News:

Precision Simulator update 10.173 (24 February 2024) is now available.
Navburo update 13 (23 November 2022) is now available.
NG FMC and More is released.

Main Menu

STAB trim cutout - still able to trim??

Started by sbergert, Mon, 7 Aug 2017 14:39

John H Watson

#20
Sorry.. crossposted again... I'll answer your last post in my next message...

Quote(There are no notes re autoland, as far as I can see. The stab cutout keeps the autopilots operative. Pitch control remains possible without stab trim.)

Surely as soon as the stabiliser doesn't move, the SRM will send a (autotrim) fault signal to the FCCs?

From my manuals...

The autotrim valid discrete from the Left SRM is transmitted to the Left and Centre FCC and from the Right SRM to the Right FCC. This discrete is used by the FCC to control autotrim and for engage interlocks.

Autotrim valid from the SRM is required for single channel to stay engaged.


If there is a single autopilot channel engaged, the loss of the autotrim valid signal causes the autopilot to disconnect...

Page 306/7
During triple-engage [and, say the Left A/P was first in command], the Left FCC uses the Left SRM. If dead trim is detected, the autotrim function will autosequence to the Right FCC and the Right SRM.



Hardy Heinlin

#21
OK, thanks.

In PSX, the A/P single channel validation checks if either SRM is powered. That's OK. But the cutout switch doesn't remove power from the SRM model. Power is set by the electrical busses and CBs, not by the cutout.

QuoteThe autotrim valid discrete from the Left SRM is transmitted to the Left and Centre FCC and from the Right SRM to the Right FCC.

Is that "autotrim valid" discrete sent continuously? If so, FCC L & C should be inop when SRM L is unpowered, and FCC R should be inop when SRM R is unpowered. In the current PSX version, any FCC can operate as long as SRM L or R is powered (referring to electrical busses and CBs, not to the cutout).

John H Watson

QuoteIs that "autotrim valid" discrete sent continuously?

I'd say that dead trim will be reported whenever a trim command is given and nothing happens (e.g. if no hydraulics to move the stab). SRM power failures should be detected immediately (I assume with a small time delay to limit nuisance messages).

QuoteIf so, FCC L & C should be inop when SRM L is unpowered, and FCC R should be inop when SRM R is unpowered.

Certainly with single A/P engagement. Here's where it gets confusing. The notes do say that the Left and Right SRM's send "autotrim valid" discretes to the FCCs (The Left SRM sends it to the Left and Centre FCCs, the Right SRM sends it to the Right FCC and the Centre FCC. I don't know what happens if only one SRM is operating with the Centre A/P in command.

"Autotrim valid from the stabilizer trim/rudder ratio module (SRM) is required for single channel to stay engaged"

During multi-engage, there is crosstalk between the FCCs. I don't know how this affects engagement if one SRM fails.

Stab position has always been a requirement for engagement. "Local" sensors are required for engagement after Service Bulletin SB 22-2197 (Local stab and local IRU for single A/P engagement)






Hardy Heinlin

In all PSX versions so far, stab position data is always crosstalked across all FCCs, even in single channel mode. So to get rid of the data, all three FCCs need to be unpowered, and then there is no FCC operative anyway :-)

John H Watson

Regarding local sensors..  There are 3 RVDTs for stab position which report to their respective Flap Control Unit (L/C/R). Each flap control unit provides stab position data to the respective FCC (ref SSM 27-51-04).

Even if the data is shared, that still contradicts the local sensor rule (during single A/P ops)

Hardy Heinlin

Quote from: John H Watson on Wed,  9 Aug 2017 11:49
The autotrim valid discrete from the Left SRM is transmitted to the Left and Centre FCC and from the Right SRM to the Right FCC.

Are these assignments of SRM L & R only related to autotrim monitoring? In PSX, before any single channel validation and multichannel validation there is an initial check when a CMD switch is pressed *; this initially checks whether a related SRM is powered. And the "related SRM" in this particular check (in PSX) is assigned like this:

FCC L requires SRM L
FCC C requires SRM L or R
FCC R requires SRM R

I can't remember whether this was based on a theory or on a maintenance manual. Maybe this part of the logic should have the same assignment like the autotrim validation, namely:

FCC L requires SRM L
FCC C requires SRM L
FCC R requires SRM R


Cheers,

|-|ardy


* This is that check which also tests the Auto Flight Warning system before it allows A/P engagement.

John H Watson

QuoteThe autotrim valid discrete from the Left SRM is transmitted to the Left and Centre FCC and from the Right SRM to the Right FCC

Sorry, that may be a misquote (unless there are parts of the manuals which disagree with each other). I did a quick search of my training manuals and I did find one part of the manual which agrees with my later message:

QuoteThe notes do say that the Left and Right SRM's send "autotrim valid" discretes to the FCCs (The Left SRM sends it to the Left and Centre FCCs, the Right SRM sends it to the Right FCC and the Centre FCC.

