Tuning the LDE to find correct RX timestamps

I have a case where anchor and the mobile tag are constantly measuring shorter distances then expected. During this time, LDE algorithm finds the first path of the signal at the index around 680. Even if I change NTM1 value to 31 which is the maximum, I can’t get rid of this. How can I force LDE algorithm to find the first path of the signal in the upper CIR region like as usual around 740-750?

Is this all the time or only in certain locations / buildings?

The only causes I’ve seen for a short distance are:

  1. poor antenna calibrations.
  2. Signals picked up on the PCB rather than the antenna, this requires the antenna to be on a cable and a strong signal level.
  3. Long lived reflections. These tend to be very environment and location specific, I’ve only seen them in very large buildings or long tunnels. Application note TN006: Effect of reflections from distant objects describes this effect. If this is happening then it’s normally fairly obvious if you look at the CIR data.

It’s happening in certain locations of certain buildings. 3rd cause that you listed might be the reason. If it’s so, how am I gonna overcome it?

That’s a tricky one, let me know if you come up with a solution. :slight_smile:

I put in a filter to reject any measurement where the leading edge was before a certain point in time. The exact location of the filter is configurable. I ran a few thousand cycles in varying conditions that didn’t show this effect, logged the leading edge locations and then set the default threshold at the point where it has a ~0.01% chance of rejecting a good range.
It’s not perfect, it doesn’t eliminate the chances of this happening and it risks rejecting perfectly good ranges. But it greatly reduces the odds of it happening, limits the impact it can have and has minimal impact on normal operation.

1 Like