Over the weekend I disassembled, cleaned, and re-assembled by MPS (022 906 051 E) with the new diaphragm rebuilt kit. When I use my Mityvac vacuum tester and draw to 15psi, the the MPS holds a vacuum for only thirty seconds.

The new diaphragm kit is installed correctly, the new rubber seals are happy. Everything looks clean an correct. I have also used a couple of different aneroid cells.

I thought about pumping a little air into it and then placing the MPS into water to look for bubbles, but I figured that would probably be a bad idea. Is there some clever way I can test to find out where/how this is losing its vacuum so quickly?

Thanks,
Marcus

Did you use a heavy-bodied grease between all mating surfaces? It is needed at any interfaces where a vacuum could be applied.

Even then, there is a possibility of leaks between the plastic connector housing to case and/or contacts.

By heavy bodied grease do you mean non-petroleum grease? Or distributor grease? I haven't applied any sealant or grease at all. I wasn't sure if it was good idea or not, and I didn't see any when I disassembled the ones I have. But it certainly seems like a good idea.

What about where the paper gasket and diaphragm meet up? I don't want to apply chemical sealant or grease where it shouldn't belong and I don't really want to experiment with the few good parts I have.

The electrical connector is a good place for leaks. I pressed it against a bit of wax to check for leaks but it seems secure. I'll put some epoxy around the seam when I'm finished to make sure.

Call Chris at Tangerine Racing, I'm sure he can help you with assembly procedure!

Did you test the mity vac to make sure it isn't leaking down on its own?

Yes. I can connect the mityvac to other items and it holds vacuum just fine. I'm sure it has little to do with the new diaphragm kit and everything to do with sealing leaks already present with a sealant. I just didn't want to goop anything without really knowing where it leaks. I would like to be able to determine the source of leak before potentially messing anything up.

But for sure the vacuum pump is not the problem.

I've never felt the need for any grease or other sealant where the diaphragm and gasket are screwed down.
Is it possible you put the paper gasket in the wrong position since you didn't have the instructions until the re-assembly was done?

It's possible. But I didn't re-assemble it until after you sent the instructions because I wasn't entirely clear on the diaphragm parts order. Here's how I have it assembled now:

paper gasket
diaphragm
sealing ring
part load stop plate

After verifying with your instructions, this is the order in which I have the parts installed.

It should be safe to pressurize the unit to no more than 10psi and check for leaks with soapy water.
You could even submerge the short end to see if it bubbles out the vent holes.

Make sure the full load stop is in and adjusted far enough to support the diaphram!

Put some dish soap in a spray bottle with water and mist the MPS when under pressure. A leak will show as a bubble(s).

When immersed in water I see a bubble or two when I apply a little air pressure, but I can't see from where. The next time about I can't get any bubbles to form at all. I disassembled the MPS to look for water in the system but found only a little moisture on the copper diaphragm. I'll run some more experiments along these lines and see what materializes.

Thanks all. I'll report back.

This is how I tested one. http://www.914world.com/bbs2/index.php?sho...=190809&hl=

Well, problem solved. But it's really inelegant.

I ended up trying all manner of tests trying to figure out exactly where the leak(s) were but this didn't work. So I tried a bit of RTV around the mating surfaces of each hemisphere. That helped only slightly, but perceptibly.

So I smeared a little Hylomar around the paper gasket and sealing ring on the internal diaphragm. I really, really didn't want to do that but there just didn't seem to be anything else visible to try. Doing this actually solved most of the problem.

At this point it took almost ten minutes to leak from 15 to 10 in.Hg, which according to the pbanders site is good enough. I've read that others maintain that no loss is acceptable but it's hard to imagine that an imperfect cast aluminum part with moving internals will ever hold a perfect vacuum. I also sealed the hemispheres with blue RTV after another thorough cleaning and letting it sit for a full 24 hours. That helped a touch.

The last big help came from carefully laying down a layer of epoxy around the edge of the electrical connector. I used long-cure gel epoxy and with careful use of a flat toothpick troweled in nice barrier around the perimeter.

Several hours later, I now have an excellent seal. It would take over an hour to drop from 15 to 10 in.Hg now.

It occurs to me that maybe the MPS I have is trashy from the get go, otherwise I don't think I should have to go around glueing it together so much. But it's 42 years old, and the best one in my box of spares. I could have bought yet another one but might have ended up with the exact same problem. The unit just isn't perfect and it doesn't have perfect mating surfaces. Somehow I thought the resolution would be sexier than a collection of sealants.

Hopefully it's good for another 42 years.

I'd like to thank Chris for making these diaphragm replacements and offering some wisdom here in the forum. Without his efforts I'd (we'd) still have a bunch of trashy MPS units in our cars.

Cheers,
Marcus

Using a grease to seal vacuum leaks at the critical mating surfaces is not inelegant, its what the factory did. Notice the original paper gasket is soaked and discolored in grease and see the medium brown dried residue caked around it? Using grease makes it a non-permanent process, an adhesive sealant like silicone would make it very challengint to remove and repeat the rebuild later.

The MPS rebuild instructions from my EFI Associates MPC Analyzer manual also indicate the use of grease. There is more to the rebuild than what they indicate but here it is for information: