On Wed, 22 Sep 2021 05:41:31 -0700 (PDT), Tabby <tabbypurr@gmail.com> wrote:>On Wednesday, 22 September 2021 at 07:35:07 UTC+1, Jan Panteltje wrote: >> On a sunny day (Tue, 21 Sep 2021 13:00:03 -0700) it happened John Larkin >> <jlarkin@highland_atwork_technology.com> wrote in >> <jndkkgh4hgtrjf8tp...@4ax.com>: >> >And this is a heat-sunk air core inductor, wound on a Sharpie pen. >> > >> >https://www.dropbox.com/s/o2hz6oi08agzdy8/T850_Inductor.JPG?raw=1 >> That soldering has me a bit worried... > >The pad looks like a lot of thickness for heat to go through. Soldering it at a few points to pads would heatsink it and make it more robust.The grey stuff is 3G thermally conductive pad, 5 w/m-K. It really works. I roasted two generations of commercial inductors. Soldering "a few points" along the coil would still leave a lot of the coil basically uncooled, and the thermal contact areas would be small. This is a 10 square inch, 7ns pulse width, 1400v, 5 MHz Pockels Cell driver. This coil, and some critical pcb traces, get hot from skin loss. The PCB is bolted to a water-cooled plate with more of the gap-pad goop. The power semis are SiC fets on AlN insulators, driven by GaN fets. Fun project. I learned a lot. -- Father Brown's figure remained quite dark and still; but in that instant he had lost his head. His head was always most valuable when he had lost it.
Turbo Royer/Baxandall in boost configuration
Started by ●September 20, 2021
Reply by ●September 22, 20212021-09-22
Reply by ●September 22, 20212021-09-22
On Wed, 22 Sep 2021 07:29:48 -0700 (PDT), "ke...@kjwdesigns.com" <keith@kjwdesigns.com> wrote:>On Tuesday, 21 September 2021 at 10:31:52 UTC-7, John Larkin wrote: >> On Tue, 21 Sep 2021 10:05:03 -0700 (PDT), "ke...@ >> <ke...@...> wrote: >> >> >... >> >> Here is a sine oscillator. >> >> >> >> https://www.dropbox.com/s/wqygdxrcr2egxsv/AGC_Sine_Osc.jpg?raw-1 >> >> >> >> The output amplitude is very stable with time and temperature. The >> >> voltage at the collector is reliably 2xV+ p-p. >> >> >> >> I invented this when I was a kid, still in college. It was used in the >> >> Boresight Alignment Kit for the C5A. >> >>... >> > >> >That configuration is usually known as a Reinartz oscillator. It has been commonly used as a self-oscillating mixer in transistor radios since the mid nineteen-fifties. >> > >> >In the common implementation the resonant tank is on the secondary rather than the collector to give some isolation from variation of transistor parameters and improve stability. >> > >> >https://www.electronics-notes.com/images/transistor-radio-self-oscillating-mixer-ferrite-rod-antenna.png >> > >> >kw >> My circuit is clearly different. It's way simpler and has a built-in >> AGC system that precisely controls oscillation amplitude. > >The circuit I linked is a complete front-end showing a common use. Most of the other components are related to that , not for the basic oscillator. The only fundamental difference is the placement of the tank on the primary rather than the better position on the secondary. > >The only way your circuit reduces the number of components is by scrimping on biasing components.The biasing is brilliant.> >Your circuit doesn't have any compensation for component variations or the operating current.It does; it's just not obvious on casual inspection.> >It also does not have AGC, it relies upon saturation of the transistor to limit the oscillation amplitude.It does have AGC. Think about it.> >As Phil Hobbs mentions his book that is not the ideal way to control the amplitude of an oscillator and comes with side effects that reduce the quality of the output.This one can have a tiny flat on one sine excursion. If that matters, it is easily fixed. -- Father Brown's figure remained quite dark and still; but in that instant he had lost his head. His head was always most valuable when he had lost it.
