I went into the CPL performance exam a few days ago expecting to fly through it, after many practice test successes, but failed. Probably by just a couple of 2 or 3 point questions.

Part of me thinks CASA purposefully writes silly and unrealistic questions, just to get another $180 out of us.

The thing I was disappointed in was that I had at least 5 questions based on performance charts, that needed answers to the nearest whole number. In all the practice tests etc. I can't remember having to put in exact numbers. The chart type questions were always multiple choice. 2 or 3 of these questions were using the linear chart.

I've booked another exam for next week, but I'm taking in some magnifying spectacles to help with those drawn chart lines.

All the 1 pointed questions were handled well, except one, which was a simple ROC @3.2 % question, that I've never gotten wrong. For the life of me I couldn't get the answer they had in the 4 choices. I think the problem was somewhere within the questions addition of density alt, finding TAS and subtracting the 20kt headwind to get GS.

Onwards and upwards

The P chart questions are causing a lot of misery lately. Everybody knows how to use the P chart. CASA reply is always to be VERY VERY VERY careful in drawing the lines EXACTLY parallel to the guide lines and interpolating between the published values VERY CAREFULLY.

I'm afraid I can't offer any other advice. Obviously the margins applied are very tight. I'd like to be assured that CASA's answer is absolutely correct in the first place!

Do they employ draftsmen with T squares and Set squares and give them all day to get an exact answer? This never was the idea behind P charts when they were first introduced.

It was simply to give a pilot a method to quickly assess whether a take-off or landing could be safely carried out with generous margins applied to allow for inevitable inaccuracies in determining such things as headwind components by reference to a wind sock 50 meters away.

Some thoughts.

The thing I was disappointed in was that I had at least 5 questions based on performance charts

A little over the top, considering that the examiner needs to cover a wide range of MOS topics ?

I can't remember having to put in exact numbers. The chart type questions were always multiple choice.

You need to run the charts to get the answer to assess the multiguess options. Surely, it's not too difficult to put your answer in to answer the question rather than tick the option ?

2 or 3 of these questions were using the linear chart.

With the fallout of the Yates Report from some years ago, you are going to see very few of the old DCA "whiz around the boxes" P-charts out in the Industry, these days. The other format is very common in the US POH market so, unfortunately, you need to practice them sufficiently to ramp up the competence. At the end of the day, both formats do much the same thing and are used in much the same way so it's just a case of some practice to get on top of them. A bit like driving Holdens or Fords.

As with any charts (performance, trimsheets, etc) attention to accuracy in plotting, especially making sure you run parallel to guide lines, is quite important. Lots of practice and you will find that you can eyeball the plots for their accuracy quite easily. Again, it is just a matter of lots and lots of practice.

I think the problem was somewhere within the questions

Keep in mind, though, that WAT gradients were established/published for nil wind - eg check one of the P-charts and you will see that there is no availability to enter wind in the carpets to determine the WAT climb gradient limit. Non-WAT limit questions, though, should be looking at wind.

Obviously the margins applied are very tight

And that's quite silly from an engineering viewpoint. Very much like measuring the distance from Melbourne to Sydney with a one-foot school rule in a doomed attempt to get high accuracy. However, that's the practical consideration out in the real world of the Industry rather than the rarefied world of theory examinations where practical considerations aren't (and I see no reason why they should be) driven by practical concerns. On the other hand, though, I see little value in going to extreme and inappropriate lengths to obtain answers using a graph way beyond the precision and accuracy which went into the preparation of said graph.

The DCA P-charts' main advantage was that DCA Performance Section published the technique in several of the old tech notes with standard equations which were applied for all the P-chart exercises. The accuracy of the output can vary quite a bit, depending on who did the work and how they went about the test work. Done well, the output, generally, is more than fit for purpose but, certainly, not dead accurate by any means. Trying to figure the charts to the nearest foot is, quite frankly, silly from a practical viewpoint. The examiner, however, is not constrained by real world practicalities and, providing that the accuracy required in the exam is compatible with the equipment permitted to be used by the candidate, it really is just an examination technique consideration.

I did quite a few of the P-charts in years past as an Industry Consultant so I have a degree of understanding about the system both as we did it and DCA folk did it. Given that we all followed the DCA tech note provisions, the differences mainly were in the detail of how the tests were done and how the charts were prepared physically.

