Jump to content

Taming a nervous NKE auto pilot


Recommended Posts

I have a Pogo 8,50 and thanks to the dual rudders it can keep perfect course autonomously for at least 10 minutes when sailing close hauled. (Perfect like in within a 3° tolerance). However, when I swith on my auto pilot it's constantly making small rudder angle corrections while there is no need for it, the boat already sails perfectly at the reference compass course or AWS. With constantly I mean multiple times per second. It means the auto pilot is wasting a lot of energy and wearing itself out (not to mention the constant noise of the pump motor). 

I already tried to tune different settings to see if they improve the behavior of the auto pilot to no avail, like: Gain on 1, Speed coefficient on 1, Damping on 9, Compass filtering on 20, fine tune the rudder angle correction although it find it difficult to determine the right setting.

Can anyone who has experience with an NKE pilot tell me if this is expected behaviour, or what else I could try or check? 

 

Link to post
Share on other sites

I don't have experience but are you sure Damping should be set to 9 (a high value?).

Does it have a separate rudder angle indicator? Can you test it to make sure the signal is good?

Link to post
Share on other sites

The higher the damping the less responsive to changes in wind the unit becomes. But doesn't matter, I just tried out all the extremes of the different parameters to understand their effect. 

The rudder angle sensor is brandnew and working perfect. There's an angle indicator on the control display, close hauled it constantly oscillates close to zero when the auto pilot is engaged.

(i did perform a recalibration of the auto pilot after replacing the rudder angle sensor). 

Link to post
Share on other sites
3 minutes ago, solosailor said:

Does it act the same under a compass, not wind mode?

 

No difference between compass and wind mode, the auto pilot constantly trims the rudder while according to the display the reference (target) = compass course or AWA. 

Link to post
Share on other sites

A small video to demonstrate the problem: NKE auto pilot in AWA mode. At the bottom you see the auto pilot is constantly making minor corrections. It's worse than you can conclude from this video, as the changes are often to small to be visible on the display.

Looks like the auto pilot is trying to keep the course accurate to 1°. For me that's to accurate at the lowest gain setting, at that setting I want it to save power. And in compass mode it would remain accurate within 1° for at least 10 sec without doing anything anyway. 

Link to post
Share on other sites

Auto pilot logic is similar to tuning a PID which can be quite frustrating to do manually.  Each setpoint works in conjunction with the others so there are always several layers of adjustment to get the target.  I would start by varifying the accuracy of your feedback. The rudder angle indicator is a resistive feedback where the span is set in calibration.  If it's new it's unlikely there are any electrical deadspots in it but the physical setup may be such that you are working over a very small span.  Alot of them have adjustable arms.  The farther out you are on the arm the more physical movement you will have and a higher degree of accuracy.  If you can see the raw data or feed back slowly run the helm all the way one way then the other and look for any dead spots.  If possible to increase the physical arm try that and go thru a new tune.  I have a friend who swears by that pilot and is a shorthanded race nut.  Have never worked on one so sorry no real other feedback.

 

Since it's a fancy new boat highly unlikely, but any physical slop in linkage between pilot and rudder will have the same effect, rudder bearings linkage and quadrants.

Link to post
Share on other sites

The rudder angle sensor comes with a fixed length arm so nothing to tune, installation instructions are clear. I don't notice any sensor dead spots (the old one had some). It's not that the system is not working, otherwise it wouldn't keep course to the degree. 

IMG_20200627_205343.thumb.jpg.02cefefe78dfd590ae315ffa6a93a59e.jpg

I'm not sure length B is accurate to the mm, but it's certainly not to far off. Unfortunately the thing is installed at a location that's very difficult to reach. 

Link to post
Share on other sites
59 minutes ago, SASSAFRASS said:

Since it's a fancy new boat highly unlikely, but any physical slop in linkage between pilot and rudder will have the same effect, rudder bearings linkage and quadrants.

Good idea. Any slop in the mechanical connections (or air in hydraulics) could create havoc. Or an intermittent problem with the wiring to motor or Rudder Angle Sensor. If the motor is commanded to move, no matter how tiny, that motion should reliably/properly affect the rudder and sensor.

Or just as likely some obscure setup/config parameter was accidentally goofed up. E.g. type of drive.

Gotta figure the NKE stuff works without solid evidence to the contrary. 

Link to post
Share on other sites

The boat is actually 14 years old (but maintained like new). But now that you mention I recall the joint that connects the hydraulic ram to the rudder arm is worn out and has a bit of play. I'm certainly going to look into that. Thanks guys. 

Link to post
Share on other sites

Mechanical slop could be a cause, also check the calibration of the rudder angle sensor. It may be 'reading' 0 angle when the rudder is not truly straight. The controller will return the rudder to 'straight' as determined by the sensor, if calibration is off the resting position is also off, and the boat will turn until the gain set point is reached & the pilot reacts again.

Calibration of RAS needs to be done when the boat is 'coasting' (propellor NOT turning & folded) and at normal sailing speeds

Link to post
Share on other sites
14 minutes ago, longy said:

also check the calibration of the rudder angle sensor. It may be 'reading' 0 angle when the rudder is not truly straight.

