MFI Pump - Open Heart Surgery

[jcge]

From the world of Mercedes Benz...cross section of a regulator



(C) Robert Bosch AG - for educational purposes only

Reglerfedern = control springs
Raumnocken = space cam

[jcge]

(C) Robert Bosch AG - for educational purposes only

[Walko]

Hopefully this threadwill help clear up some of the mystery behind the MFI pumps and the adjustments available

Michael

[jcge]

The 2 sectional images show the repositioned barometric cell and water thermostat / idle air control of Mercedes Benz applications.

On Mercedes - the barometric cell acts directly onto the compensating lever (on Porsche its via an intermediate shaft ), and the thermostat/IAC valve acts vertically (as opposed to horizontally on the Porsche)

Here's a photo of the MB pump



You probably recognise the regulator section in another form......RSR.

[jcge]

Barometric Cell test spec - 0.58 mm / 100mm Hg ( where 727mm Hg = 1 ATM = 14.5 psi) (no paper document reference...just verbals) I have no reason to doubt this but....all care..no responsibility etc etc

It's important to note that regardless of our collective findings of adjustments available within the MFI system, you still need to verify the fuel flow rates under different conditions on a pump dyno. There is very little you can adjust on a pump when fitted to an engine - only timing of the injection stroke, main rack (overall), CO (center screw) and +/- washers under the barometric cell & thermostat are about it......even then, these adjustments will only move the existing space cam around - you not actually changing the "hard" map ground on the space cam.

Taking a die grinder to the space cam ??? no way !!

John

[mobius911]

Subscribing. Great thread- help explains why also have my idle screw full lean and it's still not lean enough.

[jcge]

Here's a control diagram for the EP/RLA regulators. It provides a great visual explanation of the function of the regulator.


(C) Robert Bosch AG - for educational purposes only

A few (rough) translations might help

Raumnocken = Space Cam
Fliehkraftregler = Flyweight governor
Feder = Spring
Tastrolle = roller follower
Regellstange = Control rack
Gestanghebel = control (throttle) lever
Korretureinflusse = Correction influences (barometric cell and thermostats acting on compensation lever)

A note about the compensation lever (Korrektureinflusse) - the thermostat (not shown) and barometric cell (not shown) act on the opposite ends of the "T" lever (at the top of the image), resulting in a change or "compensation" in the actuation signal transmitted to the rack. +/- some amount depending on the prevailing control conditions. The compensation could take almost any mechanical form that delivers a linear change proportional to the control variable - an example is the replacement of the barometric cell with the boost pressure compensator on the 935 to achieve a turbo combination. I assume that in racing applications, the "temperature" end of the lever is simply held in the fixed "warmed up" position. (see Henry's famous warm up screw mod/enrichment device, or the top of an RSR pump for reference)

Also, note that the spring shown between the flyweights (holding them "together") is analogous to the spring/screw set of the capsule (low speed/ high speed and idle/CO) controlling the linear position of the space cam with changes in shaft speed.

The spring shown pushing the space cam towards the follower is the very light spring behind the space cam - its only purpose is to keep the cam against the governor flyweights.

Again - these are just my observations/interpretations - please don't take them as gospel.

[Jeff Higgins]

Well jcge, even if your conclusions are not "gospel", I arrived at the same ones looking at that diagram.

pjh69911, I'm thinking the lower part load is still controlled by the center "idle" spring. I think the outer, intermediate weight spring is the one I will make my initial adjustments with. With the main rack set where it is right now, the A/F shows a gradually leaning condition as rpm's climb over 4,000 and approach 6,000 (as high as I wanted to go on the new motor). So the current plan is to lean out the black screws that control that outer intermediate spring.

In the meantime, I've come up with an easy way to do this quickly on the dyno. Yes, the oil does drain out of the mixture screw cover; that o-ring only seals the cover. I found it is very easy to simply remove the upper right governor assy cover (the big one forward of the warm up thermostat, with the three screws). I simply snake a small rubber hose in past the flyweights and use the same pump I use for putting gear oil in the tranny to suck the oil out.

The two screws holding the small adjustment screw cover in place can be removed with a small, flat, ratcheting screwdriver. The one on the right will not come out all of the way before hitting the fan shroud, so I shortened it about 1/8" so that it would clear (actually had to loosen and raise the pump to get it out the first time). The long screwdriver normally used for the part load adjustment works great for poking through the fan and shroud to turn any and all of the five screws that I may need to turn.

Anyway, I'm "staged" for my next dyno session. I can drain the oil from the pump and easily access the screws. I'm going to try to get in later this week. I've been working on some of those other "little problems" inherent in putting a 3.0 in an early car and putting MFI on it to boot (non functioning tach, alternator/fan/shroud offset, etc. etc...). Everything found in the initial "shake down" runs has been addressed. Now it's back to the dyno...

[356RS]

This thread is full of fantastic information on the MFI pump. I'm in the process of building a short stroke MFI engine and all this information will help on the tuning end of the project. Thank you Jeff, John, Ed, and the rest of you for sharing this information.

[304065]

See also this thread:

Very technical MFI question....

