MCE not recognizing Commands with PMDG 737 NGXu

[starter 8]

5 hours ago, LecLightning56 said:

I seem to have made a little progress today. Things are a little better with the PMDG 737NGX but I am still having autopilot problems. What seems to be happening is that when a command to arm is autopilot is given, the CMD button is activated but then an autopilot disconnect warning is sounded in the cockpit, with a response by the PNF that a fault has occurred and the autopilot could not be switched on. Bizarrely, the CMD button remains illuminated and the aircraft continues to fly as if under autopilot control (with the CMD button illuminated). 

When i say "arm ap", fo is doing it, say something, and turn it off (see light is out) and when the disconnct warns sounds, then it gets on again. Don't know if she responds to my "arm ap" again.

When the sounds goes i say immeditely arm ap and then it works

Best regards

Edited February 14, 2020 by starter

[LecLightning56 33]

3 hours ago, starter said:

When i say "arm ap", fo is doing it, say something, and turn it off (see light is out) and when the disconnct warns sounds, then it gets on again. Don't know if she responds to my "arm ap" again.

When the sounds goes i say immeditely arm ap and then it works

Best regards

Not sure if I understood your logic, but are you suggesting that there is  a slightly convoluted method of engaging the autopilot? If so, this does not seem to follow how one might expect things to happen in the absence of MCE.

[charlie130 196]

Having read this thread with some interest, if I may interject here.

The FSi"panel" utility mentioned by the op, I would suspect has a bearing on the issues previously stated:

https://www.fsipanel.com/

Has this panel been used on the PMDG 747 QotS, and if so does the op find similar issues? This panel utility would appear to reset the simulation parameters to permit "training" scenarios. For example repeated take offs and landings, go arounds etc. I have no experience of the panel utility, but I believe that it's functions may result in false positives being sent to MCE. Since the prior aircraft state is altered from what MCE has previously detected. 

From my perspective, as a MCE flyer for many years and a NGX FSX flyer from some time ago, all commands work as intended. Same goes for the PMDG 747-400 QotS, which I fly currently in P3d4.5. If I had discovered a fault, I would have reported it to support. If anyone would like to buy me a copy of the NGXu, I am willing to test it and I would place a bet that the functions at issue would work correctly. I can infer no difference in MCE function between NGX, NGXu and 747-400. 

"Approach mode" will command on if AP is engaged with ILS approach entered in the arrival FMC page. 

"Auto pilot off" (admittedly never requested by PF) will work following a confirmatory request from PNF.

"Nav One track one nine six", will result in the nav one course selection of 196. The command can not be confused with a radio frequency request, "set com one, one nine six".

Hope this helps somewhat. These days we are dealing with complex simulation software running alongside other complex software. All components are required to "talk" to each other inside our Personal Computers, the latter being uniquely set up, with no two systems being identical. Ultimately, if these "bits" don't agree then the resulting arguments create strife!🥴

Regards

David

 

 

 

 

 

[LecLightning56 33]

Unfortunately it would appear that there is something amiss with how MCE is engaging autopilot with the PMDG 737NGXu. I have just tried a takeoff from KVPS and when commanded to 'engage autopilot', the PNF confirms, CMD is lit but the autopilot disengage warning is sounded in the cockpit and the PNF makes a comment that he was unable to engage the autopilot due to a fault. If I immediately manually disengage the autopilot then press the CMD button, the autopilot does actually engage correctly and no fault or warning is raised. If all that MCE is doing is effectively pressing the CMD button with certain provisos (I believe that with 'realistic' autopilot engagement with the 737NGX the aircraft has to be in a stable flight condition and perhaps above a certain altitude), then I cannot see why I am getting this fault from using MCE for this action. Please note that I am not using any other software for this exercise (e.g. FSiPanel 2017 etc.) and am simply getting the aircraft setup on the runway with correct FMC entries and NAV1 frequency set.

 

[LecLightning56 33]

11 hours ago, charlie130 said:

Has this panel been used on the PMDG 747 QotS, and if so does the op find similar issues?

Yes, I have tried FSiPanel with the PMDG 747 QOTS II and MCE with few issues. Autopilot works very well with MCE, although I do appreciate that in the real world the PNF would not be engaged to activate/disengage the autopilot but we are only in a sim world and may find it convenient to automate such a procedure through MCE.

