I also have experienced this problem. I tried basically everything you did but to no avail. When the channeling occured, the huge hole was always in the same spot. Drove me mad.

I put a restrictor in the inlet hole on the group head to reduce the flow, and added a sort-of second dispersion screen. This pretty much fixed the problem. A firm tamp is also must.

Since everyone is grasping at straws, how about this one: any chance there's some kind of scratch or flaw in the cylinder wall that one type of gasket effectively seals but the other misses (due to size or even flexibility), allowing water to circumvent? Given the high pressure, even a very small irregularity could cause a problem.

I once had a LP with a pitted (who knows why) cylinder wall, but the flaw's position was fortunate such that the seal wasn't compromised in either the open or closed position.

Now it's my school break, so I have given these seals another chance. What I've noticed is that, if I really were to live with these seals, a firm tamp(~30lb+) is a must. 5-15lb will guarantee a big hole+one side channeling big time. Just experienced it this morning. With heavy 30+lb tamp, they will work fine. But not any less.

Anvan, I will take the group head off once again to inspect it thoroughly. The water will usually come out if I lift the lever up to about 10 o'clock with the Cremina seals. But I found that with these seals, it sometimes(not always) will spurt some steam/water at 8 o'clock. So your deduction about inner cylinder could be another possibility. I will check that out.

I've checked the inner cylinder wall. It's pretty smooth with no signs of pitting or scaling. But maybe the seal with the piston isn't a completely tight fit, there's some room of wiggling occasionally, allowing pressure or vacuum escaping. For example, with the grouphead removed from the Cremina, when I pull the lever with my finger blocking the feeding tube, creating vacuum, there is about 20-30% chance that the seal will "pfft" even before the piston reach the level of the hole. This could be one potential cause/problem. I run my finger through the wall and also look at it under light and really can't find any flaw on the wall.

I have no way to circumvent this big dent effect from happening. But I've put 4 layers of coffee filter cut to the size of dispersion screen and put them in the dispersion screen before reinstalling. This has a significant effect in reducing the high pressure spray that creates the hole and that managed to dampen the big hole effect to a very minor problem. With a firm tamp, the seal could be used with no problem. For anyone with this problem, you could try this filter-paper dispersion screen trick. Good luck!

I take back what I've said. It seems that the coffee-filter thing doesn't help much.

I can't believe I am still not giving up on this seals. Haha!

If anyone is still reading this, today I found out another potential cause: the seals' diameter was a little looser/bigger than the original seals. So when it's looser than what is supposed to be,and the seals're pressed into the cylinder grouphead, then there is room for the seals to wiggle. It's not a perfect round shape sealing the piston anymore. That could explain why there are jet spraying because it's not sealing evenly and the high pressure water could find a way to escape through the seal.

Here's an image showing what I meant if the description seems abstract to you.
The top seal is OE's and the bottom is stock's. It's kinda exaggerated in the photo because I did pressed the seals purposely and run my fingers round the seals so that they can fit tightly into the groove. But they couldn't fit in 100% as it was a little looser. Instead, a portion of the seal protrudes out which is what you see in the photo.When I spread that evenly throughout the piston, there is a very small part of the seals that protrudes round the piston. This suggests that it is looser/bigger fit than the original seals. Could this be the cause of the problem?

I'm still reading. I have found that the orphan seals fit easier in the group, and that the older ones are harder, and also must be slightly bigger, since water doesn't drip out until I actually raise the lever to the right place. Replacing my seals helped me, but the problem still persists. The good news is that the shots are drinkable, but there is channeling from 12 o'clock still, and now a bit from 6 o'clock. I'm resisting putting in a paper filter because of more maintenance, and also that I haven't found a great way to take the dispersion screen out easily. My next option is to replace the dispersion screen, unfortunately (they're SO expensive!).

The frustration never ends. It's now been over a year with this machine and I want to throw it across the room, except that I know awesome shots are around the corner.

I imagine that the seal looseness you're suffering from is because they've been taken off and replaced way too many times!

bostonbuzz wrote:since water doesn't drip out until I actually raise the lever to the right place.

Somehow this is different from mine. Mine actually drips(sometimes) even before raising it to the right position with the OE seals.

The filter paper trick doesn't work(well, it works initially, for like a few shots before it fails). It goes out of place after a few usages because of the water pressure. That renders it useless until you put the filter paper back in place again.

Don't throw it! Send it here. Hahahaha! Try tamping with like 40lb and use single basket to reduce the channeling likelihood. Mine works very well with the single basket. I occasionally get the hole with bigger basket, less so with singles.

Just to clarify, did you mean you have already replaced the seal with genuine Olympia seals and it got worse? I don't think it's worth replacing the dispersion screen. I am fairly sure it's not the cause. Unless you meant you're going to add a second dispersion screen to soothe the pressurized spray

Olympia Gaskets= tighter, no water until I want it
Oprhan= dripping water at random while raising the lever

I've been trying lately to tamp the s*** out of it to reduce the impact of pressurized air/water, and it sort-of works.
I may go with some stainless steel mesh dropped above the screen, but this is an ugly fix (although you can't see it). I don't know where I'd get some fine enough to stop a jet of water. Or perhaps some sort of round screen to slow/disrupt the flow from the hole in the cylinder.

bostonbuzz wrote:Olympia Gaskets= tighter, no water until I want it
Oprhan= dripping water at random while raising the lever

Ahh, that's right. Same here.

For the mesh screen, I was thinking more along the line of Coava disk for Aeropress. I think that would work better because its shape is pretty firm. But it's hard to find something like that, at least not DIY.

Check this out if you will. With different seals, the pattern that the water sprays out is different. One is more concentrated in one jet, while the other one is spread more into 2-3 jets. Guess which is which seals. This is without dispersion screen fyi*.


If you have time, check out my other videos for OE 1-4, the water does come out randomly at times.


*Edit: Note from me (Sam) 8 years later on 4 May 2019: there was supposed to be some Youtube videos here showing the water flow from the Cremina...but the video was removed unintentionally (Forgot that I had it linked here when doing Youtube spring cleaning ).

Thanks for that aeropress filter tip. I will be ordering that and hoping it's the right diameter. I know Stainless Steel is a pain to work with when all you have is sand paper and a hammer...
This really seems like the solution to the problem. It's flat, so it won't get in the way of the piston, and it will be very easy to clean since it doesn't have 10,000 holes like the micron screen I was looking at.