Latest News

Welcome to MV Tanglewood. Please fill out the "Follow this blog by email" (to the right) to receive notifications of updates to the blog. Thanks, and enjoy!

Sunday, June 7, 2015

Simrad auto pilot bugs

This is an overdue post for people interested in Marine Electronics and their peculiarities.  Back when I started reporting all the problems I encountered with Simrad/Navico products, I promised to document the auto pilot problems.  Well, finally, here it is.  Sorry for the long delay, but I’ve been occupied with getting the boat working and getting underway for Alaska, which we finally are…  Besides, I’ve been trying to move beyond the Simrad fiasco.  But at the same time I think it’s important to hold companies accountable for their products, and in today’s world, sharing the good, bad, and ugly is the best way to do it.

I have to admit that I’m surprised by the issues I had with the Simrad pilot.  I had a nearly identical Simrad pilot on my Grand Banks and it worked great.  In general, they have a stellar reputation.  Really the only things different in the Nordhavn pilot was the addition of several Follow-up steering controls, and the NSO chart plotter which is also capable of controlling the pilot. 

Here’s what the system looked like:

  • AC12 auto pilot computer.  This is the brains that implements the steering control and drives the hydraulic steering pump.
  • AP28 controller.  This is a user control panel that lets you configure and control the pilot.  There were two of them, one in the pilot house, and one on the fly bridge.
  • RC42 rate compass.  This is an electronic compass used to keep track of where the boat is pointing and is a critical part of a good auto pilot implementation
  • RF25 rudder position sensor.  I had two of these for redundancy.  They link to the steering mechanism and report the rudder position to the pilot.
  • HS70 satellite compass.  This is a second compass that figures out your position and heading based on three GPS receivers.  It’s a very precise heading device, and is used as the primary instrument with the RC42 as a backup.
  • FU80 Follow up levers.  These are knobs/levers that are used to position the rudder instead of using the steering wheel.  It’s much like the tiller on a sail boat with it’s position directly reflecting the rudder position, except it tells the pilot where to position the rudder and the hydraulic system does the work.  These are really good for close in maneuvering where you often need to swing the rudder quickly from one extreme to the other.  It’s much faster than cranking the wheel.  And, by placing them at the wing stations, you can steer from there as well as from the pilot house or fly bridge.
  • NSO chart plotter.  This is the multi function chart plotter that, among other things, is able to act as an auto pilot controller just like the AP28.

There really is nothing unusual about this setup, hence my surprise that I encountered so many problem.  I can only guess at the causes, and won’t.  It’s really up to Simrad to figure all this out, and hopefully they will.

Here are the issues I encountered, in no particular order:

