The largest stable mark 1 i ever did was 1344 Hu, however i would prefer the more efficient 1280 hu design i made after that. i did a few mark 5 and mox aswell. The largest mox i did was at 1000 Hu with an absolutely insane efficiency (but still quite low output) i kept it between 60-75% heat. i also did a regular reactor as a mark 5 at about 1360 hu but im quite certain i could achieve much higher numbers if i spent some more time on it. just got a brilliant idea for a quite funny 5x5 design which should have extremly high efficiency and output. If things work like i think this will be an amazing design compared to what i have done previously.
There we go, this is just a theoretical build that im gonna try out soon but it should work out due to the 5x5 still cooling even while its off. The basic concent is to use the most efficient cooling available 20 venting per spot in even numbers of 100 for superheated steam. it takes about 1 minutes for the overclocked heat vents to be destroyed so as long as you make sure that their total amount of heat they can hold does not get exceeded your fine. You turn on the reactor, let it accumulate heat for a while. shut it off and let the vents remove all the heat, then once its empty again you turn it on again. As long as the cycle time is under the time it requires to overheat the overclocked heat vents it should never destroy them. So basically you use the overclocked heat vents as storage for heat which continue to produce power while the reactor is off.
I'm sure one could improve a lot on this design but it should give a indication of the basic idea.
You run it for 10 seconds, then you let it cool for another 19 seconds. this means it will produce 23200 heat (most of which is stored in the overclocked heat vents) which will dissapate over 29 seconds at which point you turn it on for another 10 seconds and so on. The issue might be that heat accumulates in some components over time which might become a problem. gonna do some practical tests tonight to see if it works out.preferably you do this with a decent buffer of hot coolant to smooth out any variation.
Should produce about 1600 hu or 1200 eu/tick at an insane efficiency. as the design puts out 4640*0,75= 3480 eu/tick (if superheated steam) over 20000s this should be an efficiency of about 1392 million eu per cycle or an efficiency of 43,5 (43,5 million eu per fuel rod).
This is the biggest one i could make, not sure if i would actually wanna use it due to the reflectors but it should produce 1350 eu/tick
The cycle of this one would be difficult thought so not sure if it works in indefinetly as you would either have left over heat or have short periods of not enough heat to keep the superheated steam going.