Panbo

Actisense & Furuno N2K junction boxes, impressions

... written for Panbo by Ben Ellison and posted on Sep 9, 2008

Actisense_Quick_Block_cPanbo

I used Actisense’s QNB-1 Quick Network Block, above and bigger here, to create Gizmo’s little NMEA 2000 network, and it pretty much lived up to expectations. While using regular tees and fixed cable lengths would have been messy in the small confines of my console, this junction box let me cut cables to length and still end up with a fairly waterproof system. The built-in fuses and LED status lights made the install easier, and those cable glands provide good strain relief on a variety of cable sizes (add Tommy Tape when they don’t). But I must say that fixing wires to the terminal strips can be tedious, and downright hard in tight quarters (I did most of the wiring on a bench).

But maybe the hard-to-wire complaint is just due to my experience with how easy it is to wire Furuno’s similar FI5002 junction box. As you can see in the photo below, and bigger here, Furuno used Wago terminal blocks which can be unplugged for easy wiring. Note, though, that the FI5002 is a lot less weather proof than the QNB-1, and lacks built in fuses and status LEDs. But it does have little jumpers (not visible in the picture) that can be set to terminate either or both ends of the backbone. If you want to use the QNB-1 as a complete backbone you have to wire in two short cables with terminators attached.
   Interestingly, the QNB-1 doesn’t delineate which two of the eight cable connections are for the backbone, and I don’t think it really matters on the FI5002 either, except that it has two extra-large terminal blocks in case you backbone with Mini-size cable instead of Micro. In other words, the architecture of these junction boxes seems more like a star than the linear backbone/tee/drop we’re used to.
  Finally, Actisense announced a reconfigured NDC-4 NMEA 0183 multiplexer today. It sounds good, but the press release description also suggests some of the 0183 issues you won’t run into using N2K. 

Furuno_FI_5002_cPanbo

Comments

If you didn't notice, its slightly easier to remove the glands and take the circuit board out of the box to get everything connected first.

Posted by: Sandy at September 10, 2008 10:01 AM | Reply

I was talking to a guy here in Seattle who is on the NEMA standards committee and I was calling this thing a hub. He seemed very confused by that. Isn't this a hub? Don't all the instruments hangoff this box like a network hub?

Very confused....

Thanks in advance for the pending enlightenment. :-)
-p

Posted by: Patrick at September 14, 2008 3:25 AM | Reply

NEMA or NMEA? Quite a difference!

As to the "hub" terminology, I wouldn't call this a hub since it contains no electronics. It's just a fancy way of connecting wires, so my vote is for the term "junction box".

Also note that unlike Ethernet, where you -must- use a hub (or switch) to connect > 2 ports, here you can pretty much connect ports any way you want as long as you keep the drops from the main cable short, and terminate the ends.

So think of these things as multiple-T-in-one-box and you're set.

Posted by: Kees at September 15, 2008 11:16 AM | Reply

NEMA vs NMEA?
Actually they are very close, less than an hour's drive depending on traffic through the District of Columbia. National Marine Electronics Association (they set standards for Gray boxes) is in Rosslyn Virginia, just across the Potomac and probably within site of the Capitol. National Marine Electronics Association (Black boxes) is in Severna Park, just north of Annapolis. Since its a long distance call, they don't talk to each other that much. [g]

Posted by: Sandy at September 19, 2008 12:45 PM | Reply

Of course its a hub, a hub is a device that simulates network wire and connectors, and that's what it does. There are no "electronics" in a hub...perhaps you're confusing a hub with a switch, or router.

A hub is the dumbest device around...;-]

Cheers,
Russ

Posted by: Russ Cooper at September 19, 2008 7:53 PM | Reply

Explain to me again why these things have terminal blocks and not NMEA2000 connectors? Kind of belies the word "Quick" in the Actisense device's name.

Also, given that the boxes cost $150 or so, what is the advantage over just stringing a bunch of Ts together? Ts are cheap or free (Garmin at least provides a T with every NMEA2000 product), relatively compact, and all the Ts I've seen have a screw hole for mounting to a bulkhead. Not to mention which, if you need to disconnect something or replace a cable, a standard plug is much friendlier.

