Here are my observations/experiences on this topic.
1) The Kadee semi-scale couplers (#58/etc.) look better.
This is not a small difference. When you look at a train
going by - even 6 feet away from you - you can spot the
cars that have the smaller couplers. Fairly easily.
And they are close enough to scale size that you don't
see the same "noticeable" difference between them and the
"true" scale couplers (the only ones I know of are the
2) They couple and un-couple - with themselves - reasonably
well. I'm talking about "intended" couplings/uncouplings
on straight level track.
3) They are more finicky about coupling/uncoupling than the
#5s (again intended" and straight level track). How
much more finicky (i.e. whether or not you care) is for
you to decide.
Much of the time it will be "no big deal". Every once
in a while it will be a considerable problem (see the
next item right below this one).
My experience is that they are noticeably more difficult
to uncouple compared to how well they couple. How much
more is a matter of personal opinion.
4) They are noticeably more finicky about coupling/uncoupling
on curves or grades (again "intended"). Even a slight
curve can make them noticeably more difficult than a #5.
It is my belief that the reason for this is that because
the overall coupler 'head' size is smaller that they
tend to 'miss' more often. And they seem to "lock with
each other" more than the #5s (a good thing when you are
going down the road but not so good when you are trying
to uncouple them).
5) They are more sensitive to any 'rapid vertical changes'
in the track ("vertical kinks") than the #5s. This
translates to more break-in-twos than you have with #5s.
Here's an interesting 'fact' ... go out and measure the
joint between two pieces of rail (in good repair) on
the real railroad nearest you. Measure how 'true' the
joint is in terms of the accuracy of both the height and
width of the rail. I suggest just taking a 12" straight
edge and holding it along the rail and look at how well
it lays along the entire 12" of the rule. Now go do the
same thing on your layout using the same ruler. Most of
the time you are going to discover that your layout is
essentially at the same accuracy as the real RR. Maybe.
You will probably find it is difficult to impossible to
find rail that you can really use to measure on the sides
(on the top is usually no problem).
But your layout is 87 times smaller. Said another way -
the real RR is 87 times "more accurate" than your layout is.
Many of the 'average' joints on your layout will be
measurably more "out of true" than the real RRs ever are.
And the worst joints you have - even the ones that you are
still willing to live with because they aren't causing any
problems - will be significantly more out of true when
compared to the real RRs.
Yes, careful attention to how you lay your track, and
how true it is, will improve things. But you are never
going to be as good as the real RRs. I've never done
any research on this but I suspect that the acceptable
variations when real rail is manufactured (the plus or
minus n thousandths) is probably the same or even smaller
than our scale rails!
Said another way - in order to get our rail to be
truly accurate in the vertical dimension - we should be
using styrene or paper shims under one rail for most (all?)
of our track joints. I don't know any layouts that have
been built this way. I'm not even suggesting that we
should be doing this. I'm just saying that if we want to
be as accurate as the real RRs are we would have to do
6) They are much less tolerant of any mis-match in coupler
height between two cars. Not hard to explain/predict -
and not an "original" observation ... but still important
to state. Said another way ... the amount of difference
you were willing to tolerate with the #5s won't work with
the semi-scale couplers - standards must be strictly
enforced. Another aspect of this is that you hear
guys say all the time that they add a shim to the
coupler box to prevent them from drooping. Why Kadee
hasn't designed and marketed a box with this on it is
somewhat of a mystery to me ... but I suspect it is
related to their desire to market to installers who
just want to swap out the coupler (and not the box
7) They are noticeably more difficult to couple/uncouple
with other size couplers/manufacturers. If you are
using the Kadee semi-scale couplers on some cars, but
not all, you will have more problems than if you use
just one coupler make/size/design. Kadee says that
the semi-scale couplers are "compatible with" the #5s.
I seem to have a different standard than Kadee does
with respect to the term "compatible". What I see
happening "all the time" (often enough that I'm not
using the term "sometimes") is that when you are
coupling they won't couple (easily) ... and when you
are uncoupling that the two couplers will "hang up
on each other" and both will deflect to the side as
a unit. (This happens whether you are using magnets or
There are times when trying to couple/uncouple a
semi-scale with a scale coupler is significantly
more difficult and can reach the frustration level.
There are other times when it will be a 'little more
This is not a 'small difference' as far as I'm
concerned. If you have a mix of brands you are going
to have more problems than if you don't. Even more
so if you are using the semi-scale couplers.
I operate on a wide variety of layouts with very different
mixes of couplers (all #5s, all 58s, a mix of mostly 5s and
some 58s, a mix of mostly 58s and some 5s, etc.). And I go
to 2 or 3 (or more) op sessions a month. (Included this so
you will have some idea of what my 'background' is.)
I have to admit that my attitudes toward the semi-scale
couplers go up and down based upon my most recent op sessions.
If the last two or three sessions have been on RRs with stellar
track and strict car standards I tend to be more in favor of
the semi-scale couplers than if my last few sessions have been
on "regular" layouts.
When I go to a "prototype meet" ... all I see are the
semi-scale couplers. No surprise - they look better and
the emphasis at these meets is on 'how accurately the models
re-produce the prototype'.
I have to say that I have similar attitudes/experiences
with respect to the same topic for wheels.
These are my observations/experiences. The thing -you- are
going to have to do is to establish for your self whether or
not you will use the semi-scale couplers - and if you do what
it will take to get operational characteristics that are
acceptable to you. I suggest that you do your own tests -
on the layout nearest you with the best track you can find -
and make up your own mind about what you consider acceptable.
And don't forget to consider how it looks as part of the
And if you do decide that they are your standard - it's
much better to "do them all" ...