On Fri, 27 Oct 2023 20:51:52 -0700 (PDT), Anthony William Sloman <bill.sloman@ieee.org> wrote:>On Saturday, October 28, 2023 at 8:54:09?AM UTC+11, john larkin wrote: >> On Fri, 27 Oct 2023 20:11:00 +0200, Jeroen Belleman <jer...@nospam.please> wrote: >> >On 10/27/23 15:01, John Larkin wrote: >> >> On Thu, 26 Oct 2023 22:03:32 -0700 (PDT), Chirantan Ganguly <chirantan...@gmail.com> wrote: > ><snip> > >> I mostly ignore him; he hasn't said anything interesting or useful in decades. Just droning, ritual insults. > >John Larkin posts here to get flattered. He doesn't post anything that deserves flattery, and he doesn't like discussing how his schemes could be improved. >The very idea that they could be is insulting. > >He's just told us that the phase detector in the 40406 is rubbish - when the circuit includes two phase detectors and the Philips HCT9046 has a sequential phase detector which has been modified so that it doesn't have flaw present in the original. Pointing this out is insulting, but John Larkin's posted misinformation was less than useful.The 4046 is great for some things and I like its zero-crossing phase-frequency detector. But zero-crossing detectors aren't always what is needed. These days, a micro is necessary if you want a multiplier phase detector that doesn't cost a lot. boB
ECE news google group (a Electrical and Computer Engineering equivalent of ML-news)
Started by ●October 27, 2023
Reply by ●October 28, 20232023-10-28
Reply by ●October 28, 20232023-10-28
On Sat, 28 Oct 2023 16:22:37 -0700, boB <boB@K7IQ.com> wrote:>On Fri, 27 Oct 2023 20:51:52 -0700 (PDT), Anthony William Sloman ><bill.sloman@ieee.org> wrote: > >>On Saturday, October 28, 2023 at 8:54:09?AM UTC+11, john larkin wrote: >>> On Fri, 27 Oct 2023 20:11:00 +0200, Jeroen Belleman <jer...@nospam.please> wrote: >>> >On 10/27/23 15:01, John Larkin wrote: >>> >> On Thu, 26 Oct 2023 22:03:32 -0700 (PDT), Chirantan Ganguly <chirantan...@gmail.com> wrote: >> >><snip> >> >>> I mostly ignore him; he hasn't said anything interesting or useful in decades. Just droning, ritual insults. >> >>John Larkin posts here to get flattered. He doesn't post anything that deserves flattery, and he doesn't like discussing how his schemes could be improved. >>The very idea that they could be is insulting. >> >>He's just told us that the phase detector in the 40406 is rubbish - when the circuit includes two phase detectors and the Philips HCT9046 has a sequential phase detector which has been modified so that it doesn't have flaw present in the original. Pointing this out is insulting, but John Larkin's posted misinformation was less than useful. > > >The 4046 is great for some things and I like its zero-crossing >phase-frequency detector.In the real CD4046B, the charge-dispensing phase-frequency detector has a huge time deadband, which makes for lots of jitter and phase noise. There are some better versions.> > But zero-crossing detectors aren't always what is needed. These >days, a micro is necessary if you want a multiplier phase detector >that doesn't cost a lot. > >boB >We usually do it in an FPGA.
Reply by ●October 28, 20232023-10-28
On Sat, 28 Oct 2023 16:48:43 -0700, John Larkin <jl@997PotHill.com> wrote:>On Sat, 28 Oct 2023 16:22:37 -0700, boB <boB@K7IQ.com> wrote: > >>On Fri, 27 Oct 2023 20:51:52 -0700 (PDT), Anthony William Sloman >><bill.sloman@ieee.org> wrote: >> >>>On Saturday, October 28, 2023 at 8:54:09?AM UTC+11, john larkin wrote: >>>> On Fri, 27 Oct 2023 20:11:00 +0200, Jeroen Belleman <jer...@nospam.please> wrote: >>>> >On 10/27/23 15:01, John Larkin wrote: >>>> >> On Thu, 26 Oct 2023 22:03:32 -0700 (PDT), Chirantan Ganguly <chirantan...@gmail.com> wrote: >>> >>><snip> >>> >>>> I mostly ignore him; he hasn't said anything interesting or useful in decades. Just droning, ritual insults. >>> >>>John Larkin posts here to get flattered. He doesn't post anything that deserves flattery, and he doesn't like discussing how his schemes could be improved. >>>The very idea that they could be is insulting. >>> >>>He's just told us that the phase detector in the 40406 is rubbish - when the circuit includes two phase detectors and the Philips HCT9046 has a sequential phase detector which has been modified so that it doesn't have flaw present in the original. Pointing this out is insulting, but John Larkin's posted misinformation was less than useful. >> >> >>The 4046 is great for some things and I like its zero-crossing >>phase-frequency detector. > >In the real CD4046B, the charge-dispensing phase-frequency detector >has a huge time deadband, which makes for lots of jitter and phase >noise. > >There are some better versions. > > >> >> But zero-crossing detectors aren't always what is needed. These >>days, a micro is necessary if you want a multiplier phase detector >>that doesn't cost a lot. >> >>boB >> > >We usually do it in an FPGA.Do you add an ADC to that ? Or just use ones and zeros as input to the FPGA ?
