have to be "at rest" when the throttle is at idle to get even flow from one throttle action to the next. With worn ball joints, this can be a problem. The whole damn thing has to be near perfect to run and idle well consistantly. When I started with mine, I found an injector that was not fully tight and allowing air past the threads.

There is probably nothing more satisfying than to overcome an MFI system and get the car to work well.

I think PMO has the ball joints.

EDIT: While I was composing my post, the others posted. If Grady says you need a flow bench device, then you need one. I don't think anyone knows these MFI's better.

[Grady Clay]

Milt,

No, a flow bench isn’t necessary. It is just one more way to do a job better and it is part of the fun. Regardless of a flow bench setting, you still need to finally set the flow and linkages on the running engine.

You bring up an important point. Even the smallest thing (not seated nozzle) can dramatically affect the settings. This is why everyone stresses CMA and extensions so much. The basics of the engine have to be near perfect for proper final adjustment.

I don’t see the MFI as something to “overcome” it is a great mechanical computer system from the 1930s. When all is OK, it is the most responsive and high performance for any early 911.

Best,
Grady

[jluetjen]

John; I'm just checking -- you did use a synchrometer -- right???

[304065]

Thanks for the advice fellas.

The concern I have over just setting the throttle plates closed is that with thermal expansion, some of them might get stuck.

Grady, why don't I want the exhaust air going over the vac's motor? I assume when the vacuum is pulled there's not enough airflow to keep the motor cool and it burns out?

John, I am using a Synchrometer from Pelican, that looks like this:



There are electronic synchronizers out there, and I could always hook up vacuum gauges to the old vacuum ports on my throttle bodies (these are plugged by screws) but I'll give this a try first.

[Grady Clay]

John,

The throttle butterflies are not likely to get stuck where the throttle pedal won’t open them. Your right foot will tell if there is a problem. With very worn throttle bodies, the engine may take a few seconds to settle back to its idle position after the throttle is closed. How far disassembled did you have your throttle bodies? If you had the butterfly plates off the shafts, there is a real art to getting them back in place.


Yes, it is keeping the shop vac electric motor cool. Most have independent cooling fans, some older ones may not.


So everyone knows; a syncrometer (for carbs, MFI and EFI throttle bodies) measures air flow while minimally restricting the air flow to a single cylinder. If you use a liquid manometer, you must use a liquid that won’t damage the engine if it somehow gets sucked in (Mercury or brake fluid are not acceptable, solvent with food coloring is OK.) It is best to have one that has very slow response, particularly with wild (S+) cams.


Best,
Grady

[jluetjen]

Thanks John for clarifying. Have you checked with Matt Blast up at Eurometrix. He is quite clearly the guy when it comes to MFI throttle bodies. When he gets done with them the butterflies will fit the bores perfectly, the shafts will be tight, passages will be clean and everything will be as it should be. He's got some special tools, jigs, etc to get everything just so.

Before you spend a bunch of time and effort it most likely would be a good idea to at least talk with him and consider just sending him your TB's to be R&R'd.

[ed martin]

Setting the correlation on a bench can only get you in the "ball park". Bench adjustments will not account for the compression differential of each cylinder. You're supposed to set up the butterflies so they are fully closed and then back it off a half turn or three quarters, whatever the Check, measure, adjust manual says. You don't want them fully closed as Grady has already pointed out, because they will stick. I pulled my throttle bodies out, cleaned them and set them up as per manuals instruction and as a check I sighted the valve openings against a light to to see that they were more or less the same. Once installed you want to set the correlation the best you can by fine tuning the butterflies and then do the really fine tuning with the turn screws on the stacks. The syncrometer pictured above is probably superior to the simpler one that I have, as the one that I have gives you a different reading depending how much pressure you apply over the top of the stack. A mechanic I knew would just put his hand over the stack to guage the flow and could always get it just right.