How to regulate aquatuner with liquid temp sensor - output pipe too cold

[justameep]

Ive read a number of guides on the forums here but I find that still, very occasionally, my aquatuner output pipe will break because of "contents too cold".

What is wrong with my liquid temp sensor setup?

I have explicitly constructed a setup where the overflow pipe should never back up because it always can empty into a liquid reservoir (and there's not enough liquid in the system to completely fill a liquid reservoir). But, somehow that doesnt appear to be working as you can see in the screenshot where there's a -16C packet of crude oil right after a valve and temp sensor set for "above 10C"

Please help!

I loaded up my autosave to the cycle just before it failed, but when I watched it run for a whole cycle on a slower speed it never encountered the problem (maybe the saves are not deterministic?).

[ExEvolution]

So to regulate the Aquatuner, you should place a liquid pipe thermo sensor set to Above x degrees, 1 tile before the input port of the AT, then run the pipe into the input as well as 1 tile above or below, and bridge from that last pipe segment into the output pipe. Run the automation wire from the liquid pipe thermo sensor into the AT. Then just make sure your piping completes a loop.

 

This will cause the sensor to turn on the AT if the liquid is above your threshold, and if its not above the threshold, it will turn it off, the bridge will act as a bypass for the aquatuner when the automation has turned it off

 

Unfortunately I don't have a good screenshot at the moment to share, but if you have questions I can help

[justameep]

hmm, that doesnt seem ideal. If I used a thermo sensor to turn on/off the aquatuner itself, then it could encounter situations where the AT is filled with hot liquid, but the thermo sensor last saw a cold packet and turned off the AT. I'd have to wait until another hot packet came along to turn the AT back on.

I really want to use the liquid valve with a thermo sensor so the AT can always run, but I dont understand why my setup sometimes (rarely) lets through a cold packet.

[ExEvolution]

No it works trust me, when the aquatuner sees a signal, it reacts on the next packet of liquid coming through. It is the same reason why you can make low powered filters using sensors and shutoff valves

[justameep]

hmm, but then shouldnt my liquid valve setup work? I think the approach is the same except ive replaced the AT with a valve. I have an temp sensor just before the valve input (which is like putting it before the AT input), and I also have an overflow pipe (similar to what you described in your AT setup)

[Steve8]

This works: https://youtu.be/OlzfMNGCb4E?t=907

Even the liquid tank is not strictly necessary. But it evens out the temperature of the incoming coolant. So you can run the system a bit closer to the freezing point and don't have to worry that a single packet is a bit too cold

 

[ExEvolution]

Its hard to tell from your screenshot, but even if your setup does work, it still uses more power and is more complicated than the setup I described. @Steve8 linked a video that shows what I was describing to an extent at 15:48, but his setup isn't using a thermo sensor because his setup only regulates the temperature by averaging it with a storage tank, but does not change the state of the AT

 

I used his video to draw what you would want to properly regulate your temperatures

 

Liquid Pipe Thermo Sensor on the red dot, insulated pipe on the orange lines

 

Oh, and by the way, unless you have a specific purpose for using oil as a coolant (such as its higher temperature ranges), you should instead use polluted water as your coolant. The AT always reduces the temperature of whatever goes through it by 14 degrees, no matter what it is, and polluted water has the highest heat capacity other than super coolant (and higher temperature ranges than pure water) meaning you move much more heat per run of the aquatuner, meaning you save power to get the same effect

 

Edit: After re-reviewing his setup, he does have a thermo sensor right after the tank, which would work since he's using the tank to average. If you don't use a tank, you want the sensor where I showed you

[Yunru]

Without something like an AND gate, the sensor and the shutoff will be a tick out of sync. Run your sensor to both inputs of an and gate, then the output of that to the shutoff. That way, the packet on the shutoff is the one that had it's temperature read, not the one preceding it. 

[justameep]

Wow, just to confirm you're saying that the automation signal has a "travel" time where it moves from one device (or gate) to the next every tick?

Then, if I buffer the signal with something (like an and gate as you recommended), then the sensor will be synced to controlling the valve?

*edit* im not sure how to reconcile that with all the guides that seem to use just a thermo sensor and valve, or just a element sensor valve, to do packet filtering

[ExEvolution]

What @Yunru is saying is that by placing an AND gate and having 1 signal be connected to both inputs, because the game has to process 1 tick to determine the state of the AND gate, the output is delayed by 1 game tick

[Yunru]

8 minutes ago, justameep said:

*edit* im not sure how to reconcile that with all the guides that seem to use just a thermo sensor and valve, or just a element sensor valve, to do packet filtering

I haven't seen such guides personally, but they probably just either wing it, or have a safety net large enough that accidentally cooling a packet once doesn't break pipes.

Alternatively, should you wish to use something smaller than an AND gate, use a NOT gate and just flip the greater than/less than on the sensor to compensate.

