DSMx telemetry

[jmentero]

Hi Mike,

In relation with the Lemon telemetry receiver, I have been trying to make sense of the noises that I hear from the Vario, but I hear chirps that do not correspond to what is happening, even with the receiver on the table.

In order to test stability, I enclosed a 4s Nicad battery and the receiver in a rigid sealed glass container, and recorded a raw telemetry file.
In every 0x40 record, Lemon is sending down six different data for Vspd, that according to Spektrum documentation correspond to data filtered (averaged ?, differentiated ?) using a time constant of 250ms, 500ms, 1s, 1.5s, 2s and 3s, trading accuracy for delay.
The data with a 250ms filter shows a poor stability. Peaks can be as high as +/- 1m/s, with a standard deviation of 0.38.
If I understand the code right, this is what ersky9x is using in the processDsmPacket function in frsky.cpp

On the other hand, data filtered with 1s, shows a much better behaviour, (max peak at 0.4m/s, and stdev of 0.15) at the expense of the corresponding delay, that I suppose, (haven't tried that yet) less annoying than chirps showing very fast changes that are in fact not happening.
Your point of view on this matter will be much appreciated.

Miguel

[MikeBland]

I recently received a Lemon DSM Rx with built in vario. I now need to find time to try this out and investigate the telemetry data.
If you have details of the 0x40 record to help me that would be helpful.

Mike

[jmentero]

Sure, whenever you can and feel like it, of course.
In relation with 0x40 ( and all other records), taken out of the Spektrum telemetry paper;

Spektrum paper

//////////////////////////////////////////////////////////////////
// VARIO-S
//////////////////////////////////////////////////////////////////
//
typedef struct
{
UINT8 identifier; // Source device = 0x40
UINT8 sID; // Secondary ID
INT16 altitude; // .1m increments
INT16 delta_0250ms, // delta last 250ms, 0.1m/s increments
delta_0500ms, // delta last 500ms, 0.1m/s increments
delta_1000ms, // delta last 1.0 seconds
delta_1500ms, // delta last 1.5 seconds
delta_2000ms, // delta last 2.0 seconds
delta_3000ms; // delta last 3.0 seconds
}
STRUCT_TELE_VARIO_S;

---

I think that ersky9x uses delta_0250ms.
From frsky.cpp around line 1210.

			if ( *packet == DSM_VARIO )
			{
				ivalue = (int16_t) ( (packet[4] << 8 ) | packet[5] ) ;
				storeTelemetryData( FR_VSPD, ivalue*10 ) ;
			}

If you could update a branch in github, I can build from there and test longer filters for the Vario.
Thank you.
Miguel

[MikeBland]

I'll be posting a test version soon that will include an option in the telemetry menu to choose which of the six, dsm vario values you wish to use.

Mike

[jmentero]

Thank you.
Just yesterday I saw a video of a Lemon receiver with Taranis. The behaviour of the vario when rising the model by hand at the end of the video (2.16sec), even with 1/sec chosen, is completely different than the unit I have at home, and I have tried with two Lemon units.
I am not talking about the tones, that are clearly different, but about the clear perception of rising.

If I raise the unit at home, I can only vaguely correlate the movement with the noise made. ¿Is this the case also with your receiver ?

Video

Miguel
Incidentally, do you prefer to discuss those matters on Openrcforums ?

[MikeBland]

Usually worth having this on Openrcforums as then others see the question and answers better.
I have put a quote from your first comment here on the forum, and added that I just posted an updated test version with an option to choose which filtered value to use for vspd.

My unit does seem to give a reasonable response when I raise it.

The video looks to be using a Spektrum Tx, not a Taranis.

Mike

[jmentero]

Thank you, and, true. it is a Spektrum Tx.
Follow up on Openrcforums.