Zippy update...



That's 4 x { 300 + 300 } = euh...

In principle, just 2.4 Kilo-Watts of pure/compact (5 x 5 x 1 cm) madness!!

Or twice the madness @ 4.8 Kilo-Watts by using both PCBoard sides...



As usual my next drawing sample was performed using nothing else than a web browser, MSPaint, XnViewMP and GIMP - the later since only a couple days ago in practical terms:

Those chips are still Ti's TPA3255 300W + 300W Class-D (in MONO mode), so in a duplex multi-mode context that's plenty of potential rated output while i managed to devise some novel trick which may help further optimize components placement, including the parts count as well:

Guess where i yet got to draw IH Energy Injection WorkCoils...



As for those 2 copper-like coils above it's about illustrating my intuition that RF Choking can be conveniently implemented in such manner without much risk to significantly interfere with the normal (presently absent) WorkCoils function culminating at a central focus point (marked by an orange cross), which would be hard-wired to add their magnetic flux constructively instead.

IMO that's as far as one can go with the simpler 1-phase Mazzilli-topology modules, possibly refurbishing some suitable ferrite powder core(s) similarily shaped as suggested next (and not necessarily plain round nor square...):
In that later case it may feel more intuitive to envision a pair of relatively large RF Choke bobins installed on the outer arms while 2 IH WorkCoils would reside each on their own middle arm, with one set of coils hard-wired to behave as the classic RF Choke previously evoked wthile the other must actually boost its field strength in between where it's marked "gap".

I'll leave the rest to the reader's own imagination.



Good day, have fun! ☮

What? Just a thumbs up!



Oh common Galvatron, don't you realize what this hypothesis could lead to if ever verified beyond that simple basic intuition:


When configuring Ti's TPA3255 Class-D HiFi "Digital-Audio" chip for hard-wired "MONO" the total rated output capacity ranges up to 600 Watts (hence no practical limitation there!) and since that's via a "Full H-Bridge" power stage this actually requires ONLY *1* RF Choke. So, i say we get rid of choices mostly imposed for no good purpose related to function, except perhaps the environmentally-wise recycling of some primitive PCBoards!



M'well at least such type of "smart" power chip clearly implies we don't wait for forbiden catastrophic conditions to take place in hope an external intervention can stop the train of events, like often fatal "shot-through" when "ZVS" is no longer "ZVS", expecially in absence of power interfacing as when Flyback Transformers used to supply very high voltage for Tesla Coil experiments i think... The prospect of a chip we can expect to just shut down during a short-circuit condition sounds appealing, and if it's still not completely up to the manufacturer's claimed ratings when transposed to an IH context then i'm sure there are many more chips (plus those coming soon) to investigate anyway. Realistically the main obstacle seems to revolve around some convenient availability of a PCBoard to populate, once again...



But i heard the DynaVap enthousiasts would like their IH Driver to be less than an inch thick too, on top of battery-operated, so it seems to me if there's got to be mechanical compatibility with 5/8" (dia.) plumbing then that IH Aperture has got to be made perpenticular to the larger surface and for this i envision the cigarette pack format, with space for a couple days supply of dry flowers + a docking bay for adapted delivery tools, presumably starting with DynaVap's GLASS VapCap IMO. Euh... Or maybe a long pack more similar in shape to a portable scale i presume, since that's a long stick to shuffle in to begin with!



For this we already know that Flux Concentration material as that used in Flyback Transformers (e.g. for Mazzilli's "ZVS" Tesla Coil "modules" or better...) actually allow multiple-parameter optimization, just as with toroïd cores for "RF Choking" to be exact. Although the choke's signals are expected to cancel each other and not reach that IH WorkLoad directly i got to wonder if it still wouldn't cause some indirect power feed-back, nor if it significantly matters considering that Energy Injection shall only last a few seconds at most.

...

In any case i feel if flux concentration material is ever put back into our application's WorkCoils then the resulting circuitry should be one more suitable for comparison with that which Mazzilli worked on over a decade ago; briefly put i predict unattented Insertion & Removal of the WorkLoad in an air-gapped twin E-cores layout as presented here won't affect inductance the same as those everage transformer-less implementations presently spreading like mushrooms after the rain...