Reply by ●September 22, 20212021-09-22
On Thursday, September 23, 2021 at 12:57:41 AM UTC+10, jla...@highlandsniptechnology.com wrote:> On Wed, 22 Sep 2021 07:29:48 -0700 (PDT), "ke...@kjwdesigns.com" > <ke...@kjwdesigns.com> wrote: > > >On Tuesday, 21 September 2021 at 10:31:52 UTC-7, John Larkin wrote: > >> On Tue, 21 Sep 2021 10:05:03 -0700 (PDT), "ke...@ > >> <ke...@...> wrote:<snip>> >As Phil Hobbs mentions his book that is not the ideal way to control the amplitude of an oscillator and comes with side effects that reduce the quality of the output. > > This one can have a tiny flat on one sine excursion.Do the Fourier transform. It's a load of high frequency harmonics - all small, but embarrassing.> If that matters,It does.> it is easily fixed.It isn't. -- Bill Sloman, Sydney
Reply by ●September 22, 20212021-09-22
On 9/22/2021 10:29 AM, ke...@kjwdesigns.com wrote:> On Tuesday, 21 September 2021 at 10:31:52 UTC-7, John Larkin wrote: >> On Tue, 21 Sep 2021 10:05:03 -0700 (PDT), "ke...@ >> <ke...@...> wrote: >> >>> ... >>>> Here is a sine oscillator. >>>> >>>> https://www.dropbox.com/s/wqygdxrcr2egxsv/AGC_Sine_Osc.jpg?raw-1 >>>> >>>> The output amplitude is very stable with time and temperature. The >>>> voltage at the collector is reliably 2xV+ p-p. >>>> >>>> I invented this when I was a kid, still in college. It was used in the >>>> Boresight Alignment Kit for the C5A. >>>> ... >>> >>> That configuration is usually known as a Reinartz oscillator. It has been commonly used as a self-oscillating mixer in transistor radios since the mid nineteen-fifties. >>> >>> In the common implementation the resonant tank is on the secondary rather than the collector to give some isolation from variation of transistor parameters and improve stability. >>> >>> https://www.electronics-notes.com/images/transistor-radio-self-oscillating-mixer-ferrite-rod-antenna.png >>> >>> kw >> My circuit is clearly different. It's way simpler and has a built-in >> AGC system that precisely controls oscillation amplitude. > > The circuit I linked is a complete front-end showing a common use. Most of the other components are related to that , not for the basic oscillator. The only fundamental difference is the placement of the tank on the primary rather than the better position on the secondary. > > The only way your circuit reduces the number of components is by scrimping on biasing components. > > Your circuit doesn't have any compensation for component variations or the operating current. > > It also does not have AGC, it relies upon saturation of the transistor to limit the oscillation amplitude. > > As Phil Hobbs mentions his book that is not the ideal way to control the amplitude of an oscillator and comes with side effects that reduce the quality of the output. > > kw > ... >The base resistor burns a lot of power for high step up ratios but you can cut that down as I posted FWIW
Reply by ●September 22, 20212021-09-22
On a sunny day (Wed, 22 Sep 2021 07:39:02 -0700) it happened jlarkin@highlandsniptechnology.com wrote in <hufmkg1gktf83ec8s0219kt6pq5eokv8vf@4ax.com>:>On Wed, 22 Sep 2021 06:33:30 GMT, Jan Panteltje ><pNaonStpealmtje@yahoo.com> wrote: > >>On a sunny day (Tue, 21 Sep 2021 13:00:03 -0700) it happened John Larkin >><jlarkin@highland_atwork_technology.com> wrote in >><jndkkgh4hgtrjf8tpe0ontrbrjuhk1kl6j@4ax.com>: >> >>>And this is a heat-sunk air core inductor, wound on a Sharpie pen. >>> >>>https://www.dropbox.com/s/o2hz6oi08agzdy8/T850_Inductor.JPG?raw=1 >> >>That soldering has me a bit worried... > >That's one of my iterations. Production units are prettier.OK>But it's perfectly reliable.Maybe I am thinking too much about G forces... Through hole at the ends. And like someone suggested, more pads along the way to support the coil. But personally I would go for a ceramic coil former with mounting bracket etc.. Yours is probably OK for just the lab table. Just spend some hours with something I build long ago that needed a new Oled display. I had ordered some, but turns out although it is the same control chip it needs one byte different in the Oled ini. But the chip was soldered in, and all programming pins used for other things. So programmed a new PIC and put in an IC socket so I can easily change code now. Works fine, but was wired in my usual way, no PCB, so needed re-wiring.
