Prev: How to compute RMS in frequency domain
Next: Calculating MFCC
From: Tim Wescott on 24 Jan 2010 17:57
On Sun, 24 Jan 2010 15:59:03 -0600, Vladimir Vassilevsky wrote:

> Tim Wescott wrote:
>
>
>> Note that both of these books only cover analog phase lock loop
>> techniques -- you'll have to extend that to an all-digital loop,
>> although that's not too big of a stretch if you understand digital
>> control system design.
>
> You should be careful with digital PLLs. The phase detector is nonlinear
> thing, and it makes DPLL different from a common digital control
> system; especially when DPLL is in acquisition mode. Besides the obvious
> effects of quantization and aliasing, there are also not so obvious
> things like false locks or limit cycle behavior so the correct lock
> could never be acquired.

OTOH, depending on the signal you're trying to lock to there are some
more or less clever things you can do with the phase detector to improve
acquisition. Mind your P's and Q's as you point out, and you may find
that the result is far better than what you could easily do with an
analog circuit.

--
www.wescottdesign.com
From: eternal_nan on 25 Jan 2010 22:14
Hi guys,

I would be very interested in getting some details regarding the
"clever things" mentioned below. Could you provide any pointers for
reading material, books, papers, whatever. By the way, I own a copy of
Tim's book - unfortunately there were no PLL examples in there.

Thanks in advance,
Ljubisa

On Jan 24, 5:57 pm, Tim Wescott <t...(a)seemywebsite.com> wrote:
> On Sun, 24 Jan 2010 15:59:03 -0600, Vladimir Vassilevsky wrote:
> > Tim Wescott wrote:
>
> >> Note that both of these books only cover analog phase lock loop
> >> techniques -- you'll have to extend that to an all-digital loop,
> >> although that's not too big of a stretch if you understand digital
> >> control system design.
>
> > You should be careful with digital PLLs. The phase detector is nonlinear
> >   thing, and it makes DPLL different from a common digital control
> > system; especially when DPLL is in acquisition mode. Besides the obvious
> > effects of quantization and aliasing, there are also not so obvious
> > things like false locks or limit cycle behavior so the correct lock
> > could never be acquired.
>
> OTOH, depending on the signal you're trying to lock to there are some
> more or less clever things you can do with the phase detector to improve
> acquisition.  Mind your P's and Q's as you point out, and you may find
> that the result is far better than what you could easily do with an
> analog circuit.
>
> --www.wescottdesign.com

From: eternal_nan on 25 Jan 2010 23:12
Hi guys,

Just a follow-up. In particular, I am interested in how to implement
the transition between the non-linear acquisition mode and linear
phase detection mode if one implements a two step locking process:

1. Lock frequency based on a binary up/down frequency indicator.
2. Once frequency is locked switch to linear phase-only detection mode
(equivalent of an xor phase detector, suffers from problems with cycle-
slipping, this the initial frequency locking).

Thanks for any responses.
Ljubisa

On Jan 25, 10:14 pm, eternal_nan <eternal....(a)gmail.com> wrote:
> Hi guys,
>
> I would be very interested in getting some details regarding the
> "clever things" mentioned below. Could you provide any pointers for
> reading material, books, papers, whatever. By the way, I own a copy of
> Tim's book - unfortunately there were no PLL examples in there.
>
> Thanks in advance,
> Ljubisa
>
> On Jan 24, 5:57 pm, Tim Wescott <t...(a)seemywebsite.com> wrote:
>
>
>
> > On Sun, 24 Jan 2010 15:59:03 -0600, Vladimir Vassilevsky wrote:
> > > Tim Wescott wrote:
>
> > >> Note that both of these books only cover analog phase lock loop
> > >> techniques -- you'll have to extend that to an all-digital loop,
> > >> although that's not too big of a stretch if you understand digital
> > >> control system design.
>
> > > You should be careful with digital PLLs. The phase detector is nonlinear
> > >   thing, and it makes DPLL different from a common digital control
> > > system; especially when DPLL is in acquisition mode. Besides the obvious
> > > effects of quantization and aliasing, there are also not so obvious
> > > things like false locks or limit cycle behavior so the correct lock
> > > could never be acquired.
>
> > OTOH, depending on the signal you're trying to lock to there are some
> > more or less clever things you can do with the phase detector to improve
> > acquisition.  Mind your P's and Q's as you point out, and you may find
> > that the result is far better than what you could easily do with an
> > analog circuit.
>
> > --www.wescottdesign.com