As a by-product users get the option to go thinner than 1 inch or so, maybe as far as 15 mm, maybe 12. This includes at least 2 layers bifilar winding for the RF Chokes (if AWG #12!...), wrapped around a flat core probably 2.5 ~ 3 mm thick of Fluxtrol (heat-cured) paste, so ultimately the real limitation may happen to be from battery size instead. But yet my point was that flux bending made possible with a solid magnetic core may give IH vaporists what Mazzilli-topology enthousiasts have been unable to envision so far for some obscure reason.

Good day, have fun!! ☮

Useful & most relevant resources NOT available direcetly right here:

...

So, how do we reconcilliate apparently incompatible camps around Mazzilli's legacy??

Well, here's one possibility i think:


It's been a while so that's hardly making sense without also posting about my "U-Core" Dipoles system:

Actually, in the creative process (...), i didn't immediately notice this is not a familiar type of winding, not even to myself! Yet if the intuition was right then such magnetic Flux-Concentration trick might easily solve a few puzzles which remain challenging otherwise. Lets imagine the implementation of a fault-tolerant battery setup where 1 single failing cell won't stop the rest to work collectively, yet electrically isolated from it...

Many years ago i was vaguely thinking of 2 cells after the Loto Labs Evoke/Lux demonstrated that just 1 sufficed for e-Cig missions:

Please admire my recent "revised" version:

Which by the way equally requires an explanation on my "2.5" dimensions thing...

The illustration's purpose is to communicate a concept of magnetic power combination (e.g. possibly providing "galvanic" isolation), in contrast to parallel and/or series connected circuits. These days i much prefer to gamble on 3 cells because if 1 goes down that's only 33 % lost while in a 2 units scenario i'd already loose 50 % with no other margin left. Plus i don't like to waste free PCBoard space when so much more awaits ahead!

While there's an appearance of IH frenzy currently going on, over some "DNA"... Well one could attribute to Steven Ward's "Self-Resonant SSTC" (1983!!) just the same, euh... Allow me to show you what makes me dream instead:

Using magnetic power combining 3 electrically isolated cells can work just the same as before except now ONLY 2 contacts are involved at a voltage that's a 3rd what it would be in typical series-connected solutions. As a trade-off for such fault-tolerance all 3 drivers must work in synchronism: TOGETHER.

Lucky me i discovered this little gem at some point, during my information quest:

No D.C. paths for a "shot-through" to cause some catastrophic failure event! And another direct consequence of this topology is to rely on much smaller RF Chokes since the junk to filter out gets somehow folded into the upper electro-magnetic band, no to mention it may be of importance to filter in a Class-D Digital (HiFi) Audio application though not necessarily for Induction Heating unless the susceptor is intended to send/receive audible/ultrasonic vibrations too... And that's the part of dream that motivates me most, to imagine a display-less/button-less user-interface offering an extended range of accessibility even for people with disabilities otherwise unable to inhale alone. So i've always been careful NOT to block any evolutive paths toward features as haptic/audio inter-action, including tap-touch command to be exact.

At that point all elements are gathered to reach new hights in portability by offering both 3-cells fault-tolerance AND Flat-Magnetics altogether. Consider a magnetic flux-concentrating core measuring 5 mm (dia.), if we use AWG#16 Litz wire (eq. 100 x AWG#35...) then include thickness for 1 single PCBoard side then it becomes 2 x 1.29 mm (dia.) or 2.58 for 1 "U-Core" dipole element, plus 1 mm of PCBoard + solid casing + clearance margin, hence we get a grand total of about ~11.6 mm thick... Not even the diameter of an "AA" cell actually! So in other words thickness is no longer a problem with circuitry: it's a power-density thing. And soon we may get AA cells just as capable as today's e-Cig monsters i bet.

And unlike e-Cig users there's no way i'll ever want to carry an electrically-stressed power pack near my nose on a chronic basis anyway, the solution being to rely on a lightweight-yet-efficient wireless heat source residing inside a delivery tool with NO other power source inside than what's been temporarily stored in a metal and only by virtue of my "Packetization" approach where the amount must match its workload, so too little residual heat is left to cause unwanted denaturation between cannabic intakes.

...

Instead of the ultimate thickness i prefer to add all the functionality that fits besides a trio of ordinary e-Cig cells. A PCBoard having 2 copper sides it's only inviting yet another evolutionary step above: the addition of a 3rd axis! Almost...

To be honest i wasn't sure it qualified as a full "Z" axis to alternate between the 2 planes of a pair of "U-Core" trios distributed on each side and combined as "Left" + "Right" nearby couples (abstracting the fact they're not on a same plane). Such layout should result in useful properties and i predict some level of 3rd-axis made available from that 2-sides strategy, and i got more ready for desert...