I find it difficult to get this right (to the single degree). When sailing straight ahead a rudder will always have a slight angle or the boat will luff. When you tack it needs to be angled to the other side. I did my best to tune this perfectly but the effect seems to be minor. I even have the impression the auto pilot behaves best with different settings on different courses.

Link to post
Share on other sites

Could be wrong but I don't think the pilot cares about anything on the rudder feedback except limits and change of state feedback IE movement amount and time.  With current and set it will hold a heading with rudder applied off zero. 

The pilot no pilot scenario and a sloppy linkage sounds like winner.

Link to post
Share on other sites

I guess the rudder angle zero setting is of some importance as the pilot will bring the rudder to neutral when the boat is on course. When the zero settings is off it will make the boat change coarse again. However a potential error will never be that large that it requires the pilot to correct the course multiple times per second. (The NKE pilot only allows a +-3° sensor angle fine tuning anyway). The video I made confirms the boat stays on course perfectly.

So I agree I probably need to focus on the sloppy linkage first. I guess it somehow messes up the pilot's feedback loop (pilot engages the ram but initially doesn't notice any change in rudder angle, so pilot keeps trying). It will be next weekend before I can test again, but I'll give an update afterwards.

Link to post
Share on other sites

Typical servo math would elegantly compensate for the 'zero' or 'neutral' position of the rudder. I don't think that it would be necessary to be set with much accuracy. More of a display issue than a servo issue: Nice if the display shows centered rudder accurately.

Yes, check the linkages.

If the heading does not change with a small motor action you may need to increase the gain. Gain too high would be indicated by too much heading change for a motor 'bump'. Sailboat A/P are somewhat different than classical servo systems because of the way the motors are controlled by 'bumping' them back and forth in an attempt to save power. Which I've never convinced myself is necessarily true......

 

Link to post
Share on other sites
22 minutes ago, El Boracho said:

Typical servo math would elegantly compensate for the 'zero' or 'neutral' position of the rudder. I don't think that it would be necessary to be set with much accuracy. More of a display issue than a servo issue: Nice if the display shows centered rudder accurately.

That might be true and explain why it doesn't seem to have much effect on how the pilot behaves.

I now also realize that the rudder angle neutral position is probably not the cause as the boat has twin rudders which are toed in a few degrees. So there is no single rudder angle setting where both rudders are perfectly neutral, in practise you will be applying an average to assure the pilot performs well over both boards. So the neutral setting will always be off for a few degrees, but that shouldn't make the pilot freak out.

Link to post
Share on other sites

You also mentioned you have twin rudders, have you verified they are properly aligned? Others have mentioned play in the system and that will result in the pilot being active as it tries to "hunt" the correct rudder reference feedback value. 

Gain at 1 is the lowest with the least amount of tiller corrections. At a 9 it is very active and is best reserved from a very uneven sea state. 

What is your wind smoothing value? At 0 it's automatic however as you increase the value you will see the deviation between reference and measured value increase. 

Link to post
Share on other sites

Align the rudders in what sense, as explained they are toed in a bit which is common with twin rudders.

Changing the wind smoothing has no effect on what I try to accomplish. The behavior is the same in compass mode, where I also already tested higher filtering (NKE terminology for smoothing) settings.

So I intend to focus on fixing the play in the system. I still hope an NKE user chimes in to share how a properly installed and tuned system should behave (which I've never witnessed as I bought the boat second hand). 

Link to post
Share on other sites
15 hours ago, Geert said:

Align the rudders in what sense, as explained they are toed in a bit which is common with twin rudders.

Changing the wind smoothing has no effect on what I try to accomplish. The behavior is the same in compass mode, where I also already tested higher filtering (NKE terminology for smoothing) settings.

So I intend to focus on fixing the play in the system. I still hope an NKE user chimes in to share how a properly installed and tuned system should behave (which I've never witnessed as I bought the boat second hand). 

What happens when you are in rudder mode, is the pilot active (what is the measured vs set value deviation)? What happens when you try to man handle the tiller, does the rudder reference sensor value change and does the ram move in and out a short distance to try and correct? 

Link to post
Share on other sites
13 minutes ago, RImike said:

What happens when you are in rudder mode, is the pilot active (what is the measured vs set value deviation)? What happens when you try to man handle the tiller, does the rudder reference sensor value change and does the ram move in and out a short distance to try and correct? 

In rudder mode the pilot keeps the rudder locked at the reference angle, it doesn't show any activity.

When you say man handling the tiller, do you mean with or without the pilot engaged? When engaged you can't move the tiller by hand, if not engaged the pilot won't correct anything.

The rudder sensor itself is brand new and it's angle is displayed fine. The pilot is doing a perfect job in keeping the boat on course, within 1° even in the lowest gain setting.

Link to post
Share on other sites
Just now, Geert said:

In rudder mode the pilot keeps the rudder locked at the reference angle, it doesn't show any activity.

When you say man handling the tiller, do you mean with or without the pilot engaged? When engaged you can't move the tiller by hand, if not engaged the pilot won't correct anything.

The rudder sensor itself is brand new and it's angle is displayed fine. The pilot is doing a perfect job in keeping the boat on course, within 1° even in the lowest gain setting.