From: Tim Wescott on 26 Jan 2010 01:20
On Mon, 25 Jan 2010 19:14:32 -0800, eternal_nan wrote:
(top posting fixed)
> On Jan 24, 5:57 pm, Tim Wescott <t...(a)seemywebsite.com> wrote:
>> On Sun, 24 Jan 2010 15:59:03 -0600, Vladimir Vassilevsky wrote:
>> > Tim Wescott wrote:
>>
>> >> Note that both of these books only cover analog phase lock loop
>> >> techniques -- you'll have to extend that to an all-digital loop,
>> >> although that's not too big of a stretch if you understand digital
>> >> control system design.
>>
>> > You should be careful with digital PLLs. The phase detector is
>> > nonlinear
>> >   thing, and it makes DPLL different from a common digital control
>> > system; especially when DPLL is in acquisition mode. Besides the
>> > obvious effects of quantization and aliasing, there are also not so
>> > obvious things like false locks or limit cycle behavior so the
>> > correct lock could never be acquired.
>>
>> OTOH, depending on the signal you're trying to lock to there are some
>> more or less clever things you can do with the phase detector to
>> improve acquisition.  Mind your P's and Q's as you point out, and you
>> may find that the result is far better than what you could easily do
>> with an analog circuit.
>>
> Hi guys,
>
> I would be very interested in getting some details regarding the "clever
> things" mentioned below. Could you provide any pointers for reading
> material, books, papers, whatever. By the way, I own a copy of Tim's
> book - unfortunately there were no PLL examples in there.
>
> Thanks in advance,
> Ljubisa
>

I need to write a DPLL primer; I haven't in part because there are so
many different applications, I haven't had time to categorize them and
sort them -- and that's not even addressing alternative synchronization
methods that make PLL's unnecessary.

If you understand the control theory part, getting a book or two on
analog PLL design is a huge help -- all the elements are there in both
digital and analog loops, so an analog PLL book is a great help.

Floyd Gardener's book comes well recommended. I like Dan Wolaver's book
"Phase Locked Loop Circuit Design", but I haven't gotten Gardener's book
to compare with it.

--
www.wescottdesign.com
From: Tim Wescott on 26 Jan 2010 01:22
On Mon, 25 Jan 2010 20:12:40 -0800, eternal_nan wrote:
(top posting fixed)
> On Jan 25, 10:14 pm, eternal_nan <eternal....(a)gmail.com> wrote:
>> Hi guys,
>>
>> I would be very interested in getting some details regarding the
>> "clever things" mentioned below. Could you provide any pointers for
>> reading material, books, papers, whatever. By the way, I own a copy of
>> Tim's book - unfortunately there were no PLL examples in there.
>>
>> Thanks in advance,
>> Ljubisa
>>
>> On Jan 24, 5:57 pm, Tim Wescott <t...(a)seemywebsite.com> wrote:
>>
>>
>>
>> > On Sun, 24 Jan 2010 15:59:03 -0600, Vladimir Vassilevsky wrote:
>> > > Tim Wescott wrote:
>>
>> > >> Note that both of these books only cover analog phase lock loop
>> > >> techniques -- you'll have to extend that to an all-digital loop,
>> > >> although that's not too big of a stretch if you understand digital
>> > >> control system design.
>>
>> > > You should be careful with digital PLLs. The phase detector is
>> > > nonlinear
>> > >   thing, and it makes DPLL different from a common digital control
>> > > system; especially when DPLL is in acquisition mode. Besides the
>> > > obvious effects of quantization and aliasing, there are also not so
>> > > obvious things like false locks or limit cycle behavior so the
>> > > correct lock could never be acquired.
>>
>> > OTOH, depending on the signal you're trying to lock to there are some
>> > more or less clever things you can do with the phase detector to
>> > improve acquisition.  Mind your P's and Q's as you point out, and you
>> > may find that the result is far better than what you could easily do
>> > with an analog circuit.
>>
>> > --www.wescottdesign.com
> Hi guys,
>
> Just a follow-up. In particular, I am interested in how to implement the
> transition between the non-linear acquisition mode and linear phase
> detection mode if one implements a two step locking process:
>
> 1. Lock frequency based on a binary up/down frequency indicator. 2. Once
> frequency is locked switch to linear phase-only detection mode
> (equivalent of an xor phase detector, suffers from problems with cycle-
> slipping, this the initial frequency locking).
>
> Thanks for any responses.
> Ljubisa
>

If your signal/noise ratio is high enough for the frequency indicator
it's pretty easy to detect cycle slips in an XOR-equivalent detector, and
use that for frequency detection.

--
www.wescottdesign.com
First  |  Prev  |  Next  |  Last
Pages: 1 2 3
Prev: How to compute RMS in frequency domain
Next: Calculating MFCC