Antenna delay Calibration


Im trying to calibrate my Decawave EVB1000 modules in an other frecuenci´s. But I dont undertand the algorithm describe in APS014.

1: (Page 7) [color=#555555][size=large][font=Roboto, arial, sans-serif] [/font][/size][/color]

The antenna delay can be determinde by minimizing the norm of the difference between the two EDM,s: What is the norm of the matrix? Is the absolut value of the result?

2: (Page 10)In the figure 5: Evaluating the quality of the candidates - flowchart

Foreach candidate: compute the time of flight matrix given the candidate delays --> Mathematimacal ecuation (Tof_candidate = [ 2Delay1 + 2Delay2 + 4*Tofmeasured]/4]).

I undertand, that I have to calculate the Tofcandidate using, the Delay candidate (that i have radomly set in the previous loop ) and Tofmeasured. But in this case, I have to use the same value for Delay1 and Delay2?

Thank´s and regards!

Hi , You may be aiming to produce a golden unit, hence the complexity of the formula.
If you just want to calibrate, as and example see description attached. Also the IC user manual will give you some info.


thanks for the reply. What’s the use of the golden unit calibration?

I’m trying to calculate chip delay for a new datarate. I have a Trek1000 module which has delays loaded to work on channel 5 at 6.8Mbps and 110kbps and I want to calculate the delay to work at 85kbps.

Thank´s and regards!

“Golden board” is basically a “source” for all other calibration to be done for example to calibrate the test setup as part of production.
This would also be described in our production test application note (APS012)

You have two basic options for delay calibration:

If you have a known good unit you can put the unit to calibrate a known distance away, measure the distance and then adjust the antenna delay until the measured distance matches the correct value.

The other option is if you don’t have a known good one then you need 3 units. You place them in a symmetrical triangle pattern so they are all at the same distance and same relative angle to each other. You then measure the distances between all three and find the combination of antenna delays that remove the errors. Since you have 3 unknown delays and three measurements this just comes down to a little bit of maths. The application note is simply one way of doing that maths.

Hi AndyA,

thanks for the reply. I dont have a referenc device for calibrete correctly my devices in 850kbps. For that I´m tryng to implement the algorithm that so you say is discribed in the “APS014: DW1000 Antenna Delay Calibration” user manual. My problem is that I dont undertand fully this algorithm, so can you try to help me with some questions please?

  1. In the page 9: Figure 4: Populate the set of candidate delay estimates - flowchart.

This algorithm is a iterative methor for generate radomly a set of numbers that you name Delay estimates. My question is, this numbers that I´m generating in this step are the itertation for the Del.Chip1, Del.Chip2, Del.Chip3 or is a random value for ToF candidate?

  1. In the page 10: Figure 5: Evaluate the quality of the candidates - flowchart.

If in the last step of “populate the set candidate…” I generate a set of Device delay´s, I dont understand the following setps.

The compute of TOF matrix given the candidate delays for each candidate? ToF candidate = [ 2Del.Chip1+ 2Del.Chip2 + 4ToFmeasured]/4
I have to use the same Del candidate to calculate the ToF candidate matrix? I dont see the meaning…

The next step is compute the norm of the difference between the ToFactual matrix and the ToF candidate matrix. What do you mean with the norm of the difference between both matrix, the absolute value of the subtraction of each value?