SIMUFLY.COM Discussion Forums

SIMUFLY.COM Discussion Forums
https://www.simufly.com/cgi-bin/yabb2/YaBB.pl
Delco Carousel IV-A INS For Microsoft Flight Simulator >> Delco Carousel IV-A INS For Microsoft Flight Simulator >> Bug in InsAutopilotControlState logic
https://www.simufly.com/cgi-bin/yabb2/YaBB.pl?num=1181487000

Message started by w6kd on Jun 10^th, 2007, 3:50pm

Title: Bug in InsAutopilotControlState logic
Post by w6kd on Jun 10^th, 2007, 3:50pm

Hi;

I'm writing an HSI that displays CIVA data, and have run into a problem.

I use the NavModeTriple gauge to select INS units, and I display the data from the selected unit using a scan of the InsAutopilotControlState values for each of the three INSs. I discovered that if I select an INS to control the autopilot with one of the switches on the NavModeTriple gauge, and then select another INS without first turning off the one that is on, both InsAutopilotControlState values are set to one at the same time, even though only the last one selected is actually controlling the autopilot. If I turn the currently selected switch off first, it'll set the control state back to zero.

Regards

Bob Scott
ATP IMEL Gulfstream II-III-IV-V
Santiago de Chile

Title: Re: Bug in InsAutopilotControlState logic
Post by Marco Ravanello on Jun 11^th, 2007, 9:04am

Bob,

Would you show me the piece of code where you manage which ins is active on the A/P? I'd like to check it.
If you don't want to post it here in the forum you could send it to ins_A_T_simufly.com

Marco

Title: Re: Bug in InsAutopilotControlState logic
Post by w6kd on Jun 11^th, 2007, 10:56pm

Marco;

I've already replaced the code...but during the debugging I wrote a short segment that displays (via FSUIPC offsets) the values returned from InsAutopilotControlState in real time (in the ADVDISPLAY window) each time the gauge is executed. essentially:

int INS1State, INS2State, INS3State;

INS1State = InsAutopilotControlState(INS_Unit_1);
INS2State = InsAutopilotControlState(INS_Unit_2);
INS3State = InsAutopilotControlState(INS_Unit_3);

IPC_Test = FSUIPC_Write(0x6D10, 4, &INS1State, &dwResult);
IPC_Test = FSUIPC_Write(0x6D14, 4, &INS2State, &dwResult);
IPC_Test = FSUIPC_Write(0x6D18, 4, &INS3State, &dwResult);
FSUIPC_Process(&dwResult);

With the CIVA NavModeTriple gauge, I turned on the #1 INS. I saw this on the display:

6D10 = 0x1
6D14 = 0x0
6D18 = 0x0

Then I selected the #2 INS (without turning off the #1 INS first). The #1 INS switch returned to off, but this is what I saw on the display:

6D10 = 0x1
6D14 = 0x1
6D18 = 0x0

Then I selected the #3 INS. The #2 switch went to off, and the display said:

6D10 = 0x1
6D14 = 0x1
6D18 = 0x1

In this case all three INSs are reporting that they are connected via the InsAutopilotControlState call.

If I turn the switch off manually, the state returns to zero as it should.

What I did to solve my problem was to write a new NavModeTriple gauge that explicitly turns off the other INS units before turning on the selected one...this keeps the state values returned by the InsAutopilotControlState call in synch with the actual state.

Regards

Bob Scott
ATP IMEL Gulfstream II-III-IV-V
Santiago de Chile

Title: Re: Bug in InsAutopilotControlState logic
Post by Marco Ravanello on Jun 12^th, 2007, 8:33am

Bob,

I'll check this evening at home, thanks a lot for the example. I'll report back tomorrow.

Regards
Marco

Title: Re: Bug in InsAutopilotControlState logic
Post by Marco Ravanello on Jun 15^th, 2007, 11:12pm

Bob,

I confirm you that it's a bug, already filed as #0021. It's limited to the SDK layer, it does not influence the INS itself which correctly excludes any other currently active INS unit, so it's not critical, but it is still a bug which will be corrected in the next release.

I'm just waiting to collect some problems then I'll produce version 1.51.
V1.51 is anyway planned to be released around the end of august/mid September.

Thanks for discovering and reporting it, I really appreciate that!

Kind regard
Marco