I have a background in CanBus design and am actually thinking of entering the marine Nmea 2000 market. what got me excercised was the comment about nema 2K up the mast
Investigating NMEA 2K leaves me concerned however.
Why ?
1. Bus based systems get unreliable as they grow large , as poor connections, bad wiring and noise all grow
2. Theres no fault isolation, one short takes the whole network down and it can be very difficult to detect where that is.
3. NMEA is expanding the concept to include whole boat control and monitoring, ie power control, dc switching, etc. This means lots of additonal devices, and no isolation between mission critical system and non critical system
4. I have seen the displayed error rates from a multi manufacturer nmea2K network, error rates were high
5. There no prioritising message system ( except fast packet)
6. Difficult to debug without expensive diagnostics equipment
7. Power specifications mean that external powered devices will be needed or multiple segrated network power supplies, which tend to be physically spread around the network, hard to fault find
8. ITs too technically underpowered, ethernet would be a simpler faster and more industry standard basis to use
I think its probably fine as a simple instrument bus , a replacement for Nmea 0183, but as a whole boat bus , no way in its current form.
discuss?
The topic has been discussed at length on this site. N2K has some ardent supporters and those like yourself who suggest Ethernet have been branded "crazies".
The manufacturers have also made their choices, essentially following your recommendation and using N2K for instruments and ethernet for higher bandwidth traffic.
I don't think additional discussion will change the respective opinions.
While I'm the guy who coined "Ethernet crazies" here, it was mostly meant in jest. But I still fail to understand why Ethernet would be better than CANbus for the purposes NMEA 2000 was designed to accomplish.
If it really is an industry standard for such purposes, please show us where Ethernet is used to data network, and sometimes power, small tiny bandwidth sensors, switches and so forth. I'm not familiar with anything like that.
Incidentally, I think system monitoring and control was in the original conception of NMEA 2000, as was Ethernet as a complimentary way to move broadband data and bridge narrowband N2K data amongst serious processors. It's definitely taking some time to define the monitoring and control messages, but it was always in the game plan.
(updated, prior entry has typo�s)
I also have an Ethernet network on Breeze Pleeze that connects the chartplotters, Sirius weather radar, and soon HD radar.
I am a former (or is the proper term "recovering"?) "ethernet crazy". Look back at some old posts with me ranting if there is any doubt. Having used N2K this year, now I am a believer and even wrote about it here:
https://www.panbo.com/archives/2009/09/adventures_in_nmea_2000_wiring_part_ii.html
I just can't see there ever being a future generation of "wired Ethernet" masthead wind sensors, depth and speed sensors at reasonable prices. If they ever become Ethernet it will be the wireless kind, but I would think more likely Bluetooth or other consumer wireless protocol that (1) uses less power, (2) potentially accessible by a cell phone / PDA application and (3) supports devices of dissimilar owners being in proximity with each other without issue, would be a better choice for wireless protocol than a/b/g/ or n.
Without those core sensors being on our Ethernet networks, and with Ethernet being so beneficial for it's core strengths, I suspect the only hope us Ethernet crazies have is for effective gateways that will get the information from the other networks to our on board PC's and cell phone / PDA's.
Clearly it will be more productive and rewarding to focus on being a wireless crazy than a wired Ethernet Crazy going forward ... but if you want to make your case, go for it below �
"The manufacturers have also made their choices, essentially following your recommendation and using N2K for instruments and ethernet for higher bandwidth traffic."
Still time to change it. In my opinion Nmea 2K will fail as an all-boat bus and will only be fixed by a hodge podge of buffers and adaptors aka nmea 0183
Do you have any evidence of failure whatsoever, "goboatingnow"? And what exactly is your alternative to N2K?
I thought CanBus was invented purely to reduce wiring and as you say ( me being an IT guy) if all devices are on the same physical wire its very hard to troubleshoot and therefore one device can take out another.
A bit like the old coax or token ring..