Forums

GHz oscillator

Started by John Larkin April 27, 2021
On Fri, 30 Apr 2021 16:19:36 -0700, John Larkin
<jlarkin@highland_atwork_technology.com> wrote:

>On Fri, 30 Apr 2021 13:26:39 -0700 (PDT), Lasse Langwadt Christensen ><langwadt@fonz.dk> wrote: > >>fredag den 30. april 2021 kl. 21.31.52 UTC+2 skrev John Larkin: >>> On Fri, 30 Apr 2021 11:10:12 -0700 (PDT), Lasse Langwadt Christensen >>> <lang...@fonz.dk> wrote: >>> >>> >fredag den 30. april 2021 kl. 19.51.58 UTC+2 skrev John Larkin: >>> >> On Fri, 30 Apr 2021 10:09:59 -0700 (PDT), Lasse Langwadt Christensen >>> >> <lang...@fonz.dk> wrote: >>> >> >>> >> >fredag den 30. april 2021 kl. 19.04.03 UTC+2 skrev John Larkin: >>> >> >> On Fri, 30 Apr 2021 19:54:36 +0300, upsid...@downunder.com wrote: >>> >> >> >>> >> >> >On Fri, 30 Apr 2021 08:45:30 -0700, jla...@highlandsniptechnology.com >>> >> >> >wrote: >>> >> >> > >>> >> >> >>On Fri, 30 Apr 2021 08:07:10 +0300, upsid...@downunder.com wrote: >>> >> >> >> >>> >> >> >>>On Tue, 27 Apr 2021 12:35:17 -0700, John Larkin >>> >> >> >>><jlarkin@highland_atwork_technology.com> wrote: >>> >> >> >>> >>> >> >> >>>>What's a good way to make a, say 2 or 3 GHz sinewave oscillator? I >>> >> >> >>>>don't need extreme accuracy, and I'd like maybe 1 volt p-p. >>> >> >> >>>> >>> >> >> >>>>I was thinking a cheap MMIC with PCB delay line feedback. >>> >> >> >>>> >>> >> >> >>>>MiniCircuits VCOs are around $20, which isn't bad. >>> >> >> >>>> >>> >> >> >>>>The output would go into a PWM modulator and then an AC-coupled >>> >> >> >>>>communications channel. >>> >> >> >>> >>> >> >> >>>hat are you trying to do ? >>> >> >> >>> >>> >> >> >>>Are you trying to transfer a very broad band DC based _analog_ signal >>> >> >> >>>(say DC .. 500 MHz) over an AC link (say 2 GHz). >>> >> >> >> >>> >> >> >> >>> >> >> >>That, or just ship a single-bit logic level over a telecom type link. >>> >> >> >>It's just a shower musing, not serious yet. >>> >> >> >> >>> >> >> >>> >>> >> >> >>>Why not simply frequency module a VCO ? Why PWM ? >>> >> >> >>> >>> >> >> >> >>> >> >> >>FM is more work on both ends. Commercial VCOs have pitiful modulation >>> >> >> >>bandwidths, kilohertz not gigahertz. PWM can be arbitrarily fast. >>> >> >> >> >>> >> >> >>For the logic level case, we'd want it to be fast with low jitter. >>> >> >> >>Whatever modulation scheme is used, the time jitter should be a >>> >> >> >>fraction of the "carrier" period. >>> >> >> > >>> >> >> >What about phase shifting keying (PSK) ? >>> >> >> How can you DC couple PSK? >>> >> >> > >>> >> >> >Just invert the output from the oscillator at input signal change. Of >>> >> >> >course, you have to establish the initial state somehow to be able to >>> >> >> >decode current state from only phase changes. >>> >> >> > >>> >> >> >To always know the state, you could use return to zero coding, i.e. a >>> >> >> >carrier at 0 degrees, a 0->1 transition as +90 degree phase change and >>> >> >> >1->0 input change as -90 degrees phase shift (QPSK). >>> >> >> > >>> >> >> >Also take look at Manchester coding. >>> >> >> > >>> >> >> Manchester is phase ambiguous! A long string of 1's looks just like a >>> >> >> long string of 0's. And the decode is clock synchronous, so has >>> >> >> massive jitter when shipping an arbitrary signal. >>> >> >> >>> >> >> PSK ditto. I want to modulate, not encode. >>> >> >> >>> >> >> AM/OOK won't