7
$\begingroup$

The 737 Classic has two different possible styles of engine-driven generators, distinguished by the generator's drive mechanism: the constant-speed drive (CSD) and the variable-speed constant-frequency drive (VSCF).

The original VSCF drives were very unreliable, failing roughly three times as often as the constant-speed drives. A 2001 modification programme by Boeing and the generator manufacturer improved this somewhat, but the generator drives remained unreliable enough for VSCF-equipped 737s, even those with modified drive units, to be subject to stricter limitations than other 737s (the United Kingdom's CAA requires them to remain within 45 minutes' flight time of a suitable airport, even if only one of the generators is VSCF-driven and the other has the reliable constant-speed drive; I don't have information on whether other regulatory authorities have similar limitations for VSCF-equipped 737s).

Why are the 737's variable-speed constant-frequency generator drives so much less reliable than the same aircraft's constant-speed drives?

Improve this question
$\endgroup$
2
  • 1
    $\begingroup$ No idea how it works, but VSCF, just from its name, sounds like a mechanical monstrosity. Usually what happens is the maker, HS in this case, will tweak this and tweak that trying to get the MTBF up without spending too much money, when the unit in reality has some fundamental weakness in its architecture that will cost bajillions to fix. So they spend years dancing around the problem until something new comes along to replace it. $\endgroup$
    – John K
    Commented Feb 23, 2020 at 19:55
  • 1
    $\begingroup$ By drive do you mean mechanical parts of the system? Constant speed system is way more complicated than a variable speed one, so I'd say this is likely to be a case of shoddy electronics. Caveat: I'm not a mechnic or an electrician... $\endgroup$
    – Jpe61
    Commented Feb 23, 2020 at 21:03

1 Answer 1

Reset to default
11
$\begingroup$

@Jpe61 is correct in his comment that it's an electronics issue, since indeed a VSCF is mechanically simpler.

VSCF is also used on the MD-90, with poor initial reliability[1] – MDC went back to IDGs for the MD-95 (717).[1] It's also used on the 777 (only for backup, also with poor initial reliability[2]), and on the F/A-18.[3]

Capacitors

The F/A-18 utilizes cycloconverters, while the aforementioned civilian applications utilize DC-links (the voltage-source inverter type based on this patent by Sundstrand, the makers of the 737 unit).[3]

... when electrolytic capacitors are used, in the case of a voltage DC-link, there is potentially a reduced system lifetime (Wikipedia: AC-to-AC converter § DC link converters).

More here: Wikipedia: Electrolytic capacitor § Operational characteristics

Your link says "the failure rate was just over 1 every 2,000 hrs". The 2,000 hours match the lifetime graphs in the article linked above (shown below). This being the culprit is also supported by a paper on aircraft power generation:

The VSCF systems, apart for some noteworthy exception previously mentioned [referring to the F/A-18 and F-117A], did not get the same level of diffusion of IDG systems. This can be safely attributed to the role played by the PEC [Power Electronics Converter] (either ac/dc or ac/ac), which processes all the generated power and represents a single point of failure. Therefore, the PEC needs to be designed for the full-power rating and with high reliability requirements.[4] [emphasis added]

enter image description here

High load and cooling

Once you lose a generator, say running at 0.35 load, the other needs to pick up the slack and run at 0.7 load. The CAA's imposed limitation despite the redundancy tells us they don't trust the VSCF at high loads.

enter image description here
Sources: airliners.net and an MD-90 paper[5]

One feature on the MD-90 is the air scoops for cooling the VSCF components (two unpressurized compartments); something the 737 doesn't have tailor-made for its VSCF. The more electronics are used, the more intricate cooling is needed (you can't just dip them in engine oil and pass fuel through them to carry the heat away).

Maintenance practices

There's an official Honeywell video on installing the MD-90's VSCF with hints at improper maintenance practices, which would be a headache for an airline since the system is very niche with different tolerances and mounting techniques to avoid vibration damage.

Summary

So whether it's a 737 or an MD-90, the issues are:

  1. Cooling
  2. Capacitors
  3. Maintenance practices

Note that the technology hasn't become obsolete, at least academically: a 2017 paper proposes a "novel DC-Link VSCF AC-DC-AC" for the Embraer 190/195, from which:

Problems existed with the older VSCF systems in the past; however, the switched power electronics and digital controllers have matured and can be now, in our opinion, safely integrated and replace existing constant-speed hydraulic transmissions powering CSCF AC generators.

  1. Douglas Jetliners, Guy Norris and Mark Wagner, page 85
  2. Boeing tackles 777 power problems, 18 August 1999, flightglobal.com
  3. Control of Power Electronic Converters and Systems, Volume 2, page 341
  4. Madonna, Vincenzo, Paolo Giangrande, and Michael Galea. "Electrical power generation in aircraft: Review, challenges, and opportunities." IEEE Transactions on Transportation Electrification 4.3 (2018): 646-659.
  5. Kressly, Arthur E., and Anthony C. Parker. "Development of the McDonnell Douglas MD-90."
Improve this answer
$\endgroup$
1
  • 3
    $\begingroup$ You are probably right. Lifetime of electrolytic capacitors does indeed drop with heat and in a VSCF the waste heat is generated directly in the electronics vulnerable to it. Additionally the issues are worst in mid-flight and increase if load increases due to alternate failing, so you have a real possibility of losing power mid-ocean. By contrast IDG generates waste heat in systems insensitive to it, probably fails when engine speeds change which tends to be near safe landing and does not particularly care what the load is. $\endgroup$
    – Ville Niemi
    Commented Mar 5, 2020 at 11:08

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .