If ...

The CAN signal does not go through the alternator!!!

I thought you claimed to know something about CAN in cars. It uses differential signalling through solid wiring with transmitters and receivers having power condition between the car's alternator/battery and the electronics.

Trying to send a high data rate signal through intermittent contacts (rails, wheels and pickups) will require a far greater level of error checking and correction than CAN currently has.

Why do you thing the DCC data rate is so slow? Why do you thing DCC packets are transmitted repeatedly?

MBQ

You don't *have* to - all CAN cares about is that there is 12v (or so) between the two wires (0 & 12V or 100 & 112V doesn't matter to CAN, though the actual hardware would have to able to live with the higher voltages of course). Data can be superimposed on power rails with the right decoder chips - there's a paper on the web describling this from the early days of CAN in industrial control, where it came from, something to do with Boeing I seem to recall. Mind you, there's also the Mitsubushi 1-wire system (uses the car body as 0V) - I've no idea if that went anywhere though - they did have a car using it some years back, but redundancy was an (percieved?) issue.

Richard

It can cope with a duff alternator, so I wouldn't bet on it.

Richard

Am I missing some thing here - always possible. I need 3 amps at 12V to power the motor on my Loco, and DCC can quite happily supply that along with the control signals over the tracks. CAN needs a SEPARATE power supply to power the target motor, it can't supply it with the data?

The electronics in a car does not run directly off the battery. There is further power condition (i.e. regulators and smoothing caps) in each module, because of the very havoc you mention above. OK this conditioning circuitry is part of the electronics, so it is running off the battery, but you know what I mean.

Can copes very well with noise due to the use of differential signalling. I am familiar with the spec and, IMHO, it would not cope so well with intermittent contact between the transmitter and receiver as you would get it you tried to send it across the rails.

MBQ

In never said it did!!! A duff alternator creates havoc on vehicle electrical systems, vehicle wide - do pay attention at the back!

You mean it runs off the battey? Yes, as does everything else on the vehicle. And while the battery does provide some smoothing, it by no means compensates for everyting - CAN copes extremely well with noise, which why it was adopted in prefernece to other proposed standards - VAN etc. I was origionaly specifed to cope with machines etc in an industral environment, which is a very noisy environment.

CAN isn't a high speed bus (even High Speed CAN has nothing like he data rates of even bog standard Ethernet).

Check out the CAN specs. There's nothing more to say.

Richard

And the circuitry is a lot simpler.

If you want CAN and significant power (e.g. as a DCC type replacement) on the same two wires then you have to be very careful. Where does the power come from when a decoder wants to send a message? It's exactly the same problem as with DCC bidirectional comms..

How many wires do you think Ethernet uses ;-)

CAN is robust due to the chosen signalling system and the error detection. It's two wire because it's designed to be cheap.

MBQ

In message Lester Caine wrote:

They don't need to be separate - CAN can run using just two wires ( the rails in this case) - the data is superimposed (an A/C element, if you like) on the 12V DC carrier. The two wires carry current and data, just like (well, except for A/C bit) "network over mains" systems. The decoder chips, on the locomotive, body controller ECU, milling machine, or whatever just monitor the A/C signals (the data), and when a the pattern matches that of the devices address it says "ah ha! That's for me" and decodes the following data (an oversimplification, but good enough). By a series acknowledgements to the sender, the decoder tells the sender when it has revieved an intelligable message. In other words, in principle, it is operates in exactly the same way as DDC, but the design of the data packets and addressing is, from what I can see of DCC, far superior. I think confusion is creeping in because CAN, in the automotive world anyway, is usually implemented with two additional wires, 0V and +12V to power the decoders, but it doesn't *have* to be - automotive systems do that for redundancy, and is has become something of a defacto standard when used elsewhere. CAN is robust (enough to be used in miltary systems) *because* it is only two wire (unlike Ethernet etc) - it only looks at the 12V difference between the two lines. Even if the voltages are varying wildly as long as the difference between the two is still 12V (+/-6V if I remember right) it doesn't matter - hence its robustness. It is also configureable for short or long packets, using short packets again aids rebustness in noisy environments.

Richard

But the current CAN bus specification is for a 120R terminated twisted pair for the data? So what you are talking about is one of the DC-BUS systems which give 10kps or so but needs a complex modem arrangement to handle it? I did not think there were readily available chips for that yet :( And it uses a high frequency carrier that has a problem meeting EMC limits when taken outside a screened box like a car :(

I suppose we'll just have to agree to disagree on that - all I can say is the two of my lecturers at Uni worked on Zero I, and they were of the opinion that CAN would have helped them had it been around, it it's fully grown form at the time. Unfortunately they would not go into details as they were still, theortically, connected with the project.

Richard

Quite probably, but if one is just buying a little black box, it's not realy an issue.

From the two wires.

As I have acknowledged - CAN is similar, but with better capabilities.

You've snipped at the wrong place, though I did put my - in the wrong place....

No, no, no, - the two wire design of CAN is a deliberate strategy to deal with noise, particularly induction, a major problem in industrial control, which was CAN's birthplace. CAN was desinged spefically to overcome problems with ethernet in a harsh environment.

Richard

Intel certainly did one a few years back - though I seem to recall they pulled out of the automotive market just before it took off - whether they pulled out of CAN chips in general I couldn't say.

As I recall from a seminar at York Uni, the EMC limits inside a car were rather more stringent that outside, due to proximity with safety critical systems (Air Bags were a hot potato at the time), but as it's now 5 or 6 years since I left the industry to do something more interesting it is quite possible that standards have changed - they were really only just coming into focus at the time with the launch of the EU marking scheme (that is, dealing with the standards imposed - the standards themselves had been being squabbled about for some time).

Richard

In message , Greg Procter writes

Wait a few years and you won't need wires at all.

I'm sure those 2 foot coils used to light a 60W bulb will look very primitive in a few years!

In 1979 in "The Fountains of Paradise", Ray Bradbury wrote about space elevators using nanofibers (another concept recently in the news), culminating in a ring of mirrors surrounding the Earth and beaming power to ground stations (light energy, perhaps, rather than electromagnetic).

But are you not describing an implementation, thats not the same as a definition.

Cheers, Simon

Ok, but I'm sure that the fact that it works allows for a definition.

Shouldnt you be in bed still ?

Not really cos thats like saying I define gravity by not floating away.

Cheers, Simon

It's 11:04 am! I'm an early riser.

What I was trying to say was more like 'gravity stops me floating away, therefore a definition is possible'. I'll leave it up to you electronics savvy people to come up with the ultimate definition, meanwhile I'm happy utilizing it.

Regards, Greg.P.

Sorry Simon but I already have my Zephyr, which was bought before Hornby released any DCC kit...

Cheers Simon C.