With it engaged to see if there is any play in the system down to the ram and see if the rudder reference sensor value changes.  I've seen on a few occasions where there was a worn joint at the end of the pilot and the ram was constantly moving as the rudder reference sensor would feed back those few mm of play as a half degree or something an would cause the pilot to "hunt". 

What is your counter rudder value? Try setting it to 1 if it's on Auto which effectively shuts it off and report back. 

Link to post
Share on other sites
10 minutes ago, RImike said:

With it engaged to see if there is any play in the system down to the ram and see if the rudder reference sensor value changes.  I've seen on a few occasions where there was a worn joint at the end of the pilot and the ram was constantly moving as the rudder reference sensor would feed back those few mm of play as a half degree or something an would cause the pilot to "hunt". 

What is your counter rudder value? Try setting it to 1 if it's on Auto which effectively shuts it off and report back. 

I know there is play in the connection between the ram and the quadrant, however not that much that you can make the rudder angle change on the display. As I don't know how sensitive the pilot is I intend to fix this first anyway (as you say a half degree might be enough to cause issues).

The counter rudder value is on auto but I tried different values. According to the manual counter rudder is off with gain settings lower than 3, in which case it shouldn't matter.

Link to post
Share on other sites

I had this on my NKE setup years ago. It's a bit of a dim memory now, but I think that the fix was related to a compensation for speed through the water. Control systems theory helps understand what's going on - basically it's an oscillation because the gain is too high and the phase margin is too low. That sucks because you can't set a Gain less than 1 and the phase margin is a physical property of the boat. 

As I recall, you can mess about with Gain settings and all that stuff, but the AP continues to saw away sucking power and generally being annoying. So, you have to dig into the menus and find a parameter that further reduces Gain. There's a parameter buried in the menus somewhere that corrects gain for boat speed. This acts like a variable control loop gain that reduces the gain with boat speed. Turn it down and you should see it stabilize the loop and prevent the oscillation.

Link to post
Share on other sites
4 minutes ago, Wet Spreaders said:

So, you have to dig into the menus and find a parameter that further reduces Gain. There's a parameter buried in the menus somewhere that corrects gain for boat speed. This acts like a variable control loop gain that reduces the gain with boat speed. Turn it down and you should see it stabilize the loop and prevent the oscillation.

I think you are referring to the 'Speed coefficient' setting. I already have that on the lowest value (and tested higher values). But I found out there might be a related option I can play with, a speed adjustment control on the pump motor which controls the oil flow to the ram. I intend to check its impact as in theory it affects the closed loop system behavior.

Link to post
Share on other sites

Think of it like this, the system uses a fast and slow channel.  The slow channel represents what you see on the display so in terms of response/accuracy it might not display the half degree incremental reading for a tenth of a second. The high speed channel is what is used for computation by the computer. If you have Toplink, you can see this value from your MHU for example. 

Wet Spreaders is referring to the rudder coefficient (is this what you mean by speed coefficient?) I believe, which automatically adjusts the value of the rudder angle to apply which is proportionally opposite to the boat speed. The faster the boat sails the lower the rudder angle. With a range of 1 to 53 the higher the value the higher greater the rudder angle. Default is 6, whats yours? 

Also what is your boat speed reference, SOG or water? 

Link to post
Share on other sites
13 minutes ago, RImike said:

Think of it like this, the system uses a fast and slow channel.  The slow channel represents what you see on the display so in terms of response/accuracy it might not display the half degree incremental reading for a tenth of a second. The high speed channel is what is used for computation by the computer. If you have Toplink, you can see this value from your MHU for example. 

Wet Spreaders is referring to the rudder coefficient (is this what you mean by speed coefficient?) I believe, which automatically adjusts the value of the rudder angle to apply which is proportionally opposite to the boat speed. The faster the boat sails the lower the rudder angle. With a range of 1 to 53 the higher the value the higher greater the rudder angle. Default is 6, whats yours? 

Also what is your boat speed reference, SOG or water? 

Yeah - distant memory resonates on the "Rudder Coefficiant". Anything labeled as a "coefficient" likely means that the nerds who programmed the interface were considering a variable that gets multiplied with the input signal as part of the control loop. Turn it down to reduce the effect.

The guys above discussing whether your linkages had slop in them were intuitively checking the phase margin in your loop - roughly speaking, that's the delay from action to outcome. If your observation to outcome is too slow, you end up oscillating as your control inputs amplify the effect rather than dampen. Same as pilot induced oscillation in an aircraft, or snake-wake when sailing downwind. 

Link to post
Share on other sites
42 minutes ago, RImike said:

Wet Spreaders is referring to the rudder coefficient (is this what you mean by speed coefficient?) I believe, which automatically adjusts the value of the rudder angle to apply which is proportionally opposite to the boat speed. The faster the boat sails the lower the rudder angle. With a range of 1 to 53 the higher the value the higher greater the rudder angle. Default is 6, whats yours? 

Also what is your boat speed reference, SOG or water? 

On my Giropilot display and in the manual it's called speed coefficient (I don't have a rudder coefficient). I had it on 5. Bringing it down reduced the activity of the pilot, but even on 1 the pilot is still to active to my liking (in light conditions). My assumption is that this is caused by play in the ram's link.

My boat speed reference is speed through water (with SOG as a backup).

Link to post
Share on other sites
33 minutes ago, Wet Spreaders said:

The guys above discussing whether your linkages had slop in them were intuitively checking the phase margin in your loop - roughly speaking, that's the delay from action to outcome. If your observation to outcome is too slow, you end up oscillating as your control inputs amplify the effect rather than dampen. 

As the pilot never stops correcting for even half a second oscillation is what it looks like. It will take some improvisation to temporary fix the sloppy link to evaluate the effect.

Link to post
Share on other sites

Hi Geert,

you really need to read the manual of your NKE Gyropilot but here is the quick rundown (I've sailed a Pogo2 for years, now a Pogo3):

1. Gain
The gain (you control it from the autopilot page where it shows you your current course) should be as low as possible. It basically defines how often the pilot will try to correct the course but it is inverse. Like if the gain is 3, it will attempt up to 3 corrections per second. If it is set to 5, it will be 5 per second (it is not exactly this formula, but you get the point). 
Important: With a gain of 4 or higher, the Gyropilot will use the internal gyro to outsteer the waves. With a gain of 3 or less, it will not activate the gyro.
As a good starting point, start with 3. 

2. Counter-Rudder
You can set this to the same range as gain or you can leave it at "Auto". In most situations you can leave it on "Auto" unless you are in heavy weather going downwind with the spinnaker with big waves. Then you may want to go Gain+2. 
For now, start with "Auto". 

3. Speed-Coefficient
This tells the autopilot how reactive you think your boat is (or how effective rudder changes are). A good rule of thumb is to set this to 1-1,5 times your average speed. On the mini, above 10 knots we need to set it to 2-3 times the speed but that is only on newer designs. A pogo 8.50 will probably behave roughly like a Pogo 2 so 1 to 1,5 times your speed should be good. 
You are referring to upwind so I guess try with a coefficient of 6-8 for now. 

These are your Autopilot settings. 
Now to your wind damping. If assume you don't have a HR sensor but the regular one (based on the age of the boat). In this case, I would recommend to set the damping to 4 or 5. 
Don't go higher than that. In very light winds it will pay to go down with the damping but then also make sure to reduce the rudder coefficient or you will lose a lot of speed because of all the rudder movements. 

The compass damping of 20 seems a bit high to me. The highest we go in the mini is 15 I think. Also, make sure you have calibrated your compass, takes 10 minutes but is worth the time. 

Your video does not match your description from my perspective. In your video, we see a Gain of 1, a pretty good wind angle steering and next to no rudder movements. That already looks fairly good to me. 
But maybe we just don't hear the pilot working all the time because of the talk on the vhf. Since the movements are small and the gain is low, my assumption is that your rudder coefficient is off, causing the pilot to not get the response from the boat it expects, causing it to keep correcting. 

Let us know if this helped!

  • Like 2
Link to post
Share on other sites
11 hours ago, scaprat said:

You really need to read the manual of your NKE Gyropilot

How did you come to that conclusion? Not only do I know it from the top of my hat, I effectively tried out the effect of all parameters.

11 hours ago, scaprat said:

 Like if the gain is 3, it will attempt up to 3 corrections per second.

As explained, mine is making 3 corrections per second with a gain of 1.

11 hours ago, scaprat said:

The compass damping of 20 seems a bit high to me. 

I agree. It's just part of attempting to averaging out all input parameters for the pilot so it's not triggered unnecessarily.

11 hours ago, scaprat said:

Your video does not match your description from my perspective.

But maybe we just don't hear the pilot working all the time because of the talk on the vhf

At 0:12 you see it making 3 corrections in 1 sec. It's constantly like that but the changes are often to small to be visible on the display. If I can't get the problem solved by fixing the sloppy link I'll make a new recording assuring the sound of the pump motor is captured. Currently you would hear it constantly engaging.

 

Link to post
Share on other sites

It seems to me that the underlying problem is to build a design (software) team that has never sailed, and told them what the machine should do: keep the boat within 0.5 degrees in all conditions, make sure you never deviate from that standard, make the interface like the iPhone that has 335 levels of submenus that are cryptically labeled and counter-intuitive (i.e. 1 is higher than 10, but it goes to 14), and make sure that "sensitivity" is buried in the "power settings" sub menu underneath "alarms".

Link to post
Share on other sites
8 hours ago, Caca Cabeza said:

It seems to me that the underlying problem is to build a design (software) team that has never sailed, and told them what the machine should do: keep the boat within 0.5 degrees in all conditions, make sure you never deviate from that standard, make the interface like the iPhone that has 335 levels of submenus that are cryptically labeled and counter-intuitive (i.e. 1 is higher than 10, but it goes to 14), and make sure that "sensitivity" is buried in the "power settings" sub menu underneath "alarms".

...and any menu item you really need to get to is down an unlit stairwell into a cellar, at the bottom of a locked filing cabinet stuck in a disused lavatory with a sign on the door saying "Méfiez-vous du léopard".

Link to post
Share on other sites
8 hours ago, Wet Spreaders said:

...and any menu item you really need to get to is down an unlit stairwell into a cellar, at the bottom of a locked filing cabinet stuck in a disused lavatory with a sign on the door saying "Méfiez-vous du léopard".

