solution for D-series poppet seat DIY?

[lexvil]

I’ve given up on this twice and yet again I’m back at it too, so far testing in the 420 using the adhesive method and only the top o ring is looking good, I have found the tightening of the nylock nut is crucial to performance, I haven’t measured exactly how far I’ve tightened it but I did go for some protrusion (this may “stiffen” the o ring a bit) and have vey smooth and easy operation. I’m also setting up a few AIR 1’s with this seat. If the angry ocean will play nice for a day I will dive them again Friday.

 

[Kupu]

I'm happy to see there is still interest and and at least some potential of a working solution.

It will be months before I have time to return to this project or any dive related activities, with a seasonal work schedule that has entered the busy time of year. Working towards retirement where the diving is easy.

 

[Kupu]

I was on the road all day, and this idea floated into my mind.

What about a custom orings? Instead of trying to work thru the issues created by oring sandwiches, use a single oring designed with ideal material and dimensions. With similar OD and adjust the ID to size needed for proper seal on the shaft and also prevent extrusion issue. Would square profile be better than round? Etc.

I would certainly be willing to put some cash towards this idea if it has merit, but have no idea of costs or minimum order size. Impractical idea, or another possible solution?

 

[rsingler]

I was on the road all day, and this idea floated into my mind.

What about a custom orings? Instead of trying to work thru the issues created by oring sandwiches, use a single oring designed with ideal material and dimensions. With similar OD and adjust the ID to size needed for proper seal on the shaft and also prevent extrusion issue. Would square profile be better than round? Etc.

I would certainly be willing to put some cash towards this idea if it has merit, but have no idea of costs or minimum order size. Impractical idea, or another possible solution?

Costs are not high, since a washer type oring is a standard shape. That's what I tried with the X-profile in a post above, but it leaked air as well. The key is getting a seal along the shaft of the poppet. Scubapro accomplished it with the tiny 1 x 2.5 mm oring at the top, and a rigid holder for the seat. I previously attempted to do this by punching out a seat from Viton and silicone sheets of varying duro, but they all leaked, because I couldn't seal hard against that 1x2.5. That's what that rigid carrier accomplishes. For D-series lovers!
The air would track up the shaft, along the top surface under the cap, and blow down the edges of the seat outside the knife edge.

I looked at a sealing washer like this:

but can't find a size to fit inside the cap. While a custom run of a standard sized oring is only a hundred dollars, a custom size piece is a whole 'nother matter.

However, the beauty of your solution is that stretching the 2-104 o-ring over the shaft goes a long way toward creating that seal.
I was messing with a few different duro's last night (see next post), and it was only when I pushed IP up into the 130's that I needed any silicone cement at all. It was fun hearing the oring "pop" followed by a hiss, as the air tracked around the seal.

No, @Kupu, I think you have a real solution here. I've got a second run of poppets curing now, so we'll know more tomorrow.
With the room temp vulcanizing silicone sealant, I think we can handle the IP now. It just remains to be seen what duro is required and whether it will withstand the flexing of thousands of cycles.

 

[rsingler]

Been working on @Kupu's poppet nonstop, and have more info. I think we're getting closer!

First experiment was on the best way to seal. You'll recall that several have noted that the cap can't be screwed down too tightly because the small lower oring extrudes. That was the past two days' experiment.

Unfortunately, I have been changing too many variables.
The previous duro 50 silicone 104 is still going strong. It's got a 1 x 2.5mm above and below the oring, and is filled inside with the room temperature vulcanzing silicone. It cracks at 0.5" but is taking a bit of a set in the oring, because it's so soft. Also, I was disappointed that I could only get to 1.1". I wouldn't run the reg any higher, but it just bothered me that I couldn't push it the full range of Scubapro's specification.

So to attack both issues, I ordered a "super-resilient" Viton oring that was duro 70-75. I was trying to get around the anti-set properties of Viton vs. nitrile, with Viton being the loser:

Anyway, the results were disappointing, because I changed too many things at once. Because the Viton was stiffer, I thought the lower oring was critical in helping to fill the gap at the shelf of the shaft. So I added 2.5mm and 3mm ID orings to the trial. As you'll note above, neither poppet with a 3mm lower oring would seal. And interestingly, the stiffer Viton oring with no lower o-ring at all, sealed just fine! So @lexvil is right - the only thing you need is that upper oring.

