.

[randgust]

Having come from HO into N a long time ago, I think the HO guys have it easier.    Everybody understands what RP25 means and has for a while.  Despite 50 years, there is a continuing evolution in N toward higher standards, which may or may not actually work for you. 

It's still a hot-button topic because, bluntly, what looks good doesn't necessarily run good.   Wheels and track are not entirely friendly to each other.   So hitting that sweet spot of getting that personal compromise of form vs. function is a very individual decision, hotly defended.

Lets just pick Micro-Trains.   Back in the 70's they were the 'fine scale' stuff, today the same material is considered mid-range to low-grade, depending on which generation of wheels are under them, as they have changed 'standard' flanges three times.   The '70's "Rapido Era" wheels are pretty much gone except for legacy stuff; high-quality metal wheels are now available, but they may or may not track OK in 70's era track.   Atlas Code 55 looks great, but original vintage N won't run on it because there's not enough flange clearance against spikeheads.    Peco Code 55 has all kinds of inside flange clearance but oversize ties..... so the rail/tie/old/new thing ain't over.   And it's not just flange depth either, tread width issues on new wheels can drop right down in a frog and stay there, too, in some cases.   But at least the potential for much better appearance exists, and it didn't back in the old days.

Coming from HO, I was actually astounded how good N was, once I standardized on track and wheels and checked gauge.   I never beat totally beat random derailments in HO, I actually have in N.  And I've still been to a lot of layouts in HO where derailments are relatively commonplace, much more so than I would accept.    So I'm what I consider 'operationally spoiled' to a zero-derailments mentality in N.  No matter how good it looks, if it derails, random uncouples, or hangs up in switch frogs, fix or sell.   I'm 48-52" above the floor with long trains.   A derailment can send a car to the concrete floor, so it just isn't tolerated.

Having tested a lot of wheels, couplers and rail combinations - with a heavy emphasis on operations over appearance.... here's what I've evolved into:

General operating specs:  13" min mainline radius (most 15"), 11" on some hidden track.   Maximum grades 2.5%    Typical train length 25-35 cars. 

Track:  Peco C55, electrofrog switches.   Lots of legacy Atlas C80 on older work.
Trucks/Couplers:   Generally, Microtrains, with either deep or current mid-range flanges.   Low-profile was too low.   Metal wheels, even with low profiles work OK as long as sufficient treadwidth.   MT couplers are used as intended with magnetic uncoupling and shortened in-track magnets.  No skewers.   'mostly' truck mounts but new equipment with body mounts generally OK.   
Exception:   Z couplers work amazingly well even mixed into general service mixed with MT N, if they fit.  Standard on lead units for appearance.
Exception:   Accumates generally get swapped out due to appearance or in any situation where magnetic uncoupling necessary.
Exception:   Bodymounted couplers with short shanks on long cars don't work well at all

Given that set of specs, I've been able to run just about everything except one BLMA car that had to have narrow wheels swapped out, Atlas 89' piggybacks due to coupler/shank issues, and mixing truck-mount and body-mount standard passenger cars.   So the 'absolutes' on things aren't really as bad as you'd think, but the preferences are another matter.   So while there's a lot of opinions here, most are at least based on experience, and frankly, I've found more wiggle room than I'd think based upon what you read.   But yeah, there's still stuff all over the map in N.

[robert3985]

ALL!

 Since this forum is the largest gathering of N scalers on planet Earth, I decided to post my
questions here.    I found some inexpensive ExactRail box cars secondhand and I want to
give N scale weathering a try...for the first time. The wheels look pretty decent but the
treads are very wide. Does N scale have semi-scale metal wheels and, if so, are they good or
bad with typical operations? The couplers look HUGE. Does N scale have semi-scale or
scale couplers? Don't some of you use Z scale couplers? If so, how do those work with
other cars that have regular N scale couplers? Please advise. These look like the two biggest
roadblocks for realistic box cars...to my untrained eyes.

 Also, if I'm making an N scale diorama to display the finished projects, what code / brand
track would you recommend? Regarding ballast, do you guys use N scale or Z scale?

 THANKS for any input.

Jeff (MrKLUKE)   

Jeff,

I replace all of my MT N-scale couplers with MT Z-scale couplers when possible.  Since (as has been previously stated) there is only one mounting option, sometimes it's not possible to replace the oversized N-scale couplers with MT 903's.