The military and other better funded groups used low tech cinetheodolites (sort of like a surveying theodolite with azimuth and elevation trace outputs, along with a video style record of tracking the aircraft (from which we could figure out tracking corrections to the marked aircraft CG. These were quite a bit down market from the high end kit used for ordnance and rocket tracking but worked real fine. I used the then RAAF kit for Nomad and UH-1 trials in the dim distant past.

The DCA approach, though, was very pragmatic and low cost. The aircraft, during the takeoff or landing tests, would have a number of still photographs taken so that its position could be estimated (quite accurately) against the runway environs background which had been accurately surveyed prior to the test program. Worked OK but, again, not super accurate when the timing was by stopwatch.

In recent years, of course, accuracy in testing has gone up in leaps and bounds. The OEMs usually run very precise computer model simulations which are then proved by the test program using fancy kit such as DGPS and the like. As an example, when the runway width test requirements came in for Australia, I did several aircraft using long lens videos for measuring lateral deviations with more than acceptable results. Boeing did the 737 using all sorts of high end tech gear for little effective improvement on the cheap approach.

I'd like to be assured that CASA's answer is absolutely correct in the first place!

Two thoughts here -

(a) if we are looking at the physical accuracy of matching the aircraft to the graph, it should be reasonable and fit for purpose but it sure enough is not in the realm of precision metrology.

(b) if we are looking at the DCA P-charts, specifically, we can use the equations which went into their preparation to give you an answer to whatever precision and accuracy you might choose. I have the relevant tech notes tucked away somewhere in the dusty filing cabinets - perhaps I should dig them out one day and reverse engineer the exam booklet P-charts ?

Just a quick tip that worked for me for the charts(third time) magnifying specs as mentioned AND quality set of calipers/ compass to measure initial spacing and transfer to other relevant section of chart

Meant to say dividers not calipers

magnifying specs

Certainly useful for us old blokes as the eyes' performance progressively deteriorates ...

quality set of dividers/compass to measure initial spacing and transfer to other relevant section of chart

Not always useful. Depends on the chart structure.

For a point which is off the printed lines, the aim is to interpolate according to the equations behind the carpets. Seeing that, in general, you don't have the equations, what you really should do (not all that practical in an exam, though) is plot several points from the bracketing lines, draw in a fair curve, and then read off the correct interpolated point. This is simple for a linear faired curve but gets a little awkward if there is a significant curvature in the faired line. The easiest way to do this is to use a rule to measure against an arbitrary scale, plot, read off the required point, and then convert that to the chart scale. Again, only takes a bit of practice to get the process under control.

Often, providing that you have done plenty of practice prior to the exam, you are better off fairing in a curve by eye on the original chart. With practice, that can be done quite easily and with more than enough reasonable accuracy for the exams. I show my students the background to the technique. They don't appear to have any difficulty understanding the process so it's not all that difficult, I suggest.

The question of whether such antics are reasonable, given the underlying accuracy and precision of the data which went into the chart's preparation, remains moot. That's the way of exams, I guess.

Thanks everyone for the responses. A couple of replies -

From John -I think this is all good, except what is the accuracy required when they ask for exact answers to the nearest whole number, instead of multiple choice. If I enter 745 and they want 750, then that is just setting someone up to fail, if their accuracy only accepts answers that are 1 point out, but I guess as you say, that's just exams and theirs no point in arguing with silly CASA exam writers.

From Marko -I do think this will help and will take to the next exam. I was previously using a torn off piece of paper to quickly mark spacing from one measured point to another point along the chart. That seemed to work ok, but I think callipers would be a little more accurate.

From John -Might be a good idea to make a video showing this.

From Bob -Exactly what I was thinking. There's margins built into the charts for ops safety, so why require exact answers in the exam. Multiple choice, I suggest, is much better at letting the examiner know if you understand how to use the chart.

with more than enough reasonable accuracy for the exams
what is the accuracy required when they ask for exact answers to the nearest whole number
If I enter 745 and they want 750, then that is just setting someone up to fail,

Only the examiner knows what accuracy/precision he is looking for.

As I suggested above, if we are looking at matching the chart to the aircraft and what went into making the chart, the chart data is a bit rubbery. Certainly OK for practical use in the field. Without knowing what actually went into the chart's preparation, any further comment can only represent idle speculation. Having said that, I would anticipate that the charts are representative of the aircraft's performance probably to around 20-30 kg and 50 ft. However, that still is only an educated guess on my part.

