MPS- delving into it's secrets?, Update January 3, 2011 Options

[jk76.914]

I collected a number of MPS over a couple of years. Most held vacuum, some did not. I measured inductance vs. vacuum on those that I could, but then took all of them apart (except one brand new one).

The is a lot of variety in the diaphrams, probably because they get replaced when virtually every MPS is rebuilt. The photo below shows four diaphrams. I'll follow with descriptions and observations...



#1- copper, 2 pleats. This came from a rebuilt MPS that was in a pond or river for some time. There was actually sand in it, along with tiny aquatic snail shells of some sort, and the copper was pretty much green. This diaphram was ruptured. The workmanship is very good, and the alloy is idential to stock (see chart below). The threaded bushing looks like stock in design and attachment.

#2- copper, 3 pleats. This is a stock diaphram from a riveted MPS. It is ruptured.

#3- brass, 2 pleats. This is from a rebuilt MPS. I have three of these diaphrams, all similar. This MPS was freshly rebuilt in a rebuilder's box. Looked like new with fresh paint and plastic cap on vacuum port, but it leaked slightly. You can see why when you look at the ripples in that flange. The O.D. looks like it was cut out by hand with tin snips, and the threaded bushing was taken from a stock diaphram and soldered by hand into place. At the bottom in the picture, you can see that the flange is pretty smooth- that's where I tapped it out with a machinist's hammer on an anvil. I have no doubt that the leakage was from around the O.D. of the diaphram, and that tapping it out would probably fix it. This material (brass, see chart below) is stiffer and thicker than any of the rest. It would add quite a bit of spring tension to the mass-spring-damper system. In general, I rate the workmanship on this as "crappy", though maybe you could get a running car out of it.

#4- stainless steel, 3 pleats. This one is a bit of an enigma. It is identical to stock except that it is made from stainless (see chart below). The workmanship is perfect. It is from a rebuilt MPS. Interestingly, it is the MOST COMPLIANT of all of them, while most steel ones are reported to be stiffer. The other part of the enigma- the aneroid cells from this MPS are also stainless- same alloy. My suspicion is that this is a late Bosch rebuilt. Who else would make stainless cells, because the cells don't fail very often, so there are lots of spares available out there, and the cells are pretty complicated to make...


*** EDITED *** Materials. I measured alloy composition at work using an X-Ray Fluorescence analyzer. This machine is very precise, but it has limits to the range of elements that it can detect. Unfortunately, it cannot detect Beryllium (Be), and it is likely that the stock (at least) diaphram contains Be to harden and strengthen it. You can see a couple of things here though- the stainless diaphram and stainless cells are of exactly the same composition- nickel-chromium stainless steel. The nickel explains why they are slightly magnetic, as many stainless alloys are not. ***



Conclusions (really opinions) - The diaphram was put in there (early VW D-jets did not have one) to provide altitude compensation and to soften transition from leaner (high vacuum) to richer (low vacuum) regions. Early D-Jets had a separate unit with a diaphram and a switch to inform the ECU to richen the mixture at low vacuum. Cost reduction may have been a third reason- eliminating the separate unit, wiring, vacuum hose, etc. Anyway, this switch was either on or off, no soft transition. The MPS had only 2 aneroid cells and no diaphram, and its inductance curve was essentially a straight line from 0" to 25". There was no mixture compensation for altitude with this arrangement.

I'm thinking that the lower the stiffness of the diaphram, the more consistency in the setup and responsiveness, while both of these objectives are met. By maximizing compliance of the diaphram, the springs acting in the system are mainly the coil spring and the leaf springs that act to locate the armature.

My stainless diaphram is the most compliant, but there is another feature of steel (if I remember correctly) that adds to the argument that this is a late Bosch design- steel has a much higher Youngs modulus than copper. The higher the Youngs modulus, the greater the fatigue resistance, and the vastly most common failure mode of the MPS is fatigue failure. Could this have been a Bosch attempt to solve a reliability problem, even as the technology was being superceded by more modern ones? Since the cells are also subject to fatigue, would they have switched them over at the same time?

I am planning on assembling my own personal MPS using the stainless parts, and seeing how close I can tune it to my engine.

There are lots of other parts in the MPS that I've formed opinions about, but I'll hold off for now.

These are my own opinions, which may not be popular, so BLAST AWAY!!