work when the receiver has AGC. >>> >> > >>> >> >wouldn't your pwm be OOK? >>> >> No. It would be a fixed frequency continuous square wave of modulated >>> >> duty cycle. That does DC coupled transmission of an asynchronous >>> >> analog or logic-level baseband signal. >>> > >>> >so the ~2GHz it self have varying duty cycle? not ~2GHz on/off ? >>> Right. It shouldn't be difficult. >>> >>> OOK has the long-term receive threshold problem. >>> > >>> >> >>> >> The duty cycle can't get extreme or the optical receivers get unhappy. >>> >> 40/60 % seems conservative for a typical telecom rosa. I might look >>> >> into duty cycle modulation over an ethernet type path, twisted pairs >>> >> with magnetics. >>> > >>> >I guess with two pairs you could send the carrier on one pair to demodulate >>> >the other pair at the far end >>> AM with synchronous demodulation? >> >>or PM >> >>>PWM only needs one channel and a lowpass filter to demodulate. >> >>but then your channel will have varying DC > >The signal on an AC-coupled channel always averages zero. If you poke >a PWM square wave through it, you get an asymmetric amplitude square >wave with zero DC average. > >Telecom-type optical signals are usually transmitted by switching a >laser on/off, following the digital input. The transmitters are >usually AC coupled, so need continuous, roughly 50% digital activity >to work right. > >That is typically received by a photodiode, an AC-coupled AGC >amplifier, and a zero crossing detector, to typically make an >AC-coupled differential CML output. > >Ethernet type signals are transformer coupled on both ends. > >That gets interesting to push a DC coupled signal through, >theoretically and experimentally.
I can speak to how optical Ethernet ensures that their signals can be transformer-coupled, regardless of what data is sent. The basic trick is that the code space of the line symbols is large enough that there are at minimum two line symbols that can be used to represent each possible data symbol (or control symbol). The line symbols are sent by OOK of the laser, it taking multiple on-off flashes to encode each line symbol. The transmitter keeps a running sum of the ones and zeros sent, and chooses which of the pair of line symbols to use depending on the running sum - if the sum is growing in one direction, the chosen symbols will drive the sum back towards zero average. There is a founding article that I read twenty years ago on the design of such codes. If I recall, this was invented at IBM. While I may have a copy of that article, I'll probably never find it now. Joe Gwinn
Piglet <erichpwagner@hotmail.com> wrote:

> On 30/04/2021 23:46, Steve Wilson wrote: >> >> You can beat the bandwidth of PWM due to the filtering required at the >> receiver. Even extensive filtering leaves some ripple, which reduces the >> accuracy of data transmission. >> > > Filtering PWM can be made less onerous using Woodward's cancellation > scheme. > > <www.dropbox.com/s/997fixfek6afwga/woodward_dac_precision_ifd.pdf?raw=1> > > SED last discussed this on 19 Feb 2021 under title "PWM Output With RC > Coupled Inverted Signal" > > I wonder if at 2GHz the inverter could be replaced with a 180deg length > delay line? Could make very simple demodulator. > > piglet
Thanks for the link. ISTR another analysis with actual waveforms. However, doing this at 3 GHz might be a problem. -- The best designs occur in the theta state. - sw
