How to restore a Diaphragm First Stage with IP Creep

[rsingler]

IP creep has been covered in multiple threads on Scubaboard.
How to restore a Piston Regulator that has IP creep
How to Restore a Knife Edge Piston

Usually, it's pistons that get the blame. Either a faulty seat

or a nick in the piston knife edge

allows high pressure (tank) air to leak into the Intermediate Pressure compartment, and we see IP creep on the gauge.

But sometimes, even a diaphragm can be afflicted.
And usually, it's the same culprits: a scratch in the movable part, or a nick in the fixed part.
Some diaphragms (Mares, Poseidon) use a movable dome or cone-shaped poppet that mates with a fixed seat, while others (Scubapro, Aqualung) use a movable flat seat that mates to a volcano orifice.
Either way, if you can replace or polish the movable part that is flawed, problem solved.
Here, a Poseidon cone (which they call a valve piston) has a sandblasting groove from tank rust that slipped past the sintered metal filter, causing IP creep. A quick polish in a drill motor (as long as the drill chuck teeth don't put a nick in the shaft) and things seal right up.

When the volcano is scratched by a ham-fisted technician, or becomes corroded, it can occasionally be a challenge to reach. The popular Conshelf first stage has its volcano buried deep in the bore of the reg body, and can be a challenge to polish. The volcano cannot be replaced.

Other regs have a removable volcano orifice. Wrapping a pencil in Micromesh and buffing out a nick is a much simpler task when you can hold the volcano in your fingers.

But what happens when you have IP creep and you replace the movable part; you polish the part that it mates with; you replace the orings...and it STILL creeps?
Well, let's look at where a diaphragm reg seals:

Courtesy of the Cressi website, here's a sealed diaphragm with a removable volcano.

There are three areas where an IP creep problem might lie:
#2, we have just discussed: the seal between the seat and whatever component it mates with. Here, we see a seat (in blue) and a volcano orifice just below it. This valve is open.
#1 is the seal between the movable seat and the reg body, via the balance chamber. This seal only applies to a balanced diaphragm, with a hole in the middle of the poppet. A leak at this seal allows high pressure air to travel up the bore of the poppet, bypassing the seat, and leak into the Intermediate Pressure chamber. If your reg is unbalanced, there's no hole in the movable part, and this area can't leak. That applies to few regulators these days, like some old Dacors and the old Poseidon 300. Most diaphragm regs today are balanced, and this area is a potential leak point.
#3 is the removable seat or removable volcano, depending upon the brand. If you have a fixed volcano, then again there's no oring seal to leak at this level, either.
In other words, if you had an old Dacor unbalanced diaphragm with a fixed volcano, there was only one place to leak! You either replaced the seat or polished the volcano deep inside and fixed the problem, or you bought a new reg because you couldn't polish out a ding in the volcano.
But most high-end diaphragms have removable volcanoes, and a seal here is key.

As the technology got more sophisticated, regs've gotten more complex. The areas where problems arise have multiplied.
And that's what happened to me.

See next post...

 

[rsingler]

In my case, I had acquired a lovely Poseidon 3960 with beautiful chrome and apparently no abuse that was another of my eBay "steals".
Alas, 'twas not to be.

IP creep on first exam was no surprise. After all, there had to be a reason why I got this reg for $50 instead of double that or more.
So I replaced the seat. It looked okay, but that had to be the problem, right?

Well, actually not. The seat looked fine.
I also looked at the "valve piston" under the microscope and while I was distressed, I wasn't worried.

This sign of abuse was disappointing, but I wasn't worried. The location of that ding was on the wider part of the cone.
The seat mates higher up the cone, where the diameter matches the hole in the seat. It couldn't be leaking there.
But, I didn't want to rebuild it twice, so I put in a new seat and valve piston anyway. That ding was too deep to polish out.

After reassembly, I checked the IP, and it was still creeping.
I took a big breath...

From the technician seminars and personal experience, I knew that the #4 factory seat mold had, in the past, had some problem output. Maybe I'd replaced with a bad seat?
Under 30x mag, even the new seat looked a touch irregular.

Perhaps I could improve on that...
So I got out some Micro mesh and worked it

The new seat edge seemed smoother

so I reassembled the reg yet again.
And it still had IP creep!

Time to go back to the theory.
Looking at our Cressi diagram, let's identify each of the components involved.

There are those three areas where a seal must be made: #1 at the balance chamber; #2 at the seat/knife edge and #3 at the seal where the removable orifice is joined to the body of the reg.

But I'd already replaced the seat twice, and the valve piston and the o-rings. What's left?
Well, if you look at the diagram more closely, the (red) dots for the orings are not the only components that make a seal.

You also have the surfaces that the o-rings contact.
As you can see with the tiny blue dots, in addition to the knife edge itself, there are also the lands where the o-rings seat. So what did I miss?

The o-ring land on the valve piston shaft in section #1 was addressed by replacing the piston with new. The land where the other side of that o-ring seals is DEEP inside the reg in a Poseidon, and might have been scratched when a previous tech removed the balance chamber o-ring. There's one place to look.