Quote from: HardyAnd the "related SRM" in this particular check (in PSX) is assigned like this:

FCC L requires SRM L
FCC C requires SRM L or R
FCC R requires SRM R

This does seem logical. During single engage, this would give the Centre A/P the ability to autosequence SRMs due to SRM failures.

Elec Book 15, page 306 shows autosequencing of SRMs during multi-engage. i.e. two or three A/Ps engaged.

I'll keep searching for more info...

Cheers
JHW






Hardy Heinlin

#27
Thanks! :-)

From the book:
QuoteThe autotrim valid signals tell the FCC if the SRM is not valid or if manual trim switches are in use.

As far as I can see, autotrim is not valid when manual trim switches are pressed or when the SRM fails. This is already modelled in PSX (manual trim or SRM failure disengages the single channel A/P). Is the SRM also considered failed (invalid) when both hydraulics are cutout? If so, I would implement this condition in the "autotrim valid" section instead of making an extra validation check for this. Or is the cutout irrelevant and just the consequence of it, i.e. "dead trim", will invalidate the SRM?

I find the "dead trim" detector more interesting :-) E.g. no stab motion for, say, 3 seconds while commanded to move. (I don't know what exact timings are used in the detection.)


Cheers,

|-|ardy


P.S.: SRM internal malfunctions which cause "dead trim" are not modelled in PSX. So autosequencing to the other SRM is a job that will never be required. "Dead trim" can only occur by electrical or hydraulic power loss, and then it's not really a "dead trim" effect anymore as the trim command itself is already invalid (probably).


P.P.S.: Re "unscheduled trim" ... now I wonder whether alternate trim actions with A/P engaged cause the A/P pitch mode to level off with a yellow line through the pitch mode word. Alternate trim with A/P engaged triggers an "unscheduled trim" alert which now also triggers the >AUTOPILOT caution (as I just learned) which typically refers to a mode fault; in this case a pitch mode fault. But perhaps the mode fault (yellow line) will not occur because the elevator still has full authority even when the stab is extremely mistrimmed.

John H Watson

Sorry, too much reading for me to do at the moment (to fully understand the system). You would need to find a diagram showing the SRM fault monitoring module logic gates.

The SRM has a lot of information going to it.
Each cutout/auto/on Stab switch has two sets of contacts. One set of contacts tells the SRM (with an earth) if a switch is in ON or CUTOUT (on separate lines). If no earth on either, it assumes the switch is in AUTO. The other set of cutout switch contacts provides P7 CB power to each MOV (which shuts off hydraulics) and the SRM also knows the position of this switch (if power is available to the switch).
The SRM also knows the position of the MOV. It also knows the position of the stabiliser if the position circuits are powered. The SRM also knows hydraulic system pressure. The SRMs also talk to each other.

It's up to the SRM fault monitoring circuits to generate EICAS messages and to output an invalid signal to the FCC's with the above (and other) data.

Cheers
JHW

Hardy Heinlin

Thank you! It's enough information now anyway, and I'm finished with this chapter.

Since SRM autosequencing is not required (both SRMs report dead trim anyway), any dead trim detection will trigger "no autoland" and never "no land 3".


Cheers,

|-|ardy

Hardy Heinlin

#30
The features discussed in this thread are now available in PSX 10.1.8 (items 1.8.006 + 1.8.007):

http://aerowinx.com/board/index.php?topic=4191.0


Cheers,

|-|ardy


Hardy Heinlin

#32
I was playing chess with various AC/DC bus configurations, and noticed a side effect of the above modifications introduced in version 10.1.8:

When bus 2 alone or bus 3 alone fails, LAND 2 mode activates. That's nice. But when bus 1 alone fails, NO AUTOLAND occurs. Not nice, is it?

During the single channel validation each FCC checks its dedicated SRM and servo control (among other things). The respective power supplies are assigned as follows. All CBs (C10..C17) are on P7.

