Another logic problem.

[Yzordderex]

I need to generate two square waves while being able to control the
phase relationship between them. Square waves at 4Mhz. Ideally I
would like to be able to control phase 0-180 degrees, but would settle
for 5-175. I need to vary at audio rate, dc to 8kHz. A phase
modulator if you will. Can a phase lock loop be configured to do the
job? I've got a function generator which will do 10Mhz to aid in
development. If not PLL, then some other off shelf component would be
desired. I have access to coworker very knowledgable with dsp, but I
would like to keep it simple.

thanks
Bob

 

[Bill Sloman]

I need to generate two square waves while being able to control the
phase relationship between them. Square waves at 4Mhz. Ideally I
would like to be able to control phase 0-180 degrees, but would settle
for 5-175. I need to vary at audio rate, dc to 8kHz. A phase
modulator if you will. Can a phase lock loop be configured to do the
job? I've got a function generator which will do 10Mhz to aid in
development. If not PLL, then some other off shelf component would be
desired. I have access to coworker very knowledgable with dsp, but I
would like to keep it simple.

A pair of Analog Devices AD9954 synchronisable DDS chips -

http://www.analog.com/UploadedFiles/Data_Sheets/31685232434714AD9954_0.pdf

would do the trick.

Their AD9854 quadrature synthesiser would produce two sine waves with
a controllable frequency difference from one package, but IIRR you'd
have to supply your own comparators to convert them to square waves

http://www.analog.com/UploadedFiles/Data_Sheets/87202050AD9854_b.pdf

You can do it more directly by dividing down a faster clock in two
parallel synchronous dividers and using a third counter to set the
phase difference between the two dividers.

Farnell stocks 120MHz crystal-controlled oscillators (but they aren't
all that cheap) which you could divide by thirty to get your 4MHz
outputs with selectable phase differences in steps of 12 degrees.
You'd have to build the counters in fast programmable logic devices -
standard TTL-compatible synchronous dividers seem to be limited to
frequencies below 50MHz.

You can buy crystal controlled oscillators that will generate
frequencies up to about 600MHz, as balanced ECL-level signals, and
Motorola ECLinPS offers six and eight bit synchronous counters that
are guaranteed to work with clocks up to these sorts of frequencies,
which could get your phase steps down to 3 degrees.

The MC100E195 or MC100E196 programmable delays could be used to take
this a little further, but the DDS chips look to be a lot easier to
use.