Try working on the HR processor which is supposed to be in set to display either English or French and choosing the English language in yet parts of it are in French and Chrome fights with its self on which language to pick. Ultimately ending up with Freglish.....

Link to post
Share on other sites
51 minutes ago, RImike said:

Try working on the HR processor which is supposed to be in set to display either English or French and choosing the English language in yet parts of it are in French and Chrome fights with its self on which language to pick. Ultimately ending up with Freglish.....

NKE making their products available to the rest of the world is not about business, it's a favor. So stop complaining ;-)

Link to post
Share on other sites

Yesterday I fixed the play in the system. In short, it didn't make any difference.

I did a lot of testing again, comparing the behavior of the pilot while the boat is docked, sailing in the flat water of the marina and on sea. The problem of the pilot being overactive only occurs at sea. The pilot's activity seems to be linked directly to the the motion of the boat, caused by the waves. The pilot's gain settings mainly determines the magnitude of rudder angle corrections and not how active is it.

My conclusion is that the pilot is over sensitive to the input of the gyroscope. Unfortunately that's the only parameter you can't tune. Regarding waves I'm sailing in very bad waters with a boat that's doesn't deal well with this type of waves, so it looks like NKE's default gyroscope tuning isn't optimal for my situation.

Guess I'm out of options, except for asking NKE if they can tune the gyroscope sensitivity.

Link to post
Share on other sites

Hi Geert,

Two questions:

- did you have this problem from the start or was your pilot working fine in the past and is now for whatever reason having this strange behavior?

- did you try to re-initialize your pilot (in the menu there is an initialize option.   You then have to set the rudder to the exact middle position,  then fully right, fully left and back in the middle again.   This re-initialization would be the first thing to try.

When sailing on the North Sea or IJsselmeer my NKE works absolutely perfect and responds well to the various tuning options via gain, rudder coefficient etc.

 

 

Link to post
Share on other sites

The gyro is not active if your gain is lower than 4. 

Quote from the Gyrographic manual:
 

Quote

Please note that for a gain between 1 and 3, the rate gyro is inactive and there is no counter rudder. For gain between 4 and 9, the rate gyro is active, and the counter rudder value is automatically adjusted according to the selected gain if the counter rudder is set on AUTO.

Is counter rudder sto to "AUTO"? 
What speed coefficient were you testing with?

 

Link to post
Share on other sites
4 minutes ago, Snowden said:

... and you're sure the gyro is tightly screwed down to a solid bit of the boat, not slopping around itself :-)

Yes, it's integrated in the pilots controller and that's fixed sturdily.

Link to post
Share on other sites
2 minutes ago, ZeeZee said:

- did you have this problem from the start or was your pilot working fine in the past and is now for whatever reason having this strange behavior?

- did you try to re-initialize your pilot (in the menu there is an initialize option.   You then have to set the rudder to the exact middle position,  then fully right, fully left and back in the middle again.   This re-initialization would be the first thing to try.

When sailing on the North Sea or IJsselmeer my NKE works absolutely perfect and responds well to the various tuning options via gain, rudder coefficient etc.

When I bought the boat years ago the pilot had difficulties maintaining accurate course. At low gain settings I didn't bother, I was actuality happy with it being silent. Last year I found out the rudder sensor had death spots so I replaced it. Now the pilot is accurate to 1°, unfortunately it also aiming for that precision at the lowest gain setting.

I probably re-initialized the pilot already 15 times. Remember it's doing a great job in maintaining course in difficult conditions, I just hoped it could turn it down a bit in the light.

So your pilot doesn't move for seconds in choppy water as long as the compass course is OK?

Link to post
Share on other sites

The movement of the pilot all depends on sea state and the gain I have set it to.

In calm water it sometime only makes a correction after 10 (or more) seconds or so.   I more rough water it may be a few seconds or maybe twice a second.  

I have 9 presets configured that define gain, rudder coeff. counter rudder and wind damping.   Then once a setting is chosen I do some fine tuning of the Gain when needed.

The 9 presets are for  UP, Reach and Downwind.   And then for High, Medium and Low winds each.

 

Link to post
Share on other sites
40 minutes ago, scaprat said:

The gyro is not active if your gain is lower than 4. Quote from the Gyrographic manual: "Please note that for a gain between 1 and 3, the rate gyro is inactive and there is no counter rudder. For gain between 4 and 9, the rate gyro is active, and the counter rudder value is automatically adjusted according to the selected gain if the counter rudder is set on AUTO".

Is counter rudder sto to "AUTO"? 
What speed coefficient were you testing with?

I also found that quote in the manual earlier today so now I'm completely lost. I might need to return the pilot's controller for inspection.

Counter rudder is on Auto but tried different settings, Also tried all possible setting for Speed coefficient, it's now on 1 as i theory that should make the pilot the least sensitive.

Link to post
Share on other sites
23 minutes ago, ZeeZee said:

In calm water it sometimes only makes a correction after 10 (or more) seconds or so. 

That's what I'm after, but mine doesn't even make it to 1 second. And that's all parameters tuned to what should make it the least sensitive (or whatever other settings).

