Originally Posted by LarryM
FWIW, I recently had some discussion with my most trusted Simrad (I know, bad word in your world) technician. I even recounted some of your prior adventures as I understood them. He remarked that they had similar problems on vessels 70' and larger due to signal loss, noise, crosstalk etc. He stated categorically that he won't use anything but Maretron
N2K cables and connectors end-to-end when building N2K networks. In much the same way that there large differences in the quality and speed of Ethernet, and other network cables, apparently, some of the really well known marine electronics manufacturers have been known to provide rather poor quality N2K cables and connectors.
I don't know if this has any bearing on the issues you have faced, but I thought I would just pass it along.
Yes, this can quickly become a problem if you use the smaller micro cables for both your backbone and your drop cables. I used the smaller cables on my Grand Banks and it was getting close to the edge out the extremes of the network.
My current boat's N2K backbone is made from the big fat "mid" cable. It's about 1/2" diameter. All aspects of the network are well within electrical specs for N2K, including voltage drops, cable lengths, drop lengths, cumulative drop lengths, etc. All cables, tees, and terminators are Maretron. The network has been reviewed and inspected by multiple people, including Simrad and Maretron and has passed with flying colors every time. It has also been tested with Mareton's test meter, and passed.
A poorly constructed network will typically result in data errors because it is electrically marginal. N2K data is all checksum protected so such errors are detected and the PGN's discarded. Many devices keep track of such errors, including Maretron's analyzer device. Seeing checksum errors or other data errors is a sure sign of an electrically marginal network. FWIW, I have never seen a single data error on my network, ever, at any time.
The problems I've seen are 100% behavior, i.e. the wrong PGNs at the wrong time, erroneous PGNs, and interactions between devices. With an analyzer that can decipher and log PGNs, the problems are mostly very visible and evident in those logs. They are all software bugs, plain and simple.
As an aside, only a few of the Simrad problems I had were N2K issues. Their N2K implementation is as good or better than most.