Although the Bowser coupler looks and operates okay too, it looks a bit oversized to me.  I could be wrong as I haven't bought any yet...HOWEVER, several years ago, I measured an actual coupler at the Utah Railroad Museum at the Ogden Union Station and dimensionally, the MT 903's are within a few thousands in five of six measurements, although their appearance doesn't look much like a prototype coupler. This means that size wise, they are extremely close to being the right size for N-scale...much closer than anything else out there with the possible exception of the Bowser coupler.  Additionally, they are readily available.

Yes, they do the "slinky effect", especially when switching, and an easy solution to that is to install one of the MT "friction springs" on every car.  This works okay for me, but I'm running trains with a maximum length of 30 cars to fit into my 9ft 7.5in passing sidings.

Here's a photo of an MT 903 coupler installed on the rear of one of my Key UP Light Mikados while coupled to the pilot coupler which is a cast, prototype sized coupler...

Photo (1) - Key UP Light Mikes double headed using MT 903 and cast-on scale-sized pilot coupler:


Photo (2) - Same models, but uncoupled for a better look:


Although I painted the couplers in these two photos a rusty color, after observing real trains and looking at photos of cars and engines in my time period (1947 thru 1956) I am now lightly brushing or airbrushing my couplers a warm medium gray, and wheels a slightly darker, browner gray along with their axles and backsides.

Photo (3) - MT 903 on UP CA-8 painted Pollyscale "Old Concrete" edited "Aged Concrete":


Photo (4) - MT 903 on UP CA-1 painted Pollyscale "Old Concrete" edited "Aged Concrete":


Additionally, I find that the MT 903's work flawlessly with cars and engines equipped with MT N-scale couplers, although I don't use the "Magenematic" uncoupling option and cut the "dongles" off of every coupler.

As for 903's "breaking" in "long" trains, in my Ntrak days, I was regularly pulling trains of 60 cars equipped with MT 903's without any breakage whatsoever.  However, I noticed that cars that were not properly secured during transport to shows would break 903's much easier when rolling around in their containers, so now I make sure all my cars have foam bits inserted to pad them and keep them in place during transport.  This has solved that problem.

As for what kind of track to use on your display diorama, I would stay away from Atlas C55 track, as the "spikeheads" are grossly oversized.  A much better option is Micro Engineering C55, as the spikeheads are noticeably smaller.  To the naked eye, the difference isn't too noticeable unless the two brands are side-by-side, but in photos...Atlas C55 flex looks glaringly toy-like.

The very BEST track to use would be old Railcraft C40 or C55, as the spikeheads on this flex were very small, but it's been out of production for decades, although you can still find it.  I have a hoard of it which I use for foreground trackage, and I'm not willing to part with any of my stash.

If you can do without spikeheads or tie-plate details, then hand-laid C40 on PCB ties (every fifth tie) looks really good, as the height of the rail is actually .044" and is the ideal height for mainline 133lb rail used by many class-1 railroads since the 1940's.  If you wish to detail out your display diorama's trackage, etched C40 N-scale tieplates and rail "joint bars" are available from Proto87 Stores here: ( http://www.proto87.com/n-scale-track.html  ).  Note that the tieplates work on C40 rail, and not on C55.  Etched turnout frogs and other turnout details are also available from Proto87 Stores. Additionally, hand-laid C40 PCB trackage will allow even the biggest pizza cutters to roll unobstructed, whereas Micro Engineering C40 flex needs to have the inside spikeheads sanded down for even medium profile wheels to run such as those found on common manufacturer's motive power such as Atlas, Kato or Athearn. Truthfully, the sanded-down ME C40 trackage looks pretty good, but it IS an added time-consuming task when laying track.

Although I have dozens of frets of these etched track details, their tiny size makes actually using them extremely difficult and I have yet to apply them to any trackage I've laid, except the closure point heel hinges, which I use on every turnout I make nowadays.

You will note that the caboose photos I've posted were taken on my hand-laid C40 Park City branchline trackage.  The ballast there is real-rock "cinder" N-scale ballast from Highball Products.  I screen my ballast so that the "grain" size is right at .020" (3.2 N-scale inches in diameter), with larger pieces being eliminated as well as most of the dust.  "Dust" will turn your fine ballast into what looks like concrete if not eliminated from real-rock ballast.  I also blend ballast colors to match the prototype I'm modeling.

Since I hand-lay all of my turnouts to a "fine" NMRA standard, I use Fox Valley narrow wheel replacements and they run without any problems on my hand-laid turnouts. My CA-1 caboose is shown with these wheels.  Yes, all of my motive power must be exactly gauged to run smoothly through my turnouts, but since I assume that every engine is out of gauge before running it on the layout, it's become easy to simply check each one and correct the problem before running.  No big deal IMO.