Now that you confirmed it should be capable to do what I want I'll send it in for inspection this winter if I can't make it work by then. That's the feedback I was after, thanks. As it is now it's burning power and wearing itself out.

Link to post
Share on other sites

I just saw this thread. I also have a Pogo 8.50 and my pilot that behaved in a similar manner but not as severe. I traced the problem ( I think) to the (standard fluxgate) compass sensor. First I managed to borrow a regatta compass sensor to test and then I bought the new 9-axis compass sensor. After getting that to work properly (I needed to upgrade to the latest Multigraphic display software which took a bit of time to find out) the pilot calmed down. 

Link to post
Share on other sites

That's interesting, thanks. I had a look at the documentation for this compass and it describes how it improves the behaviour of the pilot. However, I have a Gyrographic display and not a Multigraphic so I wonder if it supports this compass. The manual specifies the Multigrapic.

This info makes me realize the pilot probably uses the raw data from the compass, so trying to slow it down by damping the compass info was probably meaningless. 

Link to post
Share on other sites

NKE Regatta processor: 5,3k€

NKE True wind license: 0,4k€

NKE Multigraphic: 1,5k€

NKE Regatta compass + interface: 1,35k€

NKE Magnetic log sensor already ordered: 0,9k€

I think I might as well throw in a 1,2 k€ HR Wind unit to be on the safe side....

 

 

.. once I found a sponsor. 

Link to post
Share on other sites
On 7/3/2020 at 8:44 PM, Geert said:

That's interesting, thanks. I had a look at the documentation for this compass and it describes how it improves the behaviour of the pilot. However, I have a Gyrographic display and not a Multigraphic so I wonder if it supports this compass. The manual specifies the Multigrapic.

This info makes me realize the pilot probably uses the raw data from the compass, so trying to slow it down by damping the compass info was probably meaningless. 

The 9-axis needs a multigraphic or the new multidisplay I'm afraid. The older 3-axis Regatta compass works with the older displays. I'm sure that the standard compass could work OK because mine did for many years before starting to act up...I would try to borrow another compass sensor In order to see if that could improve your situation.

Link to post
Share on other sites
2 hours ago, Pelle said:

The 9-axis needs a multigraphic or the new multidisplay I'm afraid. The older 3-axis Regatta compass works with the older displays. I'm sure that the standard compass could work OK because mine did for many years before starting to act up...I would try to borrow another compass sensor In order to see if that could improve your situation.

Thanks for the info. Next week I intend to mount a magnet in front of the compass to see if it  calms down the pilot. If it does we know in which direct to search for a solution. Borrowing a compass won't be that easy, NKE isn't svery popular over here. 

Link to post
Share on other sites

Geert,

I think the NKE compass is just a NMEA 0183 compass.  The compass interface box converts it to the Topline bus.   The reason I think this is because when you use a Regatta processor, the compass is directly (without interface box) hooked up to the NMEA input of the regatta processor.

So maybe you can buy an NMEA compass (or build one with e.g. an Arduino board) and try if this solves your pilot issue.

 

Link to post
Share on other sites

The regatta compasses might be NMEA indeed as it requires an interface box. Although that's no more than a guess, they might use another protocol (e.g. I2C). The Fluxgate compass I have has Toplink out, so I don't have this interface box which I might use to connect a 3th party NMEA compass.

I'm first going to try the magnet trick I had in mind to check if improving the compass makes sense.

Link to post
Share on other sites

It's an H2183. You would then need the NKE Regatta Compass interface and the cable. I've heard of something similar with the fluxgate compass however most just use the regatta compass. Do you happen to have another compass on board you could import the NMEA0183 data from via the NMEA in on the graphic display? you might have a a delay (lag) of anywhere between a fraction of a second to a full second depending on the baud rate.

Link to post
Share on other sites

Do you have a connection to the NKE bus with for example the Box USB (90-60-538) or Box Wifi (90-60-508) or an old RS232 ProData interface (90-60-466) ?

(distance Breskens - Den Helder with a Pogo 8.50 is not that big :-)

 

Link to post
Share on other sites

I do have the Toplink bus connected to my Lowrance plotter which makes all NKE sensor info available on WIFI, I think i can capture all NMEA data in my SailGrib weather routing app.

Link to post
Share on other sites
10 hours ago, Frou said:

What nke equipment do you have ? Can you make a list of it ?

Gyrographic 

Gyropilot with remote control

Fluxgate compass

Depth sounder

Electromagnetic speed sensor

Standard masthead wind unit

Multifunction TLE25

Multifunction performance 

NMEA Output interface 

 

Link to post
Share on other sites
6 hours ago, Miffy said:

There is currently a NKE multigraphic trade in discount if you want to consider taking the opportunity to upgrade. 

I might consider that. Do you have more info on it, I can't find any reference to that offer. 

Link to post
Share on other sites
4 hours ago, Geert said:

I might consider that. Do you have more info on it, I can't find any reference to that offer. 

It expired last week and was a promotion from TEEM which is in Lorient France (Right down the road from NKE main office)

Have you reached out to NKE direct on this?  

Link to post
Share on other sites
5 hours ago, Geert said:

I might consider that. Do you have more info on it, I can't find any reference to that offer. 

Give your NKE dealer a call - it was announced in June and was supposed to run thru early July but they might cut you a break. 

