You can build a perfectly scaled condenser, but it won't function in a perfectly scaled proportional way. As you scale things down, individual dimensions shrink in a lineal fashion, but surface areas shrink as a square root function and volumes shrink as a cube root function, so that a 1/10th scale model of something has 1/10th the length, width and height, but it has 1/100th the surface area and 1/1000 the volume. So there really is no way it can function the same as its 1/1 prototype.
The best analogy I can give is using the concept of a model sailboat. If it is 1/10th scale then the mast, waterline length, beam and keel depth are all 1/10th as long as on the prototype, but the sail area is 1/100th the size of the prototype and the displacement is 1/1000th the volume (thus mass) of the original full scale boat. So what has happened in scaling it is that the sail area is now ten times greater in proportion to the displacement, and the boat won't sail without falling on its side in most any wind, unless you put a false keel extension on it that goes much deeper with the weight available (center of mass) much lower than a true scale proportion of the full sized original.
Please note that I'm not saying that a scale condenser cannot work, as they certainly can and do, just saying that it cannot work at the same level of efficiency due to the proportionality of the different scale effects. If you want a small scale steam engine to have a fully functional condenser, then the condenser will need to be proportionately larger than the actual scale of the model by some factor, and / or have the steam and water both be colder (cooler steam is likely anyway), less steam flow to water flow ratio, or some compromise such as these to achieve the desired result.
That's my take on it anyway, though I've not tried it myself nor made any experiments along these lines, so these thought are just theoretical and not empirical.