Another item which contributes to a more prototypical looking car are brake air-lines.  I use both BLMA and Precision Scale plastic brake lines, which really add to the appearance.

Here's a better look at hand-laid C40 PCB track on my Park City branch which is where I pose my trains and cars when taking close-ups...

Photo (5) - Big Boy parked on C40 PCB trackage:


Painting, weathering, superelevating and properly ballasting trackage really makes a huge difference in making it look "real" IMO, but starting with the best "base" is equally important.

Cheerio!
Bob Gilmore

[daniel_leavitt2000]

Randy made a good point with standardizing equipment. There are two schools of thought on this: consideration for legacy equipment and bleeding edge realism.

Peco C55 along with Atlas, Peco C80 will run pretty much anything and have been around since the 60's (80's for Peco C55). I would call gage a bit sloppy - wide in turnouts and even sectional track can be off by as much as 1/2mm. Wheelsets were designed wide as a result. If you run this track you need wide treads even if you want to run on this track.

Atlas and to a lesser extent Micro-Engineering C55 track can be tight in gage. They are not out of gage, but the fit is very snug. This can affect sloppy wheels as a result. I had to regage many of my Atlas engines as they were sloppy (wide) in gage. This is only an issue in switches but results in a visible hitch while traversing the frog.

With this C55 track, you can use much thinner treads. BLMA, FVM and the lo-pro MTL wheel sets work perfectly.

I have still not found a perfect coupler. MTL Z scale couplers can break on long trains and I hate the slinky affect. Full Throttle/Lee English couplers are expensive, stiff and have a tendency to uncouple. McHenry and Bachmann's are laughably too large. I do like Accumates though, and they work very well if you pull the trip pins.

[ednadolski]

Full Throttle/Lee English couplers are expensive, stiff and have a tendency to uncouple.

Being smaller than conventional N couplers makes them a bit more susceptible to vertical misalignments.  Even so, I have significantly fewer unwanted uncouplings overall with them than with conventional N scale couplers, even in long trains over 60 cars.   I agree they are less smooth than MTs for hands-off coupling, and again due to the smaller size they have less of a gathering range on sharper curves. Neither of these are much of a concern for me since my operating scenario has essentially no switching, so for me it is worth the tradeoffs to have the smaller coupler size and (especially!) elimination of the slinky.

Ed

[randgust]

One of the odd things we 'proved' here on this forum a while back is that MT coupler heads have actually gotten bigger over the years, particularly vertically.  If you go back into the early 70's and look at a body-mount 'standard' coupler, and put a micrometer on it, it is amazingly close to what a Z-scale knuckle is today.

I had an old 904 and compared it to a new 904 (short T-shank, underslung) and was really surprised how much bigger the new one was. 

I think that's one of the reasons why the Z's work as good as they do.   If they made more adaptations (including a retrofit into the 1015 box, the standard freight truck-mounted box, and a T-shank) I'd do a whole lot more replacements.   

While we're talking standards and how they keep changing, another issue along with the 'slinky' is the basic design of the legacy MT coupler head that will separate vertically under heavy load.  They fixed it with 'newer' tooling, and you can also fix it yourself, but the new head is called Reverse Draft Angle (RDA).  It really does solve the vertical separation problem, but not all couplers have it, and a lot of old stock doesn't have it, and it's annoying as the devil in N on longer trains.   I think it's probably the #1 reason for the tendency to cut off trip pins because they snag, when the problem is really that the coupler heads force the pin down due to the old knuckle design.   Solve the knuckle issue and it solves snagging trip pins at the same time.   As I do use magnetic uncoupling, this was - and continues to be - an education campaign by me on what's really going on and the underlying problem.   I didn't invent the fix, the late great Jim Fitzgerald did, and finally, MT copied it in their tooling, or at least most of it.   

Bluntly, the problem was that you could precisely set the trip pin height with pliers with the gauge, and then a car would still mysteriously bang a pin into a turnout or crossing, and cause a derailment.    A lot of members here solved the 'problem' with the wholesale slaughter of MT trip pins with wire cutters and went to bamboo skewers to uncouple.   Another one of those things you won't see in HO that stull burns up a lot of electrons here on forums.   False uncoupling on the legacy MT stuff was also caused by the knuckle design.

