Circuit signals miscalculated during brownout
Posted: Thu Oct 22, 2020 4:46 pm
I've very new to Factorio but hit on the idea of sushi belts to have a single belt containing multiple items. It kept jamming with items of a particular type so a friend suggested I used circuits to track the belt contents, for which I found a description online (arthmetic combinator memory cell etc.) All items going onto or off the belt are counted, and items are only put on the belt if there are fewer than the desired amount on the belt. Simples.
The problem was that the belt kept emptying but the circuit thinks the items are on the belt. See the screenshot below. The outer belt thinks it has 20 copper wires on it (for example) but it actually only has one.
After being stuck of four days, and almost giving up on the game, I finally worked out the answer. I had checked my logic and connections over and over and over again, but it turns out that I hadn't been watching my power levels and they were flickering at the edge of insufficient power (brownout). It seems that circuit signal counting becomes unreliable on brownout.
Now my mate (who is an experienced player) said that I should have spotted the brownout because things slow down, but I'm only starting in the game and didn't even know that things did slow down and because I was right on the edge of brownout it wasn't at all obvious. The only sign I had was the circuit counting no longer worked!
My friend then argued that this was realistic as it's exactly the kind of problem that happens in the real world, I agreed but pointed out that it's a game! I shouldn't have to spend 4 days using real world debugging skills to work out that the reason my circuits can't count properly is that my power is low.
Ideally the power levels shouldn't affect the counting, but if that's impossible it would be nice to have a graphical indication that there is a brownout e.g. a flashing brown exclamation mark over factories and inserters etc, rather than just subtly breaking the circuit logic!
Part of the reason for this bug report is in case it helps someone with same problem in the future.
Screenshot below. I've also attached a savegame. Turn the power switch beside the charactor on and watch the copper wires on the belt to the left for about 10-20 minutes.
The problem was that the belt kept emptying but the circuit thinks the items are on the belt. See the screenshot below. The outer belt thinks it has 20 copper wires on it (for example) but it actually only has one.
After being stuck of four days, and almost giving up on the game, I finally worked out the answer. I had checked my logic and connections over and over and over again, but it turns out that I hadn't been watching my power levels and they were flickering at the edge of insufficient power (brownout). It seems that circuit signal counting becomes unreliable on brownout.
Now my mate (who is an experienced player) said that I should have spotted the brownout because things slow down, but I'm only starting in the game and didn't even know that things did slow down and because I was right on the edge of brownout it wasn't at all obvious. The only sign I had was the circuit counting no longer worked!
My friend then argued that this was realistic as it's exactly the kind of problem that happens in the real world, I agreed but pointed out that it's a game! I shouldn't have to spend 4 days using real world debugging skills to work out that the reason my circuits can't count properly is that my power is low.
Ideally the power levels shouldn't affect the counting, but if that's impossible it would be nice to have a graphical indication that there is a brownout e.g. a flashing brown exclamation mark over factories and inserters etc, rather than just subtly breaking the circuit logic!
Part of the reason for this bug report is in case it helps someone with same problem in the future.
Screenshot below. I've also attached a savegame. Turn the power switch beside the charactor on and watch the copper wires on the belt to the left for about 10-20 minutes.
- belt_count_error.png (3.96 MiB) Viewed 804 times