On the other hand, if we are looking at the chart in isolation - which is exactly what the exam is doing, there is no reason why the examiner might not seek accuracy/precision to the limit of chart reading. In this case, the usual chart can be read to around 5-10 kg and 10-20 ft if one uses the appropriate techniques which I have noted earlier in the thread.

quality set of dividers/compass to measure initial spacing
I do think this will help and will take to the next exam

But do be wary, given my earlier comments, about linearity or not in the carpet line spacing. It is very easy to head off on a tangent with this and that gadget.

I show my students the background to the technique.
Might be a good idea to make a video showing this.

I haven't done that to date, although I may do so in the future. In general, there is sufficient to and fro within a class cohort to warrant doing the work on the fly to address questions which are not always predictable and warrant interruption of the basic presentation.

Do they employ draftsmen with T squares and Set squares
There's margins built into the charts for ops safety, so why require exact answers in the exam.

Bob's comment was, I suggest, a tad tongue in cheek. Those charts generated by DCA were run up on an early HP plotter/computer combination which took the basic computer generated equations output as its input. The result was, so far as the equation was concerned, accurate. Reading of the end chart was never intended to go beyond sensible Cartesian techniques.

For those charts generated by Industry consultants, we either did much the same or, having figured the equations, plotted by hand which, done carefully, gave similar output accuracy.

The design standards margins incorporated are known and published so the concern is not relevant. The question of flight test accuracy and precision remains moot unless one knows how the exercise was undertaken. The consideration of matters such as wind values, slopes, surface friction coefficients and so forth remains an error band consideration.

Multiple choice, I suggest, is much better at letting the examiner know if you understand how to use the chart.

Afraid I have to disagree. We all know why multichoice, often, is referred to as multiguess. If the answer options are well conditioned, then the accuracy of chart use still requires a competent candidate, so why not just have the answer entered ? The examiner, then, has to assign an acceptable error band for marking, whether the marking is manual or electronic. Unfortunately, moving to electronic testing probably created more problems than it solved.

All great points. Thanks.

Another accuracy thing I seem to struggle with is the ASA-E6B-CIRC calculator wind calc accuracy. I never quite get the exact headwind or crosswind component that the answer has. Although I'm only usually out by 1 or 2 knots, so I just pick the closest answer, when its multiple choice, but again, who knows what accuracy the examiner is asking for when asked to enter whole numbers? If I put 189nm to the PNR and the answer is 187, then I get it wrong probably.

ASA-E6B-CIRC calculator wind calc accuracy

I am presuming that this is the ASA implementation of the Jepp CR slide rule ? It made life a lot more complicated when some of the manufacturers started using the Dalton E6B designation to apply to the Jepp CR, I guess for some strange marketing imperative. Given that the Jepp originated in the Luftwaffe of WW2 and then through Ray Lahr (who sold the rights years ago), using a US military designation is a bit strange ....

How about you run us through an example or three with your best reading of progressive settings and we can comment. Often, a lack of understanding about what the ETAS thing is causes problems. Also helps if you are on top of the sine/cosine scales as printed. Another suggestion is to use the ETAS adjustment ALL the time and then you don't get problems with "do I ?" or "don't I" ?. Keep in mind that the 10 degrees nonsense is just a bodgie approximation to always using the ETAS adjustment which gives you the correct answer Of note, the examiner uses ETAS down to 5 degrees which is reasonable as the error in the last few degrees is negligible. Not understanding what the device is doing in the ETAS adjustment often confuses folks greatly when, really, it's all too simple, once explained.

With reasonable care, you get the same answer whether you use the Dalton or the Jepp and both answers will agree with the mathematical solution (considering reasonable physical use of the instruments.

Also, some of the CR devices aren't scaled all that accurately so you may have a dodgy unit to throw errors into the mix as well ?

If I put 189nm to the PNR and the answer is 187, then I get it wrong probably.

That's got naught to do with the wind side and is just a slide rule CD scale proportion setting. Unless you have a dodgy unit, that will come down to your technique in using the device. Also, be aware that you really can't use the 2 x TAS approximation to (G/S on + G/S home) if the W/V results in significant drift angles so that might be a source of your problem and concern ?

Hence my suggestion that you run a couple of examples and talk us through them as accurately as you can. It won't be too hard to find where the problem lies and then you can sort it out in short order.