I might be interested in a "hub" that was $80 and had regular N2K connectors, but these don't make sense to me.

Posted by: Adam at March 9, 2010 1:41 PM | Reply

Adam, these are just alternatives, particularly useful if you're running N2K cable through tight chases. I too prefer connectors over raw wires.

But I have been learning how you have to be careful about impedance on large networks. I'll write about it soon, but I did not realize until recently that a string of tees in a backbone can cause impedance slight issues. That's why Maretron encourages the use of multiport boxes, which put the little impedance breaks on a drop cable instead of on the backbone.

If you're building a big network, Adam, and it sounds like you are, you should be careful about using first class cables and connectors, and designing a clean backbone. If I were you I'd submit an N2KBuilder file to Maretron for discussion.

Posted by: Ben at March 9, 2010 2:00 PM | Reply

Ben, thanks for the heads-up. My plan is to use Maretron "mid" cable for the backbone, but in any case I'm mocking up the entire network inside the boat before running wire through chases.

I messed around a bit with N2KBuilder but found the UI slow and confusing. Some of that may be Adobe AIR on OSX's fault. I'll try again on the PC.

Posted by: Adam at March 9, 2010 4:02 PM | Reply

Yep, looking at the N2KBuilder manual I now see that the entire menu bar is in fact missing when run on OSX. That makes it kind of hard to use. :) I thought AIR was supposed to prevent problems like that.

Posted by: Adam at March 10, 2010 1:27 AM | Reply

Ben, I've discovered something interesting: Maretron offers their own junction box, the CM-CF-4.

It has one male N2K connector for connection to the backbone and 4 female drop connections. Yes, you read that right: Maretron is endorsing a network topology that is neither pure backbone/drop nor pure star. These boxes are little stars that hang off the backbone. Most convenient.

And they're only $65 each, or $16 a device, which I consider pretty much a bargain in the N2K cabling world.

Finally, one more intriguing tidbit: if you look at the photo of the box in the 2010 Maretron catalog, you can see that it's actually a generic DeviceNet junction box manufactured by Turck. That suggests that if you choose your source well you may be able to get them sans the Maretron markup. Now we just need a Turck part number...

Posted by: Adam at March 14, 2010 4:20 AM | Reply

A little more information: This wiring scheme is referred within the DeviceNet standard as DevicePort. Allen-Bradley, who originally devised DeviceNet, offer a bunch of DevicePort boxes in 4, 6, and 8 port versions. See this catalog page (under "DevicePort"):

http://bit.ly/9BsTsf

I'm seeing 8-port DevicePort boxes on eBay for around $200, but am guessing that they can be found for less.

Posted by: Adam at March 14, 2010 5:10 AM | Reply

Ben, I should've searched first as I see Dan already did a good job covering the Maretron Multiport box. But I still think it could be interesting to look at 8-port DevicePort boxes as a way to further consolidate N2K wiring. I'll ask Maretron what they think.

Posted by: Adam at March 14, 2010 5:44 AM | Reply

Right, mulitport boxes are NMEA 2000 legal and have been around for a while. They are not equivalent to daisy chaining because the data and power lines do not pass through a device. They are fairly equivalent to multiple tee drops, and the total length from the backbone -- the cable from backbone to box plus from box to device -- can not exceed the maximum drop cable length (a conservative 20 feet for N2K micro cable sizes).

What I, and many others, have never paid much attention to are the impedance issues. Keeping it more or less right is critical once an N2K network gets large. Apparently every tee in a backbone is slight decrement to its impedance, a decrement that's not as important if it's on a drop instead. This is a plus for multiport boxes. It's also another reason NMEA doesn't like daisy chaining; there's no telling what's happening to impedance when N2K passes through a device.

Posted by: Ben at March 14, 2010 9:33 AM | Reply

Well, I just bought 5 x 4-port, 1 x 6-port, and 1 x 8-port DeviceNet boxes of this sort off eBay for a grand total of about $175. We'll see how they all work.

Posted by: Adam at March 18, 2010 4:10 PM | Reply

Leave a comment