Instrument accuracy

[peterb]

There's a thread below about depth sounder accuracy. It raised a thought in my mind about the general subject of instrument accuracy, and the degree to which instruments can be believed.

Towards the end of last year I was sailing on a school yacht, and compared its hand bearing compass with mine of the same type. They differed by 5 degrees. Now I know that the specification of the compass allows a 2 degree error, so that if both compasses were on opposite limits then there could be a 4 degree error. Except that I had recently checked my own compass. Its error is only half a degree, and that is in the same direction as the difference between the pair so that the other compass had an error greater than 5 degrees. If you were using a hand bearing compass to follow a channel, then a 5 degree error could be lethal.

We all know that steering compasses should preferably be corrected and at least swung to get a deviation table. But 'should' doesn't mean that we all do it, and anyway it's not a once-for-all procedure; how often do you check yours?

The most commonly used log on small boats uses a rotating paddle wheel. The biggest error is usually a failure to spin at all, usually due to the wheel being fouled, but there are others. I once found that the log was only indicating about half the correct speed, even after using the software correction facility provided. Taking the wheel out and cleaning it had no effect. We dried out and found that a large barnacle had grown on the hull just ahead of the paddle wheel; its 'wake' was slowing the wheel right down..

I've found a radar set with a misaligned heading marker, so that something dead ahead showed up as being nearly 10 degrees off the bow. The owner had been using it for months, without once comparing the radar picture with what the Mark 1 eyeball saw.

The question really is to what degree should you believe your instruments? And what sort of checks should you apply to help increase your belief?

<hr width=100% size=1>

 

[charles_reed]

perhaps one should refer to limits of inaccuracy.

My electronic paddlewheel log reads low at low speeds and high at high speeds. At normal hull speeds it's about 2% out, at 2 indicated knots about 40% low and at 10 knots indicated 20% high. This is when the hull is clean and without the additional anomalies caused by strategically sited barnacles and partial paddlewheel fouling. The old trailing log is better, but that reads high at low speeds and low at high speeds (it's about 20% out at each end).

The best you can expect from a transducer depth sounder is about +- 100mm and that's not including spurious readings you'll get from temperature anticlines in water and fluid mud bottoms.

Radars are OK for ranges off, but range from poor to bad for bearings, dependant upon beamwidth and platform stability both of which are poor on radar in small boats.

As for wind indicators - frankly the direction indicator on most boats is a joke in poor taste, varying between 10-30 degrees out and I'm deeply suspicious of all anemometers though they, of all instruments, should be pretty nigh accurate.

Not one of the three GPS receivers on my boat, though the aeronautical one is probably the most accurate, agree on course, or speed - the only place they meet is on ETA. Even the aero receiver, faced with a turn when going slowly, persists in reflecting the old course for several minutes.

No anyone who believes in the spuriously detailed information supplied by his/her electronic toys is, to put it mildly, gullible.

Navigation is holistic, there are a lot more aids now then in the 19th century, but all that does is give you more opportunity to cross check the errors of each of the inputs.

<hr width=100% size=1>

 

[LadyInBed]

Anyone trying to read a run of the mill compass on a bouncy boat at less than 5 degrees is IMO only kidding themselves.

<hr width=100% size=1>

 

[happychap]

Re: compass accuracy

I have a Ritchie Compass, a Garmin 128 gps, an Autohelm 4000 autopilot and a hand held gps.
My readings are on the boat but from memory only the two GPS instruements were ever within 5deg of each other.
After calibrating the autohelm ( accuracy now < 3 deg) I was still getting a 15deg difference with the gps on certain points of sail.
I ended up using the GPS for passage making and managed from Barcelona to the south of Italy.

I've bought a new compass after an abortive attempt at removing an air pocket but before I start trying to get the compasses to agree I'd appreciate any tips / comments. I'm assuming that 5deg is probably as good as it gets .

The nice man at the Raytheon stall at the LBS suggested I must have the fluxgate too close to magnets ( it's about 18inches from anything) or my cables are running too close to power cables ( possible and I'm not sure if there is an alternative)
What is too close??



<hr width=100% size=1>

 

[tome]

The GPS receivers use a predictive algorithm known as a Kalman filter to calculate speed and course. This takes out a lot of the jumps you would othewise see. The filter is matched to the characteristics of the host vehicle. Aeroplanes and boats have very different filter dynamics which is why you should generally not use an aero receiver on a boat or vv.

There is usually some control on the filtering (eg on Garmin models it is known as the speed filter and is in the setup menu) and unless all GPS receivers are set to identical filtering you will see different results.

Wind instruments work pretty well for apparent wind, but when it comes to true wind there are all sorts of howlers in the way it is computed. True wind requires COG, SOG, as well as apparent wind direction and speed and vessel heading. Vessel heading has to be measured at the same time as wind speed and direction. As far as I'm aware there isn't a single instrument which does this correctly, though I'd love to stand corrected.

I'd like to see a comparison of true wind instrument performance published by the mags!

Paddle wheels are a cheap and cheerful solution and I can't believe that they are still around. The better transducers (ultrasonic) are quite large and I backed-off from fitting one when I saw the size, but it's only a matter of time before someone designs a better solution. Preferably one which is able to measure speed in 2 axes.

Echo sounders, work to around 0.01m and are probably accurate to 0.05 on shallow ranges - ie when it matters most.

Radars are fit for purpose if fitted with a good fast heading sensor - it is this, rather than platform stability, which limits their performance.

<hr width=100% size=1>