[FS++ 707]

3 hours ago, LecLightning56 said:

If all that MCE is doing is effectively pressing the CMD button with certain provisos (I believe that with 'realistic' autopilot engagement with the 737NGX the aircraft has to be in a stable flight condition and perhaps above a certain altitude), then I cannot see why I am getting this fault from using MCE for this action. Please note that I am not using any other software for this exercise (e.g. FSiPanel 2017 etc.) and am simply getting the aircraft setup on the runway with correct FMC entries and NAV1 frequency set.

 

That's exactly what Fo is doing.

Command "engage autopilot"

Fo checks no AP is on (including CMD B, CWS A and CWS B)and simulates click using PMDG SDK to engage CMD A

Good point regarding NGX AP engagement rules. In fact it may not happen even with manual clicking, this is why we get the Fo to check the AP was actually engaged after the action. If still not on after a couple of seconds delay, he'll warn you that he wasn't able to set it (something he only does with NGX).

To check he's able to read AP status, select any CMD A, CMD B, CWS A or CWS B and request "engage autopilot". He should say "autopilot already on".

You can rule out FSiPanel if it's not attempting to set MCP switches.

I assume you have "mcpmd73X.dll" V2.1.2.0 in \Prepar3D v4\MCE dlls\

Reading state and setting switches is done via that dll (for NGX and NGXu).

The audio, speech reco, program logic and things that don't need to run inside the simulator process, run in "mce.exe", and dll files present inside \Multi Crew Experience\ installation folder.

Edited February 15, 2020 by FS++

[LecLightning56 33]

2 minutes ago, FS++ said:

That's exactly what Fo is doing.

Command "engage autopilot"

Fo checks no AP is on (including CMD B, CWS A and CWS B)and simulates click using PMDG SDK to engage CMD A

Good point regarding NGX AP engagement rules. In fact it may not happen even with manual clicking, this is why we get the Fo to check the AP was actually engaged after the action. If still not on after a couple of seconds delay, he'll warn you that he wasn't able to set it (something he only does with NGX).

To check he's able to read AP status, select any CMD A, CMD B, CWS A or CWS B and request "engage autopilot". He should say "autopilot already on".

You can rule out FSiPanel if it's not attempting to set MCP switches.

 

3 hours ago, LecLightning56 said:

Unfortunately it would appear that there is something amiss with how MCE is engaging autopilot with the PMDG 737NGXu. I have just tried a takeoff from KVPS and when commanded to 'engage autopilot', the PNF confirms, CMD is lit but the autopilot disengage warning is sounded in the cockpit and the PNF makes a comment that he was unable to engage the autopilot due to a fault. If I immediately manually disengage the autopilot then press the CMD button, the autopilot does actually engage correctly and no fault or warning is raised. If all that MCE is doing is effectively pressing the CMD button with certain provisos (I believe that with 'realistic' autopilot engagement with the 737NGX the aircraft has to be in a stable flight condition and perhaps above a certain altitude), then I cannot see why I am getting this fault from using MCE for this action. Please note that I am not using any other software for this exercise (e.g. FSiPanel 2017 etc.) and am simply getting the aircraft setup on the runway with correct FMC entries and NAV1 frequency set.

 

Forgive me for sounding a little persistent, but the point I am making is that the command to engage autopilot is executed by the FO when requested, but the PMDG 737NGXu immediately displays an A/P disconnect red warning and sound and the FO in MCE responds with the fact that A/P failed to be engaged due to a fault, even though CMD A remains lit. This behaviour is not replicated if I ignore MCE and simply engage A/P manually by myself. This seems to be a quirk of MCE, not of FSiPanel which has not been used for this exercise.

[starter 8]

Made 2 circuits today.

took off and after 400ft put off the pressure of the yoke. At 800ft i say "arm autopilot", it gets armed, heard something about a fault, it got disconnected and connected again.

(maybe something to do with pressing twice the ap button to silense the siren?)

Pressed the yellow flickering a/p button.

Then tried to disconnect the ap (saying:dis arm, disconnect, disable ). FO says: ap already disabled.

As for setting course: set course/track one 196 is working.

For set course/track two 196, it needs confirmation for transponder setting.

Nav One/two track 196 is working both sides.