Next, the area where the seat or volcano o-ring sits, has a seat/volcano side (eliminated as a possible leak point in my reg by replacing that part), and a land where the seat/ volcano o-ring seals against the reg body. There's another place to look.

I went ahead and opened this reg up for the third (fourth?) time.

See next post...

 

[rsingler]

When I looked deep inside the body, the deepest o-ring land for the balance chamber actually looked pretty good!

But as I backed the focus point of the microscope up the shaft of the reg body, this appeared:

You can just imagine how this happened! A tech used a screwdriver or a straight steel pick to pull out that stiff high pressure o-ring, deep in the balance chamber. And in order to keep from dropping the oring as he pulled it out, he dragged the pick against the wall of the shaft, scoring the next critical land along the way. Arrgh!

For most shops, this reg is now a doorstop.
But in fact, it probably CAN be fixed.
What to do?
Once again, Micromesh to the rescue!

I wrapped a long, thin strip of 4,000 grit Micromesh around the right sized dowel, and twisted it in place.

After a few passes, the score looked like this:

This was one deep scratch!

The only thing to do was to keep at it.
If the land gets polished too deeply for the oring to seal, the reg is toast. But the only way to find out was to keep polishing.

After several passes with 4,000, 6,000 & 8,000 grit Micromesh on that dowel, this was the result:

You can still see a tiny gouge at the deepest part of the land, but this is below where the o-ring seals.
Upon reassembly, she sealed right up, with zero creep! Problem solved!

When you're faced with IP creep, just look closely at the schematic to figure out each of the areas where air can leak past. There are other potential leak points when air is bubbling out of your regulator, but for IP creep specifically, the choices are limited.

In a piston, there are only two, if it's balanced:

the seat & the knife edge,
(and maybe the oring and 2 lands holding the seat in place, depending upon design, if there is an air passage from the HP seat compartment to the IP compartment).

In a diaphragm, there can be up to eight:

the seat,
the shaft of the moving part (seat or valve piston), its o-ring and its land,
the orifice,
the oring that holds the orifice (if removable), and its 2 lands.

It's not hard to diagnose. You just have to be meticulous about examining each of the possible sealing surfaces.
Well, sometimes it IS hard to diagnose, if you don't inspect every sealing surface every time, and instead be like an LDS with its undertrained, ham-fisted parts changers in the back room. You can't blame the LDS. Reg service is a loss leader for most dive shops.

DIY!

 

[Bigbella]

Great post! A damaged valve piston, from tank corrosion, trashing a seat had been the usual suspect for me, though I never thought to polish it away -- so too, a couple of those wonky #4 hard seats.

Congratulations on your eBay bargain . . .

 

[Graeme Fraser]

An impressive forensic approach and, I imagine, highly satisfying. I'm just starting out on my foray into regulator self service and my takeaway from your example is to not underestimate the consequences of even the most seemingly insignificant act of clumsiness.

Fortunately I'm reasonably time rich (and cash poor) so will be even more careful during the disassembly, cleaning and assembly process. However if all else fails, keep an eye open on eBay for some Apeks bargains from the UK .

 

[herman]

Really nice presentation. I have a bit to add that this, a bite you in the rear that really screws with your mind.
You rebuild a diaphragm reg, it sets up perfectly - no IP creep- but after a while, it might be a hour or after a couple of dives it starts creeping. You pull it apart, nothing looks wrong, you replace the seat and oring anyway and it does the same thing, works fine for a while then starts creeping. A scratch does not make sense as it should start as soon as you pressurize the reg...................Any Ideas?

This is the culprit - there is a small scratch or defect in one of the seating surfaces, like the one in your example. When you replaced the parts, you lubed the oring and that lube fills the crack, sealing it off. But as pressure is applied, the lube is slowly extruded from the scratch, how fast depends on the depth of the crack and the viscosity of the lube. It may take a few seconds or it may take hours being under pressure to happen. The ones that happen quickly are not too bad but the ones that come back days later are a real pain.

 

[rsingler]

When you replaced the parts, you lubed the oring and that lube fills the crack, sealing it off.

Great point, @herman !
That was a hard lesson to learn. When I first started repairing regs, it was always "the more lube, the better!" But over time, your point has become clear. All that goo is just hiding things, and attracting grit, or plugging up your balance chamber reducing performance.
More isn't always better (except maybe in what I think of a the reservoir between the main piston head oring and the wiper ring of the Mk25 ). A little shine is all my o-rings get now.

Thanks, @herman !

 

[herman]

The other place I goop it on heavy are SPG spools.This prevents water from getting in and corroding the hose/SPG/spool together. A lot easier to clean a little goop off than to remove a corroded in spool.

 

[taimen]

ut as I backed the focus point of the microscope up the shaft of the reg body

What kind of microscope do you use and recommend for inspecting regulator parts?

 

[rsingler]

What kind of microscope do you use and recommend for inspecting regulator parts?

See post #3 here: How to restore a Piston Regulator that has IP creep