Raymarine St4000 autopilot ... Freaking out

[pgenovese]

I had an incident with my st4000+ autopilot.

I was steadily motoring on a course (COG 318deg) when suddenly the autopilot veered hard to port. I took manual control and brought the boat back on course when i noticed that the heading display on the AP was way off. The heading should have been pretty close to my cog accounting for 2.5degrees variation, some 3 degrees of deviation.it should not have been more than 10degrees off my COG. Basically with a COG of 318 degrees the AP was showing a heading of 026 degrees M ...

When i re engaged the AP the boat. Veered off course but then slowly came round to the correct COG while still indicating a wrong heading. The plotter was shoing a massive set & drift which was nowhere near reality.

I then did a heading alignment on the AP control head and brought the heading back in line and it worked fine for the rest of the trip.

On y return trip it did exactly the same freak out about 3 hours into the passage and did not repeat for the entire 11 hour passage.

Any ideas on what could be wrong and / or how to go about diagnosing this ?

Regards

Patrick

[Tymadman]

Did you happen to notice what heading your magnetic compass was showing at the time?

The basic thing to determine is whether there was an external influence affecting the compass of the AP or whether the AP itself has a problem.

[GILow]

Originally Posted by pgenovese

On y return trip it did exactly the same freak out about 3 hours into the passage and did not repeat for the entire 11 hour passage.

On the return trip... so over the same ground? Silly question maybe, but was it at roughly the same position as the trip out?

If so, as per Tymadman's comment about external influences, could it have been the same external influence at the same point? I don't know about the ST4000+ but my ST1000+ plus had an incredibly sensitive compass (some kind of flux gate circuit I think, something made by Philips) and if I moved anything metallic around in the galley of our old 20 footer, it would upset the course quite a bit. So maybe, just maybe, you passed over some big ferrous object under water.

A bit X-Files, I know...

[pgenovese]

Yes the return trip was the reverse route of the outbound track, however the positions of the incidents was about 30 miles apart.


The only correlation i can find is that the locations are cris crossed by submarine cables probably telecoms ... But in both cases the cables were very deep 400m on the first incident and 860m on the second. Also we crossed several other cables in the same area after the first and second incidents with no similar effects.

It they were power cables they would be carrying AC current which would not have any effect on the compass as the mag field is continually reversing.

I am suspecting that the flux gate compass got stuck somehow.

[GILow]

Kinda spooky. I suppose plotting the points on something handy would be interesting. Is that possible? You might find there is a line or something indicating one of the cables is giving off some kind of EMF.

I have a chip at home which has a flux gate compass in it and I was amazed at how sensitive it is when I was playing with it. (OK, so I started to make my own autopilot, silly idea, but I got close!)

[GILow]

[QUOTE=GILow;1319137] You might find there is a line or something indicating one of the cables is giving off some kind of EMF.
QUOTE]

Of course, to validate the line theory you need a third event to take place in a line with the other two....

Oh well, I guess you just have to go sailing then. In the name of science!

[Tymadman]

Next thing is to look at what might be on the boat which could influence the compass. My ST1000 had me confused one day as it was reading 50 degrees different to my regular compass. It turned out an umbrella nearby (18" away) was what was affecting it.

[CarinaPDX]

The (obsolete) RM compass design connects the flux gate directly to the controller (no apparent buffering/digitizing inside the compass); the leads are actually carrying the analog signal for 15ft or so. It would not take much corrosion or cable damage to cause problems. I recommend that you carefully check the cable and its connection at the controller. It is possible that there is corrosion inside the compass but I think that is probably less likely; still, it is pretty easy to take the compass apart and see.

The usual cause of sudden compass course changes is movement of iron or steel items near the compass. This can be a lot more subtle than you may realize. My first location was affected by the gimballed stove swinging; my second by the steel bars inside Pendaflex file folders. Anything with a speaker inside can be surprisingly effective: at the large end of the scale an 8" subwoofer messed up my compass at 6ft (solution: buy shielded speakers). Metallic influence is where to start looking IMHO, but check the compass wiring to be sure.

Greg

[pgenovese]

The strange thing is that the offset remained quite constant throughout the outbound trip and was almost nullified when i switched off the AP on arrival. Inf act he next morning I got up tear,y to try to look into the issue only to fined the heading to be very close to the correct heading.