SRMs:

FCC L uses SRM L
FCC C uses SRM L or R, whichever is powered
FCC R uses SRM R

SRM L operates if these are powered: ( C10 or C12 ) and ( C11 or C13 )
SRM R operates if these are powered: C15 and C16 and C17

C10 (C10263) "AC2L" powered by main standby bus
C11 (C10264) "DC2L" powered by main battery bus
C12 (C8782) "AC1L" powered by main standby bus
C13 (C8783) "DC1L" powered by main battery bus
C14 (C8784) "AC1R" powered by AC bus 1
C15 (C8785) "DC1R" powered by DC bus 1
C16 (C10265) "AC2R" powered by AC bus 2
C17 (C10266) "DC2R" powered by DC bus 2

Servos:

FCC L servo needs DC bus 3
FCC C servo needs DC bus 1
FCC R servo needs DC bus 2

Now ...
If bus 3 fails: Servo of FCC L fails -- I.e. just one FCC fails (LAND 2)
If bus 2 fails: Servo and SRM of FCC R fail -- I.e. just one FCC fails (LAND 2)
If bus 1 fails: Servo of FCC C and SRM of FCC R fail -- I.e. two FCCs fail (NO AUTOLAND)

A single failure of bus 2 or 3: no problem. A single failure of bus 1: problem. Can this be normal?


Cheers,

|-|ardy



Quote from: John H Watson on Thu, 10 Aug 2017 10:12
There are 3 RVDTs for stab position which report to their respective Flap Control Unit (L/C/R). Each flap control unit provides stab position data to the respective FCC (ref SSM 27-51-04).

If each flap control unit too can provide stab position data to its associated FCC, maybe I should use them when the respective SRM fails?


John H Watson

QuoteA single failure of bus 1: problem. Can this be normal?

I think it is. I don't think the aircraft was designed to carry out autolands with bus failures... only generator failures.

The battery chargers are on AC Bus #1. Wouldn't that be an issue?

I have this vague recollection that some QRHs have a special note about AC 1.

Hardy Heinlin

#34
I just noticed that at least one real-life 744 says NO LAND 3 in case of AC bus 1 failure,
whereas PSX says NO AUTOLAND (see comments above from April 2018).

Can this really be aircraft specific?


|-|ardy


Edit:

Quote from: Hardy Heinlin on Wed, 18 Apr 2018 04:05
SRMs:

FCC L uses SRM L
FCC C uses SRM L or R, whichever is powered
FCC R uses SRM R

SRM L operates if these are powered: ( C10 or C12 ) and ( C11 or C13 )
SRM R operates if these are powered: C15 and C16 and C17

C10 (C10263) "AC2L" powered by main standby bus
C11 (C10264) "DC2L" powered by main battery bus
C12 (C8782) "AC1L" powered by main standby bus
C13 (C8783) "DC1L" powered by main battery bus
C14 (C8784) "AC1R" powered by AC bus 1
C15 (C8785) "DC1R" powered by DC bus 1
C16 (C10265) "AC2R" powered by AC bus 2
C17 (C10266) "DC2R" powered by DC bus 2

I think in the red text line there should be an "or" instead of an "and", like this in blue:

SRM R operates if these are powered: ( C15 or C17 ) and C16

SRM L has an OR between two DC sources, and another OR between two AC sources.
Similarly, I think SRM R should not have an AND between an AC source and a DC source; it should combine DC with DC, and AC with AC, and not mix AC with DC.
Also, SRM L can work with a single pair of a #1 and a #2 bus; it just needs two, not three like in the red line. I think SRM R should also need just two, not three. That is, the three sources shouldn't all be combined by AND operators. There should be at least one OR. And that should be between the two DC sources C15 and C17.

Now when AC bus 1 fails, SRM R still gets power from C16 and C17. And FCC R remains valid, and "NO LAND 3" replaces "NO AUTOLAND".

John H Watson

If you still have a copy of the basic PSM (FCE power supply) diagram...

Each SRM receives (and perhaps needs) from the PSMs:

1) +5vdc and +/-15Vdc (derived from the 115Vac CBs mentioned above)
2) PSM VALID signal (an earth signal)

Looking at the schematics, each PSM has two channels (1 & 2). Two DC sources go to each PSM. One DC source powers the fault monitoring circuits in Channel 1. The other DC source powers the fault monitoring circuits in Channel 2. Each PSM needs to see both channels producing an ok signal before a "PSM VALID" (earth) signal is sent to the SRMs. If a channel is dead, I can't see how it can generate a ok signal.

The Left SRM needs 115vac C10 or C12 (it has redundancy) for +5dc and +/-15vdc
The Right SRM needs 115vac C16 (no redundancy) for +5dc and +/-15vdc

The Left SRM also needs (as far as I can see) the respective PSM to have two sources of DC to produce a "PSM valid" signal. i.e. (C11 and C13).
The Right SRM also needs (as far as I can see) the respective PSM to have two sources of DC to produce a "PSM VALID" signal. i.e. (C15 and C17)

With my current understanding, it should be:

SRM L operates if these are powered: ( C10 or C12 ) and ( C11 and C13 )
SRM R operates if these are powered: (C16) and (C15 and C17)


To me, NO AUTOLAND is more likely as a result of power source failure.

