Ultra low frequency VCO
Daku
available reference design for an ultra low frequency voltage
controlled oscillator ? I developed a 60 Hz center frequency SPICE
model using common op-amps. I was wondering if there are any reference
designs out there.
Any feedback would be very useful. Thanks in advance for your help.
Tim Williams
Replace 25k pot with control voltage input.
In a single range, this thing will go from 1Hz to 4.5MHz. That was with Ct = 680pF. Put in, say, 0.1uF and it will gladly do milihertz.
Tim
--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms
"Daku" <daku...@gmail.com> wrote in message news:0e2f803e-bc15-4688...@k8g2000prh.googlegroups.com...
Jasen Betts
Spehro Pefhany
There's one on the LM324 datasheet that will be happy enough at 60Hz,
plus it only uses 10-15 cents worth of parts.
Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
sp...@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
terryS
wrote:
> On Mon, 5 Jul 2010 00:01:47 -0700 (PDT), the renowned Daku
>
Unsophisticated and fixed one could use three components?
From a DC (or rectified AC) source of more tha 100 volts.
A resistor, a neon and a capacitor.
Size of capacitor determines rate.
I think it used to be called a 'Relaxation oscillator'?
Tim Wescott
I'd hardly call 60Hz "ultra low frequency". But it is pretty darned low.
All the suggestions you've gotten so far are good as far as they go and
may well be perfect -- but what are you trying to do? Do you need sine
wave out or square? If sine wave, how pure? Do you have any
specifications on jitter, phase noise, or frequency accuracy?
You could digitally synthesize a 60Hz sine wave with a smallish
processor -- I believe there are some TMS430 parts that could do it all
in one package with a PWM output to be filtered by a simple RC.
But "best" depends heavily on what you want.
--
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
Sjouke Burry
Buy a sine/cosine potentiometer, and put a motordrive on it
with a reduction gear .
Motor speed and gear will determine frequency, and supply voltage
to the pot. will control amplitude.
Or use a computerchip with ad channel to produce any signal you want.
With a 16-24 bit audio ad chip it should work very well.
Paul Keinanen
wrote:
>On 07/05/2010 12:01 AM, Daku wrote:
>> Could some electronics guru please help ? Is there any commonly
>> available reference design for an ultra low frequency voltage
>> controlled oscillator ? I developed a 60 Hz center frequency SPICE
>> model using common op-amps. I was wondering if there are any reference
>> designs out there.
>> Any feedback would be very useful. Thanks in advance for your help.
>
>I'd hardly call 60Hz "ultra low frequency". But it is pretty darned low.
At least for RF signals in free space according to ITU-R
MF 300 kHz - 3 MHz
LF 30 kHz - 300 kHz
VLF 3 kHz - 30 kHz
ULF 300 Hz - 3 kHz
SLF 30 Hz - 300 Hz
Thus 60 Hz would be Super Low Frequency.
In different contexts ultra low frequency might refer to something
with one cycle every second, hour or even weeks.
>
>All the suggestions you've gotten so far are good as far as they go and
>may well be perfect -- but what are you trying to do? Do you need sine
>wave out or square? If sine wave, how pure? Do you have any
>specifications on jitter, phase noise, or frequency accuracy?
>
>You could digitally synthesize a 60Hz sine wave with a smallish
>processor -- I believe there are some TMS430 parts that could do it all
>in one package with a PWM output to be filtered by a simple RC.
>
>But "best" depends heavily on what you want.
While very low frequencies could in principle be generated with
traditional RLC components, the physical dimensions would become huge
and in practice capacitor leakage resistance and coil series
resistance would finally limit how low you could go.
However, a digital circuit followed by a DAC could go as low as you
want, however common audio DAC frequency response drops at about 3 Hz,
limiting how low you can go with such components.
Tim Wescott
> On Mon, 05 Jul 2010 08:59:33 -0700, Tim Wescott<t...@seemywebsite.com>
> wrote:
>
>> On 07/05/2010 12:01 AM, Daku wrote:
>>> Could some electronics guru please help ? Is there any commonly
>>> available reference design for an ultra low frequency voltage
>>> controlled oscillator ? I developed a 60 Hz center frequency SPICE
>>> model using common op-amps. I was wondering if there are any reference
>>> designs out there.
>>> Any feedback would be very useful. Thanks in advance for your help.
>>
>> I'd hardly call 60Hz "ultra low frequency". But it is pretty darned low.
>
> At least for RF signals in free space according to ITU-R
>
> MF 300 kHz - 3 MHz
> LF 30 kHz - 300 kHz
> VLF 3 kHz - 30 kHz
> ULF 300 Hz - 3 kHz
> SLF 30 Hz - 300 Hz
>
> Thus 60 Hz would be Super Low Frequency.
>
> In different contexts ultra low frequency might refer to something
> with one cycle every second, hour or even weeks.
From a "needs huge analog components to work" it's not ultra low to me
-- but yes, from an RF standpoint it's below Ultra Low.
>> All the suggestions you've gotten so far are good as far as they go and
>> may well be perfect -- but what are you trying to do? Do you need sine
>> wave out or square? If sine wave, how pure? Do you have any
>> specifications on jitter, phase noise, or frequency accuracy?
>>
>> You could digitally synthesize a 60Hz sine wave with a smallish
>> processor -- I believe there are some TMS430 parts that could do it all
>> in one package with a PWM output to be filtered by a simple RC.
>>
>> But "best" depends heavily on what you want.
>
> While very low frequencies could in principle be generated with
> traditional RLC components, the physical dimensions would become huge
> and in practice capacitor leakage resistance and coil series
> resistance would finally limit how low you could go.
>
> However, a digital circuit followed by a DAC could go as low as you
> want, however common audio DAC frequency response drops at about 3 Hz,
> limiting how low you can go with such components.
I'd just use a general-purpose DAC for this, if I didn't PWM-and-filter.
But, how far this needs to go depends on what the OP really needs, and
he hasn't weighed in yet.
Robert Baer
RELAX and see it OSCILLATE..
Daku
> I'd hardly call 60Hz "ultra low frequency". But it is pretty darned low.
>
> All the suggestions you've gotten so far are good as far as they go and
> may well be perfect -- but what are you trying to do? Do you need sine
> wave out or square? If sine wave, how pure? Do you have any
> specifications on jitter, phase noise, or frequency accuracy?
grid.
I am concerned with the VCO as it is a crucial sub-circuit. I am
aiming for
a phase noise of approximately -100 dBc/Hz but not very sure of the
offset
frequency. Ideally, I would like to have frequency accuracy of 1 - 5%
at most.
Also, I am aware that S-parameter methods are not appropriate at these
low
frequencies.
>
> You could digitally synthesize a 60Hz sine wave with a smallish
> processor -- I believe there are some TMS430 parts that could do it all
> in one package with a PWM output to be filtered by a simple RC.
>
> But "best" depends heavily on what you want.
>
> --
>
> Tim Wescott
> Wescott Design Serviceshttp://www.wescottdesign.com
Tim Wescott
> On Jul 5, 8:59 pm, Tim Wescott<t...@seemywebsite.com> wrote:
>> I'd hardly call 60Hz "ultra low frequency". But it is pretty darned low.
>>
>> All the suggestions you've gotten so far are good as far as they go and
>> may well be perfect -- but what are you trying to do? Do you need sine
>> wave out or square? If sine wave, how pure? Do you have any
>> specifications on jitter, phase noise, or frequency accuracy?
> I am trying to design a PLL for very low frequencies, e.g., power line
> grid.
> I am concerned with the VCO as it is a crucial sub-circuit. I am
> aiming for
> a phase noise of approximately -100 dBc/Hz but not very sure of the
> offset
> frequency. Ideally, I would like to have frequency accuracy of 1 - 5%
> at most.
> Also, I am aware that S-parameter methods are not appropriate at these
> low
> frequencies.
I think that those specs would be difficult to achieve with an
all-analog oscillator running at 60Hz. Not impossible -- I could do it,
and Joerg could do it in a fraction of the time I'd take. Using some
sort of direct digital synthesis -- even if it's just a microprocessor
-- running off of a crystal reference would be almost trivial in
comparison and would probably take less board space and would be far
more repeatable in manufacturing.
If you just had to do this purely in the analog domain your best bet
might be a pair of crystal oscillators, frequency steered with
varactors, carefully built, and with their outputs mixed down to 60Hz.
But that's a solution I would expect to see in a bit of kit from the
50's through the 80's -- anything later and I'd expect to see a DDS.
I think in another response I mentioned a TMS430 -- use one of those (or
a PIC on an AVR or a Stellaris, etc.) with the right ADC, and you can
build the whole PLL application into the software, and probably whatever
measuring you're planning on doing as well.
George Herold
> On Jul 5, 8:59 pm, Tim Wescott <t...@seemywebsite.com> wrote:> I'd hardly call 60Hz "ultra low frequency". But it is pretty darned low.
>
> > All the suggestions you've gotten so far are good as far as they go and
> > may well be perfect -- but what are you trying to do? Do you need sine
> > wave out or square? If sine wave, how pure? Do you have any
> > specifications on jitter, phase noise, or frequency accuracy?
>
> I am trying to design a PLL for very low frequencies, e.g., power line
> grid.
> I am concerned with the VCO as it is a crucial sub-circuit. I am
> aiming for
> a phase noise of approximately -100 dBc/Hz but not very sure of the
> offset
> frequency. Ideally, I would like to have frequency accuracy of 1 - 5%
> at most.
> Also, I am aware that S-parameter methods are not appropriate at these
> low
> frequencies.
Have you thought about a current charging and discharging a cap? This
gives a nice triangle wave with the frequency proportional to the
current. Just make sure the voltage that defines the 'trip' points is
clean. (A mistake I made.) Or do you need a sine wave?
George H.
>
>
>
>
>
> > You could digitally synthesize a 60Hz sine wave with a smallish
> > processor -- I believe there are some TMS430 parts that could do it all
> > in one package with a PWM output to be filtered by a simple RC.
>
> > But "best" depends heavily on what you want.
>
> > --
>
> > Tim Wescott
> > Wescott Design Serviceshttp://www.wescottdesign.com
>
> > Do you need to implement control loops in software?
> > "Applied Control Theory for Embedded Systems" was written for you.
> > See details athttp://www.wescottdesign.com/actfes/actfes.html- Hide quoted text -
>
> - Show quoted text -
Paul Hovnanian P.E.
[snip]
> Also, I am aware that S-parameter methods are not appropriate at these
> low frequencies.
Why not? Actual L/C component values might be a bit larger than most people
here deal with, but the math is still the same.
--
Paul Hovnanian pa...@hovnanian.com
----------------------------------------------------------------------
Have gnu, will travel.
Paul Keinanen
wrote:
>On 07/06/2010 09:10 AM, Daku wrote:
>> On Jul 5, 8:59 pm, Tim Wescott<t...@seemywebsite.com> wrote:
>>> I'd hardly call 60Hz "ultra low frequency". But it is pretty darned low.
>>>
>>> All the suggestions you've gotten so far are good as far as they go and
>>> may well be perfect -- but what are you trying to do? Do you need sine
>>> wave out or square? If sine wave, how pure? Do you have any
>>> specifications on jitter, phase noise, or frequency accuracy?
>> I am trying to design a PLL for very low frequencies, e.g., power line
>> grid.
>> I am concerned with the VCO as it is a crucial sub-circuit. I am
>> aiming for
>> a phase noise of approximately -100 dBc/Hz but not very sure of the
>> offset
>> frequency. Ideally, I would like to have frequency accuracy of 1 - 5%
>> at most.
>> Also, I am aware that S-parameter methods are not appropriate at these
>> low
>> frequencies.
If you want to track the _actual_ mains frequency, just use a mains
driven synchronous motor. To get the noise sidebands down, use some
flywheels :-).
>
>I think that those specs would be difficult to achieve with an
>all-analog oscillator running at 60Hz. Not impossible -- I could do it,
>and Joerg could do it in a fraction of the time I'd take. Using some
>sort of direct digital synthesis -- even if it's just a microprocessor
>-- running off of a crystal reference would be almost trivial in
>comparison and would probably take less board space and would be far
>more repeatable in manufacturing.
>
>If you just had to do this purely in the analog domain your best bet
>might be a pair of crystal oscillators, frequency steered with
>varactors, carefully built, and with their outputs mixed down to 60Hz.
>But that's a solution I would expect to see in a bit of kit from the
>50's through the 80's -- anything later and I'd expect to see a DDS.
Just a few minutes ago, the Nordel AC network (Danish isles, Finland,
Norway, Sweden) was running at 50.11 Hz or +2200 ppm above nominal in
order to allow the mains synchronized clocks to catch up.
A simple fundamental frequency VXCO can be pulled about +/-100 ppm
with the load capacitance. About 1000 ppm is the maximum with
adjustable serial inductance and adjustable parallel load capacitance
at the crystal.
At 50/60 Hz, even a trivial processor can generate a variable
frequency sine wave using the NCO (Numerically Controlled Oscillator)
principle to generate a sine wave, which can be locked to the incoming
signal in some loop configuration.
Even a trivial processor might be able to generate both sine and
cosine waveforms for 49.98, 50.00. 50.92 Hz etc. in parallel and
performing a phase comparison between all these in parallel to
determine the best match.
whit3rd
> > I am trying to design a PLL for very low frequencies, e.g., power line
> > grid.
> > I am concerned with the VCO as it is a crucial sub-circuit. I am
> > aiming for
> > a phase noise of approximately -100 dBc/Hz
> I think that those specs would be difficult to achieve with an
> all-analog oscillator running at 60Hz. Not impossible -- I could do it,
An LC type oscillator is good for phase stability (multivibrator
types are less good), but C values of voltage-variable capacitors
are inconvenient for this range. So, you're stuck with an
increductor. This is an inductor with a semi-saturating core, using
a high impedance winding with DC current in it to move
the inductance.
It's an old technique (usually nowadays this kind of thing
is only used for flux-gate magnetometers) but a goodie.
Switched-capacitor filters with a 100x clock are another
approach. I think the MF10 app note has an example (figure 7).
<http://www.national.com/an/AN/AN-307.pdf>
The fast clock can be a relatively unstable CD4046
type of VCO, it'll all average out. Hopefully.
JosephKK
>On Jul 5, 8:59 pm, Tim Wescott <t...@seemywebsite.com> wrote:
>> I'd hardly call 60Hz "ultra low frequency". But it is pretty darned low.
>>
>> All the suggestions you've gotten so far are good as far as they go and
>> may well be perfect -- but what are you trying to do? Do you need sine
>> wave out or square? If sine wave, how pure? Do you have any
>> specifications on jitter, phase noise, or frequency accuracy?
>I am trying to design a PLL for very low frequencies, e.g., power line
>grid.
>I am concerned with the VCO as it is a crucial sub-circuit. I am
>aiming for
>a phase noise of approximately -100 dBc/Hz but not very sure of the
>offset
Yikes, it would take years (decades) to measure that. Your reference
standard to measure against would be problematic as well. One cycle
(about 16.667 ms) * 10^10 is over 46,000 hours, 275 weeks, 5.2 years.
How about reframing it as a jitter and wander specification?
>frequency. Ideally, I would like to have frequency accuracy of 1 - 5%
>at most.
You would blow the phase noise spec by being off by fractional
millihertz.
Tim Wescott
> On Tue, 6 Jul 2010 09:10:25 -0700 (PDT), Daku<daku...@gmail.com> wrote:
>
>> On Jul 5, 8:59 pm, Tim Wescott<t...@seemywebsite.com> wrote:
>>> I'd hardly call 60Hz "ultra low frequency". But it is pretty darned low.
>>>
>>> All the suggestions you've gotten so far are good as far as they go and
>>> may well be perfect -- but what are you trying to do? Do you need sine
>>> wave out or square? If sine wave, how pure? Do you have any
>>> specifications on jitter, phase noise, or frequency accuracy?
>> I am trying to design a PLL for very low frequencies, e.g., power line
>> grid.
>> I am concerned with the VCO as it is a crucial sub-circuit. I am
>> aiming for
>> a phase noise of approximately -100 dBc/Hz but not very sure of the
>> offset
>
> Yikes, it would take years (decades) to measure that. Your reference
> standard to measure against would be problematic as well. One cycle
> (about 16.667 ms) * 10^10 is over 46,000 hours, 275 weeks, 5.2 years.
Not if you multiplied the VCO output by a quadrature sine wave and
looked at the noise of the result. Then it would just take minutes.
> How about reframing it as a jitter and wander specification?
>
>> frequency. Ideally, I would like to have frequency accuracy of 1 - 5%
>> at most.
>
> You would blow the phase noise spec by being off by fractional
> millihertz.
I think the OP's idea is that the absolute frequency vs. command voltage
can have an offset, because the PLL will be correcting for it. It's the
random contribution to phase noise that he's trying to limit.
It's an awfully tight spec -- and one that could easily get blown with
one noisy stage in the chain, outside of the oscillator -- but I don't
think it's impossible to do on a bench top.
j
example it’s possible to have a phase noise spec of great than 100
db’s at x number of hz of the carrier and have over 10KHz or more of
frequency in accuracy. Jitter is the parameter that has a direct
correlation to phase noise.
Is the goal of the project to lock to 60 Hz line,(it's the ref for the
loop), and provide a clean 60 Hz out?
j
Michael Black
That's really the question, too many times people want a solution when
they can't define the problem. He thinks he needs XXX, but we can't know
until we know what he wants to do with it.
My guess was that he wants to feed something back into the AC line,
such as excess power from solar panels, and hence needs it to match
the existing 60Hz power grid. That is just a guess, but
obviously that actually defines the problem rather than the specs.
One thing, since he actually needs 60Hz rather than a wide range of
frequency, he actually is in better shape than some of the responses
suggest. There is a big difference between the best choice being LC
circuits at that "low" a frequency and having a wide frequency shift,
and starting with a crystal controlled oscillator and just enough
frequency control to ensure it is really at 60Hz. Since one knows
exactly what frequency one needs, you can start with something that
is highly on frequency as it is, and then just enough control to
bump it exactly where it belongs.
Michael
j
Yeah, as said, unfortunately so many of these questions are so poorly
framed, it’s almost impossible to give any meaningful help.
JosephKK
wrote:
>On 07/07/2010 04:52 AM, JosephKK wrote:
>> On Tue, 6 Jul 2010 09:10:25 -0700 (PDT), Daku<daku...@gmail.com> wrote:
>>
>>> On Jul 5, 8:59 pm, Tim Wescott<t...@seemywebsite.com> wrote:
>>>> I'd hardly call 60Hz "ultra low frequency". But it is pretty darned low.
>>>>
>>>> All the suggestions you've gotten so far are good as far as they go and
>>>> may well be perfect -- but what are you trying to do? Do you need sine
>>>> wave out or square? If sine wave, how pure? Do you have any
>>>> specifications on jitter, phase noise, or frequency accuracy?
>>> I am trying to design a PLL for very low frequencies, e.g., power line
>>> grid.
>>> I am concerned with the VCO as it is a crucial sub-circuit. I am
>>> aiming for
>>> a phase noise of approximately -100 dBc/Hz but not very sure of the
>>> offset
>>
>> Yikes, it would take years (decades) to measure that. Your reference
>> standard to measure against would be problematic as well. One cycle
>> (about 16.667 ms) * 10^10 is over 46,000 hours, 275 weeks, 5.2 years.
>
>Not if you multiplied the VCO output by a quadrature sine wave and
>looked at the noise of the result. Then it would just take minutes.
What is done at RF of 60 MHz, 600 MHz, or 6 GHz to make the measurement
easier, does not make it faster. Sure, it takes a second or so at 600
MHz for 100 dBc. And a megasecond or so at 60 Hz to measure it in the RF
traditional way. Simple scaling. That is why i recommended changing to
jitter and wander specifications, which you have a decent shot at
measuring in an hour or so.
Or maybe you are on to something, but for my and OPs sake, please flesh
it out a lot more, with some calcs please.
Tim Wescott
Who needs to calculate? Mix it down to DC in quadrature. You'll have
naught but a residual at DC, and all the phase noise will appear around
DC (amplitude noise will be nulled out, 'cause it's in quadrature).
Now amplify the snot out of it with a really low noise, low corner
frequency AC-coupled amp, and apply it to a really low frequency
spectrum analyzer (you may have to use a dynamic signal analyzer to see
anything useful, I dunno).
You'll only need to measure for hours if you want to resolve frequencies
in the mHz (that's m for milli, not M for mega).
Note that you could get total noise by notching out the carrier, but if
you just want _phase_ noise you'll be screwing your measurement with
amplitude crap.
Phil Hobbs
I don't know that -100 dBc/Hz is that hard at 60 Hz. I bet you could do
that by running a bog standard multivibrator at 1024*1024*60 Hz and
dividing down. You'd need a sine shaper, but the phase noise goes down
by N**2, so you'd get 100 dB improvement just from that. Alternatively,
you could make an LC VCO and divide that down.
You might even be able to do it with all analog--the OPA378 has 20
nV/sqrt(Hz) all the way down to DC. With a 5V sine wave at 60 Hz,
that's something like 1800 V/s, so 20 nV gives you something like 10
picoseconds per root hertz. You probably lose a factor of sqrt(2) in
there, but that ought to be good enough. Your ALC network would
contribute more than that, almost for sure.
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058
hobbs at electrooptical dot net
http://electrooptical.net
Phil Hobbs
> I don't know that -100 dBc/Hz is that hard at 60 Hz. I bet you could do
> that by running a bog standard multivibrator at 1024*1024*60 Hz and
> dividing down. You'd need a sine shaper, but the phase noise goes down
> by N**2, so you'd get 100 dB improvement just from that. Alternatively,
> you could make an LC VCO and divide that down.
120 dB. Can't count today.
JosephKK
<pcdhSpamM...@electrooptical.net> wrote:
>Phil Hobbs wrote:
>
>> I don't know that -100 dBc/Hz is that hard at 60 Hz. I bet you could do
>> that by running a bog standard multivibrator at 1024*1024*60 Hz and
>> dividing down. You'd need a sine shaper, but the phase noise goes down
>> by N**2, so you'd get 100 dB improvement just from that. Alternatively,
>> you could make an LC VCO and divide that down.
>
>120 dB. Can't count today.
>
>Cheers
>
>Phil Hobbs
Sure, you can mathematically "predict" it, but how do you measure it?
Or do you switch to another metric which can be both predicted and
measured?
JosephKK
wrote:
And that would be on the order of 30 dBc measurement of phase noise? How
long do you have to measure to get to 100 dBc phase noise?
Or do you claim that is already 60 dBc phase noise measurement, then how
many days/weeks to get to 100 dBc phase noise measurement?
Minus 100 dBc phase noise is an inappropriate measurement for the
fundamental frequency involved. Jitter and wander are more to the point.
Phil Hobbs
Let's keep the math bashing to the other thread, okay?
Although it isn't highly relevant to the OP's problem, it wouldn't be
very difficult to measure the residual FM--use MOSFET buffers to drive
two divider strings running from independent power supplies, and
cross-correlate their outputs, exchanging them periodically to get rid
of the drift in the correlator. For the correlator design, see Hanbury
Brown and Twiss, circa 1963--and they did it with discrete bipolars.
There are hard measurements, but this isn't one of them.
whit3rd
<pcdhSpamMeSensel...@electrooptical.net> wrote:
> I don't know that -100 dBc/Hz is that hard at 60 Hz. I bet you could do
> that by running a bog standard multivibrator at 1024*1024*60 Hz and
> dividing down. You'd need a sine shaper, but the phase noise goes down
> by N**2
Eh? I'd think it's N**0.5 (the multivibrator has cumulative but
random errors).
j
noise measurements. The OP never stated the offset from the carrier
nor bandwidth. Or maybe I just missed it.
It’s not clear to me why JosephKK thinks this would be either a time
consuming or difficult measurement to make. Assuming the appropriate
measurement system is in hand 100 dBc numbers are easily achievable.
Whether it’s 60 Hz or several GHz’s the global issues are the same in
making a phase noise measurement.
But having said the above, without the OP responding I guess it really
doesn’t matter. But I’d like to know more about the application and
derive solutions from there.
Phil Hobbs
The time jitter of the edges stays the same, but the resulting phase
error goes down by a factor of N due to the division. Phase is like
amplitude, so you have to square it to get the noise power--hence N**2.
JosephKK
<pcdhSpamM...@electrooptical.net> wrote:
My issue was not so much the direct difficulty of the measurement, there
are several fairly straight forward setups. But with the _time_ it would
take to make the measurement using many of those setups. The elapsed
time seriously aggravates other measurement issues, notably including
calibration.
JosephKK
OK. For a carrier of 60 MHz. Pick an instrument or test setup of your
choice, state the model[s]. Clearly explain just what is going on in the
measurement and the time it takes to accumulate sufficient data to make
the measurement. Explain why it takes that much time to reach a reliable
measurement of -100 dBc phase noise at that carrier frequency.
Now see how well it scales to one million times lower fundamental
frequency without a similar scaling in measurement time.
whit3rd
<pcdhSpamMeSensel...@electrooptical.net> wrote:
> whit3rd wrote:
> > On Jul 8, 12:29 pm, Phil Hobbs
> > <pcdhSpamMeSensel...@electrooptical.net> wrote:
>
> >> I don't know that -100 dBc/Hz is that hard at 60 Hz. I bet you could do
> >> that by running a bog standard multivibrator at 1024*1024*60 Hz and
> >> dividing down. You'd need a sine shaper, but the phase noise goes down
> >> by N**2
>
> > Eh? I'd think it's N**0.5 (the multivibrator has cumulative but
> > random errors).
>
> The time jitter of the edges stays the same, but the resulting phase
> error goes down by a factor of N due to the division. Phase is like
> amplitude, so you have to square it to get the noise power--hence N**2.
With an LC oscillator (class C transistor drive) the jitter in one
edge
(as determined by the transistor conduction) would be random, and
only a small fraction of the circulating energy would respond to the
edge error. So, the jitter in the LC output is a sequence of
random errors.
For a multivibrator, however, the internal state resets each cycle;
the jittery time of cycle N becomes the new zero, and the jitter in
cycle N+1 is the sum of those two values. This kind of timing
error is the accumulating kind. The jitter is an arithmetic (sum)
sequence of randoms.
So, for an LC oscillator you can get the N**2 behavior after
squaring; for a multivibrator oscillator only expect N**1.
I think this is why serious timing eschews the multivibrator.
Grant
I think OP disappeared because most here forgot that OP was trying to
filter out mains frequency, which varies during the day, and tries to
deliver that correct number of cycles by each midnight to stop the
clocks drifting.
IOW, you may've drifted way off-topic. What's the local short term
variation in mains frequency over there? One or two percent? or,
are you going to consider that variation as a very large phase
shift?
Grant.
Jim Thompson
<pcdhSpamM...@electrooptical.net> wrote:
>whit3rd wrote:
>> On Jul 8, 12:29 pm, Phil Hobbs
>> <pcdhSpamMeSensel...@electrooptical.net> wrote:
>>
>>> I don't know that -100 dBc/Hz is that hard at 60 Hz. I bet you could do
>>> that by running a bog standard multivibrator at 1024*1024*60 Hz and
>>> dividing down. You'd need a sine shaper, but the phase noise goes down
>>> by N**2
>>
>> Eh? I'd think it's N**0.5 (the multivibrator has cumulative but
>> random errors).
>
>The time jitter of the edges stays the same, but the resulting phase
>error goes down by a factor of N due to the division. Phase is like
>amplitude, so you have to square it to get the noise power--hence N**2.
>
>Cheers
>
>Phil Hobbs
Hey Phil! How come no comment on conservation of charge and energy?
You have a dog in this show ?:-) Weenie!
...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
Obama isn't going to raise your taxes...it's Bush' fault: Not re-
newing the Bush tax cuts will increase the bottom tier rate by 50%
Phil Hobbs
It's the modulation frequency that's relevant, not the carrier
frequency. Measurements get slower when you reduce the bandwidth.
(You can see why this doesn't work if you imagine running it
backwards--mixing or multiplying up to some very high frequency doesn't
allow you to make a measurement with 1 Hz bandwidth any faster.
Phil Hobbs
Modulation frequency isn't affected by heterodyning or frequency
multiplication and division. If you take a 60 MHz sine wave and FM it
at 1 Hz modulation frequency and 1 MHz peak frequency deviation (M=1E6),
then divide it by a million, you get a 60-Hz sine wave modulated at 1
Hz with a 1-Hz peak frequency division (M=1).
Phil Hobbs
You're moving the goal posts. We aren't talking about the phase
correlations, just the instantaneous phase noise. Phase noise sideband
power goes down as 1/N**2, period.
Phil Hobbs
> On Fri, 09 Jul 2010 14:08:28 -0400, Phil Hobbs
> <pcdhSpamM...@electrooptical.net> wrote:
>
>> whit3rd wrote:
>>> On Jul 8, 12:29 pm, Phil Hobbs
>>> <pcdhSpamMeSensel...@electrooptical.net> wrote:
>>>
>>>> I don't know that -100 dBc/Hz is that hard at 60 Hz. I bet you could do
>>>> that by running a bog standard multivibrator at 1024*1024*60 Hz and
>>>> dividing down. You'd need a sine shaper, but the phase noise goes down
>>>> by N**2
>>> Eh? I'd think it's N**0.5 (the multivibrator has cumulative but
>>> random errors).
>> The time jitter of the edges stays the same, but the resulting phase
>> error goes down by a factor of N due to the division. Phase is like
>> amplitude, so you have to square it to get the noise power--hence N**2.
>>
>> Cheers
>>
>> Phil Hobbs
>
> Hey Phil! How come no comment on conservation of charge and energy?
> You have a dog in this show ?:-) Weenie!
>
> ...Jim Thompson
I'm mainly here to talk about electronics. One-upmanship also tends to
intimidate the newbies, which I really don't want to do. I try not to
dispense Bad Info myself, and try to help other people's
misunderstandings when I can. Otherwise I just read with interest and
learn stuff.
Whit3rd seems to be talking about the phase correlations rather than the
instantaneous phase noise. Both multivibrators and LC resonators obey
equations with full locality, i.e. neither one has any memory at all.
For instance, if you have a 1 MHz resonator with a Q of a million, it
takes a second or so to get its phase to change when you put PM on the
drive waveform. OTOH, if you change the resonant frequency suddenly,
e.g. by putting 100V on a Y5V tank capacitor, the resonant frequency
changes immediately--much faster than 1/Q cycles.
Because of the switching action, multivibrators intermodulate the
switching element's noise at all frequencies, which makes their jitter
much worse; also the effective Q of a multivibrator is less than 1,
which means that there isn't any significant filtering action from the
resonator. (That's frequency-domain way of thinking about what Whit3rd
is talking about in the time domain--the conservation of energy issue is
easier to think about if there's a natural bandwidth limit to the
sqrt(t) behaviour.) The physical origin of the phase modulation doesn't
change the way it varies with division ratio, though.
j
OP giving a better definition of the problem it’s a guess at best
which measurement technique is required.
He never did state the basis for his phase noise number, nor did he
have an offset frequency.
The challenge in making –100 dBc or better measurements is a function
of the offset frequency and bandwidth. Center frequency isn’t the
issue here.
Phil Hobbs
You may not be interested, but perhaps other folks are. And how big an
offset frequency can he have on a 60 Hz carrier, anyway?
Jim Thompson
<pcdhSpamM...@electrooptical.net> wrote:
There's no one-up-man-ship involved. Larkin won't (or can't, because
he doesn't really understand it) show where the extra charge came
from. You (or Win) could put a stop to Larkin's nonsense. Larkin
displays me as a fool, and the newbies don't know any better, so
they'll never ever learn the correct solution unless someone
(politically :) respected steps in.
>
>Whit3rd seems to be talking about the phase correlations rather than the
>instantaneous phase noise. Both multivibrators and LC resonators obey
>equations with full locality, i.e. neither one has any memory at all.
>
>For instance, if you have a 1 MHz resonator with a Q of a million, it
>takes a second or so to get its phase to change when you put PM on the
>drive waveform. OTOH, if you change the resonant frequency suddenly,
>e.g. by putting 100V on a Y5V tank capacitor, the resonant frequency
>changes immediately--much faster than 1/Q cycles.
>
>Because of the switching action, multivibrators intermodulate the
>switching element's noise at all frequencies, which makes their jitter
>much worse; also the effective Q of a multivibrator is less than 1,
>which means that there isn't any significant filtering action from the
>resonator. (That's frequency-domain way of thinking about what Whit3rd
>is talking about in the time domain--the conservation of energy issue is
>easier to think about if there's a natural bandwidth limit to the
>sqrt(t) behaviour.) The physical origin of the phase modulation doesn't
>change the way it varies with division ratio, though.
>
>Cheers
>
>Phil Hobbs
Yep.
j
to phase noise measurements or something of that nature. Or stable
low freq Osc for example.
It appears that the OP wanted to discipline to line and use that as a
long term ref.. It’s not clear to me how he came up with the –100 dBc
number without an offset …
I’m not sure what you mean by “how big an offset” … offset generally
refers to the position of the measurement relative to the carrier.
The closer the offset the more difficult the measurement ... generally
do to the limitation of the measuring equipment. The interesting part
is the solution to those challenges.
I’m not trying to be a malcontent here … just seems like the
discussion doesn’t have direction.
Phil Hobbs
> For instance, if you have a 1 MHz resonator with a Q of a million, it
> takes a second or so to get its phase to change when you put PM on the
> drive waveform. OTOH, if you change the resonant frequency suddenly,
> e.g. by putting 100V on a Y5V tank capacitor, the resonant frequency
> changes immediately--much faster than 1/Q cycles.
Much faster than Q cycles, I mean. (Posted before breakfast in
Albuquerque.)
Phil Hobbs
Understood. I agree that the OP's question wasn't that well posed, but
there was a bunch of very strongly stated Bad Info here that needed
pointing out. I took the OP to be saying "I need a 60 Hz oscillator
block that's way, way quieter than I know how to build", and that the
rest of us have been making suggestions. Calculating or measuring how
good it actually is is his worry.
John Larkin
<pcdhSpamM...@electrooptical.net> wrote:
One interesting and often overlooked part is the coaxial ceramic
resonator. It's essentially a shorted transmission line formed in a
block or tube of hi-K ceramic, usually by silver or copper plating it.
They are usually treated by the RF boys as resonators or inductors,
but they really act like time-domain transmission lines. TCs are in
the single-digit PPMs and Qs in the hundreds or thousands. Dielectric
constants are in the hundreds or thousands, so they are very short for
their delay/frequency.
Remarkable parts. I use them to make instant-start/instant-stop
oscillators in the 600 MHz range. As a VCO, they will have very low
phase noise, somewhere between an LC and a quartz crystal.
John
John Larkin
<pcdhSpamM...@electrooptical.net> wrote:
>Phil Hobbs wrote:
>
>> For instance, if you have a 1 MHz resonator with a Q of a million, it
>> takes a second or so to get its phase to change when you put PM on the
>> drive waveform. OTOH, if you change the resonant frequency suddenly,
>> e.g. by putting 100V on a Y5V tank capacitor, the resonant frequency
>> changes immediately--much faster than 1/Q cycles.
>
>Much faster than Q cycles, I mean. (Posted before breakfast in
>Albuquerque.)
>>
>
>Cheers
>
>Phil Hobbs
Check out La Posada de Albuquerque. Cool old hotel. Or it was, except
they may have "upgraded" it.
John
Phil Hobbs
I don't know about that. It isn't that difficult to calculate a circuit
with two caps, an inductor, and an elf who opens and closes a switch at
the right moments. It does help to know elementary differential equations.
I haven't actually followed the original discussion closely enough to
know who made the first technical error. The larger error IMO is to
keep getting into these tiresome p***ing contests, which I decline to
do. If what you want is merely to have the correct solution posted,
post it and let's move on to some electronics.
Phil Hobbs
I confess I'm one of the ones who overlooked them...where do you get
them, and do they come in Y5V?
John Larkin
<pcdhSpamM...@electrooptical.net> wrote:
I don't think any specific problem has been clearly stated, such that
it can be analyzed. My comment, that seems to have ruffled feathers,
is that one shouldn't assume as a working tool that charge, coulombs
stored in various capacitors in a circuit, is conserved. Sometimes it
is, sometimes it isn't, sometimes the concept is silly.
The argument did make me go back and review some basics, which is
good. Messing with all this digital and software and opamp stuff can
make the old EE101 math get rusty.
John
John Larkin
<pcdhSpamM...@electrooptical.net> wrote:
Don't know. We buy ours from Skyworks/Trans-tech, and they have Er
values from 10 to 90. TCs are very good somehow. Lots of people make
these things and the little, high-frequency ones are cheap in
quantity.
John
George Herold
Yes! I've enjoyed the discussion. Say could someone explain the the
100 dBc of phase noise spec. I've been thinking of this a one part in
10^5 of jitter in the period. So for instance a 1 Hz signal the
jitter is less than 10 micro seconds aand for a 1 MHz signal a jitter
of 10 pico seconds.
Is that right?
George H.
John Larkin
The usual oscillator phase noise spec would be " -100 dBc/Hz " at some
offset frequency from the carrier. It's often given as a curve. It is
possible to convert the curve to an RMS jitter spec... I have a
program around somewhere that some s.e.d. guy posted.
John
Jim Thompson
Dodge, dodge, dodge. You specifically stated, in...
Message-ID: <3b893612tjjndo8o4...@4ax.com>
"... charge is not conserved."
Where, oh great pretend guru, where does the excess charge come from?
Tim Wescott
> Paul Keinanen wrote:
>> On Tue, 06 Jul 2010 09:52:43 -0700, Tim Wescott <t...@seemywebsite.com>
>> wrote:
>>
>>> On 07/06/2010 09:10 AM, Daku wrote:
>>>> On Jul 5, 8:59 pm, Tim Wescott<t...@seemywebsite.com> wrote:
>>>>> I'd hardly call 60Hz "ultra low frequency". But it is pretty darned
>>>>> low.
>>>>>
>>>>> All the suggestions you've gotten so far are good as far as they go
>>>>> and
>>>>> may well be perfect -- but what are you trying to do? Do you need sine
>>>>> wave out or square? If sine wave, how pure? Do you have any
>>>>> specifications on jitter, phase noise, or frequency accuracy?
>>>> I am trying to design a PLL for very low frequencies, e.g., power line
>>>> grid.
>>>> I am concerned with the VCO as it is a crucial sub-circuit. I am
>>>> aiming for
>>>> a phase noise of approximately -100 dBc/Hz but not very sure of the
>>>> offset
>>>> frequency. Ideally, I would like to have frequency accuracy of 1 - 5%
>>>> at most.
>>>> Also, I am aware that S-parameter methods are not appropriate at these
>>>> low
>>>> frequencies.
>>
>> If you want to track the _actual_ mains frequency, just use a mains
>> driven synchronous motor. To get the noise sidebands down, use some
>> flywheels :-).
>>
>>> I think that those specs would be difficult to achieve with an
>>> all-analog oscillator running at 60Hz. Not impossible -- I could do
>>> it, and Joerg could do it in a fraction of the time I'd take. Using
>>> some sort of direct digital synthesis -- even if it's just a
>>> microprocessor -- running off of a crystal reference would be almost
>>> trivial in comparison and would probably take less board space and
>>> would be far more repeatable in manufacturing.
>>>
>>> If you just had to do this purely in the analog domain your best bet
>>> might be a pair of crystal oscillators, frequency steered with
>>> varactors, carefully built, and with their outputs mixed down to
>>> 60Hz. But that's a solution I would expect to see in a bit of kit
>>> from the 50's through the 80's -- anything later and I'd expect to
>>> see a DDS.
>>
>> Just a few minutes ago, the Nordel AC network (Danish isles, Finland,
>> Norway, Sweden) was running at 50.11 Hz or +2200 ppm above nominal in
>> order to allow the mains synchronized clocks to catch up.
>> A simple fundamental frequency VXCO can be pulled about +/-100 ppm
>> with the load capacitance. About 1000 ppm is the maximum with
>> adjustable serial inductance and adjustable parallel load capacitance
>> at the crystal.
>>
>> At 50/60 Hz, even a trivial processor can generate a variable
>> frequency sine wave using the NCO (Numerically Controlled Oscillator)
>> principle to generate a sine wave, which can be locked to the incoming
>> signal in some loop configuration.
>>
>> Even a trivial processor might be able to generate both sine and
>> cosine waveforms for 49.98, 50.00. 50.92 Hz etc. in parallel and
>> performing a phase comparison between all these in parallel to
>> determine the best match.
>>
>
> I don't know that -100 dBc/Hz is that hard at 60 Hz. I bet you could do
> that by running a bog standard multivibrator at 1024*1024*60 Hz and
> dividing down. You'd need a sine shaper, but the phase noise goes down
> you could make an LC VCO and divide that down.
This actually kind of makes my point, which I didn't state clearly: if
you _don't_ use a divider it'll be hard. With a divider it gets easy,
as long as you ignore clock jitter in the divider (and clock jitter
probably isn't a big deal, given the output frequency).
> You might even be able to do it with all analog--the OPA378 has 20
> nV/sqrt(Hz) all the way down to DC. With a 5V sine wave at 60 Hz, that's
> something like 1800 V/s, so 20 nV gives you something like 10
> picoseconds per root hertz. You probably lose a factor of sqrt(2) in
> there, but that ought to be good enough. Your ALC network would
> contribute more than that, almost for sure.
Depending on how close to the carrier you want to get, you lose a factor
of up to infinity (if you get _really_ close to the carrier).
The noise gain is something like 1/(s^2 + w0^2) -- it's an oscillator.
Worse, because it's an RC, the constant you're multiplying by is greater
than one -- I get Hn(s) ~ 15/(s^2 + w0^2). That's not taking the
current noise of the part into account (which, I admit, I haven't
checked on because I'm lazy).
1Hz away your noise gain is just about 200, for 4uV/sqrt(Hz). That's
doing OK, but at 0.1Hz away the noise gain is about 2000 -- all you have
to do is measure close enough to the carrier at a wide enough bandwidth
and your noise is too high (sure would be nice if the OP specified what
he wanted, but I think we lost him).
--
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
Jim Thompson
I've been "using" them... designing them into GPS LO's since before
you were born ;-)
Joel Koltner
message news:dcim36poktoi3gc0k...@4ax.com...
[coaxial ceramic resonators]
> I've been "using" them... designing them into GPS LO's since before
> you were born ;-)
What's the advantage -- for you -- over crystals? Just lower cost?
---Joel
Phil Hobbs
Small phase jitter is the quadrature partner of small amplitude noise.
Say you have a pure carrier and add ordinary white noise, e.g. by
putting a resistor in series with the perfect oscillator's output. The
resulting RMS phase deviation in some given bandwidth is
<delta_phi> = 1/(sqrt(2*CNR))
where CNR is the carrier to noise ratio (i.e. carrier power/noise power
in the given bandwidth). The factor of sqrt(2) expresses the fact that
the noise and signal are uncorrelated, so that half the noise power
winds up in the I phase as amplitude noise, and half winds up in the Q
phase as phase noise.
You can derive this from the formula for sums and differences of sines
and cosines plus an orthogonality argument--it's quite pretty. It's in
my section 13.6 (either edition), but that derivation almost certainly
isn't original with me. One very pleasant consequence is that the phase
noise statistics are the same as those of the additive noise in the
high-CNR limit where the formula applies.
The universality of this formula is why essentially all FM and PM
detectors have equivalent performance at high SNR--where the additive
model breaks down is low SNR, where FM/PM detection schemes really
differ in performance.
Jim Thompson
Small, hi-Q, but reasonably pull-able, to phase lock.
(Cost never matters to me, I'm never paying ;-)
Joel Koltner
>>What's the advantage -- for you -- over crystals? Just lower cost?
Ah, gotcha -- I wans't thinking (the really rather obvious bit) that there's
no such thing as a 1575.42MHz crystal. :-)
Speaking of PLLs... do you happen to know why so many of them have rather high
(hundreds of MHz) lower frequency limits? E.g., this seemingly-popular part:
http://www.national.com/pf/LM/LMX2485.html ... only goes down to 500MHz on one
of the PLLs; National has a slightly different version --
http://www.national.com/pf/LM/LMX2485E.html -- that goes down to 50MHz.
What's different inside, would you think?
Discussions at work here and with Joerg have postulated:
-- AC coupling of the input amplifiers?
-- The fastest counters (dividers) are actually dynamic circuitry, and hence
they really can't count too slowly or they forget where they were. (1, 2,
3... umm? 3? 1? Where was I again? La-la-la-la-la I'm so confused!)
There aren't a lot of low-power PLL-based synthesizers in the "more than a
handful of MHz"-CD4046 (and derivative) territory and these multi-hundred-MHz+
monsters. We're after some 45MHz PLLs, and so far the ADF4001 is one of the
very few attractive-looking candidates.
Thanks for the help,
---Joel
Jim Thompson
<zapwireD...@yahoo.com> wrote:
>"Jim Thompson" <To-Email-Use-Th...@On-My-Web-Site.com> wrote in
>message news:1ekm36hmf1f3egiu3...@4ax.com...
>>>What's the advantage -- for you -- over crystals? Just lower cost?
>> Small, hi-Q, but reasonably pull-able, to phase lock.
>
>Ah, gotcha -- I wans't thinking (the really rather obvious bit) that there's
>no such thing as a 1575.42MHz crystal. :-)
Rather than think "resonator", think very-low-loss shorted-termination
transmission line stub. (Think college days :-)
>
>Speaking of PLLs... do you happen to know why so many of them have rather high
>(hundreds of MHz) lower frequency limits? E.g., this seemingly-popular part:
>http://www.national.com/pf/LM/LMX2485.html ... only goes down to 500MHz on one
>of the PLLs; National has a slightly different version --
>http://www.national.com/pf/LM/LMX2485E.html -- that goes down to 50MHz.
>What's different inside, would you think?
Don't know. Might be AC-coupling internally... makes dividers less
prone to offset voltages at those frequencies.
>
>Discussions at work here and with Joerg have postulated:
>
>-- AC coupling of the input amplifiers?
Yep. See above.
>-- The fastest counters (dividers) are actually dynamic circuitry, and hence
>they really can't count too slowly or they forget where they were. (1, 2,
>3... umm? 3? 1? Where was I again? La-la-la-la-la I'm so confused!)
Indeed ;-)
>
>There aren't a lot of low-power PLL-based synthesizers in the "more than a
>handful of MHz"-CD4046 (and derivative) territory and these multi-hundred-MHz+
>monsters. We're after some 45MHz PLLs, and so far the ADF4001 is one of the
>very few attractive-looking candidates.
>
>Thanks for the help,
>---Joel
>
>
I rarely keep track of OTC offerings, since I'm really in the
designing whole SOC business.
Occasionally a client will say, "I want all this garbage on a single
chip", and hand me a pile of data sheets.
Then I have to analyze how it all works.
That actually works out well with usually significant space and power
compaction, since I/O can use up a lot of chip area and power.
Joel Koltner
"Jim Thompson" <To-Email-Use-Th...@On-My-Web-Site.com> wrote in
message news:s5mm365lej39010hb...@4ax.com...
> I rarely keep track of OTC offerings, since I'm really in the
> designing whole SOC business.
Yeah, that's becoming a bit unfortunate for those of us doing board-level
designs where the quantities just aren't there for doing custom ICs -- many of
the RF parts that are left aren't particularly optimized for power
consumption, since they're either parts that have been around forever (big
transistors, I suppose) or they're aimed at the aerospace/military market
where they figure you've got a nuclear reactor to power the thing.
We're got a wireless handheld (battery-powered) widget that, between the
numerous synthesized LOs, mixers, and amps might double as a pretty good
hand-warmer. :-)
---Joel
Jim Thompson
Post your questions and I'll try to answer.
Larkin will certainly jump in and shovel a pile of BS in your
direction :-)
I'm thinking the way to resolve this Larkin statement, "... charge is
not conserved" in...
Message-ID: <3b893612tjjndo8o4...@4ax.com>
is for me to post the solution on my website.
BUT in a password-protected ZIP file.
White-listed individuals can contact me for the password and see how
it works, that way Larkin will be prevented from seeing it and
claiming "that's what I really meant... let him choke on his own drool
:-)
I'll let you know when it's up there.
Funny. Yesterday, while cooling my heels at a going-away swim party
for a granddaughter (they're moving to Rancho Mirage... Renee's father
is in ill-health, so she's going to run his automotive parts
business), I sat there in the shade, in 112°F heat and "played" the
circuit in my head... that's how I figured out where the missing
charge comes from ;-)
JosephKK
<pcdhSpamM...@electrooptical.net> wrote:
>JosephKK wrote:
>> On Fri, 9 Jul 2010 10:22:34 -0700 (PDT), j <jdc...@gmail.com> wrote:
>>
>>> Resolution of noise vs frequency, (as in bw), is the issue in phase
>>> noise measurements. The OP never stated the offset from the carrier
>>> nor bandwidth. Or maybe I just missed it.
>>>
>>> It’s not clear to me why JosephKK thinks this would be either a time
>>> consuming or difficult measurement to make. Assuming the appropriate
>>> measurement system is in hand 100 dBc numbers are easily achievable.
>>> Whether it’s 60 Hz or several GHz’s the global issues are the same in
>>> making a phase noise measurement.
>>>
>>> But having said the above, without the OP responding I guess it really
>>> doesn’t matter. But I’d like to know more about the application and
>>> derive solutions from there.
>>
>>
>> OK. For a carrier of 60 MHz. Pick an instrument or test setup of your
>> choice, state the model[s]. Clearly explain just what is going on in the
>> measurement and the time it takes to accumulate sufficient data to make
>> the measurement. Explain why it takes that much time to reach a reliable
>> measurement of -100 dBc phase noise at that carrier frequency.
>>
>> Now see how well it scales to one million times lower fundamental
>> frequency without a similar scaling in measurement time.
>
>It's the modulation frequency that's relevant, not the carrier
>frequency. Measurements get slower when you reduce the bandwidth.
>
>(You can see why this doesn't work if you imagine running it
>backwards--mixing or multiplying up to some very high frequency doesn't
>allow you to make a measurement with 1 Hz bandwidth any faster.
>
>
>Cheers
>
>Phil Hobbs
Now what is the equivalent bandwidth of -100 dBc for a 60 Hz carrier?
Since you said 20 log() basis 60 * 10^-5 is 600 microHz. That would have
to take some minutes, and if you wanted a proper 10 to 1 measurement
buffer, it takes ten times longer. Call it 10,000 seconds? A few hours.
And the reference stability etc., i remarked on is coming into play.
JosephKK
<pcdhSpamM...@electrooptical.net> wrote:
>JosephKK wrote:
>> On Fri, 09 Jul 2010 11:56:28 -0400, Phil Hobbs
>> <pcdhSpamM...@electrooptical.net> wrote:
>>
>>> On 7/9/2010 8:59 AM, JosephKK wrote:
>>>> On Thu, 08 Jul 2010 15:37:28 -0400, Phil Hobbs
>>>> <pcdhSpamM...@electrooptical.net> wrote:
>>>>
>>>>> Phil Hobbs wrote:
>>>>>
>>>>>> I don't know that -100 dBc/Hz is that hard at 60 Hz. I bet you could do
>>>>>> that by running a bog standard multivibrator at 1024*1024*60 Hz and
>>>>>> dividing down. You'd need a sine shaper, but the phase noise goes down
>>>>>> by N**2, so you'd get 100 dB improvement just from that. Alternatively,
>>>>>> you could make an LC VCO and divide that down.
>>>>> 120 dB. Can't count today.
>>>>>
>>>>> Cheers
>>>>>
>>>>> Phil Hobbs
>>>> Sure, you can mathematically "predict" it, but how do you measure it?
>>>> Or do you switch to another metric which can be both predicted and
>>>> measured?
>>> Let's keep the math bashing to the other thread, okay?
>>>
>>> Although it isn't highly relevant to the OP's problem, it wouldn't be
>>> very difficult to measure the residual FM--use MOSFET buffers to drive
>>> two divider strings running from independent power supplies, and
>>> cross-correlate their outputs, exchanging them periodically to get rid
>>> of the drift in the correlator. For the correlator design, see Hanbury
>>> Brown and Twiss, circa 1963--and they did it with discrete bipolars.
>>>
>>> There are hard measurements, but this isn't one of them.
>>>
>>> Cheers
>>>
>>> Phil Hobbs
>>
>> My issue was not so much the direct difficulty of the measurement, there
>> are several fairly straight forward setups. But with the _time_ it would
>> take to make the measurement using many of those setups. The elapsed
>> time seriously aggravates other measurement issues, notably including
>> calibration.
>
>Modulation frequency isn't affected by heterodyning or frequency
>multiplication and division. If you take a 60 MHz sine wave and FM it
>at 1 Hz modulation frequency and 1 MHz peak frequency deviation (M=1E6),
> then divide it by a million, you get a 60-Hz sine wave modulated at 1
>Hz with a 1-Hz peak frequency division (M=1).
>
>Cheers
>
>Phil Hobbs
I am sorry. I think i am misreading your post, are you saying you can
get a 1 MHz deviation on a 60 Hz carrier? Naw, you must be trying to say
something else and i misunderstood.
Phil Hobbs
You're confused, I'm afraid. -100 dBc phase noise in a given bandwidth
(say 1 Hz, but it doesn't matter) is 7 microradians RMS. Using a 5V
swing and a CMOS analogue gate as a phase detector, that's
dV = 7e-6 rad RMS * 5V/(pi rad) = 11 microvolts RMS,
which is trivial to measure in a 1 Hz bandwidth in a few seconds--it's
80 dB above the noise of a good op amp, so you just have to wait for the
filter to settle.
Phil Hobbs
You can put a 1 MHz phase modulation on a 60 Hz carrier, but you sure
don't wind up with anything pretty. For instance, you could put the 60
Hz on a varactor-loaded transmission line, and drive the varactors with
1.000000000 MHz. As long as the varactors were driven really
differentially, you wouldn't get any 1.000000000 MHz on the line.
That's way outside the quasistatic limit, of course, which is where
we're all used to working. It would be a nasty splattery mess, but
you'd get _something_.
But that wasn't the point I was trying to make. ;)
Phil Hobbs
Sorry to be slow responding. That's an interesting point about the
noise gain of the oscillator. I think we agree that the OPA378 is okay
down to 1 Hz offset or thereabouts, anyway. (It's a very nice part
btw--I'm using it in some millihertz things just now.)
JosephKK
<To-Email-Use-Th...@On-My-Web-Site.com> wrote:
>On Mon, 12 Jul 2010 08:33:56 -0700, John Larkin
><jjla...@highNOTlandTHIStechnologyPART.com> wrote:
>
>>On Mon, 12 Jul 2010 10:40:00 -0400, Phil Hobbs
>><pcdhSpamM...@electrooptical.net> wrote:
>>
>>>Jim Thompson wrote:
>>>> On Fri, 09 Jul 2010 14:08:28 -0400, Phil Hobbs
>>>> <pcdhSpamM...@electrooptical.net> wrote:
>>>>
<snip>
>>
>>One interesting and often overlooked part is the coaxial ceramic
>>resonator. It's essentially a shorted transmission line formed in a
>>block or tube of hi-K ceramic, usually by silver or copper plating it.
>>They are usually treated by the RF boys as resonators or inductors,
>>but they really act like time-domain transmission lines. TCs are in
>>the single-digit PPMs and Qs in the hundreds or thousands. Dielectric
>>constants are in the hundreds or thousands, so they are very short for
>>their delay/frequency.
>>
>>Remarkable parts. I use them to make instant-start/instant-stop
>>oscillators in the 600 MHz range. As a VCO, they will have very low
>>phase noise, somewhere between an LC and a quartz crystal.
>>
>>John
>
>I've been "using" them... designing them into GPS LO's since before
>you were born ;-)
>
> ...Jim Thompson
That is really good since GPS itself is not that old.
Jim Thompson
I did my first Garmin chip more than 20 years ago.
Paul Keinanen
"JosephKK"<quiett...@yahoo.com> wrote:
>
>Now what is the equivalent bandwidth of -100 dBc for a 60 Hz carrier?
>Since you said 20 log() basis 60 * 10^-5 is 600 microHz. That would have
>to take some minutes, and if you wanted a proper 10 to 1 measurement
>buffer, it takes ten times longer. Call it 10,000 seconds? A few hours.
>And the reference stability etc., i remarked on is coming into play.
To put this into perspective, national and continent wide mains
networks are typically operated so that the 24 h average cycle count
exactly match the nominal mains frequency, so that synchronous clocks
would remain correct on average for long periods of time (days,
months).
However, the instantaneous frequency error can be several thousand
ppms and while the time error is usually kept below +/-30 s so that
clocks with hour and minute displays only would not be in error more
than for the last digit, this corresponds to up to +/-650.000 degrees
total phase error during the day.
Tim Wescott
I wrote the above today. Just specifying 100dBc is practically
meaningless -- any oscillator is going to have some phase noise, and
it's going to grow the closer you get to the carrier. White noise in
the amplifying element will turn into noise that goes as 1/(f-f0), and
no matter what you do you'll have 1/f noise in the amplifying element
that'll start cutting in at _some_ point. So you can always make any
60Hz oscillator meet a 100dBc requirement -- just find a bandwidth for
the measurement that makes it work. Ditto, you can take any oscillator
and narrow up the bandwidth around the carrier to make it fail just
about any dBc phase noise requirement. Even a super-zoot cesium
fountain oscillator is going to have "phase noise" that exceeds 100dBc
if you wait around long enough.