On the return trip i immediately did a heading alignment and did not have another issue for another 7 hrs despite passing over almost the exact same spot where i had the issue on the outbound leg.

I am more inclined to thing that this is some sort of technical fault rather than external influence because i am very careful about what I stow near the Flux gate compass or in its vicinity and i also tried moving stuff around to see if the heading offset changes but to no avail.

Does anyone know how to formally test the Fgc signal to the AP? Ie voltages etc... It seems that it does not send digital data right ? It os then the AP that converts the signal into a digital data stream correct ?

[RaymondR]

I had a problem with offset of the fluxgate compass wandering about which turned out to be that the fluxgate is mounted on one of the legs for the solar panel mounting and I had rerun some of the DC cables from the panels down the same leg.

I would recalibrate the compass in the anchorage in the morning then it would go crazy later in the day and the autopilot would only keep a steady course whilst in track mode.

I guess what was happening was that as the sun rose towards the zenith and the current from the panels increased, the magnetic field around the current carrying conductor would increase in strength and affect the heading as inferred by the AP software.

I believe that the fluxgate was developed during WW2 as one of Coastal Commands anti submarine detectors for use from aircraft.

[RaymondR]

Does anyone know how to formally test the Fgc signal to the AP? Ie voltages etc... It seems that it does not send digital data right ? It os then the AP that converts the signal into a digital data stream correct ?

The fluxgate itself is a small ferite ring with two sets of windings, a bit like a transformer. One winding excites the magnetic field in the torus and induces currents in the second winding. Because the hysteresis in the magnetic material is effected by external magnetic fields small lag or lead differences can be measure in the current as the magnetic field flips direction and the direction of the external magnetic influence can be inferred from the slight phasing variations between the two currents. Or something like that.

It appears that it would be very fairly difficult to extract any meaningful information for problem solving or analysis purposes although you might be able to see the variations on a dual trace oscilloscope.

[GILow]

Originally Posted by pgenovese

The strange thing is that the offset remained quite constant throughout the outbound trip and was almost nullified when i switched off the AP on arrival.

Rats, there goes my XFiles theory. I really wanted it to be a lost submarine from the future with massive magnetic fields left over from the time travel.

[RaymondR]

My ST4000+ went haywire once after I had passed under a thunder storm, I think the program must have been corrupted by EMP.

I managed to get a Raymarine technician on the phone and asked him how to reset and reload the program.

He told me to go into Dealer Setup, choose a different model and then go out of Dealer Setup and that this would overwrite the corrupted program. I was then to go back and choose the correct program which would be reloaded when I went out of dealer setup again.

Worked and I have never had to do it again.

[Rhumbline]

I've been following this thread for some time because I've been experiencing similar problems with my ST4000+. The heading would sometimes display readings all over the map, and the boat would make random turns when running for 10 - 15 minutes. Totally unusable, although this unit used to work fine. I tried several times to calibrate the deviation but could never get a good value - always 50 degrees after several slow turns. No good. I dismounted the fluxgate compass and manually rotated it and noticed the headings didn't make sense; rotating from 270 to 360 caused the display to show lower numbers,like 100 degrees. Raymarine tech support indicated the problem was either in the fluxgate or the control head, which is now unserviceable, and they basically told me to buy a new fluxgate/control head/course computer, which is not really an option for me at this time. I ohmed-out the leads from the fluxgate as instructed, and noticed I had to press the meter leads very hard to get a reading from the leads connecting to the control head. I actually had to scratch the surface of one connector to get a good (non-open) reading. I worked the leads from the compass to the control head on/off several times to try to clean them. I then retested by manually rotating the flux gate. and now I seemed to be getting decent readings. I just took the boat out to try to calibrate the deviation, and I was able to get a good number - 4 degrees - after only a turn and a half of the boat. I then set the heading to what my ship's compass was reading. Autopilot tracked well for over a hour, so I think this is fixed. I am a happy man!

Bottom line - it is highly unlikely your course computer in the control head or the flux gate has gone bad and much more likely you just have a bad (oxidized) connection where the leads attach into the control head. You might want to reseat all the connections to the control head - especially compass to control head - as part of annual maintenance.

Bill Rathbun
s/v Vector
C&C 38 Mk II

[a64pilot]

A little silicone grease or similar may keep them from oxidizing again.
May be silly but I have had real good luck essentially packing all connections with silicone grease