Reply by ●October 28, 20232023-10-28
On Sat, 28 Oct 2023 17:15:14 -0700, boB <boB@K7IQ.com> wrote:>On Sat, 28 Oct 2023 16:48:43 -0700, John Larkin <jl@997PotHill.com> >wrote: > >>On Sat, 28 Oct 2023 16:22:37 -0700, boB <boB@K7IQ.com> wrote: >> >>>On Fri, 27 Oct 2023 20:51:52 -0700 (PDT), Anthony William Sloman >>><bill.sloman@ieee.org> wrote: >>> >>>>On Saturday, October 28, 2023 at 8:54:09?AM UTC+11, john larkin wrote: >>>>> On Fri, 27 Oct 2023 20:11:00 +0200, Jeroen Belleman <jer...@nospam.please> wrote: >>>>> >On 10/27/23 15:01, John Larkin wrote: >>>>> >> On Thu, 26 Oct 2023 22:03:32 -0700 (PDT), Chirantan Ganguly <chirantan...@gmail.com> wrote: >>>> >>>><snip> >>>> >>>>> I mostly ignore him; he hasn't said anything interesting or useful in decades. Just droning, ritual insults. >>>> >>>>John Larkin posts here to get flattered. He doesn't post anything that deserves flattery, and he doesn't like discussing how his schemes could be improved. >>>>The very idea that they could be is insulting. >>>> >>>>He's just told us that the phase detector in the 40406 is rubbish - when the circuit includes two phase detectors and the Philips HCT9046 has a sequential phase detector which has been modified so that it doesn't have flaw present in the original. Pointing this out is insulting, but John Larkin's posted misinformation was less than useful. >>> >>> >>>The 4046 is great for some things and I like its zero-crossing >>>phase-frequency detector. >> >>In the real CD4046B, the charge-dispensing phase-frequency detector >>has a huge time deadband, which makes for lots of jitter and phase >>noise. >> >>There are some better versions. >> >> >>> >>> But zero-crossing detectors aren't always what is needed. These >>>days, a micro is necessary if you want a multiplier phase detector >>>that doesn't cost a lot. >>> >>>boB >>> >> >>We usually do it in an FPGA. > >Do you add an ADC to that ? Or just use ones and zeros as input to >the FPGA ?In a simple frequency locker, we have two input pins (from, say, an XO and a VCO) and one output pin that gets RC filtered to drive the VCO. The FPGA has a couple of modes, far-from-lock search and locked. Far out, it's a frequency compare; close in, usually just an XOR is good enough. Time locking to picoseconds uses an ADC to digitize the time difference between two oscillators, and the FPGA uses that to servo one of the oscillators.
