fix: timers throttle condition on surface radio

[3djc]

Account for surface radio specific throttle.

This fixes #4499

This also fixes corresponding test scripts

[philmoz]

I was looking at this as well and took a slightly different approach.

mixer.cpp (doMixerPeriodicUpdates) - change:

    val >>= (RESX_SHIFT-6); // calibrate it (resolution increased by factor 4)

    evalTimers(val, tick10ms);

to:

    evalTimers(val, tick10ms);

    val >>= (RESX_SHIFT-6); // calibrate it (resolution increased by factor 4)

timers.cpp (evalTimers) - add at the start:

  // 'throttle' range is 0 (-100%) to 2048 (100%)
  // Scale value to range 0 - 128
#if defined(SURFACE_RADIO)
  // For surface radio throttle off position is at 0%
  throttle = abs(throttle - RESX) >> 3;
#else
  // For air radio throttle off position is at -100%
  throttle = throttle >> 4;
#endif

[pfeerick]

On both Pocket and MT12, I configured all three timers to be throttle triggered in conjunction with SD down. THs, TH% and THt (trigger). On the Pocket, it stills seems to be working fine with this PR. and on the MT12 it looks good also - needed to recalibrate though as it was a bit temperamental about the centre point to allow the THRs condition to let go. Another reason it may be handy to have a radio level setting that lets you set the THs/THt trigger point.

As a side-note, on the MT12, when I check the statistics, TH% and THR are still counting with the stick in resting position ;) Looks like

edgetx/radio/src/gui/128x64/view_statistics.cpp

Lines 83 to 87 in 795789a

	// Throttle special timers
	lcdDrawText(STATS_2ND_COLUMN, FH*0+1, "THR", BOLD);
	drawTimer(STATS_2ND_COLUMN + STATS_LABEL_WIDTH, FH*0+1, s_timeCumThr);
	lcdDrawText(STATS_2ND_COLUMN, FH*1+1, "TH%", BOLD);
	drawTimer(STATS_2ND_COLUMN + STATS_LABEL_WIDTH, FH*1+1, s_timeCum16ThrP/16);

needs to learn about surface radios too at some point... may also be worth considering how to do the throttle graph differently... maybe a line graph rather than bars?