I came across what you wrote while trying to find some test equipment for myself. I'm working with a fairly similar circumstance myself. I'm working with standard definition SDI signals, so I already have digital samples and while I'd really like to have something better to work with than a 13.5Mhz sampling rate, it's what I have and I've had to make due with it. What you'll probably want to do is either to sample at 27Mhz (if you can) or 13.5Mhz and upsample. The problem you'll constantly have is that teletext data ... roams quite a bit. The sync word can be clipped at the start of the line, so you probably don't want to test for sync on anything more than 0xAAE4 as opposed to the full 0xAAAAE4. Also, there's such a thing as inverted teletext which has 0x55551D as the sync word instead. If you're really interested in the raw data, then you'll want to get a 27Mhz signal or upsample from 13.5 to 27Mhz. You don't need a clean FIR, all you need is a clean enough low pass so that you at least compensate for the source signal being close enough to the source signal that you get around 1.9 readings per sample at 13.5Mhz. Here are the coefficients you would want to use to upsample : Coefficient Original value Scaled Value Fixed Point Value 1 0.0750243 0.07485173 153 2 0.159151 0.15899877 325 3 0.225074 0.22504441 460 4 0.249995 0.249995 511 5 0.225074 0.22504441 460 6 0.159151 0.15899877 325 7 0.0750243 0.07485173 153 They're more than good enough. I know, I just used them. I'm doing mine in software, so I've taken a bit of a round about trip to handle it, but if you upsample in hardware to 27Mhz using the FIR, then you'll receive a 1440 samples per line. You can also perform threshold testing in the FPGA giving you 1440 bits per line instead which is much more manageable. Then you have 4 possible phases to start from. Starting at bit 0, 1, 2, and 3, attempt to find the sync word (0xAAE4). You'll need to use a fractional skip between bits of 13.5/6.9375. I've done this using 32-bit integers with 16 bits for integer and 16 bits for fraction. It works quite well. Once you locate the sync word, you can decode the rest of the line. You might find the sync word in up to 3 phases, in this circumstance, choose the middle phase. In the case of finding it only on 2 phases, then your guess is as good as mine. Anyway... let me know if you have any questions. - Darren R. Starr -- View this message in context: http://old.nabble.com/Fast-locking-PLL-with-long-%22free-run%22-time-tp26932843p28816398.html Sent from the PIC - [EE] mailing list archive at Nabble.com. -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist