The one in mine is old 1981 and I think got used some and sat for many many years. The piston and sight glass gaskets had turned to stone so I'm guessing the diaphragm is quite a bit stiffer. The silicon made sense for it's temp. rating where rubber products are near, at or above their temp. limits and need the cloth insert to maintain. As long as the silicon has flexibility in an expectable range for the pstat. it should last and be consistent. The application is really quite simple and not sure a formed gasket is needed.

You are going to love this, but in the bigger stats I found that the membrane did not fail because it lost it's real flexablity but more than it scaled and the mineral build up on the surface of the membrane caused it to go out of spec. If there is a crunchy limiting buildup on the pressure side of the membrane it will not flex the way it should. OFC there is the other failure point that the seal rubber around the edge of the pressure membrane could go bad.

But strangely I have not seen that on the Siari stats. They all get hard in the middle and the seals are just fine. Scale is a worse problem for them and widens the dead band more than any material hardening ever could. I have found that a citric acid soak is enough to bring most of them right back back into line. But remember this is only for the stats that do not show greening or problems with the high pressure and low pressure sides.

If you are getting the green on the lower pressure side you have a replacement in your future.

-Cecil

Edited to make this make sense.

Mine did have some scale build up and it definitely would have effected it.

But was it on the boiler side or on the low pressure side? If it was on the boiler side a simple descale of the membrane could have brought you right back in line.

-C

After I try the silicon piece a few times I'll try putting the old original one back in to see how it performs.

I am not suggesting that, but that the more robust pressure stats may use a different material in the pressure housing to relay the destention for the stat.

That was all.

-C