Model RR applications

From: Jim King (
Date: Tue Jan 04 2000 - 20:40:45 EET

I was forwarded the initial inquiry about RP and model RR applications.
I joined this email list to respond to the inquiry but I'm not a regular
member. Here is my $0.02 worth.

I've been a model railroader for 30 years and have run my own "cottage
industry" manufacturing company since 1990. In the mid 90s, I started
producing urethane castings from RTV molds and handbuilt styrene
patterns. These are largely flat castings made by vacuum casting, open
pour methods ... 2-part molds are also possible but generally aren't
needed for rectangular objects, which is what most RR models represent.

About 4 years ago, I CAD modeled the end of a Southern Ry steel
underframe hopper, circa 1928 and had it built as an SLA in HO scale.
The intent was to make a pattern to replace traditional hand-punched
rivets, etc. While the ribbing, U-channels and overall shape was quite
accurate, the rivet detail left a LOT to be desired; in fact, the entire
experiment proved to be a failure because the rivets did not reproduce
adequately. I've since learned that the resolution of "typical" SLA
machines coupled with the "typical" build layer thickness of .004-.006"
plus the "typical" laser beam diameter of .010" all spell doom for such
fine detail. Having the part built on a fine-tuned SLA with .0015-.002"
layers and a narrow beam laser might produce a better part but the
stair-stepping on rivet heads, although reduced, still has to be

To the comment about making Keltool inserts for injection tooling, I
take exception that this an acceptable way to go. You still have to
have a pattern for Keltool to work, whether an SLA or some other means.
Until the resolution and stair-stepping issues with SLA are cured,
you're still stuck with a so-so pattern that will duplicate itself into
a so-so (or worse) core/cavity set and an even worse injection part.

And, copying Chinese-made products is a flagrant exercise in piracy!!
Aside from being illegal, it's also anything but ethical and shouldn't
even be mentioned on this list.

The current major manufacturers (Walthers, Athearn/Samhongsa, Atlas,
Life Like, Intermountain, Bachmann, etc.) use traditional CNC/EDM and/or
rapidograph (ratio arm engraving) methods to make injection molds.
Modelers expect perfection, both in running gear and injection molded
parts. Unfortunately, there are no other options (that I'm aware of)
that satisfy the customer demand for top quality at reasonable prices.
Only items made in China, Denmark and, at one time, Austria (remember
Atlas Roco?), seem to provide the quality-to-cost ratio that makes new
products sell-able.

Conclusion? No matter how good RP claims to be, there are definite
limits to what can be done on the small scale. While the Sanders
machine does have the best accuracy for small parts, it's slow speed and
related high cost per hour makes it prohibitive for production. Using
SLAs as a starting point for patterns may be an option but certainly not
the way to go for accurate detail. For me, it's back to laying out
artwork for photoetched brass sheets (rivet strips) and hand cutting
styrene shapes. My customers expect no less than top quality for their

I wish you well in your quest for state-of-the-art RP application to
"our" hobby.

Jim King
Morris Technologies (full time!)
Smoky Mountain Model Works (moonlight job!)

For more information about the rp-ml, see

This archive was generated by hypermail 2.1.2 : Tue Jun 05 2001 - 23:02:32 EEST