PPM decode with HET timer

HI Jesse,
What are the 9 different values in ppm_values? Do they make sense to you? Are they in an ascending order?

When you said 4294742144, can you clarify? Is this one single value or are they multiple values?

I'd would have expected ppm_values[0] to contain the smallest timestamp value while ppm_values[8] to have the largest value. Is there a reason to subtract as in ppm_values[0] - ppm_values[1] instead of ppm_values[1] - ppm_values[0] to get t0 and likewise for others?

Thanks for the replay Charles,

Indeed, it should be ppm_values[1] - ppm_values[0], end of the day I guess. The actual values of the spikes are ascending, and the times between them are about the same. Here are a few examples of the ppm_values:

37504
225792
224128
225664
225408
224384
225792
224512

32896
1635712
225664
224640
224640
225408
225536
224128

86144
224000
225536
225024
225280
1635072
224000
224512

The PPM signal consists 9 spikes and then a rest period before the next 9. I think my reading of out of sync with the actual signal (my 1th spike reading isn't actually the first spike). That would explain the larger value of >1600000. But still the values of ~220000 don't make sense, the real time between the spikes is around 1500uS, and with a HET frequency of 1MHz, each tick should be 1uS right? So I would expect values around 1500...

Any ideas?

Jesse

Well, figuered some things out. Re-wrote the interrupt function, this way the reading starts at the right time/spike:

void   edgeNotification(hetBASE_t * hetREG,uint32 edge){

	uint32 this_tick = hetGetTimestamp(hetRAM1);
	uint32 diff = this_tick - last_tick;

	if(diff > 1600000){
		spike = 0;
		this_tick = hetGetTimestamp(hetRAM1);
	}

	ppm_values[spike] = this_tick;
	++spike;
	last_tick = this_tick;

	if(spike >= 9){
		calculate_ppm();
	}
}


void calculate_ppm(){
	int i;
	for (i = 0; i < 8; ++i) {
		ppm_times[i] = ppm_values[i+1] - ppm_values[i];
	}

	if(ppm_times[4] < 180000){
		gioSetBit(gioPORTB, 6, 1);
	}else if (ppm_times[4] > 260000) {
		gioSetBit(gioPORTB, 6, 0);
	}
}

ppm_times[4] is from the 4th channel and connected to a switch on the RC controller. Because it is a switch I was able to figure out what the min and max values are: 160000 - 280000 (with 220000 in the middle).

The only thing that is left is the frequency of the HET timer. I've changed the HET frequency back to 110MHz, but still the results are all ~220000? Shouldn't there be more pulses/ticks compared to 1MHz, in the same period of time?

HI Jesse,

 Are you using HalCoGen? Can you please show me how you configure the Loop Time for the HET? Please show me how you configure the timebase like the below screenshot. Below screenshot is only a default setup by the HalCoGen.

The only thing changed is the HR Clock:

Alright, I will try those settings.

Thanks for the support Charles!

Hi Jesse,
I also want to remind you that the timestamp value you read out using hetGetTimestamp() will contain two fields, a loop resolution counter value and a high resolution counter value. The high resolution counter occupies the lower 7 bits of the timestamp. If you only care for the loop resolution then you will just do a right shift of 7 bits.

Let's use your earlier example where you show in below. Let's pick the 225792 value you get. If you divide this value by 128 you get 1764 loop clocks. Since each loop clock is 0.853uS as in your original setup you will get 1764*0.853 which is about 1504uS. This is close to what you were expecting.

37504
225792
224128
225664
225408
224384
225792
224512

Hi Jesse,

do you was able to read the PPM stream? Is the stream coming from a RC Receiver?

Another Question: Any idea on how to read the Futaba S.Bus stream? There are many infos for the STM32F MCUs including source code etc, but I'm looking for a solution on the Hercules MCU.

Regards,

-Jarno