My layout is high enough that it's pretty easy to do an 'eye level check' on a passing train, and I really want to see those couplers dead-center vertically under load, not hanging on with a toenail with one knuckle high and the adjacent one low.   That's the goal, and with the RDA tooling or modifications, they vertically recenter the harder they are pulled instead of popping.   

[jdcolombo]

I've used a bunch of stuff over the years, and have learned that standardization is a very, very good thing.  On my current (and probably last) layout, here's what I've decided to go with.

Track:  Atlas Code 55.  Best looking factory-made stuff on the market, but the factory turnouts are not always reliably in gauge through the switch points (especially #5's).  90% of the switches are Atlas #7's, but I have hand-built a number of turnouts (#4's and 6's) using Fast Tracks jigs and tools for switching areas.  These are so much better than the factory stuff it is laughable, and if I were starting from scratch today, I'd hand-build all the turnouts even knowing how much time that would take. 

Having said all this, I miss the overall sturdiness and reliability of Peco Code 55, which I used for nearly 15 years before building my current layout.  If you don't care about the European tie-spacing and value operation above looks, Peco Code 55 is bullet-proof.  And Peco switches have the advantage of not needing a separate switch activation linkage.  The "flick with a finger" approach is especially useful in complex switching areas, where the operator can visually see and set the turnout alignment without messing with toggle switches on a fascia or (even worse) using a DCC throttle to throw switches.  If you plan on a layout where you will invite others over to do operations, remember that simpler is MUCH, MUCH better.  You don't want to have to spend an hour each operating session teaching your operators how to run the layout.

Curves: 15" minimum on mainline.  Some industrial switching areas get as low at 10", but all they are going to see is a 40' boxcar. 

Rolling stock:  Not much discussion here about proper weighting.  Cars that are too light are a major operational problem.  I weight everything to about 1 oz (the NMRA guidelines adjust weight by car length; I don't and I've found 1 oz. to be a good compromise between operational stability and train-length capacity).  For wheels, MicroTrains lo-profile on nearly everything, though I have been experimenting with FVM wheelsets and may switch to those.  Couplers are all MT, all body-mounted.  I do not use magnetic uncoupling (I have made my own uncoupling tools out of 1/8" solid brass or copper tube), and therefore cut off the trip pins (not all the way, like Bob Gilmore, but down to about 1/8").  I use a few MT 905's on the pilots of steam locos for appearance purposes.

Like Randgust, I have gotten to the point of zero (well, 99.9% of the way there) random derailments.  I run 25-car trains behind either steam (Walthers/LL Berkshires; Kato Mikes; Bachmann 2-8-0's) or 3-unit first-generation hood diesel lashups (GP9's, RS11's, RS3's).  Basically, nothing derails unless the operator makes a mistake (open switch).

When I think back to my first N-scale layout (done in 1990 with Atlas Code 80 and snap switches), it's truly amazing how far we've come.  The one thing we desperately need is a more-to-scale coupler (like an MT 905) that can be used easily with current rolling stock, almost all of which is designed to take a  body-mount MT 1015 (Atlas' engines and rolling stock with Accumates can usually easily be converted to a 1015).  Rumors of the design and manufacture of such scale-sized couplers abound, but I'm still waiting.

John C.

[Spades]

Mr. Luke is making a photo diorama eye candy, not an operating display.  I'll add to the noise.  Couplers Bowser.  Track Micro Engineering code 40. Wheels NWSL very pleasing to the eye, but wobble when they roll.

 Removing my anal cap now.

[peteski]

For a diorama setting I would recommend the new Rapido/Vlk couplers. The are closest in appearance and size to scale couples.  They are pricey (they are only sold as replacement parts at Arnold for their locomotives (not for wide adaptation)., but they sure look good.

I am going to write a review of them - I have all the photos, now I have to sit down and write it.  Eventually..

[ednadolski]

For a diorama setting I would recommend the new Rapido/Vlk couplers.

Links?  Pics?

Ed

[wcfn100]

Links?  Pics?

Ed

Yeah, I thought the last we heard the Vlk designed couplers were DOA due to spring issues and Rapido had to do something else.

Jason

[peteski]

Yeah, I thought the last we heard the Vlk designed couplers were DOA due to spring issues and Rapido had to do something else.

Jason

Relax, this is nothing new and exciting.  Sorry if it sounded that way.

It is the current coupler used on Arnold SW1 locos. I have to dig up the post where someone quoted Charlie stating that their current coupler, while not the original Vlk design, is still partially based on his design.

