On Mon, 19 Nov 2012 09:51:44 +0100, o pere o <me@somewhere.net> wrote:
>On 11/17/2012 02:27 AM, Jim Thompson wrote:
>> On Wed, 14 Nov 2012 16:58:13 +0100, o pere o <me@somewhere.net> wrote:
>>
>>> The thread related to large signal PSpice models and an emitter follower
>>> comes from the following problem:
>>>
>>> I have an oscillator that should drive a digital part of the system.
>>> In short, what is the best way to achieve this?
>>>
>>> My first attempt has been a common base Colpitts oscillator that gives a
>>> signal riding on the +Vcc rail. This has been AC coupled to a 74AC gate
>>> biased to the point that gives square output signals.
>>>
>>> This works more ore less ok, but: the startup transient, which is
>>> important here, is different depending if the oscillator amplitude is
>>> sufficient to toggle the gate. This translates into an envelope that
>>> raises more ore less smoothly until the gate begins toggling, where the
>>> envelope raises more abruptly -and I guess that the instantaneous
>>> frequency changes.
>>>
>>> I have thought of two causes for this. The first one is feedback via the
>>> DC supply: the spikes generated by the gate switching get coupled back
>>> to the oscillator. The second one could be the change in input impedance
>>> seen by the oscillator -does this make sense? The cure for #1 could be
>>> better supply bypassing. The cure for #2 a buffer stage.
>>>
>>> So, what could be a good way to generate a digital signal from an
>>> oscillator without loading it? Ideally I would like to preserve the
>>> instantaneous frequency of the unloaded startup transient. And: power
>>> consumption should be low, say preferably (much) less than 1 mA.
>>> Operating frequency should be initially 27 MHz, but ideally scalable up
>>> to ~1 GHz.
>>>
>>> Pere
>>
>> Maybe buy the MC12148 equivalent of my mid-60's design...
>>
>> http://www.analog-innovations.com/SED/MC1648-DataSheet.pdf
>>
>> Which has a buffered PECL output, which can be easily
>> capacitor-coupled to CMOS.
>>
>> ...Jim Thompson
>>
>
>I had used the 1648 in the past! A simple and nice VCO indeed. Now it
>seems the only successor is the MC100EL1648 from OnSemi, but I may have
>overlooked something.
>
>I remember having tried to use that chip to make a quenchable oscillator
>but with modest success. Perhaps you could suggest a clever way to tweak
>the gain to allow for this? Power consumption would still not be within
>the desired range, but I might learn something from it.
>
>Pere
Pulling AGC pin low will "quench" the oscillator... anything below
~1.4V stops it.
Maybe use an open-drain or open-collector device.
...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 |
I love to cook with wine. Sometimes I even put it in the food.
Reply by John Larkin●November 19, 20122012-11-19
On Mon, 19 Nov 2012 19:15:55 +0100, o pere o <me@somewhere.net> wrote:
>On 11/19/2012 04:14 PM, John Larkin wrote:
>
><snip>
>
>>>> Here's the LLarkin oscillator, which uses a MMIC and a couple of
>>>> inductors in the gain path. This works well with coaxial resonators.
>>>> Too much power for your app, unfortunately.
>>>>
>>> <snipped circuit>
>>> Why don't you use small C's instead of L2 and L3? That should give
>>> similar impedance transformation at less expense...
>>
>>
>> The mmic has something over 180 degrees of phase lag, and the two Ls
>> can be sized to make up the rest of 360. If the gain element's net
>> phase shift is zero, the thing will oscillate at the unloaded resonant
>> peak frequency of the resonator, which should give high Q. I think.
>> Anyhow, it works. It seems to do gated/burst oscillators pretty well
>> with ceramic coaxial resonators.
>>
>
>Ok. But to take profit from the resonator Q you have to be careful with
>the loading. And you will have trouble finding inductors that achieve
>the required high reactance (for instance, in your example, 5 uH at 100
>MHz is not easy). With capacitors you don't have this problem.
>
>And, at higher frequencies, you can just add some transmission line
>length to tweak the phase.
>
>Pere
I just threw together that LT Spice thing to show the concept... the
frequency just accidentally came out to 100 MHz. We use a 600 MHz
coaxial resonator on one product, and that's about the lowest
frequency resonator that you can buy. The inductors are reasonable at
600 MHz.
I TDRd some coaxial ceramic resonators, and they really do look like
shorted transmission lines, in the 10-15 ohm range. So a good
simulation would not treat them as an LC tank, but as a real
transmission line.
We also do a lot of gated/burst LC oscillators, generally in the 50
MHz sort of range. I usually use a tiny LVDS-to-CMOS converter chip,
as the comparator and gain element all at once. Done right, the first
output clock rips through a few ns after the gate comes up, and
successive edges are right on schedule +- not too many picoseconds.
--
John Larkin Highland Technology, Inc
jlarkin at highlandtechnology dot com
http://www.highlandtechnology.com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom laser drivers and controllers
Photonics and fiberoptic TTL data links
VME thermocouple, LVDT, synchro acquisition and simulation
Reply by o pere o●November 19, 20122012-11-19
On 11/19/2012 04:14 PM, John Larkin wrote:
<snip>
>>> Here's the LLarkin oscillator, which uses a MMIC and a couple of
>>> inductors in the gain path. This works well with coaxial resonators.
>>> Too much power for your app, unfortunately.
>>>
>> <snipped circuit>
>> Why don't you use small C's instead of L2 and L3? That should give
>> similar impedance transformation at less expense...
>
>
> The mmic has something over 180 degrees of phase lag, and the two Ls
> can be sized to make up the rest of 360. If the gain element's net
> phase shift is zero, the thing will oscillate at the unloaded resonant
> peak frequency of the resonator, which should give high Q. I think.
> Anyhow, it works. It seems to do gated/burst oscillators pretty well
> with ceramic coaxial resonators.
>
Ok. But to take profit from the resonator Q you have to be careful with
the loading. And you will have trouble finding inductors that achieve
the required high reactance (for instance, in your example, 5 uH at 100
MHz is not easy). With capacitors you don't have this problem.
And, at higher frequencies, you can just add some transmission line
length to tweak the phase.
Pere
Reply by o pere o●November 19, 20122012-11-19
On 11/19/2012 05:09 PM, Jim Thompson wrote:
> On Mon, 19 Nov 2012 09:51:44 +0100, o pere o <me@somewhere.net> wrote:
>
>> On 11/17/2012 02:27 AM, Jim Thompson wrote:
>>> On Wed, 14 Nov 2012 16:58:13 +0100, o pere o <me@somewhere.net> wrote:
>>>
>>>> The thread related to large signal PSpice models and an emitter follower
>>>> comes from the following problem:
>>>>
>>>> I have an oscillator that should drive a digital part of the system.
>>>> In short, what is the best way to achieve this?
>>>>
>>>> My first attempt has been a common base Colpitts oscillator that gives a
>>>> signal riding on the +Vcc rail. This has been AC coupled to a 74AC gate
>>>> biased to the point that gives square output signals.
>>>>
>>>> This works more ore less ok, but: the startup transient, which is
>>>> important here, is different depending if the oscillator amplitude is
>>>> sufficient to toggle the gate. This translates into an envelope that
>>>> raises more ore less smoothly until the gate begins toggling, where the
>>>> envelope raises more abruptly -and I guess that the instantaneous
>>>> frequency changes.
>>>>
>>>> I have thought of two causes for this. The first one is feedback via the
>>>> DC supply: the spikes generated by the gate switching get coupled back
>>>> to the oscillator. The second one could be the change in input impedance
>>>> seen by the oscillator -does this make sense? The cure for #1 could be
>>>> better supply bypassing. The cure for #2 a buffer stage.
>>>>
>>>> So, what could be a good way to generate a digital signal from an
>>>> oscillator without loading it? Ideally I would like to preserve the
>>>> instantaneous frequency of the unloaded startup transient. And: power
>>>> consumption should be low, say preferably (much) less than 1 mA.
>>>> Operating frequency should be initially 27 MHz, but ideally scalable up
>>>> to ~1 GHz.
>>>>
>>>> Pere
>>>
>>> Maybe buy the MC12148 equivalent of my mid-60's design...
>>>
>>> http://www.analog-innovations.com/SED/MC1648-DataSheet.pdf
>>>
>>> Which has a buffered PECL output, which can be easily
>>> capacitor-coupled to CMOS.
>>>
>>> ...Jim Thompson
>>>
>>
>> I had used the 1648 in the past! A simple and nice VCO indeed. Now it
>> seems the only successor is the MC100EL1648 from OnSemi, but I may have
>> overlooked something.
