RDS "PI Mismatches"

[Russ-PA]

Here's a screen shot from of the now-infamous 'EEEE' from WTNT-94.9-FL, which is from 6/5.

[Russ-PA]

Here's WJTG-91.3-GA from 7/7. PI transmitted decodes to "KCRL".

[Sahil]

Hi Russ,

I am working on RBDS project from few months and I have been facing the PI mismatch problem. I received a signal from 104.3FM (WOW, Boise, Idaho) and decoded it using RBDS document published in 2005(http://www.nrscstandards.org/SG.asp). The expected call sign is KAWO while I'm getting KAVI (the problem is 14bits of PI match correctly while 5th and 6th bit are inverted, which gives me KAVI as new call sign). So I'm wondering if I'm making some mistake while decoding the RBDS signal or if NRSC has changed any call-sign recently?? If not, I need to prove/state the reason for why I'm getting a different PI and hence a different Call sign. So can you please give me your valuable comments on this. Your help will be highly appreciated.

Thanks,
Sahil.

[rdees]

Hi Sahil:

You will find many errors in PI codes in the US. It is common practice for stations owned by Clear Channel to change the first digit (on the left) of their PI to a 1 (and sometimes the second digit also). Here are two examples of stations in Atlanta, Georgia:

WUBL should be 899D and WKLS should be A740. But you can see from the pictures that both have changed the first digit of their PI codes to a 1.

This causes a problem because the PI code will still map to a valid "K" letter call sign in these two cases.

Hope this helps,

Ray

[k6sti]

The following DOS utility converts callsigns to PI codes and vice versa. It provides all possible callsigns for Clear Channel PI codes. It works for U.S. or full-power Canadian callsigns:

http://ham-radio.com/k6sti/pi.exe

Brian

[Russ-PA]

Hello Sahil !

I see that you've gotten some responses already. KAWO is not owned by Clear Channel, so is not subject to their revisions to the normal PI codes.

You might also want to check this site, operated by another of our members:

http://www.w9wi.com/articles/rdsreverse.htm

By entering KAWO, you get two results, one decimal, the other in hex. Since I don't know what you started with nor do I comprehend the technical details of how you're converting ( I use a pre-made converter/decoder ), this may or may not be of help in this instance.

[Sahil]

Hi Ray,

Thanks for the help. The issue though is that - the call sign I am expecting is KAWO (PI - 142A), so there is already 1 present on the first hex digit (leftmost digit) of PI code. Also when I decode entire message into bits and check for PI, I'm getting 122A which also has 1 on the left. So if we check both the PIs closely, we can see that only 6th and 7th bit is interchanged/inverted.

Expected PI(binary): 0001 0100 0010 1010 [CS- KAWO]
PI decoded from the message: 0001 0010 0010 1010 [CS- KAVI]

So, I'm wondering about the possibilities of getting incorrect PI. Is PI code entered manually by the engineer at the transmitter while tuning it for the first time?? If yes, I am guessing that he might have mistyped 1 at the 6th bit to the 7th bit.

Brian,
I checked your PI calculator again (http://ham-radio.com/k6sti/pi.exe) but it seems that it is not compatible with my OS - Windows7. Thanks for the link though. I will try to check it on my friend's system.

Russ,
I am using the same PI calculator for checking PI code and call signs. Also I cross checked it with hand calculations as per the procedure defined in the document - United Stated RBDS Standards, 2005 (page: 84-90).

So can you guys please give your expert views about the possibilities of getting different PI or is it just by human error as I guessed in the second para??

Thaks,
Sahil.

[Russ-PA]

Certainly human error is possible - we've all seen it many times. Sometimes it's mis-programmed, other times the engineer never programs it ( may not know how, may not care, etc ) so that either the result is nonsense or mimics some other real station ( sometimes because the unit was bought used or exchanged between stations ).

I can't speak to the possibility of it being something not quite right in your setup as it isn't something I've ever tried to do.

[rdees]

Sahil:

How/where are you getting the PI code?

[Sahil]

Thanks Russ..

Ray,

I received FM signal on my radio receiver and saved it on my computer using Agilent-VSA(VEctor Signal Analyzer), then I used Matlab to decode it into bits and used 'Xcorr' function in Matlab which overlaps the sequence of bits with the complete bitstream. I checked it by overlapping 16 bit PI code for call-sign KAWO, but the results showed me that only 14 bits are matching. So I kept on trying by changing single bit at a time from the PI code, finally it gave me new PI code (6th and 7th bit interchanged/inverted) which matches perfectly with 16bit sequence in bitstream. So I checked the corresponding call sign which gives me KAVI.

but if I check the next 10 bits to PI code - Checkword, it doesn't match with the one I calculated by using new PI code(for call sign KAVI) or even for expected PI code(for CS: KAWO). I used the procedure defined in RBDS document to calculate this checkword, but only 7 bits are matching instead of 10 :(. Is there any software/program to directly calculate checkwords (like we have one for calculating PI)?? Or can anybody please check/calculate the checkwords for call signs KAWO and KAVI, so I can see if I am missing something in my calculation??

Thanks.

[rdees]

Sahil:

Don't know of a program for calculating the 10 bit checkword but I get these:

KAWO >> 124A >> 1011110101
KAVI >> 122A >> 0000000101

Hope it helps..........

PS: These are the raw checkwords, add your offset!

[Sahil]

Thanks Ray..Thats exactly what I'm getting for my raw checkwords..This gives me the confidence on my calculations though..I have written a matlab code which generates raw checkword as well as corresponding final checkwords after adding offset words. I can't attach .m file due to format restrictions of attachment. I am attaching the word file though, if anyone wants to check, please copy it m-file and run the matlab code.