Bridges massively reducing throughput

Keep an eye on the electrolyzer. I suspect you are over pressuring it with hydrogen.

The pump has the input cell at the bottom left and while it can input from neighboring cells, you do mainly pump oxygen. The hydrogen will then accumulate at the top. If the high pressure hydrogen ends up on the electrolyzer, then it will stop. If it stops each time hydrogen is released, it will have a start-stop motion, meaning it only runs half the time.

The electrolyzer consumes 1 kg/s water and outputs 888 g/s oxygen. This means 88% of the mass returned as oxygen, which for one cycle is 528 kg (you mentioned 532. Rounding issue?). Now assuming the electrolyzer goes into the start-stop motion due to high pressure hydrogen, you will not consume 600 kg of water during a cycle, but only 300. That will give you 264 kg oxygen for one cycle. That's fairly close to the 240 kg you encountered.

Move the pumps up to make them take in hydrogen too. One should be enough because 500 g/s every 4th second is 125 g/s on average, which is 13 g/s more than the electrolyzer produce when running all the time.

 

Also the bridge setup isn't ideal. The gas will queue up for the merge and when it queues up on the pump output, the pump stops. An alternative merge could be like in the added screenshot. This will grant priority to one pump (the left) and the right pump will try to merge the gas dots into what the left already pumps. If the merging point backs up, there is much more pipe to fill before the pump stops due to a full pipe.

I'm not saying it's the perfect layout, but it should be a layout, which is less likely to cause the low production issue you have encountered.

Of course the electrolyzer is getting over-pressurized, I thought that was clear from my post. As I said, bridges now cause elements to pause resulting in reduced throughput. Your entire first paragraph I already referenced: "Of course a sealed room isn't going to provide 100% throughput due to pressure varying from tile to tile". The issue is not the occasional interruption of the electrolyzer, which has always existed even before this change. Pumps now register the "Pipe blocked" error ~50% of the time. I find it odd you didn't actually try to run your setup, merely screenshotting empty pipes. If you had you would've realized the only thing you changed is the ratio of time each pump will run. In my build they each run an equal amount of time, in your build the right pump will run more and the left less. You've done nothing to address element flow pausing at the merge. 

I'm not sure why you're trying to critique some very simple math. The game states electrolyzers produce 888g/s of oxygen. A cycle is 600 seconds. 888 * 600 = 532,800g. Convert to kg and you get 532.8kg. 

1 hour ago, Nightinggale said:

If the merging point backs up, there is much more pipe to fill before the pump stops due to a full pipe.

Not if, when. It doesn't matter if you provide the entire width of the map as excess pipe. The only thing you've done is delay the pipe becoming full by less than a minute. Actually run the design and feel free to come back with the results. 

5 minutes ago, situpc said:

It doesn't matter if you provide the entire width of the map as excess pipe. The only thing you've done is delay the pipe becoming full by less than a minute.

It may be caused by the bridge just after the output of the left pump.

Bridges don't really like to merge two pipes just after an output.

 

Try to merge the two ouputs like this :

 

It will not be perfect, because the pumps output 2 gases in the same pipe, but it should be a bit better.

If you really need a fluid flow, then the only way is to set up 2 filters : one for each pump.

 

So it's not really a bug, because it's essentially due to a pipe with two different gases inside it.  

1 minute ago, Gwido said:

 

 

A T connection is exactly the behavior the bridge is mimicking, why would you think that's a solution? If you don't know how bridges worked before and after the pipe optimization changes why are commenting about them in a bug forum?

2 minutes ago, situpc said:

A T connection is exactly the behavior the bridge is mimicking, why would you think that's a solution? If you don't know how bridges worked before and after the pipe optimization changes why are commenting about them in a bug forum?

A T section means gas will wait to enter the "merging cell". This will create a queue to enter the merged pipe. If the queue reach the pump output cell, then the pump stops. By placing at least one cell between the merging point and the pump, the queue will normally not reach far enough back to stop the pump.

12 minutes ago, Nightinggale said:

the queue will normally not reach far enough back to stop the pump.

Your build. Pipe is permanently backed up just like I said it would be. Now I know why you uploaded a screenshot of empty pipes.

The problem is that different gases can't merge. If you add a filter for each pump and then merge the output of the filters, then you have just one type of gas in each merge, hence no problem merging.

13 minutes ago, situpc said:

Now I know why you uploaded a screenshot of empty pipes.