Reply by ●September 22, 20212021-09-22
On Tue, 21 Sep 2021 21:28:43 -0400, bitrex <user@example.net> wrote:>On 9/21/2021 9:10 PM, Anthony William Sloman wrote: >> On Wednesday, September 22, 2021 at 2:16:41 AM UTC+10, jla...@highlandsniptechnology.com wrote: >>> On Mon, 20 Sep 2021 17:45:12 -0400, bitrex <us...@example.net> wrote: >>>> Is it possible to take your standard Baxandall and tap the capacitor, >>>> take an aux winding off the secondary and feed a somewhat higher DC >>>> voltage to the cap (through the usual two-diode supply handoff >>>> arrangement), and use it to feed the gate drive to the transistors as >>>> well, which would be clocked rather than self-oscillating. >>>> >>>> The goal would be to have a quiet step-up converter that could do say 5 >>>> to ~150 in one step, or maybe with a single multiplier stage. That seems >>>> hard to do with anything off the shelf as compact pulse transformers >>>> with the appropriate turns ratio don't seem to be really available, you >>>> get into CFL-type transformers whose ratios are too large, but there >>>> might be something appropriate with a third winding to bootstrap the >>>> primary swing >>> Here is a sine oscillator. >>> >>> https://www.dropbox.com/s/wqygdxrcr2egxsv/AGC_Sine_Osc.jpg?raw-1 >>> >>> The output amplitude is very stable with time and temperature. The >>> voltage at the collector is reliably 2xV+ p-p. >>> >>> I invented this when I was a kid, still in college. It was used in the >>> Boresight Alignment Kit for the C5A. >> >> For a rather undemanding understanding of "very stable". None of Vbe, saturation Vce, and current gain aren't all that stable. >> >> Seriously stable stuff demodulates the output - carefully - compares it with a good quality voltage reference and twiddles the loop gain to keep the output amplitude where you want it >> > >For high step-ups at low powers with isolation I like the idea of an >isolated Cuk too. Not as quiet as driving sines, but has the benefit of >a continuous input current. And the transformer turns ratio doesn't need >to be large to get a large step-up so off-the-shelf components should be >easy to find. > >Lots of variations of the non-isolated Cuk online but seemingly not as >much about the isolated kind. >What Sloman can't understand, he insults. That happens a lot. -- If a man will begin with certainties, he shall end with doubts, but if he will be content to begin with doubts he shall end in certainties. Francis Bacon
Reply by ●September 22, 20212021-09-22
On Wednesday, 22 September 2021 at 07:57:41 UTC-7, jla...@highlandsniptechnology.com wrote: ...> >> My circuit is clearly different. It's way simpler and has a built-in > >> AGC system that precisely controls oscillation amplitude. > > > >The circuit I linked is a complete front-end showing a common use. Most of the other components are related to that , not for the basic oscillator. The only fundamental difference is the placement of the tank on the primary rather than the better position on the secondary. > > > >The only way your circuit reduces the number of components is by scrimping on biasing components. > The biasing is brilliant.That's a matter of opinion. It doesn't temper component or temperature variations.> > > >Your circuit doesn't have any compensation for component variations or the operating current. > It does; it's just not obvious on casual inspection. > > > >It also does not have AGC, it relies upon saturation of the transistor to limit the oscillation amplitude. > It does have AGC. Think about it.I know how it works - but the bias generated on the capacitor that reduces the conduction angle doesn't do enough to prevent saturation. It's not automatic level control in the same way that good quality oscillators utilize which keep the active element in its linear region. Most BJT oscillators do self-biasing, whether they want to or not. Often JFET oscillators add an external diode to avoid forward biasing the gate junction but still provide the automatic biasing. The circuit may fulfill its requirements very well, I'm not saying that it doesn't but to claim that it is a novel invention when the same design had been around for decades and was in common usage is stretching things. ... kw
Reply by ●September 22, 20212021-09-22
On Wed, 22 Sep 2021 11:24:41 -0700 (PDT), "ke...@kjwdesigns.com" <keith@kjwdesigns.com> wrote:>On Wednesday, 22 September 2021 at 07:57:41 UTC-7, jla...@highlandsniptechnology.com wrote: >... >> >> My circuit is clearly different. It's way simpler and has a built-in >> >> AGC system that precisely controls oscillation amplitude. >> > >> >The circuit I linked is a complete front-end showing a common use. Most of the other components are related to that , not for the basic oscillator. The only fundamental difference is the placement of the tank on the primary rather than the better position on the secondary. >> > >> >The only way your circuit reduces the number of components is by scrimping on biasing components. >> The biasing is brilliant. > >That's a matter of opinion. It doesn't temper component or temperature variations.It's physics, not opinion. The AC amplitude is extremely stable over temperature, which was one requirement, and all the transistor needs is some minimum beta.> >> > >> >Your circuit doesn't have any compensation for component variations or the operating current. >> It does; it's just not obvious on casual inspection. >> > >> >It also does not have AGC, it relies upon saturation of the transistor to limit the oscillation amplitude. >> It does have AGC. Think about it. > >I know how it works - but the bias generated on the capacitor that reduces the conduction angle doesn't do enough to prevent saturation. It's not automatic level control in the same way that good quality oscillators utilize which keep the active element in its linear region. > >Most BJT oscillators do self-biasing, whether they want to or not. Often JFET oscillators add an external diode to avoid forward biasing the gate junction but still provide the automatic biasing. > >The circuit may fulfill its requirements very well, I'm not saying that it doesn't but to claim that it is a novel invention when the same design had been around for decades and was in common usage is stretching things.Got a link? I suppose it's simple enough that someone invented it before I did. -- If a man will begin with certainties, he shall end with doubts, but if he will be content to begin with doubts he shall end in certainties. Francis Bacon