On Sat, 01 May 2021 11:15:57 -0400, Joe Gwinn <joegwinn@comcast.net>
wrote:

>On Fri, 30 Apr 2021 16:19:36 -0700, John Larkin ><jlarkin@highland_atwork_technology.com> wrote: > >>On Fri, 30 Apr 2021 13:26:39 -0700 (PDT), Lasse Langwadt Christensen >><langwadt@fonz.dk> wrote: >> >>>fredag den 30. april 2021 kl. 21.31.52 UTC+2 skrev John Larkin: >>>> On Fri, 30 Apr 2021 11:10:12 -0700 (PDT), Lasse Langwadt Christensen >>>> <lang...@fonz.dk> wrote: >>>> >>>> >fredag den 30. april 2021 kl. 19.51.58 UTC+2 skrev John Larkin: >>>> >> On Fri, 30 Apr 2021 10:09:59 -0700 (PDT), Lasse Langwadt Christensen >>>> >> <lang...@fonz.dk> wrote: >>>> >> >>>> >> >fredag den 30. april 2021 kl. 19.04.03 UTC+2 skrev John Larkin: >>>> >> >> On Fri, 30 Apr 2021 19:54:36 +0300, upsid...@downunder.com wrote: >>>> >> >> >>>> >> >> >On Fri, 30 Apr 2021 08:45:30 -0700, jla...@highlandsniptechnology.com >>>> >> >> >wrote: >>>> >> >> > >>>> >> >> >>On Fri, 30 Apr 2021 08:07:10 +0300, upsid...@downunder.com wrote: >>>> >> >> >> >>>> >> >> >>>On Tue, 27 Apr 2021 12:35:17 -0700, John Larkin >>>> >> >> >>><jlarkin@highland_atwork_technology.com> wrote: >>>> >> >> >>> >>>> >> >> >>>>What's a good way to make a, say 2 or 3 GHz sinewave oscillator? I >>>> >> >> >>>>don't need extreme accuracy, and I'd like maybe 1 volt p-p. >>>> >> >> >>>> >>>> >> >> >>>>I was thinking a cheap MMIC with PCB delay line feedback. >>>> >> >> >>>> >>>> >> >> >>>>MiniCircuits VCOs are around $20, which isn't bad. >>>> >> >> >>>> >>>> >> >> >>>>The output would go into a PWM modulator and then an AC-coupled >>>> >> >> >>>>communications channel. >>>> >> >> >>> >>>> >> >> >>>hat are you trying to do ? >>>> >> >> >>> >>>> >> >> >>>Are you trying to transfer a very broad band DC based _analog_ signal >>>> >> >> >>>(say DC .. 500 MHz) over an AC link (say 2 GHz). >>>> >> >> >> >>>> >> >> >> >>>> >> >> >>That, or just ship a single-bit logic level over a telecom type link. >>>> >> >> >>It's just a shower musing, not serious yet. >>>> >> >> >> >>>> >> >> >>> >>>> >> >> >>>Why not simply frequency module a VCO ? Why PWM ? >>>> >> >> >>> >>>> >> >> >> >>>> >> >> >>FM is more work on both ends. Commercial VCOs have pitiful modulation >>>> >> >> >>bandwidths, kilohertz not gigahertz. PWM can be arbitrarily fast. >>>> >> >> >> >>>> >> >> >>For the logic level case, we'd want it to be fast with low jitter. >>>> >> >> >>Whatever modulation scheme is used, the time jitter should be a >>>> >> >> >>fraction of the "carrier" period. >>>> >> >> > >>>> >> >> >What about phase shifting keying (PSK) ? >>>> >> >> How can you DC couple PSK? >>>> >> >> > >>>> >> >> >Just invert the output from the oscillator at input signal change. Of >>>> >> >> >course, you have to establish the initial state somehow to be able to >>>> >> >> >decode current state from only phase changes. >>>> >> >> > >>>> >> >> >To always know the state, you could use return to zero coding, i.e. a >>>> >> >> >carrier at 0 degrees, a 0->1 transition as +90 degree phase change and >>>> >> >> >1->0 input change as -90 degrees phase shift (QPSK). >>>> >> >> > >>>> >> >> >Also take look at Manchester coding. >>>> >> >> > >>>> >> >> Manchester is phase ambiguous! A long string of 1's looks just like a >>>> >> >> long string of 0's. And the decode is clock synchronous, so has >>>> >> >> massive