So I guess that all of the suggestions that have been given will work.
Or none of them. Or some, if only the OP would tell us the rest of his
requirement.
Geoff C
>
> So I guess that all of the suggestions that have been given will work.
> Or none of them. Or some, if only the OP would tell us the rest of
> his requirement.
>
The OP seems to be interested in syncing his PV solar system to the grid,
at least thats what I infer from reading some other of his posts. Kind of
makes the 100dBc spec look silly if so.
Tim Wescott
No kidding! If he's within 10 degrees one way or another that's
probably plenty good.
Of far greater concern with PV usage is making sure that putting what is
essentially a negative resistance on the line won't cause instability,
or at least knowing exactly what conditions will lead to instability so
that you may avoid them during installation.
Particularly if you're going to move from your lab with one or two PV
panels attached to a good solid grid, to some solar farm out in the
boonies where your PV array is the biggest power source for miles.
Phil Hobbs
Genuine phase noise sidebands have flat tops, so they aren't as
sensitive to modulation frequency as FM noise. Various authors go to
various lengths in trying to identify regions where the noise goes as
1/f, 1/f**2,.... I expect that the OP just wanted a 60 Hz oscillator
that was quiet enough that he didn't have to worry about it, due to
being hip deep in alligators of another species.
(I actually have a confirmed seat on the one flight from ATL to JFK that
looks like making it this afternoon--all the LGA and EWR (La Guardia and
Newark) folk are upschkrauen. Gotta run.)
j
> sensitive to modulation frequency as FM noise. Various authors go to
> various lengths in trying to identify regions where the noise goes as
> 1/f, 1/f**2,....
What in the world are you saying? Sounds kind of ignorant to me … but
I’ll reserve judgment until you answer.
I made a living at designing multiloop uw / rf synthesizers and taking
this statement as fact sure wouldn’t have helped.
regards
k...@att.bizzzzzzzzzzzz
<To-Email-Use-Th...@On-My-Web-Site.com> wrote:
Is anyone here younger than 20? 40? I know a few aren't yet 60. ;-)
Grant
>On 07/12/2010 11:42 PM, Geoff C wrote:
>>>
>>> So I guess that all of the suggestions that have been given will work.
>>> Or none of them. Or some, if only the OP would tell us the rest of
>>> his requirement.
>>>
>>
>> The OP seems to be interested in syncing his PV solar system to the grid,
>> at least thats what I infer from reading some other of his posts. Kind of
>> makes the 100dBc spec look silly if so.
>
>No kidding! If he's within 10 degrees one way or another that's
>probably plenty good.
Not quite so good, phase lead to pump energy to mains, lag a little to
suck energy, fully controlled bridges to the mains is scary stuff if
things go wrong.
Lock onto local mains, connect power, advance phase lead until desired
energy flows. 'Somehow'[1] detect loss of mains while you're driving
power into it so you don't create a powered island if the mains goes off.
Not something one designs without reference to local regs?
[1]Dithered local mains reference? I dunno, read about it, unlikely to
go there.
>
>Of far greater concern with PV usage is making sure that putting what is
>essentially a negative resistance on the line won't cause instability,
>or at least knowing exactly what conditions will lead to instability so
>that you may avoid them during installation.
>
>Particularly if you're going to move from your lab with one or two PV
>panels attached to a good solid grid, to some solar farm out in the
>boonies where your PV array is the biggest power source for miles.
That's for sure :) Why older systems had the big battery banks, some
now are using the mains as their battery, give energy during day, wind
meter back, take energy at night. Only works if a few do it?
Grant.
Phil Hobbs
Which statement? You don't think that pure phase noise is white? Or
that different authors say different things?
I plead guilty to ignorance of many things.
j
Yeah, I suspected you were talking about white noise.
Unfortunately for folks that design low noise freq synthesizers white
noise isn’t the tough spot. We typically live in those 1/f places.
The whole process is about shaping that noise profile. Without
targeted system spec’s, one can see why it’s virtually impossible to
select loop components such as a VCO’s, amps, etc., for this type of
job.
Btw, I apologize for the ignorant comment … didn’t mean to sound so
nasty. I regret that.
Paul Keinanen
wrote:
>On 07/12/2010 11:42 PM, Geoff C wrote:
>>>
>>> So I guess that all of the suggestions that have been given will work.
>>> Or none of them. Or some, if only the OP would tell us the rest of
>>> his requirement.
>>>
>>
>> The OP seems to be interested in syncing his PV solar system to the grid,
>> at least thats what I infer from reading some other of his posts. Kind of
>> makes the 100dBc spec look silly if so.
>
>No kidding! If he's within 10 degrees one way or another that's
>probably plenty good.
>
>Of far greater concern with PV usage is making sure that putting what is
>essentially a negative resistance on the line won't cause instability,
>or at least knowing exactly what conditions will lead to instability so
>that you may avoid them during installation.
>
>Particularly if you're going to move from your lab with one or two PV
>panels attached to a good solid grid, to some solar farm out in the
>boonies where your PV array is the biggest power source for miles.
In remote areas, the electricity distribution network is often like a
tree (not a ring) and becomes weaker when approaching the leaves of
the tree. Connecting one or more 1-3 MW wind turbines at the edges
will cause problems.
Ideally, it is assumed that the turbine should provide power to the
customers at the local branch. However, when operated close to the
cut-in wind speed, the power output will vary significantly, causing
voltage variations and the lights will flicker at nearby customers in
the weak net.
With wind turbines at different branches of a weak net, some operated
below average, some above average power, power is routed long
distances along weak lines through the common distribution point of
the original distribution net. The direction of power transfer in the
weak lines varies constantly, depending on the local wind variations.
In effect, the wind turbines in different branches have a slightly
different phase compared to each other and the main power grid.
When connecting individual PV or other small scale power sources to a
weak net, the control loops must be able to follow much faster phase
variations in weak nets with local generation, compared to a strong
network.
JosephKK
<pcdhSpamM...@electrooptical.net> wrote:
OK. I have your "Making it All Work" and AoE 2nd Ed and more. Where do
i go to get less confused? This phase noise measurement is twisted.
JosephKK
<pcdhSpamM...@electrooptical.net> wrote:
I understand your post NOT. Please explain more thoroughly, or point me
to texts. I still do not understand the way you are using units for
phase modulation. I have trouble understanding phase modulating a 60 Hz
carrier with a 1 MHz signal. What is the p-p angle you are achieving?
How do you know?
Sorry about sounding like a fool, but i cannot find common frame for us
to understand each other.
JosephKK
<To-Email-Use-Th...@On-My-Web-Site.com> wrote:
Larkin only wishes he were that young. No matter, he posts just like
such a brash young punk.
JosephKK
If that is the case, a simple twist on the standard PFC circuit will do
it. Though that may not meet all of the safety requirements, which are
rather difficult.
JosephKK
Actually it works even if a majority of residential and commercial
customers do it. The utility gets a distributed peaking plant without
capital investment and maintenance, and residential and commercial
customers get reduced energy costs, tax rebates, and feel good. Saves
the utility companies a bundle. Even makes the nutty regulators happy.
keit...@gmail.com
Except that it's not a "peaking" plant. The utility has no control
over this power. At best it's a (sometimes) baseline plant. A
significant share of the electricity generated this way would be
difficult to control.
John Larkin
"JosephKK"<quiett...@yahoo.com> wrote:
Of course I'd like to be young. Wouldn't you?
As far as behaving like a "brash young punk", I plead guilty, with
pleasure. I hope to keep designing better and faster, and skiing
steeper and faster, for a good long time. Electronics is not something
you have to give up designing after the age of 30.
I knew people who gave up electronics when transistors replaced tubes,
and people who refused to learn how to use programmable logic or
microprocessors or whatever. May as well move into a managed-care
facility and take up miniature golf.
John
Phil Hobbs
If you read the derivation in Section 13.6 and do the math, which isn't
difficult--just sums and differences of trig functions--we should be
talking the same language.
The main point is that we discuss small-amplitude phase noise using the
small angle approximation, i.e. sin theta ~= theta, so that it's just
like amplitude noise except that it's in phase quadrature with the
carrier. That makes it a bog-standard propagation-of-errors
calculation: you take all the noise sources, multiply them by the
relevant partial derivatives, and compute the RMS sum.
If you add white noise, half winds up in the I phase, which looks like
amplitude variations, and half in the Q phase, which looks like phase
variations. The small angle behaviour makes the statistics and
frequency spectrum of the resulting phase and amplitude noise equal to
those of the original additive noise. It's quite pretty.
When the SNR is below about 20 dB, we have to start being a lot more
careful mathematically.
Phil Hobbs
> On Jul 13, 7:41 pm, Phil Hobbs
> <pcdhSpamMeSensel...@electrooptical.net> wrote:
>> j wrote:
>>>> Genuine phase noise sidebands have flat tops, so they aren't as
>>>> sensitive to modulation frequency as FM noise. Various authors go to
>>>> various lengths in trying to identify regions where the noise goes as
>>>> 1/f, 1/f**2,....
>>
>>> What in the world are you saying? Sounds kind of ignorant to me … but
>>> I’ll reserve judgment until you answer.
>>
>>> I made a living at designing multiloop uw / rf synthesizers and taking
>>> this statement as fact sure wouldn’t have helped.
>>
>>> regards
>>
>> Which statement? You don't think that pure phase noise is white? Or
>> that different authors say different things?
>>
>> I plead guilty to ignorance of many things.
>>
>
> Yeah, I suspected you were talking about white noise.
>
> Unfortunately for folks that design low noise freq synthesizers white
> noise isn’t the tough spot. We typically live in those 1/f places.
