Auto throttle oddities B737-300/400
 

Autothrottle response oddities, anomalies and curiosities. (Plus some rambling as I get my thoughts down.). Quite likely these aren't that odd, and have been encountered many times before.

I'm curious about the operation of the autothrottle and interaction with the thrust levers on the 737 classic, with a couple of scenarios in mind. The scenarios being actual events, ongoing. The aircraft are B737-300 and -400.

Scenario 1.
Autothrottle armed for departure.
TOGA selected at or above 40% N1.
Target 'reduced' N1 in the region of 86-88%.
The thrust levers begin to advance as normal, but at about 75/80% N1, one thrust lever momentarily physically halts or even retards a fraction, requiring manual intervention. There is associated yaw until the thrust lever catches up.
Setting thrust at an initial higher N1 before selecting TOGA, e.g. Upto 50% prior to selecting TOGA has little effect.

Same aircraft, with the Autothrottle not armed. There is normal, free thrust lever operation, and normal thrust response while setting reduced take off thrust manually. I can't say specifically whether the various parameters such as N1, N2, EGT, fuel flow are matched and linear with the autothrottle off while setting thrust, but there is no obvious yaw.

Scenario 2.
Autothrottle armed for departure.
Setting the N1 to 40% (ish), a lag in acceleration- It manifests as a delay of about 2 seconds before the acceleration. The acceleration rate being similar between both engines, but with a lag.
Engines stable at 40/45%, then selecting TOGA. Right engine lags behind left, for a few seconds, and up to 20% N1 difference. Seems an eternity. Same effect with the A/T off. Once above 55% N1, the thrust lever response is normal.
Setting thrust initially at higher N1, e.g. 50/55% reduces the effect, but does not completely overcome it.

Maintenance checks of the A/T function invariable come back with no fault, suggesting (as opposed to proving) that it's not a fault of the autothrottle system itself.
In the above, on a rare occasion, intervention has continued as the FMA goes to 'Arm'. Hence why I've considered it appropriate to turn the A/T off.
Of course, a further exacerbation of the situation may come when the other fella selects TOGA early. It still happens quite frequently, especially with some more than others, and not necessarily low time pilots.

What might be the source of these anomalies?
Fuel control? Blade issues? Fan or turbine?
Operation of the autothrottle because it is controlling both throttles/engines?

As far as I recall, the CFM-56 had a propensity to exhibit different acceleration rates between engines on the same airframe. I remember seeing effects similar to your 'scenario 2' on a few occasions. When the engineers were informed, the response was along the lines of, "that's within tolerance".

The only way to mitigate this was to be absolutely sure that the engines were stabilised at the same RPM before pressing TOGA. As you already noted, the higher the initial RPM, the less the effect.

BTW, I had a low speed RTO in an A320 with IAE engines caused by the 'other fella' selecting TOGA too early and having an apparently uncontrollable swing, so it's not just a CFM problem!

I have never seen your 'scenario 1', where one thrust lever momentarily stops or retards.

Yes, the second scenario is the more common, and naturally I've encountered it in NG as well.
I've been both guilty of, and witness to, hitting the T1T early in the past, and for most it serves as a lesson to not be so hasty. The concern there, being those who don't learn from the event. But that's another story.

When you have one engine which is slightly slow to accelerate with the autothrottle engaged, but both appear to accelerate normally without the autothrottle- what function does the autothrottle play that might cause this? Do both engines accelerate independently to the target N1 based purely on physical action of the autothrottle or some rough control/matching?

Flying the good old Classic was always a surprise... Especially with different engines on the wings, derated at the same power, or with an old engine and a "less" old engine on the other side, the spool up time had a difference of a few seconds... Staggering was pretty common and written in the tech log with the stated tolerances...
I think that this problem is caused by the fuel control on the engines and differences between the engines...

To my knowledge, we've never had mismatched engines on ours, here. Many years ago, I was introduced to some nasty -600s that ran intermixed engines, which had some of the mentioned anomalies. Except they weren't anomalies; the ops manuals and notices stated they were to be expected, and therefore 'normal'.

Another odd thing is when an autothrottle induces a greater stagger than manual thrust. Old kit, I guess.

I recall a Boeing bulletin or similar document. In effect it said there is a tolerance of three seconds allowed from the rate of spool up between idle N1 of 22% to 40%. Thus if both throttles are pushed evenly (same rate of movement) from 22 % to 40% and one of the engines decided to take up its 3% tolerance then a lag and a possible strong yaw will occur.

However, from 40% to rated thrust for both throttles at the same rate of movement, the tolerance is only one second. Using normal autothrottle procedure of manually pushing the throttles to approximately the 40% position then selecting TOGA, the throttles are programmed to advance to within 8% N1 of the selected take off thrust - then a slight hesitation - then they advance to the selected take off power.

The Bulletin cautions not to try to hurry the throttle levers by manually pushing them as TOGA moves them. This can damage clutch motors in the throttles. Damaging the clutch motors can cause erratic autothrottle operation.

In the simulator we have often seen the PM "backing up" the throttles and pushing them under the captain's hand and not allowing the autothrottles system to do its thing. This is what Boeing warns about. It is also irritating to have the PM "riding" the throttles under the PF's hand.