jitter when shipping an arbitrary signal. >>>> >> >> >>>> >> >> PSK ditto. I want to modulate, not encode. >>>> >> >> >>>> >> >> AM/OOK won't work when the receiver has AGC. >>>> >> > >>>> >> >wouldn't your pwm be OOK? >>>> >> No. It would be a fixed frequency continuous square wave of modulated >>>> >> duty cycle. That does DC coupled transmission of an asynchronous >>>> >> analog or logic-level baseband signal. >>>> > >>>> >so the ~2GHz it self have varying duty cycle? not ~2GHz on/off ? >>>> Right. It shouldn't be difficult. >>>> >>>> OOK has the long-term receive threshold problem. >>>> > >>>> >> >>>> >> The duty cycle can't get extreme or the optical receivers get unhappy. >>>> >> 40/60 % seems conservative for a typical telecom rosa. I might look >>>> >> into duty cycle modulation over an ethernet type path, twisted pairs >>>> >> with magnetics. >>>> > >>>> >I guess with two pairs you could send the carrier on one pair to demodulate >>>> >the other pair at the far end >>>> AM with synchronous demodulation? >>> >>>or PM >>> >>>>PWM only needs one channel and a lowpass filter to demodulate. >>> >>>but then your channel will have varying DC >> >>The signal on an AC-coupled channel always averages zero. If you poke >>a PWM square wave through it, you get an asymmetric amplitude square >>wave with zero DC average. >> >>Telecom-type optical signals are usually transmitted by switching a >>laser on/off, following the digital input. The transmitters are >>usually AC coupled, so need continuous, roughly 50% digital activity >>to work right. >> >>That is typically received by a photodiode, an AC-coupled AGC >>amplifier, and a zero crossing detector, to typically make an >>AC-coupled differential CML output. >> >>Ethernet type signals are transformer coupled on both ends. >> >>That gets interesting to push a DC coupled signal through, >>theoretically and experimentally. > >I can speak to how optical Ethernet ensures that their signals can be >transformer-coupled, regardless of what data is sent.
Optical ethernet is almost always SFP modules. Their i/o is capacitive coupled differential CML, no transformers.
> >The basic trick is that the code space of the line symbols is large >enough that there are at minimum two line symbols that can be used to >represent each possible data symbol (or control symbol). The line >symbols are sent by OOK of the laser, it taking multiple on-off >flashes to encode each line symbol. The transmitter keeps a running >sum of the ones and zeros sent, and chooses which of the pair of line >symbols to use depending on the running sum - if the sum is growing in >one direction, the chosen symbols will drive the sum back towards zero >average.
That's 8b10b or longer versions. The idea is to keep the 1/0 balance exact long-term. I want to send PWM, which doesn't, which is why it's interesting.
> >There is a founding article that I read twenty years ago on the design >of such codes. If I recall, this was invented at IBM. While I may >have a copy of that article, I'll probably never find it now. > >Joe Gwinn
A typical FPGA serdes engine natively does 8b10b and longer balanced codes. We have used the serdes engines to make PWM and programmable pulses too. You have to get inside the serdes blocks and work near the end. -- John Larkin Highland Technology, Inc The best designs are necessarily accidental.
