sine wave as Sampling clock

HI all,

I am trying to explore the effect of using sine wave clock from sampling
instead of conventional square wave. I understand that the major problem
will be due to slow rise time of sine wave as compared to square wave but
at high frequency sine wave still has sufficiently good rise time. Or may
be we can make the slew rate of both sine and square wave same.

All this apart, i have following questions:
1) at high frequency sampling clock , so we conventionally use sine wave
or square wave? As i feel , at high frequency sine wave is same as square
wave.

2)what is the state of the art technique of sampling for high frequency?

thanks

STUPIDENT

dkumar wrote:

> HI all,
> 
> I am trying to explore the effect of using sine wave clock from sampling
> instead of conventional square wave. I understand that the major problem
> will be due to slow rise time of sine wave as compared to square wave but
> at high frequency sine wave still has sufficiently good rise time. Or may
> be we can make the slew rate of both sine and square wave same.
> 
> All this apart, i have following questions:
> 1) at high frequency sampling clock , so we conventionally use sine wave
> or square wave? As i feel , at high frequency sine wave is same as square
> wave.
> 
> 2)what is the state of the art technique of sampling for high frequency?
> 
> thanks
> 
> 

i understand this is a bad idea. but my adviser is hell bend on doing this.

so, if you could tell me some drawbacks of doing this then may be i can
convince him. 

the idea actually goes like this:

In conventional sample and hold with bottom plate sampling, we use a sine
wave to clock the bottom plate and this sine wave is at high frequency (2G
around) and the top plate is clocked with low frequency square wave of say
1G with the input signal begin below nyquist. Now as the bottom plate is
sampled using sine wave, it inherently  has low jitter as compared to a
square wave of same frequency. So if the held samples of bottom plate are
taken at the instant with the top plate cuts off, that would mean we are
still sampling at 1G and the effect of jitter in 1G clock will not matter
as it is sampling the held signal. Since sine wave has less jitter, the
total output will have better performance from the jitter point of view.


thanks

Please.
thanks
>STUPIDENT
>
>dkumar wrote:
>
>> HI all,
>> 
>> I am trying to explore the effect of using sine wave clock from
sampling
>> instead of conventional square wave. I understand that the major
problem
>> will be due to slow rise time of sine wave as compared to square wave
but
>> at high frequency sine wave still has sufficiently good rise time. Or
may
>> be we can make the slew rate of both sine and square wave same.
>> 
>> All this apart, i have following questions:
>> 1) at high frequency sampling clock , so we conventionally use sine
wave
>> or square wave? As i feel , at high frequency sine wave is same as
square
>> wave.
>> 
>> 2)what is the state of the art technique of sampling for high
frequency?
>> 
>> thanks
>> 
>> 
>

dkumar wrote:
> i understand this is a bad idea. but my adviser is hell bend on doing this.
> 
> so, if you could tell me some drawbacks of doing this then may be i can
> convince him. 
> 
> the idea actually goes like this:
> 
> In conventional sample and hold with bottom plate sampling, we use a sine
> wave to clock the bottom plate and this sine wave is at high frequency (2G
> around) and the top plate is clocked with low frequency square wave of say
> 1G with the input signal begin below nyquist. Now as the bottom plate is
> sampled using sine wave, it inherently  has low jitter as compared to a
> square wave of same frequency. So if the held samples of bottom plate are
> taken at the instant with the top plate cuts off, that would mean we are
> still sampling at 1G and the effect of jitter in 1G clock will not matter
> as it is sampling the held signal. Since sine wave has less jitter, the
> total output will have better performance from the jitter point of view.

Bottom plate? Top plate? What are you talking about?

Why do you believe that locating the zero crossings of a sine wave has 
"inherently ... low jitter as compared to a square wave of same 
frequency"? The jitter than matters it the uncertainty of the instant 
that the hold begins.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Just take Vlad with a grain of salt -- admittedly a 10^large grain

Side question to V.V. --- Did you *NEVER* ask a naive question?

Side comment to a fellow former Boston resident - no not off my meds ;/

On Apr 20, 9:40 am, Vladimir Vassilevsky <antispam_bo...@hotmail.com>
wrote:
> STUPIDENT
>
> dkumar wrote:
> > HI all,
>
> > I am trying to explore the effect of using sine wave clock from sampling
> > instead of conventional square wave. I understand that the major problem
> > will be due to slow rise time of sine wave as compared to square wave but
> > at high frequency sine wave still has sufficiently good rise time. Or may
> > be we can make the slew rate of both sine and square wave same.
>
> > All this apart, i have following questions:
> > 1) at high frequency sampling clock , so we conventionally use sine wave
> > or square wave? As i feel , at high frequency sine wave is same as square
> > wave.
>
> > 2)what is the state of the art technique of sampling for high frequency?
>
> > thanks

You should ignore anything that VV says.  He has a reputation for
being rude, obnoxious and often adding nothing to a thread.

Has you adviser explained why he wants to clock with a sine wave?
Also, what is being clocked?  I have seen some ADC chips that use
complementary sine waves as clocks rather than digital clocks.  I
assume this has something to do with noise since a sine wave has no
high harmonic content like a square wave does.  For an ADC, I would
expect jitter to be a bigger problem, but square clocks are no panacea
for jitter either.

I'd like to hear your adviser's rational for sine wave clocking.

