MPS Autopsy

[mkinne]

A few months ago the MPS failed in my '73 1.7L. Took a long time to figure out which was the problem part. Checked / traded / replaced everything on the engine. Nothing worked until I dropped in a different MPS (from a '73 2.0L). Started instantly. The 1.7 MPS checked out fine (holds vac, ohms OK), the car simply would not run with it. I've been driving 'round with that 2.0 MPS ever since. Runs OK, but rich. Worst is on hot days. Idle drops very low, engine blubbers.

Anyhow... Last weekend I completely took apart the old 1.7 MPS. Cut the rivets and split the shells. Spotless inside. Digging out the potting on the adjustment screw cap was a b!*@h. Removed the solder joints and pulled the core. Interesting. Also spotless. Nothing obviously wrong so I put it back together. RE-solder joints. Tap rivet holes and use screws to close it up. Might as well see if it works. Put it back in the car. Nope, still wont run. But now I have access to the adjustment screw. 1/4 turn one way, no good. 1/4 turn the other way, hey! It almost ran! Another 1/4 turn and the car is idleing on its own.

So here are the questions:
1. Any ideas how the MPS could have gone out of adjustment?
2. How can I set the MPS correctly? (think I saw an article somewhere, can't find it now)

For now, I am still running the 2.0 MPS.

TIA
Matt

[pbanders]

I've done some fairly extensive analysis of the MPS, I've disassembled a number of them and have developed some mechanically-based numbers for initial adjustments for the 2.0L MPS.

I cannot think of a way that the MPS would go out of adjustment. The two adjusting screws are tightly held by plastic sections in the threads to keep the unit vacuum tight.

It might be possible that something else in your FI system went out of adjustement or failed. The 1973 MPS is a bit of an unusual unit, designed to work with 022.906.021E ECU and the 022.906.041A head temperature sensor with the 270 ohm ballast resistor. In comparison to your car's original pressure sensor, a 022.906.051E, it is probably adjusted richer. BTW, for 914's the only difference between the pressure sensors is the adjustment, the internals are all the same (as related to me by the owner of Bret Instruments, the rebuilder).

If a richer sensor made your car run, maybe something on your car has caused it to go lean. Do you remember which way you turned the inner adjustement screw? Clockwise makes the car leaner, counter-clockwise makes it richer.

Also, you may not have noticed, but there are two adjustment screws. The inner screw (the obvious one) is actually screwed into the middle of a second screw, which screws into the flange of the full-load diaphragm. It is very important to hold this outer screw and keep it from turning when adjusting the inner screw. Otherwise, you will change the amount of full-load enrichment. Additionally, the plug you removed is also an adjuster. It is a mechanical stop that limits the movement of the full-load diaphragm. To readjust it to the correct setting, use a veriner caliper depth probe and measure the depth from the top of the flange where the plug screws in to the shoulder on the outer adjusting screw. Now measure the thickness of the plug. Screw the plug back in so that it pushes down on the shoulder of the outer adjusting screw by 0.2 mm. This will prevent the diaphragm from quick failure due to stress.

Brad Anders

[mkinne]

Wow! Thanx for all the info Brad. Had no idea there was more than one adjustment. I'll take a closer look soon as possible.

Can't remember which way I turned the screw. Should be a ble to figure out with a little fiddling.

I realy think the MPS was the culprit. No other part made any difference at all.

How about that sealed double can inside the MPS? Vac assist or temperature compensation? They seem to be still sealed, but one can appears slightly collapsed compared to the other. Any dimensional info available at some temperature / pressure?

The plug may not be useful as an adjuster any more. Threads were pretty messed up by removal. How critical is it?

Matt

[pbanders]

As you said, it may be that one of the bellows failed, and that's why readjusting the unit brought it back to an idle condition. The bellows convert the vacuum level to an electrical signal by expanding under vacuum and moving the linear transformer core (which changes the mutual inductandce and modulates the basic injection quantity calculated by the ECU). But if one of the two bellows has failed, the unit is no good, as you need both bellows to cover the full range of part-load operation. You can check the bellows by making your own vacuum chamber. I used a Klausen pickle jar. I cut a hole in the lid, and put a sealed fitting on the jar. Use a Mityvac or other hand-held vacuum pump to evacuate. You should get about 2.5 mm of expansion when you go from atmospheric pressure down to 15 in.Hg of vacuum. If you only get 1.2 mm, one of the bellows has failed.

The full-load stop is critical for long-term operation. Every time you floor it, the full-load diaphragm is activated. Eventually, even with the mechanical stop, the flexing cracks the diaphraghm out around the flange, and the unit leaks. Without the stop, this will happen a lot sooner.

I'd suggest that you package it all up as good as you can, and exchange it as a core for a rebuilt unit. With a possible failed bellows and a stripped full-load stop, it's ready for retirement.

I think "autopsy" was the right word...