On Sat, 1 May 2021 15:45:43 -0000 (UTC), Steve Wilson <spamme@not.com>
wrote:

>Piglet <erichpwagner@hotmail.com> wrote: > >> On 30/04/2021 23:46, Steve Wilson wrote: >>> >>> You can beat the bandwidth of PWM due to the filtering required at the >>> receiver. Even extensive filtering leaves some ripple, which reduces the >>> accuracy of data transmission. >>> >> >> Filtering PWM can be made less onerous using Woodward's cancellation >> scheme. >> >> <www.dropbox.com/s/997fixfek6afwga/woodward_dac_precision_ifd.pdf?raw=1> >> >> SED last discussed this on 19 Feb 2021 under title "PWM Output With RC >> Coupled Inverted Signal" >> >> I wonder if at 2GHz the inverter could be replaced with a 180deg length >> delay line? Could make very simple demodulator. >> >> piglet > >Thanks for the link. ISTR another analysis with actual waveforms. > >However, doing this at 3 GHz might be a problem.
Exactly. Just LC lowpass filtering was what I had in mind. -- John Larkin Highland Technology, Inc The best designs are necessarily accidental.
Steve Wilson <spamme@not.com> wrote:

> Piglet <erichpwagner@hotmail.com> wrote: > >> On 30/04/2021 23:46, Steve Wilson wrote: >>> >>> You can beat the bandwidth of PWM due to the filtering required at the >>> receiver. Even extensive filtering leaves some ripple, which reduces >>> the accuracy of data transmission. >>> >> >> Filtering PWM can be made less onerous using Woodward's cancellation >> scheme. >> >> <www.dropbox.com/s/997fixfek6afwga/woodward_dac_precision_ifd.pdf?raw=1> >> >> SED last discussed this on 19 Feb 2021 under title "PWM Output With RC >> Coupled Inverted Signal" >> >> I wonder if at 2GHz the inverter could be replaced with a 180deg length >> delay line? Could make very simple demodulator. >> >> piglet > > Thanks for the link. ISTR another analysis with actual waveforms. > > However, doing this at 3 GHz might be a problem.
As a side note, everyone knows you can't string two 8-bit dacs together and make one 16-bit dac. The high byte dac would have to have the accuracy and linearity of a 16-bit dac, so you might as well use a 16-bit dac in the first place. Jim Willims managed to string two 16-bit dacs together to make one 20-bit dac, as shown in Application Note 86F,"A Standards Lab Grade 20-Bit DAC with 0.1ppm/C Drift", at https://www.analog.com/media/en/technical-documentation/application-notes/a n86f.pdf Jim used a LTC2400 with external reference and +/- 2 ppm nonlinearity: https://datasheet.octopart.com/LTC2400CS8%23PBF-Linear-Technology-datasheet -11536201.pdf That was in 2001. Today, we have a 32-bit adc with 24.6 noise free bits at 5 SPS, and on-chip 2.5 V reference (&#2013266097;2 ppm/&#2013266096;C drift). 2^24.6=25,429,504, or about 7 1/2 digits: https://datasheet.octopart.com/AD7177-2BRUZ-Analog-Devices-datasheet-388995 01.pdf This is used in a 8 1/2 open source voltmeter by Marco Reps at https://hackaday.com/2021/02/26/homebrew-metrology-the-cern-way/ The CERN project is described at: https://ohwr.org/project/opt-adc-10k-32b-1cha/wikis/home The AD7177 is USD $25.74 at arrow: https://octopart.com/search?q=AD7177-2&currency=USD&specs=0 Now is a really good time for those interested in electronics! -- The best designs occur in the theta state. - sw
On Tuesday, April 27, 2021 at 12:35:27 PM UTC-7, John Larkin wrote:
> What's a good way to make a, say 2 or 3 GHz sinewave oscillator? I > don't need extreme accuracy, and I'd like maybe 1 volt p-p. > > I was thinking a cheap MMIC with PCB delay line feedback. > > MiniCircuits VCOs are around $20, which isn't bad. > > The output would go into a PWM modulator and then an AC-coupled > communications channel.
Hand-wavy, seems totally reasonable. It is one way of measuring the phase noise of amps. https://dl.cdn-anritsu.com/ja-jp/test-measurement/reffiles/About-Anritsu/R_D/Technical/95/95-08.pdf