[Steve8]

1 hour ago, ExEvolution said:

but his setup isn't using a thermo sensor because his setup only regulates the temperature by averaging it with a storage tank, but does not change the state of the AT

As you noticed, it's there. He initially forgot it. So he added it above, which is not necessarily the best position

The tank alone wouldn't work for temperature control. You have to turn of the AT at some point. It just means that the packets will all be the same temperature. Without a tank you can get stray packets that are a bit colder than others. So you need to leave yourself with a bigger margin for the the temperature setting. Which can still work perfectly fine in many applications. A base cooling loop for example doesn't need really low temperatures, whereas you want a great deal of accuracy for liquid hydrogen.

[Cybeon]

From my experience the most reliable setup which allows pretty accurate control of the temperature is to use combination of AT bypass, liquid pipe thermosensor and liquid tank

1 Bypass ensures that even if AT is stopped liquid continues to flow

2 Sensor sets desired temperature

3 liquid tank contains some amount of extra coolant (I prefer to use half full tank, or at least 1t of coolant). It evens out spikes and provides steady output. Senor is set to desired temperature of output, but not less than 14 degree above coolant freezing

 

 

 

[KittenIsAGeek]

3 hours ago, Cybeon said:

From my experience the most reliable setup which allows pretty accurate control of the temperature is to use combination of AT bypass, liquid pipe thermosensor and liquid tank

Yes, I do this as well.   I can get a lot closer to the freezing point of my coolant without causing problems. with my piping.  Generally I run the output of the AT directly into the buffer tank, and the output from the buffer tank goes to whatever I need cooled before returning through the AT.  The build you pictured is close to what I do, except that I put the thermal sensor one pipe segment before the AT.  Either way works.

[justameep]

5 hours ago, Yunru said:

Without something like an AND gate, the sensor and the shutoff will be a tick out of sync. Run your sensor to both inputs of an and gate, then the output of that to the shutoff. That way, the packet on the shutoff is the one that had it's temperature read, not the one preceding it. 

I tried this, but after another 30 or so cycles I hit the same problem (output pipe too cold). In the screenshots you can see that somehow a very cold packet has gone through the filter.

I followed the advice of using a not gate to buffer the liquid themo sensor output one tick for the valve

As a note, I've been playing this game mostly with debug mode's ultra game speed (ctrl-u). Are there any known liquid behavior issues related to that?

What's really frustrating is every time I see this, I load up my autosave just before the problem and watch it run. But, ive never been able to see the problem - I watch for a few cycles, but it always works when im watching it. Then like 30 cycles later the problem will reoccur.

[babba]

8 hours ago, justameep said:

Ive read a number of guides on the forums here but I find that still, very occasionally, my aquatuner output pipe will break because of "contents too cold".

What is wrong with my liquid temp sensor setup?

I have explicitly constructed a setup where the overflow pipe should never back up because it always can empty into a liquid reservoir (and there's not enough liquid in the system to completely fill a liquid reservoir). But, somehow that doesnt appear to be working as you can see in the screenshot where there's a -16C packet of crude oil right after a valve and temp sensor set for "above 10C"

Please help!

I loaded up my autosave to the cycle just before it failed, but when I watched it run for a whole cycle on a slower speed it never encountered the problem (maybe the saves are not deterministic?).

On a little side note...

Try not to use bridges for pipes/gas as they leak a lot of temperature.

It requires building more pipeline stuff and takes more space as one avoids bridges, but its often worth it.

Just as a little tip

Also keep in mind that storage tanks temperature adjust their content with the environment.

 

Cant wait till the DLC is out !

[Cybeon]

19 hours ago, KittenIsAGeek said:

Yes, I do this as well.   I can get a lot closer to the freezing point of my coolant without causing problems. with my piping.  Generally I run the output of the AT directly into the buffer tank, and the output from the buffer tank goes to whatever I need cooled before returning through the AT.  The build you pictured is close to what I do, except that I put the thermal sensor one pipe segment before the AT.  Either way works.

Well it depends on the task you need to solve. If your task to cool as much as possible without breaking pipes then placing sensor right before AT is the best option. If you need precise control of the cooling loop temperature and your working temperature is more than 14 degrees above coolant freeze point (e.g. Liquid hydrogen production or any other operation which requires pretty narrow temperature range) then placing at the sensor right after tank output is the better way

[Yunru]

20 hours ago, justameep said:

As a note, I've been playing this game mostly with debug mode's ultra game speed (ctrl-u). Are there any known liquid behavior issues related to that?

IIRC there's quite a few problems that become more evident the faster the games runs, most noticeably with buffer and filter gates.

[ExEvolution]

I'd really just try it the way I recommended.

My base is currently on cycle 1100 and I've never had a single pipe break on my AT using the bypass bridge with the liquid pipe thermo sensor, and most of my coolers are set to cool down to up to -19c (thermo sensor set to -5) and I don't even use the tank method.