Reply by ●October 28, 20232023-10-28
On Sunday, October 29, 2023 at 10:22:55 AM UTC+11, boB wrote:> On Fri, 27 Oct 2023 20:51:52 -0700 (PDT), Anthony William Sloman > <bill....@ieee.org> wrote: > >On Saturday, October 28, 2023 at 8:54:09?AM UTC+11, john larkin wrote: > >> On Fri, 27 Oct 2023 20:11:00 +0200, Jeroen Belleman <jer...@nospam.please> wrote: > >> >On 10/27/23 15:01, John Larkin wrote: > >> >> On Thu, 26 Oct 2023 22:03:32 -0700 (PDT), Chirantan Ganguly <chirantan...@gmail.com> wrote: > > > ><snip> > > > >> I mostly ignore him; he hasn't said anything interesting or useful in decades. Just droning, ritual insults. > > > >John Larkin posts here to get flattered. He doesn't post anything that deserves flattery, and he doesn't like discussing how his schemes could be improved. > >The very idea that they could be is insulting. > > > >He's just told us that the phase detector in the 40406 is rubbish - when the circuit includes two phase detectors and the Philips HCT9046 has a sequential phase detector which has been modified so that it doesn't have flaw present in the original. Pointing this out is insulting, but John Larkin's posted misinformation was less than useful. > The 4046 is great for some things and I like its zero-crossing > phase-frequency detector. > > But zero-crossing detectors aren't always what is needed. These days, a micro is necessary if you want a multiplier phase detector that doesn't cost a lot.A proper analog phase detector built around a decent analog multiplier will probably perform better, but a fast and precise analog multiplier isn't cheap. Zero crossing phase detectors throw away a lot of information, and are vulnerable to spikes. They are good enough for a lot of applications. The Barker and Hart long-tailed pair based phase detector effectively multiplies the sine wave being looked at by a sine wave, which isn't a good as proper product detector. but adequate in many more applications. -- Bill Sloman, Sydney
Reply by ●October 28, 20232023-10-28
On Sunday, October 29, 2023 at 11:59:19 AM UTC+11, John Larkin wrote:> On Sat, 28 Oct 2023 17:15:14 -0700, boB <b...@K7IQ.com> wrote: > >On Sat, 28 Oct 2023 16:48:43 -0700, John Larkin <j...@997PotHill.com> wrote: > >>On Sat, 28 Oct 2023 16:22:37 -0700, boB <b...@K7IQ.com> wrote: > >>>On Fri, 27 Oct 2023 20:51:52 -0700 (PDT), Anthony William Sloman <bill....@ieee.org> wrote: > >>>>On Saturday, October 28, 2023 at 8:54:09?AM UTC+11, john larkin wrote: > >>>>> On Fri, 27 Oct 2023 20:11:00 +0200, Jeroen Belleman <jer...@nospam.please> wrote: > >>>>> >On 10/27/23 15:01, John Larkin wrote: > >>>>> >> On Thu, 26 Oct 2023 22:03:32 -0700 (PDT), Chirantan Ganguly <chirantan...@gmail.com> wrote:<snip>> Time locking to picoseconds uses an ADC to digitize the time difference between two oscillators, and the FPGA uses that to servo one of the oscillators.That implies the use of a short acquisition sampling ADC to sample the output of one oscillator at time determined by the other oscillator. It's output is naturally noisy. There are better schemes. -- Bill Sloman, Sydney
Reply by ●October 29, 20232023-10-29
On 2023-10-28 05:23, Jeroen Belleman wrote:> On 10/27/23 23:53, john larkin wrote: >> On Fri, 27 Oct 2023 20:11:00 +0200, Jeroen Belleman >> <jeroen@nospam.please> wrote: >> >>> On 10/27/23 15:01, John Larkin wrote: >>>> On Thu, 26 Oct 2023 22:03:32 -0700 (PDT), Chirantan Ganguly >>>> <chirantanganguly01@gmail.com> wrote: >>>> >>>>> Hello, >>>>> Inspired by the thriving ML-news google group community >>> >>> [Snip!] >>> >>>> >>>> Take Bill Sloman. Please. >>>> >>> >>> Why should you care? Just killfile him, and also those who keep >>> arguing with him. >> >> I mostly ignore him; he hasn't said anything interesting or useful in >> decades. Just droning, ritual insults. >> >>> >>> Jeroen Belleman (who kill-files about 90% of s.e.d by now) >> >> Kill filters confuse me, given the followups. But yes, s.e.d. is >> pretty bad now. >> >> I think we can make a TPS562208 into a pretty good +5 to -5 volt >> converter. We're using an LM2662 charge pump now and it's soft and >> noisy. You know the trick, ground the switcher output and its ground >> becomes the negative out. > > > I once had a board with lots of PECL and some +3V stuff. I used the > +3V also to terminate the PECL, which dominated the +3V current, so > I used a negative regulator to make +3V. The layout guy was puzzled. > > Jeroen Belleman;) Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
Reply by ●October 29, 20232023-10-29
On 2023-10-28 19:48, John Larkin wrote:> On Sat, 28 Oct 2023 16:22:37 -0700, boB <boB@K7IQ.com> wrote: > >> On Fri, 27 Oct 2023 20:51:52 -0700 (PDT), Anthony William Sloman >> <bill.sloman@ieee.org> wrote: >> >>> On Saturday, October 28, 2023 at 8:54:09?AM UTC+11, john larkin wrote: >>>> On Fri, 27 Oct 2023 20:11:00 +0200, Jeroen Belleman <jer...@nospam.please> wrote: >>>>> On 10/27/23 15:01, John Larkin wrote: >>>>>> On Thu, 26 Oct 2023 22:03:32 -0700 (PDT), Chirantan Ganguly <chirantan...@gmail.com> wrote: >>> >>> <snip> >>> >>>> I mostly ignore him; he hasn't said anything interesting or useful in decades. Just droning, ritual insults. >>> >>> John Larkin posts here to get flattered. He doesn't post anything that deserves flattery, and he doesn't like discussing how his schemes could be improved. >>> The very idea that they could be is insulting. >>> >>> He's just told us that the phase detector in the 40406 is rubbish - when the circuit includes two phase detectors and the Philips HCT9046 has a sequential phase detector which has been modified so that it doesn't have flaw present in the original. Pointing this out is insulting, but John Larkin's posted misinformation was less than useful. >> >> >> The 4046 is great for some things and I like its zero-crossing >> phase-frequency detector. > > In the real CD4046B, the charge-dispensing phase-frequency detector > has a huge time deadband, which makes for lots of jitter and phase > noise. > > There are some better versions.But none that are worth the extra money, seeing as a 1-M resistor from the PD2 output to ground pulls the servo point out of the deadband. (We have this discussion every few years.) The metal gate 4046 has quite a nice oscillator, apart from its inevitably high phase noise and its very poor settability. Once you get its center frequency right, it'll easily work over a 100:1 range with decent linearity. None of the silicon-gate versions can do that--their oscillators up and quit at the low end, and have slope variations of up to 5:1 or so. That makes it hard to build decent PLLs. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
Reply by ●October 29, 20232023-10-29
On Saturday, October 28, 2023 at 10:38:50 AM UTC-4, John Larkin wrote:> On Sat, 28 Oct 2023 11:23:21 +0200, Jeroen Belleman > <jer...@nospam.please> wrote: > > >On 10/27/23 23:53, john larkin wrote: > >> On Fri, 27 Oct 2023 20:11:00 +0200, Jeroen Belleman > >> <jer...@nospam.please> wrote: > >> > >>> On 10/27/23 15:01, John Larkin wrote: > >>>> On Thu, 26 Oct 2023 22:03:32 -0700 (PDT), Chirantan Ganguly > >>>> <chirantan...@gmail.com> wrote: > >>>> > >>>>> Hello, > >>>>> Inspired by the thriving ML-news google group community > >>> > >>> [Snip!] > >>> > >>>> > >>>> Take Bill Sloman. Please. > >>>> > >>> > >>> Why should you care? Just killfile him, and also those who keep > >>> arguing with him. > >> > >> I mostly ignore him; he hasn't said anything interesting or useful in > >> decades. Just droning, ritual insults. > >> > >>> > >>> Jeroen Belleman (who kill-files about 90% of s.e.d by now) > >> > >> Kill filters confuse me, given the followups. But yes, s.e.d. is > >> pretty bad now. > >> > >> I think we can make a TPS562208 into a pretty good +5 to -5 volt > >> converter. We're using an LM2662 charge pump now and it's soft and > >> noisy. You know the trick, ground the switcher output and its ground > >> becomes the negative out. > > > > > >I once had a board with lots of PECL and some +3V stuff. I used the > >+3V also to terminate the PECL, which dominated the +3V current, so > >I used a negative regulator to make +3V. The layout guy was puzzled. > > > >Jeroen Belleman > That still confuses me, using a linear regulator upside-down.Fig 17 page 24 https://www.mouser.jp/datasheet/2/308/1/fairchild_semiconductor_ka7805ae-1191496.pdf It's only drawn upside down, nothing about it is upside down. Kinda clever actually. Configuring the input unregulated supply could be a challenge.> > In a recent case, we had to plan for a buck switcher being driven > backwards, such that it would blow up the big caps in its input > supply.