But this is a long text and i ain't explained quite a few details requiring attention, for example the NOT "Fault" signal on Allegro's ACS733 current-sensor chip might serve to control the NOT "MUTE" or NOT "ENABLE" pins of Infineon's MA12070 Class-D Digial Audio driver - while i'd have prefered to simply lower its sound level (via an hypothetical "VOL-" pin...) so a closed-loop item of the analog days is put to good use again: the "AGC" function. Briefly put, "VOC" adjustment ("adj") would set the maximum safe operation limit; events violating this threshold otherwise sending a pulses stream to force the amplification gain down until its resolved, or a watchdog ultimately cancel the show via the NOT "ENABLE" pin, eventually going into High-Z state if i'm not mistaking.

In any case it's supposed to be about RESONATOR circuits where intense currents only need to circulate between the "Tank Capacitor" and associated magnetic inductor, swapping energy which passes from the electrical field to its magnetic alter-ego, respectively. Which brings the crucial point of measurement and that's where this concept might prove successful. Because having access to data reflecting both fields using magneto-resistivity + simple induction (from "secondary" windings) it shall allow a whole range of possibilities just not imaginable with a Mazzilli so-called IH "module". The later still making some sense i suppose if only it came with an ACS733 chip installed, for basic protection (or more, a lot more even!)...

More "classic" Royer alternatives though are directly supported as a short/affordable lengths of copper wire (AWG#26 in my last drawing) will suffice, with a bonus: i figured maybe it can equally operate in "duplex" mode: DC for LEDs and AC for listening to the secondary. Specs for Allegro's ACS733 announce excellent support for the 38 KHz "Flyback" band and i feel optimistic past 100 KHz as well. I wonder about "scanning" susceptors too, quite especially Curie alloys for obvious reasons; thanks to having 2 planes 3rd-axis feedback i go as far as to envision operator-guiding feedback to facilitate ideal magnetic alignement.



Little else is needed for a micro-controller to manage the rest, paving the way to Synthetic Radio technologies as synchronous quadrature demodulation, since we can extract/rejuvenate the fundamental resonator frequency to define a master local clock reference. With pulsed Infra-Red scanning through the glass of a delivery tool it should be convenient enough to measure (and perhaps even hear) cannabic "fumet" opacity, pressure and speed (+turbulence) of flow. Making it all to tempting to envision a device capable of recording "sessions" for later sharing on karaoke-like "social" forums where newbies can obtain quick guidance and acquire operator skills painlessly.

In conclusion if i ever come as far as to experiment it myself then "microphonics" rank as a top priority, because nothing can be safer in absence of electricity than to only listen at a dynamic earphone in hope permanent magnets and coils are sufficient to excite it... If negative then i'll know that was a mirage, but i really doubt so.

Make your bets! It's already been done before, just not yet...

Good day, have fun!! ☮

Salutations,

Well it seems i didn't feel like talking to myself in this previous post linked below, actually intended to express unfinished views on my "LAVACapsule" ideas (or yet-undefined fantasies...):

Egzoset said:

Click to expand...

Observation enforces a most intuitive principle that 3-D airflow geometry, or simply "Path Layout", actually impacts on a session's ritual and/or a device's "personality" - and reciprocally. In other words the consumption method shapes a person's consumption "profile", with variable/individual influence and long-term persistance.

Anyway here's my evaluation in retrospective of 3 vaporist "tools" carefully selected among ready-made, patented and commercially-available products going to get transformed (customized) into "Prototyping Platforms" at the cost of implicitely voiding its manufacturer waranty. So, briefly put that DiY experimention quite reminescent of my 1st attempt to find a one hundred buck$ pipe for vaporists.

The outcome of such venture boils down to these VaporGenie configurations:

#1) Above all, my fully-customized (PinHole-modified) VG aluminium Classic-like pipe-handle equiped with its associated Full-Metal "Top" (e.g. half-a-capsule).

#2) My custom Bronze Sherlock handle with short neck (e.g. a "Mating Junction" doubling as cotton holder, plus "Optical Probe" accesses as an option...), also featured with 1 custom Full-Metal "Top", though preferably not fitted with exactly the same Hybrid Core configuration installed - as an effort to support "Near-Symetrical" Reversible self-maintenance/reclaim for further/future evaluation.