Reply by ●September 22, 20212021-09-22
On Mon, 20 Sep 2021 17:59:22 -0700 (PDT), sea moss <danluster81@gmail.com> wrote:>On Monday, September 20, 2021 at 2:45:16 PM UTC-7, bitrex wrote: >> Is it possible to take your standard Baxandall and tap the capacitor, >> take an aux winding off the secondary and feed a somewhat higher DC >> voltage to the cap (through the usual two-diode supply handoff >> arrangement), and use it to feed the gate drive to the transistors as >> well, which would be clocked rather than self-oscillating. >> >> The goal would be to have a quiet step-up converter that could do say 5 >> to ~150 in one step, or maybe with a single multiplier stage. That seems >> hard to do with anything off the shelf as compact pulse transformers >> with the appropriate turns ratio don't seem to be really available, you >> get into CFL-type transformers whose ratios are too large, but there >> might be something appropriate with a third winding to bootstrap the >> primary swing > >I was gonna take a look at your Spice model, but I am currently digesting the active-clamp forward topology (off the clock) so why look at yet another power converter circuit? > >But then I thought, maybe the active-clamp forward would work for you too. Try one of the transformers Lasse recommended. I cannot immediately see anything wrong with doing that... the active-clamp forward is nice since there is no reset winding needed, and the clamp capacitor soaks up the leakage energy making it one of the quieter isolated topologies. > >But like I said I am new to the topology so I cannot tell you if there are any subtleties to be careful with. Does anyone here have any insight into the active-clamp forward topology that I can't find in all the app notes? (not trying to hijack this thread)If you want to compare an older stone tools and animal hide version to the newer part simulations, you can fool with this one, which uses the LTC version of jellybean UC3842 and self-driven mosfet synchronous rectification. http://ve3ute.ca/query/active_clamp_forward_syncrec_sim.zip It illustrates all of the issues of control limits, dead-time, leakage,stability (chaotic attractors) etc etc etc, when applied to the forward topology. The self-driven syncrec was notorious for self-oscillation on application of voltage to it's output terminals (ie no input power applied). Could be good, could be bad, depending on what you want it to do . . . It's always important to establish environmental design goals and limits, when working with any topology. None will do everything, but some can have distinct cost and performance advantages in specific applications, at specific times in history. Some older articles: http://ve3ute.ca/query/active_clamp_-_switched_snubber_articles.zip Don't forget also - this is a simulation and so are those for models supplied by vendors of newer control chips. RL
Reply by ●September 22, 20212021-09-22
On Wed, 22 Sep 2021 11:44:13 -0700, John Larkin <jlarkin@highland_atwork_technology.com> wrote:>On Wed, 22 Sep 2021 11:24:41 -0700 (PDT), "ke...@kjwdesigns.com" ><keith@kjwdesigns.com> wrote: > >>On Wednesday, 22 September 2021 at 07:57:41 UTC-7, jla...@highlandsniptechnology.com wrote: >>... >>> >> My circuit is clearly different. It's way simpler and has a built-in >>> >> AGC system that precisely controls oscillation amplitude. >>> > >>> >The circuit I linked is a complete front-end showing a common use. Most of the other components are related to that , not for the basic oscillator. The only fundamental difference is the placement of the tank on the primary rather than the better position on the secondary. >>> > >>> >The only way your circuit reduces the number of components is by scrimping on biasing components. >>> The biasing is brilliant. >> >>That's a matter of opinion. It doesn't temper component or temperature variations. > >It's physics, not opinion. The AC amplitude is extremely stable over >temperature, which was one requirement, and all the transistor needs >is some minimum beta. > >> >>> > >>> >Your circuit doesn't have any compensation for component variations or the operating current. >>> It does; it's just not obvious on casual inspection. >>> > >>> >It also does not have AGC, it relies upon saturation of the transistor to limit the oscillation amplitude. >>> It does have AGC. Think about it. >> >>I know how it works - but the bias generated on the capacitor that reduces the conduction angle doesn't do enough to prevent saturation. It's not automatic level control in the same way that good quality oscillators utilize which keep the active element in its linear region. >> >>Most BJT oscillators do self-biasing, whether they want to or not. Often JFET oscillators add an external diode to avoid forward biasing the gate junction but still provide the automatic biasing. >> >>The circuit may fulfill its requirements very well, I'm not saying that it doesn't but to claim that it is a novel invention when the same design had been around for decades and was in common usage is stretching things. > >Got a link? > >I suppose it's simple enough that someone invented it before I did.It kinda reminds me of a TV video DC restorer circuit from tube TV days - it used grid leak to generate the needed bias. Joe Gwinn