[robert3985]

Truthfully, the subject of realistic looking, reliable track is a fairly complex subject to talk about, especially when listing individual reasons for every aspect of it.  This post will give some of my reasoning for what I do and recommend as well as repeat some things I've talked about in my earlier post, but limit the majority of my comments to the OP's question about what track and ballast to use on a display diorama used to display cars and engines he's done some work on to make them more realistic.

OP is requesting what track to use for a "display diorama" which I assume may also be used for photos.  Being that he's displaying N-scale rolling stock and motive power, photos will naturally be close-ups...which means that "looks" are paramount.

So, being that "looks" is important, this eliminates almost all N-scale track except Micro Engineering C55 & C40, and maybe Atlas C55.

Since I consider my entire layout to be a "photo-diorama", my recommendations are based on my own experiments as well as having photos published in various model railroad mags.

Getting down to brass tacks, if you compare ME flex, either C55 or C40 with Atlas C55...you eliminate Atlas C55 also, since it is definitely NOT the " Best looking factory-made stuff on the market,.." with its hugely oversized thingies that pass for spikeheads on each and every tie.

Atlas C55 turnouts suffer visually (and functionally) from being much too short between the frog casting and the heels of the closure points, and with a shiny silver plating on the cast closure points and frog, which soon rubs off to reveal copper plating underneath, then when that rubs off...whatever metal those parts are made from...which obviously isn't Nickel-Silver since it has to be plated to look more like the NS rails on the rest of the turnout model.

As for turnouts, the very best looking N-scale turnouts on the market are ME #6's in C55, being made with properly proportioned rails, closure rails and points, with frog castings made from actual Nickel-Silver (as opposed to plated, mystery-metal on the Atlas turnouts), with spikeheads MUCH smaller than Atlas's offerings. 

ME turnouts, even though they offer only #6's are within a few fractions of an inch of being as long as Atlas' "#7" turnouts, which should be, if properly proportioned, nearly 1/3rd longer than #6 turnouts. Comparing ME #6's to prototype drawings, they are properly proportioned as far as rail placement and length is concerned as well as tie-size and proportion.  Distant second best is Atlas C55, with everything else not to be considered.

However, the "very best" track for a photo and/or display diorama will be hand-laid C40 with PCB ties every fifth tie and wooden ties in between.  And, the "very best" turnouts will be hand-laid C40, using Proto87 Stores tri-planed closure points and closure point heel hinges.  Additionally, for the "very best" track, turnout and rail details, Proto87 Stores etched turnouts details, tieplates & joint bars should be used.

Okay okay okay...  ...not very practical as far as time and effort is concerned for the "very best", but practicality isn't one of the OP's parameters.  A small, super-detailed scene or diorama lends itself to a super-detailing project and really sets off super-detailed cars and locomotives, whereas the same realistically detailed equipment sitting on grossly misproportioned Kato Unitrack with a big ol' rail/ballast joint in the middle just doesn't make a lot of sense to me.  I know it does to some of you, but really...track IS a model too, isn't it?

The truth is that it'd probably be a good idea if the trains (or cars and engines) will at least roll on the diorama track, which eliminates ME C40 flex unless you want to take the time to sand down the inner spikeheads (which I do).  Luckily, ME C55 #6 turnouts have an over-center spring mechanism which eliminates the need for a Tortoise underneath or other mechanism to hold the closure points in place...Atlas C55 turnouts don't have this handy feature.  Frankly, the height of the rail isn't very obvious as far as close-up photos are concerned unless it's oddly shaped like Atlas C80 and Kato Unitrack.  I've got photos of ME C70, C55 and C40 which have been painted, weathered and ballasted and you don't notice the rail height...you notice the really nice spikeheads and the tie size and spacing...which look better than any other commercial offering...including Atlas C55.

Sooo...for photogenic manufactured track on a photo-diorama, the choice is ME C55 flex, with ME #6 C55 turnouts. 

For ultimate, "very best" track, hand-laid C40 ( C44) is the only way to go, with wooden & PCB ties, etched rail, track and turnout details.

But, it doesn't stop there.  All that good lookin' track needs to be placed upon properly proportioned and contoured roadbed and subroadbed, depending on the type of track it represents.  This influences the ballast too, as does the region, and the road.

Painting and weathering is essential for realistic looking track and ballast too, and we can have a discussion again on what color to paint the rails is the most "real".  My suggestion is to choose a road, location and era you like...and do a little research, because that will make a big difference in the colors you choose.

