NMEA ground vs boat earth

skyflyer

skyflyer

Well-Known Member
Joined
26 Jan 2011
Messages
1,433
Location
Worcester, UK
Visit site
Is there any reason these can not be the same?

If not, how do you keep the NMEA ground separate?

All my normal 12v electrics are referenced to a common "earth" with everything connected back to a big busbar that is connected to battery negative.

My instrument network is the older two wire NMEA0183 type so basically a TX line an RX line and a common ground line.

Obviously the instruments also have an electrical supply, so there is a +12v and 0v (earth) terminal on each.

I am trying to minimise the wiring runs (not for economy but to save space in tight conduits) and wondered if i can reference everything to the instrument supply/boat earth or must i run a separate earth wire for the instrument signal network?

If I do run a seperate ground wire around the circuit then I assume it needs to be connected to something somewhere else how is it defined as an earth?

I'm sure I am overthinking this, but have now tied myself in knots!

Thanks
 
Hopefully, this will help.

You cannot really think about "earth" on a boat dc system. Earth is something whose electrical potential is conventionally taken as zero. You need the loop from battery positive to negative to complete a DC circuit and the non-switched side can be run to a common bus as you describe. Grounding, for various reasons, to protect from static for instance, is outside this loop. Does not seem sense or logical to mix this with current carrying circuits.
 
Ok, so IIUC the view is that the 'ground' or 'common' wire of the NMEA circuits should be separate from the normal electrical 'negative'.

But as someone says above, if within an instrument, the electrical power ground is linked internally to the nmea ground theres not much point!
 
skyflyer said:
Ok, so IIUC the view is that the 'ground' or 'common' wire of the NMEA circuits should be separate from the normal electrical 'negative'.

But as someone says above, if within an instrument, the electrical power ground is linked internally to the nmea ground theres not much point!
Click to expand...

The original NMEA 0183 spec called for a signal Hi &signal Lo encased in a shielded cable with the shield connected to equipment cabinet at sending unit only.
The cabinet was usually internally connected to the sender unit's negative DC supply,but not always,especially in higher quality products & Loran C.
The idea was that any external electrical interference could not penetrate the shield & corrupt the data on Hi & Lo.
Mfgrs were not long in discovering that they could save a penny by not using shielded cable on 90% of vessels-especially those small vessels using DC only.
They also discovered that the data return Lo could be connected to sender's DC neg.
The only time us techs would encounter interference was on usually metal ships with poor AC wiring,non-shielded hi power
eqpt,fluorescent lights,old TV's & other non-marine eqpt.

Today,for practical purposes,just use the 2 wire without shield & let the mfgr decide how to treat the data Lo inside their eqpt.
Do not attach any external "ground" or "Neg" to this 2 wire. Use the factory cable(or equivalent) & connect to both ends per the manuals.

The old shield was never intended to carry current-hence the connection at sending eqpt only. The worst thing you can do is have current flowing in a Radio Frequency (RF) shield.

The term "ground" is greatly abused by many :)

Cheers/ Len
 
Last edited:
Len Ingalls said:
The original NMEA 0183 spec called for a signal Hi &signal Lo encased in a shielded cable with the shield connected to equipment cabinet at sending unit only.
The cabinet was usually internally connected to the sender unit's negative DC supply,but not always,especially in higher quality products & Loran C.
The idea was that any external electrical interference could not penetrate the shield & corrupt the data on Hi & Lo.
Mfgrs were not long in discovering that they could save a penny by not using shielded cable on 90% of vessels-especially those small vessels using DC only.
They also discovered that the data return Lo could be connected to sender's DC neg.
The only time us techs would encounter interference was on usually metal ships with poor AC wiring,non-shielded hi power
eqpt,fluorescent lights,old TV's & other non-marine eqpt.

Today,for practical purposes,just use the 2 wire without shield & let the mfgr decide how to treat the data Lo inside their eqpt.
Do not attach any external "ground" or "Neg" to this 2 wire. Use the factory cable(or equivalent) & connect to both ends per the manuals.

The old shield was never intended to carry current-hence the connection at sending eqpt only. The worst thing you can do is have current flowing in a Radio Frequency (RF) shield.

The term "ground" is greatly abused by many :)

Cheers/ Len
Click to expand...

When connecting devices where nmea - & power neg are internally connected, isn't there a danger of creating a cobweb of parallel neg power paths?
(scuse the lack of drawing skills..)

bdde13eb-eae0-42aa-b055-8a291e5d5ff0.png
 
Older Garmin and maybe newer stuff as well had a counterintuitive use of the grounding cable which baffled most of us who had ever wired up NMEA gear - Google "Garmin NMEA" and many hits emerge of confused yachties being talked through it. Of course if you only had Garmin stuff it was fine. Shame, they made damn fine GPS units.
 