>>
>> I remember having tried to use that chip to make a quenchable oscillator
>> but with modest success. Perhaps you could suggest a clever way to tweak
>> the gain to allow for this? Power consumption would still not be within
>> the desired range, but I might learn something from it.
>>
>> Pere
>
> Do you need to quench the oscillator itself, or just the buffered
> output? The output is simple PECL.
>
> ...Jim Thompson
>
It is the oscillator that has to be quenched. I recall fiddling around
with ramping up the Vcc of the oscillator part (with the last original
1648's that I had left) and perhaps I even tried something else -it is
quite a long time ago. Do you see a possibility via the AGC pin, for
instance?
Pere
Reply by Jim Thompson●November 19, 20122012-11-19
On Mon, 19 Nov 2012 09:51:44 +0100, o pere o <me@somewhere.net> wrote:
>On 11/17/2012 02:27 AM, Jim Thompson wrote:
>> On Wed, 14 Nov 2012 16:58:13 +0100, o pere o <me@somewhere.net> wrote:
>>
>>> The thread related to large signal PSpice models and an emitter follower
>>> comes from the following problem:
>>>
>>> I have an oscillator that should drive a digital part of the system.
>>> In short, what is the best way to achieve this?
>>>
>>> My first attempt has been a common base Colpitts oscillator that gives a
>>> signal riding on the +Vcc rail. This has been AC coupled to a 74AC gate
>>> biased to the point that gives square output signals.
>>>
>>> This works more ore less ok, but: the startup transient, which is
>>> important here, is different depending if the oscillator amplitude is
>>> sufficient to toggle the gate. This translates into an envelope that
>>> raises more ore less smoothly until the gate begins toggling, where the
>>> envelope raises more abruptly -and I guess that the instantaneous
>>> frequency changes.
>>>
>>> I have thought of two causes for this. The first one is feedback via the
>>> DC supply: the spikes generated by the gate switching get coupled back
>>> to the oscillator. The second one could be the change in input impedance
>>> seen by the oscillator -does this make sense? The cure for #1 could be
>>> better supply bypassing. The cure for #2 a buffer stage.
>>>
>>> So, what could be a good way to generate a digital signal from an
>>> oscillator without loading it? Ideally I would like to preserve the
>>> instantaneous frequency of the unloaded startup transient. And: power
>>> consumption should be low, say preferably (much) less than 1 mA.
>>> Operating frequency should be initially 27 MHz, but ideally scalable up
>>> to ~1 GHz.
>>>
>>> Pere
>>
>> Maybe buy the MC12148 equivalent of my mid-60's design...
>>
>> http://www.analog-innovations.com/SED/MC1648-DataSheet.pdf
>>
>> Which has a buffered PECL output, which can be easily
>> capacitor-coupled to CMOS.
>>
>> ...Jim Thompson
>>
>
>I had used the 1648 in the past! A simple and nice VCO indeed. Now it
>seems the only successor is the MC100EL1648 from OnSemi, but I may have
>overlooked something.
>
>I remember having tried to use that chip to make a quenchable oscillator
>but with modest success. Perhaps you could suggest a clever way to tweak
>the gain to allow for this? Power consumption would still not be within
>the desired range, but I might learn something from it.
>
>Pere
Do you need to quench the oscillator itself, or just the buffered
output? The output is simple PECL.
...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 |
I love to cook with wine. Sometimes I even put it in the food.
Reply by John Larkin●November 19, 20122012-11-19
On Mon, 19 Nov 2012 10:05:29 +0100, o pere o <me@somewhere.net> wrote:
>On 11/16/2012 04:22 PM, John Larkin wrote:
>> On Fri, 16 Nov 2012 12:04:00 +0100, o pere o <me@somewhere.net> wrote:
>>
>>> On 11/16/2012 01:25 AM, George Herold wrote:
>>>
>>> <snip>
>>>
>>>> Nice thread, (wiki calls the tickler the Armstrong.)
>>>> The only RF oscillator I know is used to drive a Rb discharge lamp.
>>>> It's a Hartly basically copied from an efratom lamp circuit.
>>>>
>>>> To Opere, I don't quite get your problem.
>>>> To sense the circuit you're going to have to take a bit of energy
>>>> out.
>>>> This must change the Q and (thus) the resonant frequency.
