O.K. I just bought an A/B pair of BLI F7s, installed DH123s in the A and B units-- pretty simple actuallty, remove front coupler on the A, lift body off, plug-in the 123, put body back on, remount coupler and go. Both ran fine at factory defaults, and sounded pretty cool. Off they went to the paint shop.
A will come back as #1537 and when I go to change its address will all heck break loose? If I want to consist them as per prototype and set B as #1537 as well, does everything blow-up?
Help me before I harm myself!
[edited by selector]
*Warning* I'm frustrated with Broadway Limited Locomotives and their silly MULTI-DECODER setup. But I've also included some suggestions for resolving the issues.
First the good:
My wife got me some of those Broadway Limited "Blueline" locomotives for Christmas....
On the one hand they are *REALLY* nice. They look great, pull great, 30+ cars up a 1.5% incline! They SOUND absolutely fantastic!!! BLI got that part REALLY REALLY right! I *LOVE* them. I wouldn't be so upset about the next part if I didn't like all this about them. I'd just skip them entirely and stay with the Proto2K's...
Then the bad:
On the other hand they can be a royal pain-in-the-behind too. Anybody who's ever tried to program the decoders in the thing knows EXACTLY what I'm talking about. It all works pretty good until something happens to mess up the factory programming -- and then they get all hosed up. And getting them UN-hosed is NOT FUN. [edited by selector]
Then the ugly:
BLI claims they do the dual-decoder thing to "save money" but I question how much they're really saving when the retail price for the typical decoder that they themselves recommend is only around $20 bucks or so. (That means the wholesale cost is actually lower-- or if they made their own decoder, it could be even cheaper still)
So what does it actually take to make the typical DCC motor decoder anyway?
The typical DCC motor decoder is a pretty cheap rig. Its comprised of a very small, simple (often one-sided) circuit board, so the copper and phenolic cost is exceptionally small. There are very few holes to drill (some have no holes) that cost is very small. They typically have only a small handful of parts (around 8 or 10 SMD style components) which can be assembled easily using standardized automated "pick-n-place" assembly techniques, so again very cheap. They typically use some variation of a 16Fxxx type PIC microprocessor which is very cheap in typical mfgr run quantities. And finally a quick run through the wave solder machine (or hot-air machine, however they do it these days), and maybe a stop at the QC (quality-control) bench before heading out the door. Aside from some programming for the CPU, that's about all there is to it-- overall, pretty cheap.
If they farmed the whole thing out to some company in Taiwan or China its even cheaper still. Heck, even if they just out-and-out bought the decoders directly from another mfgr, a custom run would be very cheap. Or _even_ if they bought 'em straight off the shelf at their local hobby shop and paid the full retail price, they would still be cheap and only add about $20 bucks or so to the overall cost. So saving money doesn't really appear to be the real motivation here, or if it was, it doesn't seem to me that they were doing a very good cost/benefit analysis)
Would $20 extra bucks make THAT much of a difference !?!?!?!
Maybe I don't know there market, but on the surface of it, it seems doubtful to me that the majority of the train-buying public would be put off from buying an otherwise fabulous locomotive because of a $20 difference in price. Besides, that's about what they turn around and pay a third-party installer to stick a motor decoder in the loco anyway, maybe even less.
So how did BLI come up with this mess?
I don't understand BLI's thought process here...if they're _gonna_ put out a locomotive with a multi-decoder setup, _why_ make it so difficult to reach the blessed things? You have to take the _whole_ locomotive apart (at least on the SD7/9 and F7 models, don't know about the others) to get at the decoders. Then you have to _remove_ the motor decoder and put it into a _SPARE_ locomotive!! (SPARE locomotive !?!?!?! SPARE locomotive !?!?!?! Who has SPARE locomotive !?!?) So, the upshot is that I have to disassemble TWO locomotives to get ONE locomotive set up...
THATS NUTS!
If they have to make it that way, why not just put a jumper or two on the bottom so I could selectively enable/disable the decoders?
Giving BLI the benefit of the doubt
So, okay-- let's give them the benefit of the doubt-- they say cost saving was their goal-- let's choose to believe them. So here we are, BLI has this crazy setup-- or maybe they thought it would be nice to let the consumer pick the motor decoder. I can go along with that, let's pretend that's the real reason. Whatever the reason-- there were low-cost (maybe even zero-cost) alternatives to the method they selected.
Ways to program a multi-decoder loco
Joe Fugate has a nice posting on his site about programming multi-decoders. Its a good read with useful points so I'm not going to duplicate it here (here's the link: http://model-trains-video.com/e107_plugins/forum/forum_viewtopic.php?150.0)
But that's all _workaround_ stuff-- what to do about the actual _issue_ of putting two decoders in the same locomotive?
What exactly *IS* the problem anyway?
The problem is whenever you reset the decoders back to their factory settings, they end up with a short address of 03 and the value 00 written into CV16. This effectively disables the multi-decoder locking scheme (where each decoder is assigned a different address in CV16 and selected via CV15) and now the decoders having the same address and no locking are free to _both_ interpret the programming instructions, or interpret them partially, or not interpret them at all. This can quickly lead to decoders with the wrong (or possibly even harmful) information programmed in them.
The whole thing is a mess. Seriously.
Although, in BLI's defense they do use "Direct Mode" for programming their sound decoder while most motor decoders use "Paged Mode", which means its possible (though tricky and with much key-pressing) to actually talk to both decoders if your DCC is capable of utilizing both forms of programming... but its still a big pain in the behind, I can attest to this personally.
Suggestions for Broadway Limited to resolve the problem entirely
So, here's a suggestion for Broadway Limited (or anybody else considering a multi-decoder arrangement):
Why not make the default setting for CV16 be two (02) instead of zero (00) ?!?!?!?!
Then when the thing gets reset it will still be addressable using the standard locking mechanism?
The people who are savvy enough to want to do it Joe Fugate's way can reprogram the things however they want, and for the rest they won't get tied up in knots, won't have to disassemble their whole locomotive or dig out their spare locomotive (I still love that-- SPARE locomotive, it kills me :) just to get the things set back up and under control.
For DigiTrax and the other motor decoder manufacturers, why don't they put a one (01) in CV16 instead of zero (00). Then their stuff would then follow the NMRA-approved convention as well?
Certainly the workarounds would be there for people who already have decoders manufacturered before these standards-- but _why on earth_ do the manufacturers insist on perpetuating nonsense?-- especially when its as simple as pre-loading a SINGLE BYTE into their little gizmos????
COMMENT from BROADWAY LIMITED ???
Will someone from Broadway Limited _please_ answer at least the last question?