And that brings me to my next surprising finding. For some reason (too many variables?) the stiffer 104 would not seal at a higher cracking effort with full spring pressure than the softer silicone oring. To me, that is completely counterintuitive. To me, the higher the spring pressure, the deeper the knife edge digs into the oring. That increases the total area of contact, and depending upon how dull a knife edge you have, potentially creates a leak, because the "pounds per square inch" may not increase, despite increasing spring pressure, because the area of contact increases so much. Hence my surprise that the stiffer 104 wouldn't seal as high.

So that's my current experiment.

I'm using no lower oring at all, for any poppet. The only thing I'm changing is the composition of the 104. I've got duro's ranging from 50 to 75.
I'm using a plain nitrile 1 x 2.5 for the top o-ring, because I found that the small silicone orings I ordered tended to break with much stretch, as confirmed in the chart above. So now we'll find what duro gives what peak cracking effort. I just don't know why those stiffer 104's wouldn't crack higher than the D50. Hopefully this experiment will tell us that it's some other variable that caused that.

I'm getting better at building these poppets, with practice.
The RTV silicone is a mess to work with, but using loupes and a brass pick, I can be fairly discrete.
But I'm running through ALL of my wife's Q-tips cleaning the excess off the oring surface. That's critical, because the RTV cures to a firm bead that disturbs the smooth sealing surface of the 104. It's now taking about 10 min per poppet, including inspection. The RTV cures very quickly, so my loss rate was about 10-20% until I got my technique down.

Anyway, check back tomorrow! The cool part is that those stiffer Viton's often got me down to 0.4" without a leak! I'd like it if we could make a stiffer o-ring work for us, because I'm frankly a bit concerned about component separation after several thousand cycles. The soft D50 silicone probably flexes a lot more than the D75, so without a brass disc that the seat is sealed to like the Scubapro seat, we need either very flexible silicone sealant, or a stiff seat.

 

[lexvil]

Nice methodical approach, while my main concern has been with improving the D420, I have started putting the 104 in my AIR 1’s which is similar to the other D series with the same type of knife edge.
Right now I’m only working with the 50 duro silicone 104, i silicone the 104 onto the shaft by running a bit around the shaft and installing the 104, turning it so it goes onto the fat part of the shaft, put the 2.5 on top and then some more adhesive on top and sides of these, I put the plastic centering guide on next and turn it to evenly coat the adhesive, install the nut and snug it so that there is slight pressure on the parts.

on the 420 I have to tighten the nylock nut down so that a little more than 1 full thread shows on top. So far this is making for very low cracking effort (not measured) and smooth stabile flow.

 

[Kupu]

@rsingler I am certainly impressed by the amount of time and effort you are investing in this idea. Great stuff.

I just had a look at the 3 AIR1s that were set up with D50 -104 plus single 1 X 2.5 at the bottom, and then put away with seat saver keys a couple months ago. As expected, they are in same working condition as before. No problem getting them down to .3" cracking effort and one of them is set just a hair above .2". They all have original anti friction washers installed between spring and poppet, and spring adjusting pad set for minimum spring pressure.

Speaking of variables, I wonder what differences might come into play based on the regulator the poppet with oring seat is installed in. If I am not mistaken, the D400, D420, and AIR1 all have different springs.

 

[rsingler]

Well, here's the next installment on the experiments.
To respond to @Kupu's question, I don't think the spring that backs the poppet is as much of a factor as the style of knife edge. Some are razor sharp, and some are quite rounded. After restoration, all are rounded to a certain degree. That markedly changes the surface area of contact with the poppet, which is the primary determinant of the spring pressure required to seal. To the extent that the Air 1 and D420 have longer springs than the D400, there might be a greater range of adjustment if the longer spring can also go higher on the force applied at its max compression. The bigger factor will be the range of adjustment of their respective screws, which is quite limited with both regs. But you can add to that with shims, which is what I did with the D400 below. That makes adjustment more sensitive, but also raises the max compression before the adjustment screw bottoms out.

