Predicting cycle times is difficult. This is because the theoretical cycle time assumes 100% efficiency. You never reach 100% efficiency because you have to move cars on a conveyor, you have to sometimes wait for parts and sometimes you have to wait for cars. All of the waiting is waste.
The conveyor movement can be up to 2-10% of the downtime, and waiting for parts can be a huge downtime creator, I’ve seen as much as 75% in Production Line. Of least concern in a well balanced line is the time spent waiting for the next unit, which could be significant but is also a waste because you never get the time back.
Even if you had 100% up-time, if your cycle time to complete that step was 20 seconds and all previous steps were 18 seconds you’d still have a bottleneck there. If you had a cycle time at that step of 20 seconds but all prior steps had a cycle time of 40 seconds, then you’d have a massive inefficiency while that step waits for work.
If I understand your goals right to have a tact/cycle time of 20 seconds then all you have to do is build enough flexibility in to your design both on the floor space utilization to allow changes to accommodate the different cycle times of each component slot along the path, and with regards to the number of pathways to completion of the operation.
Does that make sense?
Depending on the size of your available floor space, when you setup your initial layout try, to leave yourself with some inventory space between your slots, say 10-20 conveyor spaces, if possible. In the beginning you have slots which have the greatest cycle time, some of them will take upwards of 30 minutes even at 100% efficiency.
This can help eliminate bottlenecks, which is where cars pile up behind a particular slot.
If you use two slots to perform these operations that are inefficient or have a high cycle time, you will cut the time of the processing time down by as much as 50%. I don’t say exactly 50% or more than 50% because you are subject to the wastes for each slot, waiting on parts, waiting for cars, and waiting to move the completed car. As your production proceeds, you will of course have some slots which process faster than your average cycle time. The whole process will be organic and dynamic, in that things are dependent on both the up-steam and down-steam conditions.
Once you have a layout that you think is efficient, watch or time-study each station, to determine what is actually occurring to understand where your bottlenecks occur and where you need to redeploy the floor space. In the beginning my chassis assembler slots massively outpace my fit body slots. Requiring 3-4 fit body slots to keep pace. I’ll build these in parallel, and as I research better fit body assembly I’ll be able to re-deploy my floor space to process faster, even if I am not achieving the same uptime at each step or runtime, simply because the cycle times are lower because the process is broken into specialty steps.