(eye candy below)

[pbanders]

Wow, really great info! With your permission and credit to you, can I use some of the photos and data on my MPS page (see link in sig)?

You mention the full-load diaphragm provides altitude compensation. This is only true under heavy and full-load conditions. The only MPS's that provided part-load altitude compensation were the 0 280 100 100 series that was used on the MB's. These units lacked the full-load diaphragm (full-load was sensed by the throttle switch) and had an altitude compensation cell that was connected to the atmosphere through a holow screw.

I agree with your observation that the lower the "stiffness", or spring constant of the diaphragm, the less effect it will have on the transition characteristic. For any material used, the thickness and pleat design will vary. Reliability and resistance to fatigue failure is also an important consideration in the material selection.

Crafting the aneroid cells out of a different material and/or a different pleat design would be tricky. The issue here is to identically duplicate the pressure vs. displacement characteristic of the OEM cells. It would interesting to compare the SS cells to the OEM cells for this characteristic.

Positioning of the full-load stop is also critical to defining the transition region, setting the proper full-load mixture, and reducing the mechanical stress on the full-load diaphragm. Bosch sets most MPS to a 2 in. Hg engagement, though I have seen some units set to 4 in. Hg.

Again, great stuff, thanks for sharing it with us.

[jk76.914]

Wow, really great info! With your permission and credit to you, can I use some of the photos and data on my MPS page (see link in sig)?

You mention the full-load diaphragm provides altitude compensation. This is only true under heavy and full-load conditions. The only MPS's that provided part-load altitude compensation were the 0 280 100 100 series that was used on the MB's. These units lacked the full-load diaphragm (full-load was sensed by the throttle switch) and had an altitude compensation cell that was connected to the atmosphere through a holow screw.

I agree with your observation that the lower the "stiffness", or spring constant of the diaphragm, the less effect it will have on the transition characteristic. For any material used, the thickness and pleat design will vary. Reliability and resistance to fatigue failure is also an important consideration in the material selection.

Crafting the aneroid cells out of a different material and/or a different pleat design would be tricky. The issue here is to identically duplicate the pressure vs. displacement characteristic of the OEM cells. It would interesting to compare the SS cells to the OEM cells for this characteristic.

Positioning of the full-load stop is also critical to defining the transition region, setting the proper full-load mixture, and reducing the mechanical stress on the full-load diaphragm. Bosch sets most MPS to a 2 in. Hg engagement, though I have seen some units set to 4 in. Hg.

Again, great stuff, thanks for sharing it with us.

On the data and photos- you're welcome to them, with one caveat. We're using beryllium in a project at work, and one of the characteristics is that it is transparent to X-rays. So I need to verify that our XRF is capable of detecting Be in alloys. If not, those numbers may not be right. I'm travelling until the end of the week, so it'll be Friday before I can get with the lab guy. We have other means of measuring alloys that I can pursue that don't use X-Rays. The brass and stainless alloys should be correct.... Sorry to everyone if my info was premature. I'll be sure to confirm or correct it shortly.

That makes sense about the full load only altitude compensation. The diaphram has to lift off the stop for any compensation to occur.... makes sense. I have a 0 280 100 120 MPS from a Cosworth Vega that is like the MB MPS you described... but then it is, as you said, part of the 100 series... Interesting that Chevrolet licensed the technology directly from Bendix, not Bosch, but all the comonents are Bosch.

I measured several (maybe 6) aneroid cells sets using the same MPS, but with the diaphram locked. I set their inductance to a common value at the same vacuum, and then measured and plotted the curves. Their slopes are identical. I plotted them on the same axis, and they are as coincident as I can imagine. These 6 included a set from a 0 280 100 001 (early type 3, no diaphram, probably the simplest MPS out there) AND the stainless steel set.

So now I'm setting up to measure the vacuum/L curves for diaphrams alone. I sacrificed a set of perfectly good aneroid cells by cutting them open and filling them with gorilla glue to set them. I let them harden in my vise with the jaws set to original cell dimension (gorilla glue expands when it hardens, and would have pushed them open). First thing I found was that the cells ARE MANUFACTURED WITH PARTIAL VACUUM INSIDE!! As soon as my cutoff blade broke through, they expanded out and became flacid. Anyway, now that they're hardened up, I'm going to measure inductance curves, which will represent the response of the diaphrams alone.

I'm thinking that I can combine the diaphram and cell responses using superposition and get pretty close to what the finished MPS should be.... getting interesting.