  • Like 1
Link to post
Share on other sites

I didn't reach out to NKE yet as I think NKE expects you to contact a dealer first. However I have no dealer nearby and I don't believe this is an issue that's easy to fix over the phone. That's why I'm doing some troubleshooting first.

I also don't know if the dealers I could contact are technically capable. A box mover isn't going to be much of a help, as I'm an electronic engineer and computer expert myself.

You also need to take into account the system isn't broke, this is more about fine tuning. That's why I appreciate input from experienced users.

Link to post
Share on other sites
  • 2 weeks later...

Geert, 

Not sure this helps (directly) but I have seemingly the same problem - except that my pilot is a Raymarine EV 100. In light air (or when motoring)it make incessant and wide corrections, whereas In normal conditions (chop or waves, speed > 4kts), it corrects every 20 secs or so,  as I expect.

Drives me nuts, and it sounds like the pilot will safe-destroy ! 

My diagnosis at this point is that the pilot reacts to very minute heel movements. I sail a trimaran, and on completely flat sea and very light wind,, the boat is balanced - litterally - on the main hull, so it keeps rocking ever so slightly, nearly imperceptibly, and it seems that this triggers the crazy behavior.  I did mange to calm it down a bit through tuning to extreme values, but this is not satisfactory as it alters the reponse in normal conditions, and it clearly does not address  the root of the problem anyway. This is my 4th year with this equipment, I do remember it getting slightly « nervous » in those conditions, but this year it really become totally crazy, so it might be degenerative, which would indicate that it is  due to hardware and not to Parameters.

The EV100 has a 3 axis compass + 3 axis gyro + accelerometer, says the brochure. Certainly not as sophisticated as NKE but maybe  similar enough for us to understand what going on ? I am curious to hear how your story ends. As for myself I might open the discussion with Raymarine, after having confrimed more of the symptoms.

 

 

 

Link to post
Share on other sites
On 6/27/2020 at 1:22 AM, Geert said:

I have a Pogo 8,50 and thanks to the dual rudders it can keep perfect course autonomously for at least 10 minutes when sailing close hauled. (Perfect like in within a 3° tolerance). However, when I swith on my auto pilot it's constantly making small rudder angle corrections while there is no need for it, the boat already sails perfectly at the reference compass course or AWS. With constantly I mean multiple times per second. It means the auto pilot is wasting a lot of energy and wearing itself out (not to mention the constant noise of the pump motor). 

I already tried to tune different settings to see if they improve the behavior of the auto pilot to no avail, like: Gain on 1, Speed coefficient on 1, Damping on 9, Compass filtering on 20, fine tune the rudder angle correction although it find it difficult to determine the right setting.

Can anyone who has experience with an NKE pilot tell me if this is expected behaviour, or what else I could try or check? 

 

Geert,

- First of all, you must make sure that the rudder sensor is correctly mounted, as per the instructions manual (basically the pivot must be 100mm).

- Second, you must do the pilot initiation (under the pilot settings menu). the wizard will walk you through the steps. It takes about a minute.

- Remember that your autopilot will do as well as the input from the sensors, so before playing with the pilot settings, make sure that you have calibrated the AWA, the speed and the compass. None of the calibrations are difficult.

- The nke autopilot is extremely performant and also easy to use. You need to know the basics though:

1) the gain: it is adjustable from 1 to 9. Remember that the nke autopilot has a gyrometer, and most importantly that the pyrometer is NOT ACTIVE when the gain is 3 or less. Therefore in calm seas, or when motoring, it is ok (and best) to keep the gain low. This being said, as soon as you have wind, best is to be at 4 and higher if needed. I never had my gain over 6. The slower the boat speed, the higher the gain. This being said, that's not always true. But basically, the gain is "how much the helm will move to keep up with your reference". Imagine going fast downwind, you don't want huge correction of the helm, so the gain will be fairly low. In high speed downwind, my gain will be 5 (not very high, definitely nowhere close to 9) but combined with a high rudder coefficient).

2) The rudder coefficient (RC). it is adjustable from 1 to 50 (I think). The RC is basically at what speed the autopilot will correct the changes (wind, heading) to keep up with the reference. It is different than the Gain because it is not how MUCH it moves, it is how QUICKLY moves. You can ask the pilot to move very little (low gain) but extremely fast (high RC). Downwind, if you are sailing deep on the TWA mode for example, you will ask the autopilot to react quickly to any changes in the TWA (so you don't gybe accidentally)...your RC will therefore be high. On a 35 foot boat, light displacement, so relatively fast, the RC can definitely be set between 20 to 30.

3) The Counter Rudder. This is basically what prevents the boat from rounding up. if you have too much weather helm, increase the counter rudder. Usually, I set it to 2 and it works well. If you are reaching hard, increase it. I recommend you ease your sail, travel down the main or even reef rather than using too much counter rudder, which will use more power on your battery but also more wear and tear on your steering gear.

4) The Wind Damping: filter the wind will help as well. One thing about the sensor with nke is that you have 2 channels (on the high resolution sensors). one is unfiltered and goes straight into the autopilot computer. this allows the autopilot to be very reactive because it receives the data frequently. that's the case for the wind, speed and heading sensors. the other channel is filtered and that's the one you see on the display (slower). If there is lots of seas and the boat's moving back and forth, front and aft, increase the damping (filtering) so there is less wind data coming into the autopilot so it doesn't work as hard trying to correct course.

