All gear drive does not improve the pulling power of a steam locomotive, it is just adding more gears and therefore more complication for no added benefit. Rod couplings do exactly the same job and have to be there for prototypical appearance anyway. And there is an additional risk that the gears can fight the rods if the mesh is not perfect. As Terry points out, gear backlash means that the rods are already doing all the work anyway.
Driving through to the tender would add little to the performance either, as the additional losses to the motor drive train would absorb any benefits gained. Tender drives are also problematical, particularly as these are bogie tenders, so you would need the equivalent of a BO-BO drive in the tender. Imagine what that would add to the cost. And it is not likely to improve performance much as you cannot fit much weight in the tender for adhesion when it is full of decoders, speakers and such.
And in answer to Terry, the SEM R Class model that weighs in at 400 grams has been working just fine for many years, the weight is pretty ideal for obtaining good adhesion, as the model can pull prototypical length trains. There is little penalty in having that sort of weight in a model, given properly designed bearings and a carefully selected motor and gearbox.
I am personally not able to dispute the aspect regarding a steam loco fitted with gears to all wheels as per the TOR 32cl, & proposed with future Eureka models, simply because I had thought the gears would have been an improvement for small locomotives that we have here in Oz, such models like the 32, standard goods engines & several of the VR steamers are the candidates for such a set up.
My understanding is that the all geared system in the 32cl was suggested by the then engineer at SDK who was charged with its production, it was part of the reason why TOR at the time also accepted the move to metal for the model owing to the very light weight of them, therefore presenting problems with pulling power. I understand that the same engineer is in charge of the factory that is now used by Eureka as well as Austrains, so its likely a reason why the K & other models are heading down that direction.
I made the comment re the tender more as tongue in cheek as others have suggested this idea in the past when similar discussions have arose regarding the light weight models & pulling power. For me the biggest blite on the tender motors is when the drive shaft is seen going through the cab to the engine, just something the modeller has to accept was the inference.
As we look into the prototypical length loads & other aspects, its once again a problem for the modeller, particularly for the steam era modeller, as its one thing to quote the prototypical nature of a load but are we talking empty or loaded wagons? Likewise whether or not the model is working on what grade & curve conditions.
Its one thing when we look at open wagons with nothing in them & easy to say its empty, but when we then consider the grade of the layout as well as the radius of the curve it brings an equation into that basically defies the prototype. The prototype also had not just the load of the train as part of its make up, but also was based on the prevailing track conditions especially the curves over the ruling grades.
In most cases we operate our models with closed type vehicles & consider them loaded, simply a result that any steam model would not pull a prototypical load & length load up a ruling grade that the real thing did. Using a Garratt as an example on a 1:40grade, & using the TOR BWH hoppers as an example which are very good rolling models or even try it with BCH's of the same brand. The load for a garratt, was dependant on which actual 1:40 it was working on so going for the best load being on the Short North from Hornsby it was 650tonnes. That load is more than the 615 tonnes for 29 hoppers & van, but I could not see any HO model of a garratt pulling that many of the hoppers on a full 1:40 grade, doubtfull if it would even take the equivalent length load in S wagons.
Usually as a modeller we run very light loads as close to the prototype as possible & when we come to grades, even the ruling or heaviest grade on a layout, would have much tighter curves than the prototype, & this is where all the importers have to bring in compromises, & many of them do not sit well with the end user.
Looking at the Rcl, I would say in the prototype version would have not had great pulling power for goods trains, which was the same with the 38cl, even on passenger trains there would have been limitations based on the grade & also I would think on the curvature & other track conditions. As such a model needs every little bit of an advantage the engineer in charge of production can wring out of it. In the same way, the importer has a real need to listen to him but also have people on his team that can work that sort of thing out with the engineer, & advise accordingly the best option for the importer.