There is no need for the PM to have his hand anywhere near the throttles even when the pilot says "Set take off thrust" or whatever terminology is SOP for the operator. A visual observation of the N1 needles is sufficient to confirm correct power is set. According to the FCTM, once take off thrust is set automatically by the A/T system, only then the PM should, if need be, adjust the final setting to -0% +1% target N1.

The habit of some pilots to "back up" by placing their hand behind the throttles in the 737 and no doubt other types, may stem from previous general aviation aeroplanes where hand operated throttle friction nuts were worn or not effective and throttles were apt to slip back once the hand as released from the throttles.

The big danger of unnecessary "backing up" is if a sudden decision to reject the take off happens and the captain whips the throttles closed, trapping the PM's hand between the aft edge of the levers and the start levers directly behind the throttles. Serious injury can occur.

Of course, the captain is always entitled to politely warn his PM to "kindly remove your hairy mitt from behind the throttles as I wish to make a rapid abort decision..":E

Thats useful-

Coincidence?
The 8% that you mention may well be the point where that one thrust lever feels like it has completely stopped, or feels as if it is pushing back. It certainly holds there for what seems a very long time. Longer runways where we are never limited, maybe not such a concern. Some of our more limiting cases are another matter.
I shall Have to look into this further.

I've conducted several air tests on 300s where asymmetric engine acceleration on the ground was apparent but both engines performed as they should during the airborne slam accelerations which was a bit odd. Also vice versa which was a bit odder.

The solution is a bit engineery and initially required fettling of the MEC and various other adjustments which I can't remember. Ultimately, if that didn't work, changing several of the engine sensors (P6 rings a bell) was the next step. In one case, after about 5 days of engine runs, a complete engine change was required (which is actually quicker and easier than a MEC change). In most, if not all, cases the AT wasn't the issue.

Also it transpires that fitting a new or reconditioned MEC to an older engine can produce similar symptoms.

On the point of not 'helping' the AT achieve the required N1 - no, it shouldn't be necessary but quite often is required to ensure that thrust is set prior to 60kts. If it isn't then your perf. is invalid and you should reject.

On Airbus A320 with CFM-56 engines, we go to 50% N1 and wait for each engine to stabilise at that before advancing to Flex/TOGA thrust.

Perhaps this extra 10% over the Boeing SOP is to better mitigate mismatched engines? The lag of one over the other can be several seconds.

Uplinker, the Airbus and the NG are FADEC equipped. The aircraft types in question here use mechanical fuel metering and throttle cables and I suggest that arrangement could likely be the source of the oddity.

Lancelot:

When TOGA is selected, you should see the throttles move smartly in unison to a coarse "predicted" position. Once there, N1 feedback takes over, and the movement will be slower. The former (coarse) movement is dependent on position feedback from the mechanical fuel control in the engine and if rigged within spec, L/R levers should be fairly close to one another.

Cable rigging through the servo mechanism has been known to be a potential cause of stagger on the classic, but this is a sledgehammer approach that likely has already been eliminated by your mechs.

I may be out of date in my knowledge in the classic, but I thought the A/T system would be disabled on aircraft with intermixed engines.

On both aircraft? By a few seconds, I am guessing less than 3 or 4?

Are both engines close in N1 when this occurs?

Vapilot. (I'll make a small edit to the initial post)
In the case of asymmetric acceleration- from idle to roughly 40%, it's probably less than 2 seconds lag, before acceleration. The acceleration rate being similar between both engines, but with a lag.

For the aircraft with the thrust lever holding at 80%. I'd say both N1 are within 2% upto 80%, The right thrust lever then continues to advance smoothly, while the left is still hesitating.


As for adjustment by maintenance- one engineer has commented that the most reliable function on the aircraft is the autothrottle self test- it's used often, and works every time.

Not a pilot ...from maintenance view there are two posibilities how to improve scenario 2 :
Use highest stabilised N1 seting approved by company SOP before pushing TOGA - I remember companies which normally used 60% N1 in such cases.
Ask engineers to improve acceleration on slow engine - old MEC has function for fuel specific gravity setting - it can help, but should be used with caution. Correct (dynamic) riging of VSV can help too.

Lancelot, I only worked the 737-3/4/500 series for a short time, and it was over 20 years ago, so the memory is a bit foggy.
That being said, I'd be suspicious of the PMC. The CFM56-3 PMC was very crude - an analog device (rather than the digital PMCs on the CF6-80A and CF6-80C2). In fact, it wouldn't surprise me if the CFM PMC is the last time Boeing put analog electronics on a new aircraft. Anyway, it could be the PMC is acting up since it only comes into play at higher N1s.

All Nippon airline in Japan operates the Boeing 767. I understand from talking to pilots recently returned from their contract the company SOP requires brakes to be held by pedal pressure and thrust levers opened to 70% N1 (still with brakes on).


When N1 stabilised at 70% brakes are abruptly released and TOGA set. The resultant aircraft acceleration from brakes released is a real neck snapper.


Certainly a weird SOP IMHO and need to have your medical insurance current:E

Difference in bleed state?