> The whole process is about shaping that noise profile. Without
> targeted system spec’s, one can see why it’s virtually impossible to
> select loop components such as a VCO’s, amps, etc., for this type of
> job.
>
Understood. My first engineering job was designing 2/3 of the time and
frequency reference boards for the first direct-broadcast satellite
system (the Spacetel system from AEL Microtel), including the VHF
synthesizer that controlled the 14 GHz local oscillator on both the
central station and the remotes. The noise spec was 7 Hz RMS in 5-100
Hz bandwidth around a 14 GHz carrier, i.e. after being multiplied up by
120 times from the output of my board. I was allocated half of this
budget, i.e. 5 Hz RMS at 14 GHz, or 0.041 Hz at 115 MHz, and my
synthesizer had to be tunable over about 5 MHz in steps of 8.3333... kHz
(1 MHz at the LO frequency). This was in 1981-83, remember, which was
well before DDS.
I had no idea how hard that was before I started--I had a brand new
astronomy and physics B.Sc., and only a hobby background in electronics
(though I did start when I was 10 years old). I knew about PLLs, but
I'd never seen one, let alone designed one. Talk about being chucked in
the deep end of the pool! (I eventually made a fairly novel
fractional-N synthesizer with an 833.33... kHz comparison frequency,
using a MC12013 ECL 10/11 dual modulus prescaler(*) with the modulus
controlled by a 74LS163 synchronous decade counter, whose carry input
was driven by a string of CD4527 CMOS BCD rate multipliers. The rate
multiplier jitter was pretty well outside the LO PLL's bandwidth, and
certainly wasn't in the 5-100 Hz band that they mostly cared about. One
nice feature was that you didn't need a switch setting table--because it
was BCD and used a synchronous divide-by-12 to get the reference, you
just set the BCD DIP switches to the desired 14 GHz LO frequency.
It all eventually worked fine, thanks to Floyd Gardner's book and the
Mini-Circuits catalogue. Its main wart was that even with lead-lag
compensation, I _still_ couldn't get enough loop gain to control the
noise of an LC VCO with a 5 MHz tuning range. I eventually had to
retreat and use a VCXO, meaning that you had to change a crystal as well
as set the DIP switches. (Changing crystals was SOP in those days, so
nobody minded too much.) The oscillator was a one-transistor Colpitts
with self-limiting, just like in the ARRL Handbook. If I'd known how to
design better oscillators, or had had coaxial resonators, I could
probably have kept real tunability. Such is life.
> Btw, I apologize for the ignorant comment … didn’t mean to sound so
> nasty. I regret that.
No worries. The plus side of having large areas of ignorance is the
opportunity to learn lots of new things, which is one reason I like SED
in spite of the spam, troll baiting, and flame wars. I think that
standing at a white board arguing about technical stuff with a few smart
people is the most fun you can have standing up. (I have a directory
full of pithy Usenet posts that I refer to periodically--some of the
stuff in my book came out of things I learned here.)
Cheers
Phil Hobbs
(*) Did you design that one, Jim?
George Herold
> On Mon, 12 Jul 2010 23:23:56 -0400, Phil Hobbs
>
>
>
>
>
> <pcdhSpamMeSensel...@electrooptical.net> wrote:
> >JosephKK wrote:
> >> On Mon, 12 Jul 2010 09:37:07 -0400, Phil Hobbs
> >> <pcdhSpamMeSensel...@electrooptical.net> wrote:
>
> >>> JosephKK wrote:
>
> - Show quoted text -
Hi Joseph, I'm trying to get my head around this too. (I like Phil's
intro to section 13.6, "We live in a fallen world, so the signals we
process are never free of noise, distortion, and extraneous
interfering signals.")*
I think it would help me if I understood how one measures the phase
noise. My simple minded approach would be to trigger my digital
'scope on the carrier zero crossing, and then look 'down stream' 100
or 1,000 periods later and see how much 'jitter' there was in the
delayed zero crossing. Seems like there must be a better way.
George H.
*Does this mean there is no noise in heaven? (all R's have zero
temperature)
Joel Koltner
> Understood. My first engineering job was designing 2/3 of the time
> and frequency reference boards for the first direct-broadcast
> satellite system (the Spacetel system from AEL Microtel), including
> the VHF synthesizer that controlled the 14 GHz local oscillator on
> both the central station and the remotes.
If you don't mind my asking, Phil, how long did that all take?
With the synthesizers-in-a-chip (PLL+VCO), off-the-shelf VCOs, plus the design
tools from Analog Devices, etc., I have a suspicion that many such RF
generators are now given all of perhaps a day or two of design time. :-)
---Joel
Phil Hobbs
I joined Microtel in about June of 1981, and left in August 1983 to get
married and go to grad school. The first few months were working on the
system demo, and the rest of it was spent getting the Pilot Tone
Generator and Timing & Frequency Unit designed, breadboarded (no SPICE
either), laid out, and tested with the 14 GHz LO setup. Plus
miscellaneous other stuff.
So I'd say a year, give or take. I had a fair few false starts in
there, too, of course--such as trying to use anything with an 8 kHz
comparison frequency!
Cheers
Phil Hobbs
Phil Hobbs
> On Jul 14, 12:49 am, "JosephKK"<quiettechb...@yahoo.com> wrote:
>> OK. I have your "Making it All Work" and AoE 2nd Ed and more. Where do
>> i go to get less confused? This phase noise measurement is twisted.- Hide quoted text -
>>
>> - Show quoted text -
>
> Hi Joseph, I'm trying to get my head around this too. (I like Phil's
> intro to section 13.6, "We live in a fallen world, so the signals we
> process are never free of noise, distortion, and extraneous
> interfering signals.")*
>
> I think it would help me if I understood how one measures the phase
> noise. My simple minded approach would be to trigger my digital
> 'scope on the carrier zero crossing, and then look 'down stream' 100
> or 1,000 periods later and see how much 'jitter' there was in the
> delayed zero crossing. Seems like there must be a better way.
One approach is to phase lock a quiet oscillator to the unknown one,
with a very narrow loop bandwidth, and look at the noise at the phase
detector output.
>
> *Does this mean there is no noise in heaven? (all R's have zero
> temperature)
In _The Screwtape Letters_, C. S. Lewis quotes George Macdonald on the
subject of Heaven: "...the regions where there is only life, and
therefore all that is not music is silence". (Unspoken Sermons, Vol 1)
So, no noise there. ;)
John Larkin
That's pretty much what the Allan Variance does,
http://en.wikipedia.org/wiki/Allan_variance
which is to characterize the jitter as a function of the time, or
equivalently the number of skipped edges, between the edges you
measure. Cheap crystal oscillators may have a few ps RMS between
adjacent clock edges (aka cycle-to-cycle jitter) but may have many
nanoseconds of jitter if you measure edges that are a second apart.
You've got to be careful about the scope jitter, too. It may well be
worse then the oscillator you're trying to measure.
The "real" way to measure phase noise in the frequency domain is to
get two of the things you want to test, set them to slightly different
frequencies, and mix their outputs, then analyse the mixer output.
That's about the only way to characterize really quiet sources. Of
course you have to make sure they aren't injection locking or showing
the same line-hum jitter or any other sneaky correlation.
We have a couple of 10 MHz atomic clocks around here, one rubidium and
one cesium. If you trigger a scope from one and look at the rising
edge of the other, you could swear that the scope is triggered
internally. At, say, 5 ns/div, it looks rock steady. Come back a half
hour later and the trace has drifted a little left or right.
That test is the time-domain equivalent of the mixer thing. That's how
we test the timing system stuff we did for NIF: trigger a scope from
one unit, observe the rising edge of another.
John
Bob Masta
<ghe...@teachspin.com> wrote:
>*Does this mean there is no noise in heaven? (all R's have zero
>temperature)
All the Real terms are zero... heaven is purely Imaginary!
<g>
Best regards,
Bob Masta
DAQARTA v5.10
Data AcQuisition And Real-Time Analysis
www.daqarta.com
Scope, Spectrum, Spectrogram, Sound Level Meter
Frequency Counter, FREE Signal Generator
Pitch Track, Pitch-to-MIDI
DaqMusic - FREE MUSIC, Forever!
(Some assembly required)
Science (and fun!) with your sound card!
George Herold
OK that sounds easy enough... you then have to add a third oscillator
and measure each againts the other if you want to really know any
individual bandwidth.
>
> We have a couple of 10 MHz atomic clocks around here, one rubidium and
> one cesium. If you trigger a scope from one and look at the rising
> edge of the other, you could swear that the scope is triggered
> internally. At, say, 5 ns/div, it looks rock steady. Come back a half
> hour later and the trace has drifted a little left or right.
Fun! I did the same trick when comparing digital function
generators. This cheap protek one had a 'hick up' every second or so
and the transistion would slowly march across the screen.
>
> That test is the time-domain equivalent of the mixer thing. That's how
> we test the timing system stuff we did for NIF: trigger a scope from
> one unit, observe the rising edge of another.
>
> John- Hide quoted text -
>
> - Show quoted text -
Thanks again John and Phil.
George H.
John Larkin
Usually just dumping the mixer output into a spectrum analyzer is good
enough. One assumes the units are identical. You can use three or more
DUTs, in pairs, to demonstrate that.
>>
>> We have a couple of 10 MHz atomic clocks around here, one rubidium and
>> one cesium. If you trigger a scope from one and look at the rising
>> edge of the other, you could swear that the scope is triggered
>> internally. At, say, 5 ns/div, it looks rock steady. Come back a half
>> hour later and the trace has drifted a little left or right.
>
>Fun! I did the same trick when comparing digital function
>generators. This cheap protek one had a 'hick up' every second or so
>and the transistion would slowly march across the screen.
One of my Tek scopes has a few ps of time shift that correlates to a
blinking led on the sampling head.
John