  • The first and most frustrating issue is that I was unable to build a redundant auto pilot system as desired, even though Simrad told me in no uncertain terms that it would work.  The idea was to have two computers, multiple control panels, dual rudder sensors, two heading sensors, such that if anything broke I could keep on going with little fan fare.  The Pilot controller has menus where you can select which rudder sensor to use, which heading sensor, and even which computer.  When I was planning the system, I detailed what I was thinking of doing and asked Simrad if it would work as I expected.  They told me in writing that it would do exactly what I asked.  So that’s what I bought and built.  Well, after hours and hours of beating our heads against the wall dealing with odd configuration issues, lost configuration information, etc., we finally gave up on the idea of dual pilots and rewired everything so that I had a single system with pre-installed spares that I could switch over to quickly in a pinch.  Simrad later told me that it wouldn’t work (duhh) and that it was never designed to be used that way.  Wow, thanks guys.
  • The AP28s and FU80s are powered by the N2K bus, and switched on via a button on the AP28.  Press the AP28 power button and the AP28 comes alive along with the FU80s.  Power down the AP28 and the FU80s power down too.  It’s pretty slick, except it doesn’t work.  This is one area where the introduction of the NSO into the mix may be the cause, but I’m not sure.
    • When you power on the NSO, it wakes up the FU80s, but it doesn’t power up the AP28.  That’s inconvenient, but not a lot worse.
    • But, when you power down the NSO, it does not also power down the FU80s.  This places all the FU80s into an alarm state making a beep beep beep sound, and there is no way to turn it off other than killing power to the N2K bus.  That’s a real problem.
    • The NSO has an option to disable its internal pilot controller, which I tried.  But even with the pilot controller disabled, it still powers up all the FU80s when you turn it on.
    • If you power up the AP28, it turns on the FU80s.  If you powering down the AP28, it also powers down the FU80s.  At least it does this most of the time, but not always.  A number of times I powered down the AP28 and the FU80s remained on, and promptly went into an alarmed state.
    • Here’s a real doozy.  If the NSO is turned on before the AC12 breaker is turned on, the AC12 becomes permanently undiscoverable, locked up, and dead.  Even after the AC12 power is on, the NSO will alarm saying there is no AP computer.  Turning on either of the AP28s also generated alarms saying there was no AP computer.  Shutting everything off and then powering up the AC12 first, followed by the NSO and/or AP28s had the same results – no auto pilot computer.  It is forever lost.  But, I finally tripped over how to unlock it.  If you cycle power on the N2K bus, it would reset.  Cycling power on the AC12 didn’t help – it had to be the N2K power.  Once N2K power had been cycled, the AC12 would again be discovered as expected provided you always apply power to the AC12 before you turn on the NSO..  How’s that for a screw-ball problem?
  • One mode you can operate the AP in is called Nav mode, where it follows a route provided from your chart plotter.  In this mode, it displays the name of the current waypoint that it’s steering towards.  Except it doesn’t properly interpret the name and always includes two random junk characters after the waypoint name.  So something like “Bald Point” gets displayed as “Bald Point&#”.  This one is harmless, but having such an in-your-face bug reflects so poorly on product testing and overall quality.
  • The rudder sensor reports the rudder position via a message on the N2K bus.  This message has the provision for including a rudder number so you can report two or more rudder positions.  Due to vagaries in the N2K specs and resulting different interpretations of what’s legitimate behavior, Simrad reports 255 as the rudder number which they think means “not supported”.  But Maretron interprets 255 as an invalid rudder number and does not display the data, leaving me with no good way to verify that both rudder sensors are operating and in agreement – a basic check for any redundant system.  Simrad provided me with new firmware to work around this, and it solved that problem.  But it thoroughly broke their own AP product so had to be abandoned.
  • I never figured out how to make this happen, but several times the pilot got into a state where it said it was using the RC42 compass, but was displaying the heading from the HS70.  Each reported a slightly different heading, so it was clear which was being used.
  • The pilot would also get into a state where even though I had selected the HS70 as the heading source, every time I would power down and back up, it would revert back to the RC42.    Under all other circumstances and for all other sensors, it would remember the last selection across all power cycles and restarts.
  • I also encountered a case where after selecting the HS70 as the heading source, it would display “-40” instead of “HS70-40”.  It basically lost the first part of the device name.

By the time the last few problems were surfacing, I had decided to get rid of the Simrad auto pilot too.  The pilot always steered the boat very well, so I had no complaints on that front.  But all these interfacing and control problems were yet another huge time sink for me, and I frankly had no confidence that Simrad placed any priority on fixing them.  And as you can see, some of these are real problem that reflect fundamental operation errors.


Wil S said...

Peter, yet another well-written, clearly explained post. I really appreciate how you make complicated subjects much easier to understand.

It occurred to me that a lot of factory advertising literature these days should be written this way. Just simply, completely, and clearly summarize what their products do. Not.

Andrew Hickman said...

Having hand steered for the last day and a half due to a "no auto pilot computer" message I am thinking too if replacing this unit. Turning the system and main power off hasn't helped me! So that's now the ais module, radar and now autopilot that have died on our 2.5 year old boat.

Anonymous said...

...I spend several hours chasing ghost like that in all marine electronic.. it's not simrad.. furuno, garmin, raymarine... all have problems when you want to build something a little more complex than a simply installation.
When I design redundant system, I design completely separated and isolated in order to avoid this kind of problems.
A second autopilot (and sometimes a 3) it's a wise decision... I think you're gonna clean almost all ghost if you change your connection building a second N2K parallel.. autopilot generate a lot of data (and AIS and radar..) so if you check the flow on the N2K system you're gonna find conflict between direction here and there... I installed a HALO radar system and I'm pretty happy with the result... Just my 0,2 cents