LCC - Layout Command Control

[C855B]

Gary mentioned in the RR-CirKits thread that Dick (Bronson) has just released LCC products. It's not much - just a bus power injector and CANbus terminators - but if he's doing that already then (IMO only) then he's probably getting ready to jump ship from his Simple Serial Bus to be first in line with LCC encoders or gateways, and decoders.

Is anybody else besides NMRA talking about LCC? Did anybody go to the convention and go to one of the LCC clinics? Are there hints of anything on the horizon for LCC? I couldn't find any discussion in the JMRI Yahoo group about LCC, but frankly I wasn't looking very hard.

Obviously my bringing this up and asking about LCC is self-serving. My game plan for the GC&W was a mixed bag of DCC hardware and networks, with three separate networks - track/train power, signals (etc.), and then Tam Valley LocoNet decoders for turnout control. If LCC is close enough to be thinking there will be actual product within, say, about six months or so, I can start construction under a simpler plan with a single DCC network, then breakout the non-train components on an LCC bus later.

However, I have that nagging feeling that this is one of those "next great thing" technology situations where it's mostly talk at this point and usable product might show-up in some nebulous future. In my work I delayed more than one big deployment based on something supposedly just around the corner which never came. It's very frustrating in fast-moving technology markets.

S'pose my next move is to correspond with Dick and ask outright what he has on the bench, and how soon.

[jagged ben]

This is the first I've heard of it.  So what does that tell you?   

Your last sentence is very sensible.

[C855B]

Already sent the note.

My entire career was spent pushing the practical application of new tech. My catchphrase was "You can always tell who the pioneers are - they're the ones with the arrows in their backs." So here I am again.

LCC seems to solve issues I expected to have with DCC as the layout grows. If DCC can just run the trains I can worry less about network performance thresholds. One concern I see right away is how much power can the bus pass. It appears to be based on CAT5/6 cabling and RJ45 connectors, which is great from an availability standpoint, but not exactly a robust power conduit. What if I'm addressing, say, 8-10 stall motors in a yard ladder? If I have to inject power at every significant client node, the advantages start to fade a little.

[C855B]

If anybody is interested, the tech docs for LCC are here: http://www.nmra.org/sites/default/files/standards/sandrp/OpenLCB/lcc-2015-02-17-complete-set.zip

500mA is the bus limit. I haven't found mention (...yet...) of any spec regarding isolation of the power bus between injection nodes. I find it more than a little annoying that here we have the new, better, runs-faster-jumps-higher version of DCC for trackside use, but with extremely stingy ability to supply power. And wait, there's more - the spec calls for current limiting by the power supply at each injection point, so no parallel power bus with a big power supply unless you supply limiting electronics to each injector. Do any of these guys actually design layouts on something bigger than a 4x8 sheet of plywood?

[GaryHinshaw]

A few thoughts from someone who has not thought about this deeply: 

* The LCC selling point seems mainly to be bandwidth.  What exactly is the bandwidth available on Loconet?  I'm sure it's written down somewhere but I have not seen it (nor looked very hard).  Is there any easy way to monitor how close one is to capacity?

* The c/mri crowd is still alive and well and they are actively developing Arduino based applications.  I think they claim that LCC has already been done and they seem mostly to be detractors.  It would be really helpful to all if there were not two rival approaches to layout automation.

* An example of where LCC might shine: I was corresponding with Dick over the weekend about my jmri panel problems, and I discovered that, because of programming defaults in the RR CirKits boards, I had two boards competing for the attention of Loconet sensors 1-8.  This was causing my panel to think that blocks were occupied, when in fact it was my Motorman board reporting those sensors active.  When I reported this to Dick he made the point that this would never happen with LCC because each board (node) would be uniquely identified under LCC.  This is just an example of where you might not need to be as vigilant about configuration control under LCC.

* RE power: I don't know how it is supposed to work with LCC, but the RR CirKits SSB components probably provide a good example of how Dick's components will work.  First, the SSB adapter has its own 15 V, 700 mA power supply, and it distributes power to each board on the SSB bus.  The bus itself is any commercial 3-wire cable you choose: I am using 18-ga stranded, but that is probably overkill.  Second, the Motorman board for controlling switch machines requires a separate power supply for switch machine power, which is simply passed through the board with the appropriate polarity.  So switch machines are not being powered over the bus, and I'm sure the same will be true for Dick's LCC version of this board.