I hope this helps.

I am in the USA but you can reach me with whatsapp if you want. PM on SA and we can get in touch over the phone which will be easier.

 

Link to post
Share on other sites

Yes I mentioned that. Everything is working fine now, except that even when you try to tune it down as much as possible it will still target 1° accuracy. It gets triggered by the slightest riple in the water. I think it behaves 'as designed', maybe a more sophisticated compass might help. 

Link to post
Share on other sites
  • 1 month later...

Did you ever solve this @Geert? I'm still learning to work with my system, I have a grid of "calibrated" settings to use in differing conditions to make it work, but I've yet to satisfactorily check them in differing conditions (due to generally being solo when needing to check the heavy weather settings, which er, is problematic when you can't leave the pilot alone long enough to change them :lol:)

Link to post
Share on other sites

No I didn't, as other parts of the boat also asked for attention. I'm afraid I'm out of options and need to contact a dealer to check out the system (if he believes it doesn't behave as designed, which I'm still not sure off). 

Link to post
Share on other sites
  • 1 month later...

I can't remember if this has been asked before, but what compass are you using? I see that on the NKE website there is the following text in the 9X Compass page:

Quote

A standard fluxgate compass does not provide the boat’s acceleration data. Therefore, there is a damping for compass heading data in order to get a stable value; otherwise it would not be useable. The integrated gyrometer in the autopilot partly compensates for this but with clean data coming from the 9X Compass, the performance is even better.

Could your compass not be dampening the date enough to account for acceleration data? 

Link to post
Share on other sites

I'm using the standard fluxgate compass. If have the impression the damping configuration is only used to average the data being displayed on the instruments, it has no effect on the behavior of the pilot. The integrated gyro meter also doesn't help as according to the manual it's disabled at the lowest pilot gain settings. 

Upgrading the compass has been mentioned before. It's a valid option but as it's pretty expensive I want to be sure if effectively fixes the problem. 

I had some other issues to attempt lately, but next season I'll probably get in touch with a dealer. 

Link to post
Share on other sites

Yes, NKE has a new compass out, the 9X it has a list price of €1,010 (ex VAT). It has replaced the Regatta Compass and Interface. 

The only downside is that you need a MultiGraphic with V2.5 or higher to calibrate it. 

Link to post
Share on other sites
On 6/27/2020 at 4:22 AM, Geert said:

I have a Pogo 8,50 and thanks to the dual rudders it can keep perfect course autonomously for at least 10 minutes when sailing close hauled. (Perfect like in within a 3° tolerance). However, when I swith on my auto pilot it's constantly making small rudder angle corrections while there is no need for it, the boat already sails perfectly at the reference compass course or AWS. With constantly I mean multiple times per second. It means the auto pilot is wasting a lot of energy and wearing itself out (not to mention the constant noise of the pump motor). 

I already tried to tune different settings to see if they improve the behavior of the auto pilot to no avail, like: Gain on 1, Speed coefficient on 1, Damping on 9, Compass filtering on 20, fine tune the rudder angle correction although it find it difficult to determine the right setting.

Can anyone who has experience with an NKE pilot tell me if this is expected behaviour, or what else I could try or check? 

 

 I’ve never used your brand auto pilot 

but I have used autopilots for 40 years 

typically control functions like yaw , delay , rudder response, weather helm  are adjusted in two locations

course adjustment is at the autopilot brain and is typically set up by the technician who installs the system 

fine adjustment, the one you use everyday ,  is on the autopilot controller 

clarify this with your equipment , identify course adjustment and fine adjustment location 

Also check that your rudder angle feedback unit is precisely aligned 

all commands reference the rudder feed back unit 

its this rudder angle feedback unit that causes most problems 

 

 

Link to post
Share on other sites
4 hours ago, RImike said:

Yes, NKE has a new compass out, the 9X it has a list price of €1,010 (ex VAT). It has replaced the Regatta Compass and Interface. 

The only downside is that you need a MultiGraphic with V2.5 or higher to calibrate it. 

So it hasn't replaced the Regatta Compass, although it's cheaper if you already have a multigraphic...

(There's plenty of gyrographic only systems around who will continue to have to buy the Regatta Compass and it's interface box).

Link to post
Share on other sites
On 10/23/2020 at 5:23 PM, JonRowe said:

So it hasn't replaced the Regatta Compass, although it's cheaper if you already have a multigraphic...

(There's plenty of gyrographic only systems around who will continue to have to buy the Regatta Compass and it's interface box).

No, It has in terms of what compass NKE offers (the only other compass option is the fluxgate) as the Regatta Compass is no longer available. Performance wise it's the same as the regatta compass (on paper anyways) You've brought up the point about a GyroPilot Graphic with the new 9X and it's inability to calibrate it. I have a question about it to them (NKE). I know when the first 9X compass first was released on the Figaro 3's the Multigraphics needed to be doublechecked they could perform the calibration it and if not I had to update them. I have yet to have any time on the actual compass first hand though to monitor it. 

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...