Brad Anders


[This message has been edited by pbanders (edited 04-28-2001).]

[mkinne]

Brad, thanks for the extra info.

Did some more playing today. Pretty sure I turned the screw in 1/2 turn in (CW) before. If I turn the screw farther in (about 1 more turn) the engine begins to hunt. Another 1/2 turn and it dies. Turn the screw out (CCW) and the idle slowly drops until the engine dies. About 2-1/2 turns from dead to dead. Runs nice in the middle. Drove up and down the street. Seems normal. All of this was done with the engine already warm.

Pulled the 1.7 MPS out to work on the bench. (I'll keep using the 2.0 for now.) Yes, the second screw is turning with the middle screw. I'll have to make a special tool to turn them seperately.

Question still stands... Is there any way to properly bench set the MPS?

New question... Wher can I exchange the core for a rebuilt?

Thanks

matt

[pbanders]

Matt, if you turned the screw in a full 1/2 turn to get the mixture right for idle, I'll bet you're right and one of the aneroid cells has failed. Adjustments to this unit are on the order of 1/8 turn or less. A "flat" cell lets the core move into the coil deeper, causing the mixture to go rich. Turning the adjuster in leans it out, which is what you've done. But if the cell is dead, you have very limited part-load operation. The engine is likely to go very flat (lean) once you get to half-load and beyond. Full-load will also be too lean. Did you try any moderate to full load operation?

You really don't need a special tool to hold the outer screw - just take a second small screwdriver and use it to hold the screw.

There is no simple way to bench calibrate the MPS. I took a couple of the 2.0 MPS's apart and carefully noted the mechanical setup of the inner and outer screws, by measuring their positions with a vernier caliper. But that only provides a starting point. A bench setup replicates the D-Jetronic system, including the ECU. You measure the injector pulse widths at idle, part load, and full load, and adjust the MPS to the values from "the table". Except that Bosch won't release "the table" with all the values for all the D-Jet setups. The rebuilder (Bret Instruments) reverse-engineered the process by buying a collection of new Bosch units, and measured the pulse widths of each of these new units. He's not giving away his table, either!!

You can do rebuilt exchange with Automotion and other 914 aftermarket places. I don't know if Pelican sells rebuilts, but if they do, go there first!!!

Brad Anders

[Blu Flag Chicane]

Brad, nice work on the MPS research.

Do you have any references on the MPS' operation? I'm interested in locating a more generic part that might be made to work with the D-jet.

I was about to start researching this but it looks like you probably already know more about how it operates than I'd be able to figure out in a few months.

-BFC


[This message has been edited by Blu Flag Chicane (edited 04-30-2001).]

[mkinne]

Brad -

Outer screw looks like it takes an allen wrench to turn, but inner screw is in the way. Would have to be a hollow allen wrench to allow adjustment of both with MPS assembled. Does that sound right?

Matt

[pbanders]

Originally posted by Blu Flag Chicane:
Brad, nice work on the MPS research.

Do you have any references on the MPS' operation? I'm interested in locating a more generic part that might be made to work with the D-jet.

I was about to start researching this but it looks like you probably already know more about how it operates than I'd be able to figure out in a few months.

-BFC


[This message has been edited by Blu Flag Chicane (edited 04-30-2001).]

Blu, below is the best reference you will find that describes the MPS.

http://www.conservatory.com/vw/manuals_nyx/EFI-DL.html

This is a Bosch reference with excellent diagrams and description of the operation of the MPS at part-load and full-load. I also have a page at:

http://members.nbci.com/zubeneschemali/MPS.HTM

Please view the information on my page with some skepticism - it's my analysis of the MPS and I'm sure others out there would disagree with my conclusions. Additionally, I'm not sure how long it will remain up on nbci.com - and it takes a while to load.

As for replacing the MPS with a simpler device, such as a solid-state MAP sensor, I think that's unlikely. The operation principle of the MPS is based on mutual inductance, which is very hard to duplicate through any other kind of circuitry you might interpose between a MAP sensor and the ECU. You would be much better off replacing the entire D-Jet system with an aftermarket FI system or full engine management system. SDI systems model EM-3 has been successfully used for replacing D-Jet and it will cost about $1K. Search this forum for the term "SDI", there was a guy who recently posted who has done the conversion.

Brad Anders

[pbanders]

Originally posted by mkinne:
Brad -

Outer screw looks like it takes an allen wrench to turn, but inner screw is in the way. Would have to be a hollow allen wrench to allow adjustment of both with MPS assembled. Does that sound right?

Matt

Just about. It's actually a hollow star socket, something like 7 mm. I looked around through some tool catalogs and wasn't able to find anything like it. You could probably make a tool yourself with a piece of aluminum or steel tube - or plastic, for that matter. It doesn't take much torque to turn it.

Or, you could get some smaller screwdrivers! I haven't had much trouble using screwdrivers, and the rebuilder told me that's the way he does it.

Brad Anders