For the VOR/LOC and APP mode i get the message" you must either engaged the ap/fd". (both were engaged and everything was setup).

Best regards,

PS forgot to mention that i have disabled and self enabled the ap/fd. Still FO saying ap already disabled

Edited February 15, 2020 by starter

[LecLightning56 33]

2 minutes ago, starter said:

Made 2 circuits today.

took off and after 400ft put off the pressure of the yoke. At 800ft i say "arm autopilot", it gets armed, heard something about a fault, it got disconnected and connected again.

(maybe something to do with pressing twice the ap button to silense the siren?)

Pressed the yellow flickering a/p button.

Then tried to disconnect the ap (saying:dis arm, disconnect, disable ). FO says: ap already disabled.

As for setting course: set course/track one 196 is working.

For set course/track two 196, it needs confirmation for transponder setting.

Nav One/two track 196 is working both sides.

For the VOR/LOC and APP mode i get the message" you must either engaged the ap/fd". (both were engaged and everything was setup).

Best regards,

 

Thanks, starter. We seem to be singing from the same hymn book with similar experiences. I hope that the MCE crew can verify/duplicate this behaviour and provide a fix if possible. 

[FS++ 707]

2 minutes ago, LecLightning56 said:

Thanks, starter. We seem to be singing from the same hymn book with similar experiences. I hope that the MCE crew can verify/duplicate this behaviour and provide a fix if possible. 

I assume you have "mcpmd73X.dll" V2.1.2.0 in \Prepar3D v4\MCE dlls\

Perform this test.

manually engage any of these CMD A, CMD B, CWS A, or CWS B, then ask co-pilot "engage autopilot"

Unless he replies "autopilot already on", it could be the root cause of the issue.

For NGX data reading, it's also important the dll clearly identifies which variant is currently loaded (NGX versus NGXu).

That's why I asked you to perform the test regarding GPU connection.

Could you re-do it and confirm GPU is connected every time, be it with NGX or NGXu.

That test will confirm the dll knows exactly which variant is currently loaded.

[LecLightning56 33]

57 minutes ago, FS++ said:

I assume you have "mcpmd73X.dll" V2.1.2.0 in \Prepar3D v4\MCE dlls\

Perform this test.

manually engage any of these CMD A, CMD B, CWS A, or CWS B, then ask co-pilot "engage autopilot"

Unless he replies "autopilot already on", it could be the root cause of the issue.

For NGX data reading, it's also important the dll clearly identifies which variant is currently loaded (NGX versus NGXu).

That's why I asked you to perform the test regarding GPU connection.

Could you re-do it and confirm GPU is connected every time, be it with NGX or NGXu.

That test will confirm the dll knows exactly which variant is currently loaded.

I have performed your test manually engaging CMD A and then requested the PNF to "engage autopilot". This resulted in the PNF commenting that the autopilot was already engaged. If then I manually disconnect the autopilot and then request the PNF to engage the autopilot, this results in a failure along exactly the lines discussed in this thread and, although the CMD A button may be lit, as far as MCE is concerned, the autopilot is effectively disengaged although the PMDG model carries on as if the autopilot is fully functional with the CMD A button lit.

I do have the correct DLL and the ground test is successful.

Edited February 15, 2020 by LecLightning56

[FS++ 707]

2 hours ago, LecLightning56 said:

I have performed your test manually engaging CMD A and then requested the PNF to "engage autopilot". This resulted in the PNF commenting that the autopilot was already engaged. If then I manually disconnect the autopilot and then request the PNF to engage the autopilot, this results in a failure along exactly the lines discussed in this thread and, although the CMD A button may be lit, as far as MCE is concerned, the autopilot is effectively disengaged although the PMDG model carries on as if the autopilot is fully functional with the CMD A button lit.

I do have the correct DLL and the ground test is successful.

OK, a new dll will be posted in the middle of next week.

 

[LecLightning56 33]

21 minutes ago, FS++ said:

OK, a new dll will be posted in the middle of next week.

 

Many thanks for your investigations and for providing a fix for the issues brought to your attention by the user community. 

[charlie130 196]

And all this for a single payment that covers all platforms dating right back to fs9! 

That's what I call value for money!

But I think you should charge for any upgrades to MSFS 2020...you deserve a Ferrari or two!😁

 

[goggi 9]

I concur wholeheartedly!
Giorgio