Peak EGT?

[n4gix 4,947]

I need to determine when EGT has "peaked" during the leaning process.I have a digital display that will show the EGT for cylinders 1 through 6, and the "average EGT." Since on most Lycoming engines, cylinder #5 will nearly always be the 'hottest,' I'm tracking the delta of #5 during the "LeanFind" process.I've tried the following logic, but it isn't working... :(case PANEL_SERVICE_PRE_UPDATE:if ( leanfind == 1 && stopfind == 0 ) { peak_egt = cyl5_egt ; } if ( peak_egt > cyl5_egt ) { stopfind = 1 ; } The above should record the current cyl5_egt, store it to peak_egt, and then STOP whenever the highest temp is found, leaving the 'peak value' stored in peak_egt.Then for the digital display, I use this in the callback...FLOAT64 FSAPI callback3( PELEMENT_STRING pelement){float rwert=pelement->source_var[0].var_value.n;if ( step == 0 && stopfind == 0 ) { rwert = (float)cyl1_egt ; } if ( step == 1 && stopfind == 0 ) { rwert = (float)cyl2_egt ; } if ( step == 2 && stopfind == 0 ) { rwert = (float)cyl3_egt ; } if ( step == 3 && stopfind == 0 ) { rwert = (float)cyl4_egt ; } if ( step == 4 && stopfind == 0 ) { rwert = (float)cyl5_egt ; } if ( step == 5 && stopfind == 0 ) { rwert = (float)cyl6_egt ; } if ( step == 6 && stopfind == 0 ) { rwert = (float)eng1_egt ; } if ( stopfind == 1 ) { rwert = (float)peak_egt ; } if ( GENERAL_PANEL_CIRCUIT_ONvar.var_value.n == 1 ) { LIGHT_IMAGE(pelement) ; SHOW_IMAGE(pelement) ; } else { HIDE_IMAGE(pelement) ; } sprintf(pelement->string,"%4.0f",rwert);return 0;}The above should display the values for cyl1_egt through cly6_egt, then the 'average value' eng1_egt.If the variable stopfind is 1, then it should display only peak_egt, which will LOCK once the peak is found.Can anyone explain why this isn't working the way I intend? :)

[Guest jk]

the first thing I see, is that during the leanfind, peak_egt can never be greater than cyl5_egt, because you are making every time peak_egt equal to cyl5_egt.

[n4gix 4,947]

>the first thing I see, is that during the leanfind, peak_egt>can never be greater than cyl5_egt, because you are making>every time peak_egt equal to cyl5_egt.Duh! Why could I not see the obvious... I simply need to move the comparison to one line *before* the assignment... :)Thanks for finding my silliness... That's what I get for working until 3am on a project...

[ddawson 35]

I think you want:case PANEL_SERVICE_PRE_UPDATE:if ( leanfind == 1 && stopfind == 0 && peak_egt < cyl5_egt) { peak_egt = cyl5_egt ; }if ( peak_egt > cyl5_egt ) { stopfind = 1 ; }Doug

[n4gix 4,947]

Doug, that doesn't seem to have solved the problem either...Let me restate the goal, and see if anyone can think of a different approach...Whenever you adjust the mixture control, the EGT will rise as you slowly lean mixture, until it reaches a "peak value," and then will begin decreasing until you've leaned too much, at which time the engine will suffer from fuel starvation and stop.This is, of course, because the excess fuel that is fed into the cylinders acts as a 'coolant,' resulting in a lower exhaust gas temperature.What I need to do is "capture" the peak temperature, and then display it on a digital readout. This will provide a base value from which the pilot can either adjust the mixture to run LOP or ROP, as he chooses.Ideally, when the 'peak' is found and locked into the peak_egt variable, the display will briefly flash "PK FND," and then several seconds later will display the contents of peak_egt.Any ideas, anyone? :)

[ddawson 35]

Bill,It occurs to me that FS variables are seldom 100% rock solid stable at any value. I'm wondering if your trap (the second if statement) is being triggered by a trivial fluctuation in EGT, while the mixture is still at 'full rich'. Have you tried this same rountine while rounding the temperatures to full degrees, or, alternatively, a variation like this:if ( peak_egt > cyl5_egt + 2 ) { stopfind = 1 ; } Doug

[n4gix 4,947]

>Bill,>>It occurs to me that FS variables are seldom 100% rock solid>stable at any value. I'm wondering if your trap (the second>if statement) is being triggered by a trivial fluctuation in>EGT, while the mixture is still at 'full rich'. Have you>tried this same rountine while rounding the temperatures to>full degrees, or, alternatively, a variation like this:>>if ( peak_egt > cyl5_egt + 2 ) { stopfind = 1 ; } No. When I get back this afternoon, I'll change the variables to int type instead of float, and give it another whirl... Thanks for the suggestions!