#3) A 3rd more recent semi glass-based hack (shown in past quote) only depends on VG's stainless steel "MonoBloc", as i tagged it. Still using one of those 2 reconfigured Hybrid Cores which cost me their fair share of personal involvement in quite multiple manners. The glassware itself is bare-bones, all i ask of it is transparency and relative food-grade neutrality over a whole range of operating temperatures, silicone would do find if transparency loss is compensated by bio-feedback - which relies on IH Driving and a special type of capsule.... For example, wouldn't that prove great to finally apply electromagnetic Induction Heat to my LAVACapsule concept? A top without it MonoBloc in place has significant solid volume inviting more efficient use of material, like creating possibly sealed room for 1 pick-up coil aligned to a common central axis, completed by a semiconductor, a LED (or even 3 for nicer symetry!), and also some other discrete component(s) as "passives" (zener, power resistor, capacitor). In retrospective, i've already considered removing excess material but that was with lightweight in mind, while now that gain could also come as a tiny "secondary" power source sufficient enough to diver just what it takes to feed LED lighting *IN SEQUENCE* with IH Driver frequency.

Each half-capsule or "Top" will conveniently hold at least 1 auxiliary inductor coil, no need to bother deciding if it can be done. What's on my mind instead is the need to decide between 2 types of light driving, depending on IH Driving modes. Preliminary impressions suggest maybe Pulse Heating is equally capable of emulatin the "Slow & Steady" thermostatic behaviour of popular Hot Dry Air Ovenizing Portable, go figure. Yet, assuming a capacitor will effectively extend a charge's useful/brief duration then it seems continuous lighting can probably get generated long enough to enjoy a toke in full Light & Sound style for a reasonable suplement. Paradoxically i might foresee continuous beam(s) of light as an alternate substitute for the lact of visual hints reflecting variations in cannabic cloud opacity, though i'd favour pulse-mode/time-coherent bio-feedback transducing for precision and freedom to prefer zero displays. Consequently a user-interface based on sound and/or vibrations makes more sense than dump/blind sensor-less thermostats working more like rheostats in open-loop fashion, all in denial of development potential relying on Curie aloys, with a proper scanner at hands... Semi-automation offering at least a fair degree of "haptic" return would enable human operators by easing control over closing a pleasurable regulation loop. Think of musical instrument play as in a flute!

What's great with a no capacitor(s) pulsing LEDs alternative is that i can still imagine nice Light 'n Sound entertainment but only while IH energy is actively injected, so the glass bong setup will seem more appropriate. As a trade-off this light source gets implicitely synchronized to the driver's own "clock" source, directly accessible to double as "reference" for synchronuous demodulation, hence conveniently providing the means to extrapolate a pair of "I/Q" analog signals from cloud opacity, since turbulence inter-actions can be safely expected to show up in audible conversion as well. So IMO it may suffice to run a Software-Defined Radio under some free GNU Linux environment connected via a bare-bones interface. For this i'd want to explore opportunities from the use of galvanically-isolated current-sensing chips as Allegro's ACS733; it has 1 analog comparator-adjust input and 2 outputs - 1 analog + 1 digital. Looking at the manufacturer's documentation i'd predict a lot of pleasurable prototyping, assuming basic fail-safe security has been integrated to the whole project. Besides of it i'm also dreaming of CMOS "timer" chips, including programmable ones.

The initial goal revolves around handling A.C. power and this implies energy transfers based on field reversals, ideally withing an L-C resonator circuitry to conceal high-intensity currents locally. In any case A.C. designs shall intrinsically involve the notion of finite-duration events refered to as "pulses"... Perfect simplistic "KISS" type reasoning should evoke the need to do more with less, e.g. presumably to "enslave" the control-gates of each power device (Si, SiC, GaN) in a way NOT denying higher functions while simultaneously implementing a power protection via elementary timer-based hardware destined to ALWAYS self-extinguishes (or "Reset") unless the cycle gets shortened/synchronized by some regenerating "resfesh" signal. The trick being to keep sensitive input(s) electrically isolated from the most powerful output of an amplifier, basically.