>>>> If it's a changing Q when you switch in your circuit.. (?)
>>>> then you have to balance it out.
>>>>
>>>> George H.
>>>
>>>
>>> Current oscillator version is
>>>
>>> Vcc Vcc Vcc
>>> | | |
>>> C L R1
>>> | | |
>>> | C-*--*-----C---*----Dig. Gate
>>> *-B C1 |
>>> | E-*-* R2
>>> R | | |
>>> | Re C2 gnd
>>> ctrl | |
>>> gnd
>>>
>>> If the Gate impedance were constant for small and large signal, this
>>> would be ok (controlled loading). What I am seeing is that when the Gate
>>> sees enough signal to begin toggling the transient changes. Hope this
>>> sketch helps explaining:
>>>
>>> envelope: *************
>>> *
>>> *
>>> *****
>>> ***** |
>>> 0****** | --------------------------0
>>> <smooth rising> | <saturated>
>>> |
>>> start toggling-> (very) fast rising
>>>
>>> As mentioned in my original post, I see 2 possible causes for this: 1.
>>> Coupling through the Vcc rail (although more ore less well bypassed)
>>> and 2. Different behavior of the gate for small and large signal.
>>>
>>> The cure for 2 is what I am looking for. Cure for 1 would be better
>>> bypassing.
>>>
>>> The other configuration I have used (at ~900MHz) is
>>>
>>> Ctrl-
>>> Rb C---Vcc
>>> --C-*---B
>>> | | E
>>> L C1 |
>>> | | |
>>> | *----*
>>> | C2 Re
>>> | | |
>>> gnd gnd gnd
>>>
>>> And a similar one, with a transmission line instead of L has been used
>>> at 2.4 GHz.
>>>
>>> A minicircuits gain block with a hairpin resonator has also worked ok at
>>> 2.4 GHz, although this was not low-power (tens of mA):
>>>
>>> �---Amp---�
>>> | |
>>> �-� �---� <-Phasing lines
>>> || ||
>>> || || <--Resonator
>>> || ||
>>> �-�
>>>
>>> (Maximum simplicity)
>>
>>
>> Here's the LLarkin oscillator, which uses a MMIC and a couple of
>> inductors in the gain path. This works well with coaxial resonators.
>> Too much power for your app, unfortunately.
>>
><snipped circuit>
>Why don't you use small C's instead of L2 and L3? That should give
>similar impedance transformation at less expense...
The mmic has something over 180 degrees of phase lag, and the two Ls
can be sized to make up the rest of 360. If the gain element's net
phase shift is zero, the thing will oscillate at the unloaded resonant
peak frequency of the resonator, which should give high Q. I think.
Anyhow, it works. It seems to do gated/burst oscillators pretty well
with ceramic coaxial resonators.
--
John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
Reply by o pere o●November 19, 20122012-11-19
On 11/16/2012 04:22 PM, John Larkin wrote:
> On Fri, 16 Nov 2012 12:04:00 +0100, o pere o <me@somewhere.net> wrote:
>
>> On 11/16/2012 01:25 AM, George Herold wrote:
>>
>> <snip>
>>
>>> Nice thread, (wiki calls the tickler the Armstrong.)
>>> The only RF oscillator I know is used to drive a Rb discharge lamp.
>>> It's a Hartly basically copied from an efratom lamp circuit.
>>>
>>> To Opere, I don't quite get your problem.
>>> To sense the circuit you're going to have to take a bit of energy
>>> out.
>>> This must change the Q and (thus) the resonant frequency.
>>> If it's a changing Q when you switch in your circuit.. (?)
>>> then you have to balance it out.
>>>
>>> George H.
>>
>>
>> Current oscillator version is
>>
>> Vcc Vcc Vcc
>> | | |
>> C L R1
>> | | |
>> | C-*--*-----C---*----Dig. Gate
>> *-B C1 |
>> | E-*-* R2
>> R | | |
>> | Re C2 gnd
>> ctrl | |
>> gnd
>>
>> If the Gate impedance were constant for small and large signal, this
>> would be ok (controlled loading). What I am seeing is that when the Gate
>> sees enough signal to begin toggling the transient changes. Hope this
>> sketch helps explaining:
>>
>> envelope: *************
>> *
>> *
>> *****
>> ***** |
>> 0****** | --------------------------0
>> <smooth rising> | <saturated>
>> |
>> start toggling-> (very) fast rising
>>
>> As mentioned in my original post, I see 2 possible causes for this: 1.
>> Coupling through the Vcc rail (although more ore less well bypassed)
>> and 2. Different behavior of the gate for small and large signal.
>>
>> The cure for 2 is what I am looking for. Cure for 1 would be better
>> bypassing.
>>
>> The other configuration I have used (at ~900MHz) is
>>
>> Ctrl-
>> Rb C---Vcc
>> --C-*---B
>> | | E
>> L C1 |
>> | | |
>> | *----*
>> | C2 Re
>> | | |
>> gnd gnd gnd
>>
>> And a similar one, with a transmission line instead of L has been used
>> at 2.4 GHz.
>>
>> A minicircuits gain block with a hairpin resonator has also worked ok at
>> 2.4 GHz, although this was not low-power (tens of mA):
>>
>> �---Amp---�
>> | |
>> �-� �---� <-Phasing lines
>> || ||
>> || || <--Resonator
>> || ||
>> �-�
>>
>> (Maximum simplicity)
>
>
> Here's the LLarkin oscillator, which uses a MMIC and a couple of
> inductors in the gain path. This works well with coaxial resonators.
> Too much power for your app, unfortunately.
>
<snipped circuit>
Why don't you use small C's instead of L2 and L3? That should give
similar impedance transformation at less expense...
For mmic and coaxial resonator I have used a tapped C at the input and
output:
--C--*----*-----C--*----
| | |
C res C
| | |
-----*----*--------*-----
Pere
Reply by o pere o●November 19, 20122012-11-19
On 11/17/2012 02:27 AM, Jim Thompson wrote:
> On Wed, 14 Nov 2012 16:58:13 +0100, o pere o <me@somewhere.net> wrote:
>
>> The thread related to large signal PSpice models and an emitter follower
>> comes from the following problem:
>>
>> I have an oscillator that should drive a digital part of the system.
>> In short, what is the best way to achieve this?
>>
>> My first attempt has been a common base Colpitts oscillator that gives a
>> signal riding on the +Vcc rail. This has been AC coupled to a 74AC gate
>> biased to the point that gives square output signals.
>>
>> This works more ore less ok, but: the startup transient, which is
>> important here, is different depending if the oscillator amplitude is
>> sufficient to toggle the gate. This translates into an envelope that
>> raises more ore less smoothly until the gate begins toggling, where the
>> envelope raises more abruptly -and I guess that the instantaneous
>> frequency changes.
>>
>> I have thought of two causes for this. The first one is feedback via the
>> DC supply: the spikes generated by the gate switching get coupled back
>> to the oscillator. The second one could be the change in input impedance
>> seen by the oscillator -does this make sense? The cure for #1 could be
>> better supply bypassing. The cure for #2 a buffer stage.
>>
>> So, what could be a good way to generate a digital signal from an
>> oscillator without loading it? Ideally I would like to preserve the
>> instantaneous frequency of the unloaded startup transient. And: power
>> consumption should be low, say preferably (much) less than 1 mA.
>> Operating frequency should be initially 27 MHz, but ideally scalable up
>> to ~1 GHz.
>>
>> Pere
>
> Maybe buy the MC12148 equivalent of my mid-60's design...
>
> http://www.analog-innovations.com/SED/MC1648-DataSheet.pdf
>
> Which has a buffered PECL output, which can be easily
> capacitor-coupled to CMOS.
>
> ...Jim Thompson
>
I had used the 1648 in the past! A simple and nice VCO indeed. Now it
seems the only successor is the MC100EL1648 from OnSemi, but I may have
overlooked something.
I remember having tried to use that chip to make a quenchable oscillator
but with modest success. Perhaps you could suggest a clever way to tweak
the gain to allow for this? Power consumption would still not be within
the desired range, but I might learn something from it.
Pere
Reply by Jim Thompson●November 16, 20122012-11-16
On Wed, 14 Nov 2012 16:58:13 +0100, o pere o <me@somewhere.net> wrote:
>The thread related to large signal PSpice models and an emitter follower
>comes from the following problem:
>
>I have an oscillator that should drive a digital part of the system.
>In short, what is the best way to achieve this?
>
>My first attempt has been a common base Colpitts oscillator that gives a
>signal riding on the +Vcc rail. This has been AC coupled to a 74AC gate
>biased to the point that gives square output signals.
