I use Redding Type S Match Die sets. If I need an additional sizing die, I will just buy one separately - no need for multiple seating dies for the same cartridge.
Initially, I used the Type S bushing dies as the sole sizing step in my brass prep method with un-turned necks. I remove the decapping pin and expander ball from all my resizing dies prior to use. I found that one bushing size seemed to work pretty well with a given brand of brass, even across several different Lot #s. For example, I used a 0.248" bushing with Lapua .223 Rem brass and a 0.336" bushing with Lapua Palma .308 Win brass, both of which yielded very close to .002" neck tension (interference fit) in my hands. This approach worked quite well for me for a number of years.
Eventually, I came to the conclusion that my approach left something to be desired in terms of generating consistent neck tension, likely because I do not turn necks. This notion was supported by the observation that a small percentage of cases in every prep would exhibit noticeably different seating force than the majority when seating bullets. For that reason, I decided to add a mandrel sizing step as the 2nd (final) sizing step in the process. The premise is that sizing necks from the inside with a mandrel tends to transfer neck wall thickness variance/inconsistency to the outside, whereas a bushing die will move such inconsistency to the inside.
During testing with various bushing and mandrel diameters, I found that using a mandrel diameter of ~.0015" under bullet diameter will generate a neck tension of pretty close to .002". That is because spring-back works in the opposite direction when opening necks up with a mandrel than it does when squeezing them down with a bushing. With a bushing die, a neck opens up slightly (~.0005") when it is withdrawn from the die. In contrast, a neck will close up slightly (~.0005") when the mandrel is withdrawn. However, it is necessary to use a bushing of ~.001" or so smaller than would be used if the bushing die were to be the sole sizing step when a subsequent mandrel sizing step is incorporated into the process. This is so that the mandrel will actually do work on all the case necks, something that may not be very consistent if the necks are too close in size to what the mandrel will leave them. Thus, I would use a 0.247" bushing with Lapua .223 Rem brass and a 0.335" bushing with Lapua Palma .308 Win brass (unturned necks) prior to sizing with a mandrel (0.2225" or 0.3065", respectively).
It is also worth noting that bushings do not always generate a final neck diameter that exactly matches the value stamped on them. It is a good idea to have a few different sizes in .001" increments within a useful range for a given cartridge/caliber. Likewise, having a few different mandrel sizes can be useful. I use the 21st Century mandrels, which come in .0005" increments.
Whether one chooses to use a bushing die only, or a bushing die followed by a mandrel can only be decided by the end user, and their intended level of precision should be part of that decision-making process. Adding the mandrel step effectively doubles the amount of work/time required to resize the brass, so the use of a mandrel may not be desirable in situations where any small improvement in precision may not be noticeable. I believe that switching to a bushing die followed by a mandrel has led to a slight but noticeable improvement in neck tension consistency, which in turn translates to slightly better ES/SD values. The difference is not huge, but it is not zero, either. IMO - even little improvements in consistency can potentially help one improve their scores in F-Class matches, even though it can be very difficult if not impossible to statistically document such improvements. For that reason, I believe that answering the "bushing die only versus bushing die plus mandrel" question may need done primarily on faith, rather than a statistical guarantee, which may not be forthcoming.
