|
Seuss
Error: divide byzero



Registered: 04/27/01
Posts: 23,480
Loc: Caribbean
Last seen: 3 months, 8 days
|
Frequency data analysis
#8545173 - 06/20/08 05:09 PM (15 years, 10 months ago) |
|
|
I am trying to find a way to analyze some data. More or less, I have a wire that I have tapped into. The wire carries groups of pulses followed by a pause, followed by another group of pulses, and so on. I have sampled the data, and need to write a program that can count the number of pulses in a group. Not a problem. Unfortunately, the complex part follows...
It turns out that there are two wires running close together and there is a lot of crosstalk from one of the wires to the other. There is nothing I can do about this, the problem domain is fixed. The main thing in common is that both wires get groups of pulses at about the same time. Because of the crosstalk, I tend to pick up both signals on the signal I am interested in. This causes my pulse count to double or more.
Even worse, the signals do not carry the exact same data and are not in phase or have a common frequency. The period of the start time drifts as do the number of pulses in a group as do the periods of the pulses in a group.
I can tap into both wires. Again, ignore the physics as I am simplifying the domain. It is the analysis of the data that is important to me. I have no idea where to begin... (well, I have an idea or two, but they are weak...) Any ideas are greatly appreciated!
|
entheomandotcom
Stranger

Registered: 03/28/08
Posts: 282
Last seen: 15 years, 4 months
|
Re: Frequency data analysis [Re: Seuss]
#8545304 - 06/20/08 06:23 PM (15 years, 10 months ago) |
|
|
Are both wires operating on the same frequency,,im thinking something like a dsl filter,,
|
trendal
J♠



Registered: 04/17/01
Posts: 20,815
Loc: Ontario, Canada
|
Re: Frequency data analysis [Re: Seuss]
#8545352 - 06/20/08 06:48 PM (15 years, 10 months ago) |
|
|
Even worse, the signals do not carry the exact same data and are not in phase or have a common frequency.
That's the part that gets me 
About the crosstalk...it's not at the same voltage as the actual signal, right? Is there any way you could use the difference in voltage?
--------------------
Once, men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them.
|
entheomandotcom
Stranger

Registered: 03/28/08
Posts: 282
Last seen: 15 years, 4 months
|
Re: Frequency data analysis [Re: trendal]
#8545361 - 06/20/08 06:53 PM (15 years, 10 months ago) |
|
|
Sorry missed that,,
|
Diploid
Cuban



Registered: 01/09/03
Posts: 19,274
Loc: Rabbit Hole
|
Re: Frequency data analysis [Re: Seuss]
#8545415 - 06/20/08 07:13 PM (15 years, 10 months ago) |
|
|
More info: what's the nature of these pulses? Are they analog or digital? Sharp voltage ramp envelope or gentle?
Generally, the cross-talk signal will be weaker than the primary signal. If so, you might be able to use a filter to reject the weaker pulses?
Or... if you sample fast enough, you might be able to apply some DSP to invert the cross-talk signal (sampled from the other wire) and add it to the (cross-talk + primary) signal on the wire you're interested in. If it's analog, an op-amp has an inverting input you can use in the same way to get the difference as Trendal suggested.
-------------------- Republican Values: 1) You can't get married to your spouse who is the same sex as you. 2) You can't have an abortion no matter how much you don't want a child. 3) You can't have a certain plant in your possession or you'll get locked up with a rapist and a murderer. 4) We need a smaller, less-intrusive government.
|
Seuss
Error: divide byzero



Registered: 04/27/01
Posts: 23,480
Loc: Caribbean
Last seen: 3 months, 8 days
|
Re: Frequency data analysis [Re: Diploid]
#8546510 - 06/21/08 04:51 AM (15 years, 10 months ago) |
|
|
> what's the nature of these pulses? Are they analog or digital?
Analog. Sample rate is 10KHz, fixed.
> If it's analog, an op-amp has an inverting input you can use in the same way to get the difference as Trendal suggested.
Everything I do has to be at the data level, as I am working with the data that somebody else already gathered.
> you might be able to apply some DSP to invert the cross-talk signal (sampled from the other wire) and add it to the (cross-talk + primary) signal on the wire you're interested in.
Interesting idea. Close to what I was looking at doing. (I was thinking along the lines of a FFT, drop the lows, inverse FFT) ... but a subtact... hmmm...
-------------------- Just another spore in the wind.
|
|