* I am sure Dick is going to produce a line of boards for LCC, and my guess is that they will largely parallel his SSB line.  If that is the case, it should be pretty straightforward to convert my system over to LCC if I ever feel a need to.  At the present time, the only reason I can see wanting to do that is bandwidth, but I have no idea how close I might be to needing it...

-gfh

[C855B]

First, product info from Dick:

  http://www.rr-cirkits.com/description/LCC-Power-Point-flyer.pdf
  http://www.rr-cirkits.com/description/LCC-Terminator-flyer.pdf
  http://www.rr-cirkits.com/description/LCC-usb-flyer.pdf
  http://www.rr-cirkits.com/description/Tower-LCC-flyer.pdf

He expects production quantities by the Springfield show next month.

Yes, the selling point is mostly bandwidth. LCC runs at 125K bps. However, the NMRA gloss-over seriously shortchanges the programmability of the system, and that configuration and state information are stored in each node, non-volatile. No more wringing hands over startup states and hoping that JMRI (or whatever) initializes the layout properly. Also, there are 2-way communications, where a change in state initiated by the controlled device - such as track detection - can launch events such as signal aspects, turnouts and operator panels, even without a supervisor such as JMRI.

Dick said the biggest shortcoming at the moment is the JMRI work is not ready. The bootloader for the "decoders" (logic nodes) is not finished, and he described the current state of the JMRI programming and configuration tools as "very cumbersome".

I sent him the link to the GC&W website, which elicited a comment about a LCC-to-DMX512 interface on his radar. @Blazeman made a half-serious comment in my layout thread about interfacing DCC to my lighting system... well, there you go, it's already in early design.

As to c/mri... I know nothing about it, and need to educate myself, I s'pose. An established supplier jumping behind LCC with plug-ready product might let a bit of air out of their balloon. But you're right, we don't need another VHS vs. Betamax situation.

I'm standing by to hear back from Dr. McRee about any plans he may have with LCC. It would be great if he could do a servo controller with direct LCC.

[C855B]

I heard back from Duncan McRee (Tam Valley)... "No plans for LCC." Don't blame him, he has 90% of what LCC accomplishes for most MRR'ers in his QuadLN product.

[C855B]

...
* The c/mri crowd is still alive and well and they are actively developing Arduino based applications...

The crowd may still be alive and remain die-hards to the cause, but C/MRI itself appears to be turning into a footnote in MRR history. It seems to have degraded into a hobby-within-a-hobby, not a tool for integrating something bigger. So far everything I've found online are references to roll-your-own or dead links. Oh... I think I found one site selling boards clearly of 1980's designs.

Speaking of researching LCC and its progenitors such as OpenLCB and C/MRI, I am... frankly... horrified by the number of dead links, abandoned domains, circular references and products that vanished after one shot, or didn't even make it beyond prototypes. NMRA backing the LCC project seems to be one primarily of herding cats. Maybe Dick Bronson can make hay with it. I have to hitch my cart to somebody's horse sooner or later.

EDIT: I just joined the "Official NMRA LCC" Yahoo group. You can hear the crickets chirping - no posts since Nov. 8. The supposed official website, nmranet.org, links to an error message. This is looking worse and worse the deeper I get into it.

[eja]

I always thought a bus was something you rode to work .......

   

[peteski]

I always thought a bus was something you rode to work .......

   

Oh, is that why some people (incorrectly) call the electronic version "buss"?   

[C855B]

No, 'cause "Buss" was a brand of fuses. (Now known as "Bussmann".)

And I thought "CAN bus" was mass transport for our friends up north.

[GaryHinshaw]

Thanks for the follow up.  I feel pretty comfortable forging ahead with the Simple Serial Bus, and if Dick comes out with something even better I'll have a look.

[John]

Thanks for the follow up.  I feel pretty comfortable forging ahead with the Simple Serial Bus, and if Dick comes out with something even better I'll have a look.

his SSB is still Loconet compatible .. so you should be ok

[GaryHinshaw]

Yep, the boards I have installed now are working great, once I got them programmed correctly.

[cgw]

LCC is a knock off of the Merg Can-bus that was developed by that group.     Check out the model electric railways group on the web  They are based out in the UK.    They are very active and extremely helpful if you wan to get into the Can-bus system.       Another system that has can-bus architecture is the ROC rail  group.