[304065]

Originally Posted by jcge

Barometric Cell test spec - 0.58 mm / 100mm Hg ( where 727mm Hg = 1 ATM = 14.5 psi) (no paper document reference...just verbals) I have no reason to doubt this but....all care..no responsibility etc etc

Hmm. . . 100mm hg =1.93367747 psi. So you're saying that when the pressure drops, the elements inside expand, pushing the pin further out of the cell. This makes sense to me, insofar as if you remove the cell entirely, the engine runs excessively rich. So when you climb in altitude the pressure drops and the pin extends, leaning out the mixture.

[jcge]

Originally Posted by john_cramer

Hmm. . . 100mm hg =1.93367747 psi. So you're saying that when the pressure drops, the elements inside expand, pushing the pin further out of the cell. This makes sense to me, insofar as if you remove the cell entirely, the engine runs excessively rich. So when you climb in altitude the pressure drops and the pin extends, leaning out the mixture.

John - Correct - consider the function of both control elements acting on the correction lever - the thermostat and barometric cell. (What i've called the correction lever is not strictly true - the "real" correction lever has the elongated hole and is actuated by the short "H link" - but you get the idea.)

Thermostat - when it's cold (richer mixture required), the lever is at its highest position. When the engine's warm, the thermo discs have expanded and pushed (via the small 90 degree cam) the correction lever down until it bottoms out its end of the lever (leaner)- . The critical setting here is where the compensation bottoms out - i.e. define a "warm" engine. If the compensation did not stop at some point (bottoming out) a hot engine fed leaner mixture gets hotter until some major damage is caused.

Barometric cell - at higher atmospheric pressure eg sea level (more 02 to burn) a richer mixture will be required - so at increased atmospheric pressure, the pin in the barometric cell retracts 0.58mm / 100mm Hg of barometric pressure increase. Hence the rich mixture when you remove the cell from the unit. (do a web search on "aneroid barometer" or "aneroid cell" to see how it works inside) The critical setting here is for 1 standard atmosphere - achieved by washers under the baro cell.

Several engine builders of repute have replaced the baro cell and/or thermo units with simple set screw devices to meet the requirements of their particular application(check out Henry's 2.8SS motor for example - no baro cell Short Stroke 2.8 , and RSR applications - no thermo compensation)

Pushing the correction lever down at either end will result in a leaner mixture. The various distance ratios (lever rule) determine the total degree and proportion of compensation signal ultimately delivered to the rack. So the thermo and baro compensation base or "standard" settings must be set together during a pump rebuild, to ensure their correct relativity. (refer to dimension "c" in VDT-WJP 711/1 B Suppl. 2 Ed. 2 Figure 1 http://www.pelicanparts.com/techarticles/911_MFI/Bosch%20_More_MFI_Repair_Instructions.pdf)

It shouldn't be too hard to map the actual delivered values for each control signal - just need to put a dial gauge on the end of the rack and replace the control in question with a screw and dial guage and create an XY plot - the "transfer function" of the control element.....(hours of fun for all the family)

A logical extension of this mechanical arrangement could allow for 3 compensation variables (as opposed to 2 used by Bosch in this application), by using a triangular plate with articulated pin in the center - each control acting on the 3 corners of the triangle, and the sum of the inputs delivering compensation at the articulated pin (by superposition) ...but lets learn to walk before we run...and stick with what Bosch gave us !!

Regards
John

[David E. Clark]

I just got back into town and had a chance to look more carefully through my resources. These pages give some additional insight about adjusting the internal partial load adjusting screws (especially the first two pages). Note the statement "no adjustment is generally provided in the field for this operating mode on Porsche systems" on page 15. It appears that Porsche thought that these adjustments were beyond the capability of most repair shops (and certainly a backyard mechanic) so they never included this information in the Factory Service Manual or CMA. Apparently, Bosch has discussed these adjustments in their technical manuals, especially MFI Pump Test Instructions VDT-WPP 711/1B (ref Section 3.0 Inspection & Repair), but this instruction is next to impossible to find!





© Harold T. Glenn, Glen's Diesel and Gasoline Fuel-Injection Manual, pages 15-18 (1973). Posted here for educational purposes only.

[jcge]

David - interesting reference - I didn't know Alfa Romeo's used MFI..

Regards,

John

[David E. Clark]

The BMW 2002tii also used MFI on the early 1970's models. They used Kugelfisher pumps and apparently Gus Pfister is about the only guy who can fix those pumps too!

A scan of the Kugelfisher manual is available here.

[304065]

Alfas used SPICA, a form of MFI.

[jcge]

Interesting John - compare the toothed drive wheel to the Porsche "baby" pump on ebay....

[Jeff Higgins]

I just had another thought. Maybe I should buy an LM-1 instead of more dyno time. Since my base motor is an '83 3.0, it requires the thick flange SSI's. Those come by default with the O2 sensor bung in the left collector, which I plugged, so it would be no trouble at all to hook one up. With dyno time running at $150/hour, I could pay for the thing on one Saturday morning drive.

[304065]

Jeff,

They tell you different things. I personally think the LM-1 is a required, not optional, accessory for tuning MFI on the car. Between that and the data acquisition box I have enough to do my own dyno runs on a flat road at full throttle under load-- you can get WOT AFR graphs against the RPM and log the whole thing to see the differnce.

[Jeff Higgins]

Thanks John; that settles it. LM-1 on the way...