Here's what I found (somewhat disappointing, and definitely counterintuitive).
I tested two poppets of each 2-104 oring type on the same D400, with the same spring, and same lever. I used two graphite impregnated shims to give max compression at the high end, since this is an old spring. I also used an Air 1 poppet shim 01.060.100 to further increase spring compression, as well as decrease friction as I screwed down the adjustment. Even with those shims, you could see the poppet want to rotate as I added turns to the adjustment, so all changes were made with the lever depressed, to raise the poppet off the knife edge. As noted above, the poppet consisted of a standard metal shaft, no lower oring, a 2-104 of varying material, a 1x2.5mm 50 Duro nitrile top oring under the cap, and enough RTV102 to fill any voids. However, every poppet shaft had a microscopically different length antifriction cap (11.012.107) on the end, which meant that the lever had to be adjusted for each poppet change. More on that below. I also re-tested the original D50 silicone poppet with a 1x2.5mm bottom o-ring from the original trial.

Here are the results: (104 type/poppet #, min crack, max crack)
Orig D50 Silicone, 0.4-1.4"
D50 Silicone #1, 0.4-1.2"
D50 Silicone #2, 0.3-1.3"
D65 Nitrile #1, 0.45-1.1"
D65 Nitrile #2, 0.4-1.1"
D70 Silicone #1, 0.4-1.1"
D70 Silicone #2, 0.4-0.95"
D75 Viton #1, 0.4-1.0"
D75 Viton #2, 0.3-1.1"

In other words, going from squishy to hard orings didn't make a difference. If anything, the softest oring was able to generate the highest cracking effort (not that I have any use for a hard breathing second stage).

I did not expect this.
Which brings me to my next planned trial.
Knife edges come in a range of sharpness, especially after restoration.
In fact, the last example is a new knife edge from my stock, from late production. Notice the broad cone and the very rounded edge.

I think that the physics of sealing for each of these is quite different. The test above was done with a knife edge closer to #1.
So I'm going to compare similar poppets on three different knife edges, and we'll see if duro makes a difference there.

That will bring us to the last worry: whether our adhesive will keep constantly flexing orings sealed together, so the poppet is safe to dive.
I've just reached out to Pelseal Technologies, LLC for information on a fluoroelastomer adhesive/caulk sealant that sticks to both rubbers and metal. Since I only need an ounce, I'm hoping it won't cost an arm and a leg. They have a PLV2036 2-part adhesive with the consistency of molasses which might bond better than our RTV silicone. There are caulks that are even thicker, but I'm not sure of their adhesive properties.

One important tuning discovery when working with a multitude of different poppets:
Sometimes the new poppet would have a longer nib at the end, and the lever was already putting unseating pressure on the poppet as soon as I inserted it.

This valve leaked even before I added the diaphragm, until I dropped the lever.

But what was more important is the amount of lever movement when I screw down the adjustment screw with the softest 104. I would recommend that you set your lever height with the adjustment cap fully screwed in, so the poppet is fully depressed into the knife edge by the spring, before you decide that your lever is correct.

Otherwise, you'll find as I did that when you try to "adjust" your cracking effort, when you add a bit more spring pressure, the poppet hits the lever fulcrum, and you won't increase cracking effort at all. In fact, with the counterforce placed on the nib, the valve may leak. Dropping the lever a bit more allows it to seal right up. Perhaps that's why later Scubapro OEM poppets seemed to be made with firmer seat material that the early ones. Additionally, we saw this very same instability with the D420, when the lever was too close to the back of the diaphragm.

Anyway, I've added this to my D-series tuning routine now: Screw in the adjustment fully before setting lever height. Minimize lever play at this point, recognizing that it will get slightly floppier when you pressurize. Then unscrew the adjustment to back off to the cracking effort you want.
I suppose if you wanted max performance, you could tweak the lever height just for the cracking effort you use, since you'll never be honking down the spring that far in actual use. But I was surprised by how much the lever rose as I added spring pressure to the softest 104.