>
>This works more ore less ok, but: the startup transient, which is
>important here, is different depending if the oscillator amplitude is
>sufficient to toggle the gate. This translates into an envelope that
>raises more ore less smoothly until the gate begins toggling, where the
>envelope raises more abruptly -and I guess that the instantaneous
>frequency changes.
>
>I have thought of two causes for this. The first one is feedback via the
>DC supply: the spikes generated by the gate switching get coupled back
>to the oscillator. The second one could be the change in input impedance
>seen by the oscillator -does this make sense? The cure for #1 could be
>better supply bypassing. The cure for #2 a buffer stage.
>
>So, what could be a good way to generate a digital signal from an
>oscillator without loading it? Ideally I would like to preserve the
>instantaneous frequency of the unloaded startup transient. And: power
>consumption should be low, say preferably (much) less than 1 mA.
>Operating frequency should be initially 27 MHz, but ideally scalable up
>to ~1 GHz.
>
>Pere
Maybe buy the MC12148 equivalent of my mid-60's design...
http://www.analog-innovations.com/SED/MC1648-DataSheet.pdf
Which has a buffered PECL output, which can be easily
capacitor-coupled to CMOS.
...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 |
I love to cook with wine. Sometimes I even put it in the food.
Reply by Bill Sloman●November 16, 20122012-11-16
On Nov 17, 12:39=A0am, o pere o <m...@somewhere.net> wrote:
> On 11/16/2012 12:16 PM,BillSlomanwrote:
>
>
>
>
>
>
>
>
>
> > On Nov 15, 2:58 am, o pere o <m...@somewhere.net> wrote:
> >> The thread related to large signal PSpice models and an emitter follow=
er
> >> comes from the following problem:
>
> >> I have an oscillator that should drive a digital part of the system.
> >> In short, what is the best way to achieve this?
>
> >> My first attempt has been a common base Colpitts oscillator that gives=
a
> >> signal riding on the +Vcc rail. This has been AC coupled to a 74AC gat=
e
> >> biased to the point that gives square output signals.
>
> >> This works more ore less ok, but: the startup transient, which is
> >> important here, is different depending if the oscillator amplitude is
> >> sufficient to toggle the gate. This translates into an envelope that
> >> raises more ore less smoothly until the gate begins toggling, where th=
e
> >> envelope raises more abruptly -and I guess that the instantaneous
> >> frequency changes.
>
> >> I have thought of two causes for this. The first one is feedback via t=
he
> >> DC supply: the spikes generated by the gate switching get coupled back
> >> to the oscillator. The second one could be the change in input impedan=
ce
> >> seen by the oscillator -does this make sense? The cure for #1 could be
> >> better supply bypassing. The cure for #2 a buffer stage.
>
> >> So, what could be a good way to generate a digital signal from an
> >> oscillator without loading it? Ideally I would like to preserve the
> >> instantaneous frequency of the unloaded startup transient. And: power
> >> consumption should be low, say preferably (much) less than 1 mA.
> >> Operating frequency should =A0be initially 27 MHz, but ideally scalabl=
e up
> >> to ~1 GHz.
>
> > One weird and expensive approach would be to use something like an
> > AD834 as your gain stage; set it up with enough initial gain to get
> > the oscillator to start up respectably fast, then drop the gain back
> > to a level that sustains the oscillation when the digital logic
> > detects an edge. If you can live with a little bit of clipping, the
> > "sustain" gain wouldn't have to be too well defined.
>
> >http://www.analog.com/static/imported-files/data_sheets/AD834.pdf
>
> > The power consumption is rather higher than you want.
>
> > Multipliers are designed to have the same gain over a respectable
> > range of input amplitudes, so the start-up ought to be well-defined.
>
> A multiplier would offer a better control of gain, which could translate
> into a bigger linear range of operation. Of course, at the price of cost
> and power consumption but it could serve as an idealized prototype.
> However, when I played with analog multipliers in the past (at
> frequencies much much lower than the 500MHz of your suggestion), they
> did not perform as well as announced in the datasheets -it could have
> been my fault...
The Analog Devices multipliers all have added extra Barry Gilbert. He
invented the concept and has run with it for a long time now. There
are a variety of copies available, all much cheaper, but not as good.
--
Bill Sloman, Sydney