What type of obstacles do you want your device to "see through". Some materials are transparent to infrared, others opaque. Do you want it to see through a human, for example? In general, the longer the wavelength the harder it is to block, but the more difficult it is to measure distances accurately.
Also, some objects might introduce a diffraction effect, causing spread in the arrival time of the signal.
I think you might need to do a bit of research to help clarify your requirements, especially when it comes to not being affected by objects in its path.
I wouldn't and you wouldn't, but the OP seems to want to. First cut would be to see it on an oscilloscope.
Years ago I had a HP rack mounted oscillator/clock that emitted a 1 second pulse. I connected that to my scope horizontal trigger and the vertical of the scope to a short wave receiver tuned to WWV. The one second pulse from WWV moved all around, time wise, due to propagation changes. Similar to what I proposed for the OP to try.
Paul
And therein lies a fundamental problem, you want to measure the distance between say a transmitter and receiver when there are objects in the way.
Radio waves are not like lasers. If there is an object in the way, the receiver will most likley still receive the signal but the signal will not arrive by a direct route but by a series of reflections which increase the distance travelled.
The Semtech 'sensor' that measures time of flight in this way does suffer from this problem, outdoors in a large open spaces distance readings tend to be fairly stable, in a typical street or urban area the plentiful reflections can cause significant variations in the 'distance'
Well it can be done, but its not trivial and requires special coding on the signal transmitted and relative phase measurements (rather like what GPS signals use). A standard RF transceiver is not really geared up to do this, and may not have enough modulation bandwidth (not a problem for optical frequencies).
The key insight is that if you can measure relative phase accurately then you can measure to a small fraction of a wavelength - but you also have to know which wave cycle, relative to some marker, and that marker needs to survive modulation without being smeared out too much in time (this pretty much requires a wide bandwidth and low-noise to achieve).
In the real world however multi-path reception renders this mostly useless for RF, so typically a real ToF approach is done using UWB (ultra-wideband) which permits very short modulation pulses. Optical is just another UWB approach.
GPS relies on line-of-sight reception to prevent multi-path interference corrupting the signal. It also uses a coding technique to greatly increase the effective signal/noise ratio.
So I can't measure the time taken to reach one rf module to second .
But if I do this action cyclic (I means multiple times)
Until I start to get time in micro second.
Let's assume 1000 times sending that rf signal takes 10 micro second
Then for distance [d = 1000 x speed of light x time in second (by deviding micro second to 10^6 ]
Somewhere in there is controller processing time.
For example.
1 start timer in master controller
2 Store timer val.
3 master controller send Tx command.
4 Tx transmits.
5 Rx receives.
6 Rx sends data to slave controller.
7 Slave controller sends Tx command.
8 Tx transmits.
9 Tx sends data to master.
10 Master checks how many cycles.
11 If less than 1000 cycles go back to step 3
12 If equal to 1000 then store timer value.
13 Master computes result.
Between and including step 5 to 8 and 11 back to 3 is possibly a SIGNIFICANT PART of your experiment.
How would you calculate that OUT or your results, your hardware is too slow.
Your hardware and code processing time would be adding microseconds alone.
Tom...
PS. Counting cycles until you get a microsecond result is also futile, your hardware is not that fast.
In the Semtech 'sensor' that does most of that processing in silicon, so you would think its well controlled, the small variances in processing time lead to the time through iterations varying by maybe 10-20nS. To get sub cm accuracy you would need to cut the variation to 10-20pS, a big ask.
There are some suggestions that the accuracy\resolution of the Semtech 'sensor' can be improved by reducing the frequency shift between master\slave, but I have yet to try it.
1000 pulses in 10µs means each pulse is 10ns in duration, required a modulation bandwidth of at least 1/10ns = 100MHz. You don't have a modulation bandwidth of 100MHz...