I think there's a bit of overthinking going on here. I usually complete NMEA circuits by using a hi-lo pair where possible but if this isn't convenient/possible (e.g. common ground garmin) I'll wire the NMEA low to the battery negative busbar. My 0183 system is a mix of the two as it includes some Garmin gear which forces this connection.

I've never found it necessary to shield any installation on a 30ft GRP boat, although the manufacturer's pre-installed cables for the longer runs are shielded (the shield is, and always was, connected to nothing so I assume it purposeless).

The key trick is to wire the system up and check that it works. If it works then it works, QED. Sail.
 
As I understand the situation NMEA 0183 specifications were written to specify the 'physical layer' (the electrical part) to be 2 wire balanced with no connection to 'ground'. Usually connected by twisted pair wires with or without a screen which could be grounded. Balanced (and screened) circuits have the advantage of greater noise immunity, specially over long cable runs. But balanced I/O circuits are more expensive to manufacture than 'single sided/unbalanced' where one line is ground, so we saw the higher end manufacturers producing 'pure' balanced I/O circuits and lower cost units being unbalanced. They can be mixed (with exceptions) but when they are the noise immunity is negated.
In practice, if a NMEA 0183 'talker' has + and - wires they connect to + and - of the 'listener', but if either the talker or the listener is un-balanced, the + will connect to + and the -ve will be connected to ground/earth.
The short runs for recreational boating systems usually do not need the benefits of the 'pure' balanced circuits so the mixing of the two doesn't cause problems (exceptions noted on this forum).
But beware, some NMEA un-balanced 'talkers' operate with TTL o/p circuits which only give 0-5V whilst some 'listeners' expect +/- 12V data! However in my experience 90% of the balanced-to-unbalanced connections work without any special attention.
 
GHA said:
Or when you're not sailing find out more about how things work. Cos you might just have got lucky on a good day.
And because it's fun and interesting :cool:
Click to expand...

+ 1

'trial and error' might better be called 'trial and execution'; it reminds me of my father in law whose method of wiring 240v appliances was to try any combination until the fuse didn't blow. When i gently pointed out that having live connected to the appliance earth would achieve this result, (and the pitfalls!) he thankfully decided to give up home elctrics
 
The specification for the talker end has changed during the life of NMEA0183. It used to be single ended - the negative line being ground of the talker, similar to RS232. This changed somewhere around version 3 to differential, the negative line no longer being connected to talker ground, similar to RS422. The listener end has not changed through the versions. It has always been opto-isolated. Either talker configuration works with the listener circuit.

A problem can come when connecting a later version differential talker to a non-isolated single ended listener. In one of the binary phases an asserted signal line of the talker is connected to the ground of the listener. This can damage the talker output circuit, although most driver chips are tolerant of this sort of abuse. Manufacturers, like Garmin, making non-compliant equipment without the listener isolation is penny pinching and in my opinion inexcusable.

Gypsy said:
As I understand the situation NMEA 0183 specifications were written to specify the 'physical layer' (the electrical part) to be 2 wire balanced with no connection to 'ground'. Usually connected by twisted pair wires with or without a screen which could be grounded. Balanced (and screened) circuits have the advantage of greater noise immunity, specially over long cable runs. But balanced I/O circuits are more expensive to manufacture than 'single sided/unbalanced' where one line is ground, so we saw the higher end manufacturers producing 'pure' balanced I/O circuits and lower cost units being unbalanced. They can be mixed (with exceptions) but when they are the noise immunity is negated.
In practice, if a NMEA 0183 'talker' has + and - wires they connect to + and - of the 'listener', but if either the talker or the listener is un-balanced, the + will connect to + and the -ve will be connected to ground/earth.
Click to expand...
 
Last edited:
GHA said:
When connecting devices where nmea - & power neg are internally connected, isn't there a danger of creating a cobweb of parallel neg power paths?
(scuse the lack of drawing skills..)

bdde13eb-eae0-42aa-b055-8a291e5d5ff0.png
Click to expand...

Yes-your drawing could definitely cause problems. If one piece of eqpt in your web loses it's DC bat. Neg connection,then it will use the NMEA Lo wire as a Neg power source. Problems could range from corrupted data to frying the Lo wire as data wires tend to be small gauge & not capable of carrying several amps of DC current. I've seen this happen especially in eqpt that had a fuse in the Neg bat. supply lead.View attachment 55705

That is where the opto isolator specified in the original NMEA 0183 specs comes into play. According to original specs :
(1) If a 2wire plus shield cable was used,the shield was only grounded at the sending eqpt-never at the receiver.
(2) Both the data Hi & data LO were connected to an opto isolator at receiver & never to receiver's pos or neg bat. supply.
The 2 signal wires should not be called dat pos & data ground or neg.