The real answer to that one is it was faster. I could have just given you something I have used myself (a very different design with a much different room shape), but I'm not going to tell people how to play. Instead I decided to try to adjust your design into something which could solve the issue. In fact it did, but apparently it caused a different issue.

45 minutes ago, situpc said:

If you don't know how bridges worked before and after the pipe optimization changes why are commenting about them in a bug forum?

Are you saying that your setup worked before the pipe optimization? If so, then that's something very important, which you didn't include in your bug report. I would however be somewhat surprised if you could manage to get the original setup to work with full efficiency with any version of ONI, but I'm not going to spend the time testing old versions. It might have worked in the past with a single type of gas, but unlikely with multiple gas types.

Not only did you have two people replying, both people spent enough time to make screenshots for you. However based on your replies you are openly not interested. Now you can get it like you want. I won't write another thing about this bug report.

And in case anybody is interested: I just looked up the source code for gas bridges. It look very much like it did in QOL3. Sure ConduitFlow is optimized and might send gas around the pipes in a different way, but the bridge building component (liquid and gas bridges) itself appears to be unchanged. The old rules regarding how pipes can merge with various bridge setups should still be valid.

• "Join" intersections for conduits (where more than one pipe tries adding to the intersection) will now merge similar elements if possible, resulting in much better throughput.

Maybe this entry in the newly released patch will fix your original problem. Interestingly that was what I was aiming for with the bridge merge. Also this has nothing to do with the bridge in the first post. It's the merge of two pipes.

The layout, which is most likely to improve the most from this change is this one:

Adding a cell on each side of the merge point will make the system much less likely to stall if hydrogen and oxygen tries to enter at the same time.

6 hours ago, Gwido said:

 

10 hours ago, Nightinggale said:

I could have just given you something I have used myself (a very different design with a much different room shape), but I'm not going to tell people how to play. Instead I decided to try to adjust your design into something which could solve the issue. In fact it did, but apparently it caused a different issue.

Are you saying that your setup worked before the pipe optimization? If so, then that's something very important, which you didn't include in your bug report. I would however be somewhat surprised if you could manage to get the original setup to work with full efficiency with any version of ONI, but I'm not going to spend the time testing old versions. It might have worked in the past with a single type of gas, but unlikely with multiple gas types.

Not only did you have two people replying, both people spent enough time to make screenshots for you. However based on your replies you are openly not interested. Now you can get it like you want. I won't write another thing about this bug report.

And in case anybody is interested: I just looked up the source code for gas bridges. It look very much like it did in QOL3. Sure ConduitFlow is optimized and might send gas around the pipes in a different way, but the bridge building component (liquid and gas bridges) itself appears to be unchanged. The old rules regarding how pipes can merge with various bridge setups should still be valid.

Please don't as this is a bug reporting forum, no one here is seeking your unsolicited advice. People reporting bugs shouldn't be subjected to you in every thread saying "I don't know about the bug you're experiencing but want to see my SPOM build?" Now you're claiming you solved the bug when your build demonstrates the exact same thing?

I don't know if you're being purposely obtuse but "Bridges now cause" and "There have been other reports of this as early as 7/10/19" clearly tell you when this bug started. And yet again, I never claimed to have full efficiency. Allow me to quote this same sentence again: "Of course a sealed room isn't going to provide 100% throughput due to pressure varying from tile to tile". 

I think you're confused about the purpose of a bug reporting forum. This isn't general discussion. This is where players bring potential bugs to dev attention. Not for wannabe ONI devs to fail at diagnosing them. 

5 hours ago, Nightinggale said:

• "Join" intersections for conduits (where more than one pipe tries adding to the intersection) will now merge similar elements if possible, resulting in much better throughput.

Maybe this entry in the newly released patch will fix your original problem. Interestingly that was what I was aiming for with the bridge merge. Also this has nothing to do with the bridge in the first post. It's the merge of two pipes.

After you just got done saying you were done posting here and nothing changed with bridge merges you can't help yourself but try to defend how clearly wrong you were. According to you this has nothing to do with bridges, how would this fix bug? Is it perhaps you're not an ONI dev and have no idea about the bug you're trying to diagnose? Didn't know when it started, didn't know about the multitude of other reports, didn't know what causes, didn't know how to solve it, didn't even test a supposed work-around,...maybe let the actual ONI devs handle the bugs? 

By the way, the new update increased throughput to ~80%. So much for having nothing to do with bridges, right? Now radiant pipes aren't working properly, I'm sure that also isn't a bug according to you.