Pipe damage

[Mastermindx]

Hello guys.

Let me introduce you to Mr. Chilly!

Mr. Chilly and I have been long time friend.  I can always rely on him to cool down my base and turn the steam of those cool steam vent back a warmish 95 degrees water.

However, with current (vanilla) version of the game, he's been feeling a bit under the weather.  He's been busting pipes and I can't figure out why!</storytime>

So, for some reasons, the output pipe of the right aquatuner is getting damage.  Since I haven't witnessed it , I don't know what's the reason (would be great to have a damage reason history!). But my best guess would be cold damage. Here's how the piping looks

I seriously don't know what could be causing this.  Both aquatuner are set exactly the same. Both thermo sensor are set to "Above 22C". And yet, the output pipe on the right AT got damaged quite a few times already. My best hypothesis at this point would be that, in some circumstances, the aquatuner can cool the same packet of water twice.(I know, I'm grasping at straws here...)

Left AT uptime : 30%
Right AT uptime : 10%

You guys can think of anything I might have overlooked?

[JRup]

Hi!

Consider the following and let me know your thoughts on this:

• When the second AT activates the second bridge will be blocked eventually. So if a packet of water is rejected it will still remain in the input tile of the shutoff valve because the path is blocked.
• This gives way to an opportunity to feed cold enough water to the AT when the blockage is resolved...
• The pipes outside your steam collection area are not insulated. So I guess the environment must have helped avoid this issue for some time but it may now be cool enough to allow this oddball situation to happen.

There are some cookie cutter solutions for AT loops that can be found on the forum, let's hold back on that for a bit and see if the diagnostic procedure helps before revamping your design.

In order to troubleshoot:

I would suggest temporarily testing with a liquid reservoir. This can be easily done without changing your design that much. Just add a line coming out of the bridge shown by the arrow and feed it into said reservoir:

 

Feeding the output back into the loop should be a given... if you do it directly on the bridge output you'll notice some alternating happening between the AT and the reservoir tho' -only when the AT blocks the bridge. Another bridge to feed it on the loop will help.

 

[zach123b]

the way the automation is set is the problem.  when the bridge backs up, it can pass a packet of water that's too cold into the aquatuner cause it stops with a warm packet under the sensor but a cold packet on the valve

to fix it?  either move the temp sensor to control the aquatuner like most cookie cutter builds so the cold water passes through the aquatuner.  or rework the water flow, instead of prioritizing the water output from the aquatuner to prioritize the water passing by.  or put a liquid reservoir between the bridge and the shutoff valve.  with such low uptime, maybe just remove one, unless you're using gold amalgam which then i'd split the throughput so they each chilled 10kg every other packet then combine the line back together afterwords

at least those are my ideas to fix it

[Mastermindx]

I just went to sandbox and did some testing... And my design is just inherently flawed.

Taking this setup with top pump pumping 95 degree water and bottom pump pumping 4 degree water, it's not possible with this setup to split the cold and hot water if either "output" can block. If top blocks, both temp water will take bottom route at 100% throughput.  If bottom blocks, they'll take top route at 66% throughput. 

And this design : 

Will allow everything to pass at 66% throughput.

Based on this,  I'm tempted to believe it's not possible to make a reliable automation with the shutoff valve based on pipe content if any output can back up, even for just 1 tick. Which is... Kindda sad. And makes the shutoff valve an expensive paperweight for most design.

But thanks for your reply.  It made me challenge the assumption I had that a packet was "committed" to the output that was "available" to it at the moment it entered the input tile.  It's apparently not so.

[JRup]

15 minutes ago, Mastermindx said:

And makes the shutoff valve an expensive paperweight for most design.

Older AT loop designs did include the shutoff to make it function as a bypass. But because the AT can be controlled by automation what is actually done is that the liquid thermo sensor is usually located right before the AT and a bypass bridge arrangement is used to keep the loop running.

Actual applications for the shutoff are different for the most part, but each design will have its own quirks and nobody says it shouldn't be used for an AT loop.

Because the AT does not receive anything if the automation disables it: Let's think of it having an internal shutoff of its own. Filling the loop is usually done with the AT off. I'll leave some pictures in the spoiler along with the link to the cheat sheet...

Spoiler

Single bypass loop... This will leave an empty spot in the pipes when the AT is turned on.

 

A double bypass loop accounts for that (Not necessary if a reservoir is in the loop.)

Automation... Filter is optional.

A good read for building AT bypasses:

 

 

38 minutes ago, Mastermindx said:

But thanks for your reply.  It made me challenge the assumption I had that a packet was "committed" to the output that was "available" to it at the moment it entered the input tile.  It's apparently not so.

That's ONI for you... It does take a while to "deactivate" common sense assumption mode when imagining a build, the game still packs a bunch of fun though.

Glad it helped.

[Yobbo]

4 hours ago, Mastermindx said:

Based on this,  I'm tempted to believe it's not possible to make a reliable automation with the shutoff valve based on pipe content if any output can back up, even for just 1 tick. Which is... Kindda sad. And makes the shutoff valve an expensive paperweight for most design.

You can do this if you ensure that the liquid passing the shutoff never stops flowing. This can be achieved with a circular loop, which you feed into with a bridge. If the exit of the loop (shutoff and/or bypass depending on what you are trying to achieve) backs up and blocks, the loop simply cycles forever. This causes the input bridge to also block and back up. So long as the liquid in the loop keeps cycling, the shutoff valve will continue functioning perfectly.

Something like this:

[degr]

I suggest to deconstruct mr Chilly, no matter how close you are. Reason is energy consumption rate - you need much more efforts and energy to chill water. Instead of it chill oxygen. You may build radiant liquid pipes, radiant air pipes and copper tiles, and chill oxygen from 95C to 22C with passing it through 2 tiles of heat exchanger (only one loop).

 

Spoiler

If you still want to chill water, in such schemas I prefer to use water tank and AT bypass. How it work - water going straight to AT, and cold water going into watertank. Watertank should be placed on door, so, if there is less then 500kg door is open, and watertank can only "consume" water. Place termosensor on first pipe after watertank output, and if temperature will be too hot, enable AT. With such design, you will have water with temperature +-3 degrees. If water cold enough, AT is disabled by automation, and AT bypass will just do it work. if you still have too hot water, send it into loop again, without any weird pipes crossing.