Using the opto isolator prevents the data lo from being in contact with receiver's Neg bat or chassis ground.

However,as I & other posters have said,the original specs have been "bastardized" to save pennies & 90% of the time a one wire data Hi & a Neg bat data Lo connection will work-as long as you have a simple small boat,short run cables,clean DC system. & nothing goes wrong.

We really don't have much choice as the majority of small boat eqpt mfgrs only supply what they supply & we are forced to work with it as is,or hand build our own "adapter" ckts-which isn't practical for the avg. user.
I've never done it,but if you are concerned enough,I suppose you could use 1amp inline fuses in your data Hi & Lo leads.This would help eliminate possible frying of data wires if they are called on to provide DC bat current to a unit that has lost it's normal bat Neg connection.

Most of my career was on comm.fishing & shipping vessels. I would occasionally encounter "loop" problems due to hi powered fish finders,radios,etc. as well as crappy AC stuff & other things caused by poor practice.
I encountered very few problems on fiberglass comm. boats relatively speaking.
One problem that was common to both is "neat" wiring-the tendency to run lengths of cables from various eqpt in long parallel runs in or out of conduits. High current/voltage cables such as fishfinder transducer cables,etc. could cause induced crosstalk in other wires. Keep this in mind-though most yacht stuff is not that powerful.
A common complaint was noisy VHF reception & sometimes TX. 99% of the time,the antenna coax braid was corroded/broken allowing electrical interference to enter the center conductor because shield braid was now open.

Just some of my experiences.

Cheers/ Len
 
Len Ingalls said:
Yes-your drawing could definitely cause problems. If one piece of eqpt in your web loses it's DC bat. Neg connection,then it will use the NMEA Lo wire as a Neg power source. Problems could range from corrupted data to frying the Lo wire as data wires tend to be small gauge & not capable of carrying several amps of DC current. I've seen this happen especially in eqpt that had a fuse in the Neg bat. supply lead.View attachment 55705

That is where the opto isolator specified in the original NMEA 0183 specs comes into play. According to original specs :
(1) If a 2wire plus shield cable was used,the shield was only grounded at the sending eqpt-never at the receiver.
(2) Both the data Hi & data LO were connected to an opto isolator at receiver & never to receiver's pos or neg bat. supply.
The 2 signal wires should not be called dat pos & data ground or neg.

Using the opto isolator prevents the data lo from being in contact with receiver's Neg bat or chassis ground.

However,as I & other posters have said,the original specs have been "bastardized" to save pennies & 90% of the time a one wire data Hi & a Neg bat data Lo connection will work-as long as you have a simple small boat,short run cables,clean DC system. & nothing goes wrong.

We really don't have much choice as the majority of small boat eqpt mfgrs only supply what they supply & we are forced to work with it as is,or hand build our own "adapter" ckts-which isn't practical for the avg. user.
I've never done it,but if you are concerned enough,I suppose you could use 1amp inline fuses in your data Hi & Lo leads.This would help eliminate possible frying of data wires if they are called on to provide DC bat current to a unit that has lost it's normal bat Neg connection.

Most of my career was on comm.fishing & shipping vessels. I would occasionally encounter "loop" problems due to hi powered fish finders,radios,etc. as well as crappy AC stuff & other things caused by poor practice.
I encountered very few problems on fiberglass comm. boats relatively speaking.
One problem that was common to both is "neat" wiring-the tendency to run lengths of cables from various eqpt in long parallel runs in or out of conduits. High current/voltage cables such as fishfinder transducer cables,etc. could cause induced crosstalk in other wires. Keep this in mind-though most yacht stuff is not that powerful.
A common complaint was noisy VHF reception & sometimes TX. 99% of the time,the antenna coax braid was corroded/broken allowing electrical interference to enter the center conductor because shield braid was now open.

Just some of my experiences.

Cheers/ Len
Click to expand...

Len

In that case surely its better to NOT have a separate 'signal ground' but use the general boat ground, and then there is never any chance of a meaty bit of kit using a thin signal wire as a return?

The general view seems to be that as somewhere on the circuit an instrument/transducer manufacturer will have connected the "power ground" to the "signal ground" anyway, there's little point trying to isolate the two
 
You must log in or register to reply here.

Other threads that may be of interest

Lightwave395
Wiring mess
Replies
18
Views
809
thinwater
thinwater

Share this page

Top