I seem to remember pulling FCE CBs to see what messages would be generated (on the real aircraft), but I can't remember the results.

Hardy, ref: Blue Test forum post 13850

John H Watson

QuoteCan this really be aircraft specific?

Well, we know there are hundreds of options when it comes to autopilot stuff... Where do I begin?  ;)

John H Watson

Next theory...  Maybe I misunderstood the function of the PSM VALID signal...

From the books:

Quote- VALID: PSM internal monitors output (send) a valid discrete to the YDM,
SRM and FCU. Loss of all valid discretes will inhibit fault
annunciating by the operating modules resulting from an out of
tolerance power supply output. Receipt of a valid power discrete
is required prior to microprocessor initialization.


Assuming "loss of all valid discretes" can be produced by loss of  DC power to either channel in the PSM ...  and assuming "operating modules" means YDM, SRM, FCU.... then the loss of one or more DC voltage sources might simply stop the YDM, SRM and FCU generating a fail message (?).

However, on power up, the "operating modules" (YDM, SRM, FCU, etc) must receive a valid signal from the PSM. Otherwise they might generate fail messages...  and even fail to operate (?).

So, in theory, if both sources of PSM DC power is lost in flight (and an SRM is operating at the time), the SRM carries on with what it was doing. However, if the PSM were to lose AC power, the SRM would lose +5Vdc and +/-15Vdc and the SRM, etc, would definitely fail.

The PSMs also provide 26Vac (to various flight control systems). If both DC power sources are lost from the left PSMs, the backup source for 26Vac (in PSM 2L) won't be available (as the PSM internal switching relays and logic use DC voltage).

I don't really know how to prove this without a real aircraft to play on.

Hardy Heinlin

#38
I haven't read post #37 yet. Here's my reply to post #35:

Quote from: John H Watson on Sat, 25 Aug 2018 01:07
Two DC sources go to each PSM. One DC source powers the fault monitoring circuits in Channel 1. The other DC source powers the fault monitoring circuits in Channel 2. Each PSM needs to see both channels producing an ok signal before a "PSM VALID" (earth) signal is sent to the SRMs. If a channel is dead, I can't see how it can generate a ok signal.

Problem A (Are always two DC sources required for a PSM-valid signal?)
Can we be sure that a PSM needs to see -- during preflight and inflight -- both channels to produce a VALID signal? I see no point in having a second source as a backup if the PSM will never work with just a single source anyway. The box promises to provide double safety, but it will stop operating when double safety is no longer assured?

Could it be that the validation logic implies more than just a power source check? Maybe it checks whether the DC failure occurs preflight or inflight. It may be a MEL item before takeoff, but may still allow autoland once the aircraft is in the air?

Problem B (Does SRM L need two PSM-valid signals? Not to confuse with problem A!)
Regarding your new suggestion that SRM L should get two FCE valid signals, i.e. two DC pairs (double-double). Since SRM R is able to operate with a single FCE, that is, with just one valid signal from FCE 2R, why should SRM L be different and require more than one FCE? SRM L gets a valid from FCE 1L and another valid from FCE 2L. Either valid signal already incorporates a pair of DC channels. I guess SRM L will work with the one pair or with the other pair; an SRM doesn't need 4 DC channels. 1 pair, i.e. one FCE is sufficient. The other pair is a backup. Only SRM L has a backup.


Cheers,

|-|ardy


P.S.: Re problem B -- just in case my code text is being misunderstood ... just in case ...
"SRM L operates if these are powered: ( C10 or C12 ) and ( C11 or C13 )"
This means: "SRM L operates if these are powered: Fce1L or Fce2L"

Fce1L = ( Dc.C11 or Dc.C13 ) and Ac.C12
Fce2L = ( Dc.C11 or Dc.C13 ) and Ac.C10

Either Fce gets the same blue stuff. So in the programming language it needs to be written just once. This line ...

( Ac.C10 and ( Dc.C11 or Dc.C13 ) )  or  ( Ac.C12 ) and ( Dc.C11 or Dc.C13 ) )

... is equal to this line ...

( Ac.C10 or Ac.C12 ) and ( Dc.C11 or Dc.C13 )

Note that this P.S. refers to problem B :-)

Hardy Heinlin

Quote from: John H Watson on Sat, 25 Aug 2018 04:09
Next theory...  Maybe I misunderstood the function of the PSM VALID signal...

From the books:

Quote- VALID: PSM internal monitors output (send) a valid discrete to the YDM,
SRM and FCU. Loss of all valid discretes will inhibit fault
annunciating by the operating modules resulting from an out of
tolerance power supply output. Receipt of a valid power discrete
is required prior to microprocessor initialization.

I interprete this as "a single valid discrete is sufficient to provide fault annunciating".