Our recent note on the Embraer E2 flight test program showed a discrepancy between what we could track using Flightradar24 and what the OEM told us. This difference was substantial and warranted further research. Here is what we have found.
The airport where Embraer‘s flight tests are being undertaken is AQA (Araraquara), which a rather remote area of Brazil. This is probably ideal for Embraer. But for anyone tracking flight tests independently, it’s tough. The screen capture below shows the aircraft on the ground at the airport at this writing. Clearly Embraer is very busy doing its various program testing.
The apparent reason we showed fewer test hours than Embraer is because it is difficult to track flights since there is no local ADS-B reciever generating a feed to Flightradar24. As Flightradar24 explained, “…it is likely that the algorithm won’t have matched the departure and arrival airports due to the flights only being detected above a certain altitude”. The outcome of this measuring difficiency means that there are readings that appear as “Unkown” in the data we are looking at.
So, for example on for PR-ZEY we see the following:
The first impression is that there was no activity on these dates. But when downloading the CSV file, we do actually see flight test data. For example on June 7 there was some activity but it looks like no flight took place since the rows are empty. In fact there were two flights; one of about 3:15 and another of about 2:38. We say “about” because the Flightradar24 data shows the first flight starting at 11,150 feet and a speed of 300. The second flight starts at 11,675 fee and a speed of 246. The tracking data ends equally abruptly.
The explanation is therefore that while the Flightradar24 data is tracking, it has very limited feed in the region around AQA. This is why the numbers we posted were far off from what Embraer shared. In reviewing the data it looks like the 900+ hours Embraer shared to date is the credible number.
Two further examples of the Flightradar24 data we are using to track flight tests support this. The C919 tests in Shanghai are well tracked – see below. The aircraft has not flown since its first flight. But we have good data on even its taxi tests.
The final example is the MC-21. Here we seem to have a similar combination of factors as we see at AQA making tracking complicated. The aircraft’s tail number is RA-73051.
But as we can see, there is no tracking data. This is almost certainly because in Irkutsk there is no coverage to generate tracking.
In conclusion, we accept the Embraer flight hour numbers. We also better understand the limits of tracking in remote areas. Doing test flights in remote areas is not what Airbus or Boeing – nor even COMAC – can do. There is great ADS-B coverage where they fly. But for Embraer and UAC they are able to do their tests without the same level of tracking.
Co-Founder AirInsight. My previous life includes stints at Shell South Africa, CIC Research, and PA Consulting. Got bitten by the aviation bug and ended up an Avgeek. Then the data bug got me, making me a curious Avgeek seeking data-driven logic. Also, I appreciate conversations with smart people from whom I learn so much. Summary: I am very fortunate to work with and converse with great people.
It is interesting that your did not mention BBD as the vultures were all over their test data good or bad.