Rick

On Apr 20, 2:34&#4294967295;pm, "dkumar" <joi...@gmail.com> wrote:
> i understand this is a bad idea. but my adviser is hell bend on doing this.
>
> so, if you could tell me some drawbacks of doing this then may be i can
> convince him.
>
> the idea actually goes like this:
>
> In conventional sample and hold with bottom plate sampling, we use a sine
> wave to clock the bottom plate and this sine wave is at high frequency (2G
> around) and the top plate is clocked with low frequency square wave of say
> 1G with the input signal begin below nyquist. Now as the bottom plate is
> sampled using sine wave, it inherently &#4294967295;has low jitter as compared to a
> square wave of same frequency. So if the held samples of bottom plate are
> taken at the instant with the top plate cuts off, that would mean we are
> still sampling at 1G and the effect of jitter in 1G clock will not matter
> as it is sampling the held signal. Since sine wave has less jitter, the
> total output will have better performance from the jitter point of view.
>
> thanks
>
> Please.
> thanks
>
> >STUPIDENT
>
> >dkumar wrote:
>
> >> HI all,
>
> >> I am trying to explore the effect of using sine wave clock from
> sampling
> >> instead of conventional square wave. I understand that the major
> problem
> >> will be due to slow rise time of sine wave as compared to square wave
> but
> >> at high frequency sine wave still has sufficiently good rise time. Or
> may
> >> be we can make the slew rate of both sine and square wave same.
>
> >> All this apart, i have following questions:
> >> 1) at high frequency sampling clock , so we conventionally use sine
> wave
> >> or square wave? As i feel , at high frequency sine wave is same as
> square
> >> wave.
>
> >> 2)what is the state of the art technique of sampling for high
> frequency?
>
> >> thanks

Why does a sine wave have less jitter than a square wave?  A sine wave
has a slow rise time compared to a square wave.  So there is a wider
time window in which a bit of noise will cause jitter in the trigger
point.

However, with 1 and 2 GHz square clocks, they will look a lot like
sine waves anyway.  At 1 GHz you will need to pay careful attention to
the board layout and SI issues.  At 2 GHz you will need to treat this
as an RF design.

Do you have much experience with digital signals at these high
frequencies?

Rick

rickman <gnuarm@gmail.com> wrote:
(snip)
 
> Why does a sine wave have less jitter than a square wave?  A sine wave
> has a slow rise time compared to a square wave.  So there is a wider
> time window in which a bit of noise will cause jitter in the trigger
> point.
 
> However, with 1 and 2 GHz square clocks, they will look a lot like
> sine waves anyway.  At 1 GHz you will need to pay careful attention to
> the board layout and SI issues.  At 2 GHz you will need to treat this
> as an RF design.

It will not be easy to tell a 2GHz sine wave from what otherwise
might be a 2GHz square wave.   A 2GHz sine wave does not have
a slow rise time on the scale of most circuits today.

If you have an unknown amount of second harmonic in the sine then
I suppose it could be called jitter.  

-- glen

dkumar wrote:
> HI all,
> 
> I am trying to explore the effect of using sine wave clock from sampling
> instead of conventional square wave. I understand that the major problem
> will be due to slow rise time of sine wave as compared to square wave but
> at high frequency sine wave still has sufficiently good rise time. Or may
> be we can make the slew rate of both sine and square wave same.
> 
> All this apart, i have following questions:
> 1) at high frequency sampling clock , so we conventionally use sine wave
> or square wave? As i feel , at high frequency sine wave is same as square
> wave.
> 
> 2)what is the state of the art technique of sampling for high frequency?
> 
> thanks
> 
> 
You need a reply without the insult appended.

0) By definition a true square wave has an infinite slew rate.  What 
passes for "square" gets awfully round at high frequencies, though.

1) At very high frequencies it's difficult to maintain squareness to a 
signal, because of all the higher-yet harmonics that must be maintained.

2) I dunno.  But the basic principal that you can't lose sight of is in 
the name -- sampling.  The longer the on-time of the switch compared to 
the off-time the less sampling you're doing and the more mixing.  At 
some point (possibly when all you have available is sine waves) you 
should probably consider downconversion with a regular ol' mixer, 
followed by real sampling, if necessary.

2a) Generating a squared-up sampling clock in a _very_ local part of 
your board from a sine wave may make sense, if you've got some really 
good board to use (i.e. not FR-4 at the GHz frequencies you mention 
elsewhere) and power to burn in the conversion.

-- 

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" was written for you.
See details at http://www.wescottdesign.com/actfes/actfes.html

On Apr 20, 4:21&#4294967295;pm, glen herrmannsfeldt <g...@ugcs.caltech.edu> wrote:
> rickman <gnu...@gmail.com> wrote:
>
> (snip)
>
> > Why does a sine wave have less jitter than a square wave? &#4294967295;A sine wave
> > has a slow rise time compared to a square wave. &#4294967295;So there is a wider
> > time window in which a bit of noise will cause jitter in the trigger
> > point.
> > However, with 1 and 2 GHz square clocks, they will look a lot like
> > sine waves anyway. &#4294967295;At 1 GHz you will need to pay careful attention to
> > the board layout and SI issues. &#4294967295;At 2 GHz you will need to treat this
> > as an RF design.
>
> It will not be easy to tell a 2GHz sine wave from what otherwise
> might be a 2GHz square wave. &#4294967295; A 2GHz sine wave does not have
> a slow rise time on the scale of most circuits today.
>
> If you have an unknown amount of second harmonic in the sine then
> I suppose it could be called jitter. &#4294967295;
>
> -- glen

Second harmonic will not cause jitter.  Jitter can be caused by other
signals that are *not* harmonic.  The shape of the signal is not
important until the harmonic is so strong that it causes multiple
transitions through the threshold of the receiver.  Even then that is
not really jitter, that is distortion.

Rick