I have finished two installations of the Tsunami medium steam decoders in the USRA 12,000 gal. tender that comes on both the USRA light 2-10-2 and heavy 4-8-2.  They are wired differently though they are the same castings.  I concluded that the easiest solution was to splice a short Digitrax decoder wiring harness onto the shortened Tsunami wires.  It was then "plug and further modify".  The coal load casting comes right out of the coal "bunker" which was fully formed, making a perfect speaker chamber.  I drilled many small holes (#55) in the coal load.  I made a baffle to mount the speaker and glued it (with the speaker mounted to it) to the inside of the coal load as high up as possible, drilled two small holes in the bottom of the coal bunker for the wires and used latex caulk to seal the wires in the holes.  What I found is that, even with the very flat Digitrax plug, there is not a whole lot of room for the decoder when you reassemble it - be very careful.  I found that even double stick foam tape can add too much thickness.  Put the decoder on its side against the engineer's side of the tender, as far back as it will go and tape it in with the textured scotch tape that wound nurses use.  The Tsunami is VERY programable and the instruction manual on their web site is both readable and thorough.  Its  hundreds of pages and following the instructions you can really make your locomotives "sit up and bark". 
I really miss the good old days of simple two wire DC.  My fun lies elsewhere, not in cutting edge electronics.

I did not see in your original question any mention of DCC so I will not go there.  In trouble shooting I have learned that when a problem occurs at one place with one unit, car or engine it is a condition of compound problems.  If there is a switch or a crossing at the point of problem I would look to a low coupler trip pin.  If it is not a switch or crossing I would look for a track nail that has backed out above rail height, or a bad rail joint at the point of stopping, though this is far less likely.  Loading of an engine can cause the chassis to "motorboat" which is to say raise in front and squat to the rear.  Combining two units can have one pushing or pulling the other.  The Bachman truck suspension is very supple to allow the units to traverse really bad track, well.  It is a function of their trainset history.  However, the downside is that the high truck pivot point can exaserbate the "motorboating" problem with vertical curves, lack of curve easements, track irregularities caused by temp or humidity changes, or pulling steep grades or heavy loads.  It can cause one coupler to rise up on another, thus driving the coupler trip pin even further too low.
Of course, when you reverse the unit the offending trip pin curves in the other direction and rides right over what had been an obstruction.  So my advise is to check the coupler afixing clip to be sure it is holding the coupler in the correct alignment.  Then check to see if the coupler has too much vertical play.  Any is too much but excessive amounts cause real problems.  And then bend the coupler trip pins up just a little bit more than the manufacturers reccommend.  (use the Kadee special trip pin pliers if you have them or have a friend who does)  The precise location of the trip pin is only important if you are using magnetic uncouplers.  And so few of us do any more.  So I always make sure they have a little excessive curve to make the trip pin clearance greater, as a preventative measure.  Its simpler than improving my trackwork.

Well,y'all, I do thank you for your responses.
However no-one has been addressing my question, which was whether anyone could tell me the wiring color protocol that Bachman uses on these locomotives and tenders.
As a friend finally explained to me, and as most of your comments alluded to, my mistake has been in expecting that there IS a wiring color and location protocol, just because the NMRA has one.  Since Bachman's Chinese assemblers aparently use whatever color wire they care to on any locomotive and vary them from run to run, then my only solution is to mechanically and electrical trace the wires in the tender, on the boards (which also vary by locomotive type and run to run), and to the plugs, and diagram them.  Then "all I have to do" is make sure that the tender I want is wired just like the tender that came with the locomotive, being careful to note that the orientation of the plugs can also vary.  I have no fear of soldering.  But what started this whole mess was trying to hard wire a tsunami sound decoder into a heavy USRA Mountain tender only to discover that the color wires did not follow the NMRA standard protocol and that I, therefore, did not know which wires went to the plugs in what arrangement.
Now I am a lot smarter than I was.  And this has become a quest.  I will get these locomotives and tenders to all work as I want them to.  3 out of 6 are up and running and with careful workmanship I believe I will have them all arranged and working properly.  Model Railroading is Fun.  model railroading is fun.  model. . . .

You could do what the prototypes did:
1.) The actual wheelbase is the critical issue.  Engines were often placed on the turntable VERY carefully, overhanging on both ends. Or. . .
2.) You could replace the tender with a much shorter one, such as the USRA 10'000 gal that comes on the Light Mountain to replace the 12.000 gal tender that comes on some heavy mountains.  Usuall, historically, as the railroads mixed and matched tenders they went to larger ones for greater distance between water stops.  However many railroads couldn't afford to replace turntables after about 1929 so kept shorter tenders through to the end of steam, even with larger more modern 4-8-4's.  Spectrum tenders are available seprately.

I am trying to use the C&O and SP hicken Spectrum tenders on the Spectrum 2-10-2 and both the light and heavy 4-8-2 (mix and match) with DCC.  Are the wireing protocols different between the tenders and the engines?  The harneses apear the same and fit from one to the other.  However, I get very strange results - the lights don't work, the engines will not run in reverse,  and the engines won't run except when the lights are "on", which of course they never really are!  I could be into a bad batch of Digitrax decoders, but that doesn't seem to be the likliest choice.  If the wiring harness protocol is the the problem, can anyone tell me how to reconnect the tender (male end) of the harness to the light board?
I do so wish that Bachman had used the standard DCC wire color coding in these tenders and in the stock tenders of the locomotives.  My old eyes have trouble differentiating between all the browns, tans, and oranges which are yet to make any logical sense to me.