We won't get into other trackside details, or how to make realistic grass/weeds/brush...

Like several of the other contributors to this thread, I also have standardized my track, and most likely, its appearance is more important to me than most model railroaders.  However, reliability is equally important to me, and my standards are for both reliability AND appearance, and include using ME C55 and C40 flex (or Railcraft flex from my hoard)...all of which are really the " Best looking factory-made stuff on the market.." by a long-shot, and hand-laying every turnout on my layout and/or on my modules, and hand-laying my Park City Branch with C40 PCB trackage.  I build and use #6 turnouts on branch & industrial sidings, #8's on mainline sidings, and #12's for mainline passing sidings.  I hand-cut every C55 and C40 rail joiner so they fit in the space between ties, so all my ties you see on the layout (except turnout ties) have spikes, and the rail joiners virtually disappear.  Every turnout frog is isolated and live, every closure point and rail is also live...and every single piece of rail has a 22ga. solid copper feeder soldered to its railfoot near its center.  Most of my mainline curves are superelevated, and all of my flex is laid on sanded Midwest Cork Products roadbed.  Minimum visual radius is 24" on mainlines but with many areas of much larger radii exceeding 100 inches, 18" on branchlines and where mainlines can't be seen, and 16" in industrial trackage. Every mainline curve also starts and ends with a spiral easement.

I do not have track-induced derailments, unless something breaks during transporting to shows, which rarely happens. I do not have dead-spots, all of my longest engines and cars operate flawlessly on trackage they're designed to operate on. All my engines will creep slowly over every turnout or combination of turnouts without a wobble or a hiccup.  All of my cars are equipped with lo-profile wheelsets, as they have been since I started machining my own in the early '80's...with zero lo-pro flange induced derailments.

I simply do not accept the premise that to have "bullet-proof" trackage, track realism must be compromised.  Ya only have to take a bit more time to have both reliability and realism...and when a modeler decides to take less time, realism suffers...and sometimes so does reliability. 

It's up to you to decide what's "good enough"...and even though I consider myself to a "rivet-counter" in many aspects, I have not yet opted for full etched details on hand-laid track...but I eventually will do a test to see if adding them is worth it...and post photos here so others can make a well-informed decision too.  Nor will I be building a "proto-160" layout simply so I can run scale-sized flanges and wheel widths on everything.

Some things in N-scale, for example the rail proportions, are a compromise I have decided to live with. Railfoot contour, web thickness and railhead width are seriously compromised in N-scale, simply because the rail is made for HO scale, so it's wider than it should be for N-scale, and only two heights of rail are readily available...with only C40 (actually C44) being low enough to truly represent prototype 130 lb rail.  I consider using C55 for mainline rails and C40 for sidings and branchlines to offer the best visual representation of different weights of prototype rails for differently trafficked trackage, even though C55 is markedly too tall for prototype rail.  Ah well....  I am also not prepared to machine down the width of the railheads on my layout, but I know at least one model railroader in this forum who does exactly that!!

I guess it comes down to what bugs you, and what you choose to live with...or what you consider to be "fun"...

Cheerio!
Bob Gilmore

[MrKLUKE]

.

[jdcolombo]

Bob is correct that ME Code 55 is better-looking than Atlas; I had forgotten about ME track, with it's much-better looking spike heads.

But for a small diorama that one wanted to be as realistic-looking as possible in N, hand-laid Code 40 would be the way to go, and for ultimate realism, add track details from Proto87's N-scale selections (e.g., individual tie plates).  I'd probably pick Fox Valley wheelsets; and use MT905 couplers on everything (if properly mounted, the 905's will work with "regular" MT N-scale couplers, so the operating character of these cars wouldn't be impaired; or a new owner could relatively easily swap them for MT1015's if he/she wanted).

For an operating model railroad, Bob is also correct that given enough time (money really isn't an issue, since hand-laid track is often cheaper than commercial track; detail parts are not very expensive, etc.), one can make a "bullet-proof" operating model railroad that also looks incredibly realistic.  As Bob notes, it all boils down to an individual's choices and what compromises one is happy with or willing to live with.  I think Bob's work is amazing, but I'd never have the patience to replicate his attention to detail.  I know that, so I've made my peace with it . 

John C.

[crrcoal]

JEFF!
After seeing your AMAZING weathering skills in HO scale, I cannot wait to see what you do with N scale!!!! Please post how-to's as you do this.

Thanks!!

Andy