 

[lexvil]

Did a long (80ish minute) dive this morning, D420 with the 104 seat, super smooth and stabile. Testing to continue when the conditions improve.

 

[rsingler]

We’re reaching the end of our investigative saga regarding the D-series poppet.
This time, I took two radically different D-350 orifices, and ran our poppets through them.

To cut to the chase, the results weren’t that different!! ANY orifice, and ANY duro 2-104 o-ring gave a good seal in the normal working range for the D-series (0.8-1.0” cracking effort). ALL would seal as low as 0.5”, which is spectacular news. About the only thing that was statistically significant was that a sharp knife edge could reach 1.6” max cracking effort, while a dull one would only reach 1.2-1.4”.

Here are the two knife edges:

Here’s the data:
DULL knife edge -
D50 Silicone, 0.4-1.3”
D70 Silicone, 0.4-1.2”
D75 Viton, 0.45-1.4”

SHARP knife edge -
D50 Silicone, 0.4-1.6”
D70 Silicone, 0.4-1.6”
D75 Viton, 0.35-1.6”

Interestingly, with the sharp knife edge, getting to the low end of cracking effort required returning the valve to its factory configuration: no Air 1 shim on the poppet and no graphite impregnated shims under the adjustment cap. Any shim at all would make minimum cracking effort around 0.7”. Conversely, getting to max cracking effort with the sharp knife edge required all three shims. Otherwise, max cracking without shims was 1.25” (D50S), 0.9” (D70S), 1.0” (D75V).

And required spring pressure at low cracking effort was universally higher with the dull knife edge than with the sharp, which also makes sense. A sharp knife edge has a thinner area of contact with the seat. Thus, the total area of that seat seal is lower. Intermediate pressure trying to sneak under the edge of that contact ring requires a “pounds per square inch” on the sealing area greater than the IP. If the contact area is larger (dull knife edge), the spring pressure will have to be greater to create the same sealing pressure.

What have I learned with all of this?
Well, keep in mind that we were working with pristine knife edges, and if you create a seal with enough spring pressure, the minimum cracking effort is not significantly affected over 50-75 Duro. They all work, within the range of our measuring precision. Now if your knife edge was getting ragged, things might change. In that case, you’ll have areas of better and worse sealing, depending upon the nicks. You might expect that a softer duro would conform to the irregularities better, and you still might get low cracking, but that’s just a guess.

Between @lexvil ‘s D420 diving and these experiments, I think we’re getting to a point that we can consistently create a poppet that works. So which Duro to choose?

I got a phone call back from Bill Ross, President of PelSeal, Inc. He and I had a wonderful 20 minute chat about adhesives. He was honest enough to say that his Viton adhesive would certainly adhere to our Viton o-ring, if that’s what we chose. But adhesion to the plastic cap was another matter, as it is with ALL adhesives, including our silicone RTV102. He points out that guaranteeing “no leak” with the constant flexing of the seat is impossible. What is required is what SP engineered: a rigid carrier that seals to an o-ring in the cap. Our flexible components cannot guarantee a seal at the shaft/seat/cap interface. It was very generous of him to spend the time teaching me about the factors when he admitted right off that his product couldn’t give me a solution.

I’m going to keep diving this workaround for the fun of it. I think it’s great that we have a solution that will give us the supremely low cracking effort that few second stages can handle. I think I’ll make a stash of poppets with the D70 Silicone (for lower flexion), and return to adding the lower 1x2.5mm oring (of stretchable Buna, instead of silicone), to decrease the total gap that must be filled with RTV sealant. I'll still use the critical upper o-ring between the cap and the shaft. I think I’ll dive with an inline shut-off, so that in the event of freeflow, I can just switch to my necklace and preserve my gas for the ascent. We’ll see how big a duty cycle of we can get. The problem is, to safely advocate this workaround to our SB brethren, we need a duty cycle of 250,000. That’s 100 dives x 2 hours x 20 breaths per minute. I just don’t think I can trust my RTV glue-up that long. But as a fun experiment, it’s very cool that @Kupu ‘s solution worked!