15 Volts CMOS programmable timer chips would offer relative noise immunity for a potentially-challenging electric environment. I even suppose this can spare my old KISS concept of an "OverRide" jumper, to force falling-back into "dumb" autonomous mode (without the bells & whistles) when a controller goes eventually defective - which is better than zero output at all. It's important NOT to block evolutive paths as synchronization via SDR PLLs, etc.; long-term features could emerge from exploiting data generated in the process, like skill-acquisition helpers (teaching) applications to tame down negative aspects of uncontrolled consumption. Rendering a consumption method eventually more acceptable both to cannabis enthousiasts and their rivals, the moderate prohibitionists - e.g. those capable/willing to welcome "Harm Reduction". Personally my 1st data-gathering scenario was about socializing applets for iphones and tablets, something like karaoke toking with real-time animated curves directly echoing cannabic stream opacity in the participant's favorite tools.

Etc., etc.

In my dreams, a couple decades from now maybe!

Good day, have fun!!

Well, since i've mentioned Infineon's MA12070 Class-D Digial Audio driver before...

It's low-voltage enough and being "filterless" (no heavy RF Choke) is some blessing, but a need for large capacitors remains considering i wanted all of it to match the same 3-D volume of a King Size cigarette pack... Add the glue logic and it should be fun playing with .MP3 media, litterally! Only guessing. But large 13 mm thick capacitor take up so much space in a tiny 16 mm thick box there can't even be a PCBoard under it and more remains to be done that will consume PCBoard space until more trade-offs become necessary. And talking of which...

In this scenario the trade-off was made at the cost of loosing a storage compartment for its consumption tool, consumable, etc. The gain is an additional "AA" size power cell and the Mazzilli/Royer ZVS topology in mind, in this case tamed down by 2 high-speed voltage-programmable comparators provided in Allegros ACS733 chip equally mentioned before i think. Take note the 3rd current monitor is reserved for the LC Tank Resonator itself...

Using USB-C and cooperative magnetic merging the 2 power sources should give confidence of implementing a 25 Watts IH Driver which truly fits the pocket, while the cable can withstand 20 Volts @ 5 A (e.g. 100 Watts), as i recall.

Merry Chistmass, happy New Year!!

Observationist said:

What does this have to do with growing?

I don’t know what the duck y’all talking about.

Click to expand...

Bongs, Pipes & Other Smoking Tools​

You may want to ignore this section completely.

More hints:

Salutations,

Almost 3 years have past and the two dominant 14500 power cell types remain LiMn or LiFePO4, one rated around 3.7 Volts @ 1000 ~ 1250 mAh (3.60 Wh ~ 4.6 Wh) and the other 3.2 Volts @ 800 mAh (2.96 Wh) or so, while the later offers some more conservative performance in a slightly lower energy density format (e.g. a lightweigtht ~20 g per cell), its tradeoff being additional stability and safety. Lifespan would range between 300 ~ 500 charge cycles.

Assuming 3 cells and an energy budget of 100 Joules for 1 single Hybrid Core or 200 Joules in a LAVACapsule configuration this can be evaluated to require some approximative pre-heat time varying from 8.3 to 18.2 seconds, which still rivals with what was announced out there:


In short, 3250 Joules ÷ 240 Watts = 13.5 seconds, more if we allow a margin of errors.

The difference would be that one fits the « King Size » cigarette pack format and not the other, and although both still imply a powerful generation of RF noise in the A.M. radio band only one may succeed in addressing my challenge to reflect temperature on top of microphonic modulations. Guess which it is while the other only gambles, based on a timer...

Anyway, years ago i made an audio recording of my IH cooker RF parasitics, which i happened to revisit very recently:


Salton ID-1081 of 2014-Jun-26 via Measure by Sugar Labs Community as FlatPack 2024-Oct-1

It illustrates how owners of an IH driver could acquire the means to compare signatures from different brands/models: all this requires is an A.M. radio, found in emergency lamps equipped with a manual dynamo, for example.

One aspect i'm most curious about is related to power throttling vs heat-soaking delays. This particular piece of free Linux software is from Sugar Labs Community, it's called 'Measure' and can be obtained as a FlatPack compatible with Mint, Manjaro, etc. The recording was captured on June 26, 2014 (better late than never!)... One main advantage here is real-time visualization, this means if my intuition that predicted microphonics ever verifies then the "clicks" shall also be seen, not only heard. And if it's matching a pattern then we can expect automated detection to enable a simple LED in response to insertion/removal, clicks and why not tap-commands as well if the tool has been magnetically polarized by a tiny magnet, implying audible feedback may become feasible too.

The reader will probably perceive it all as science fiction, but:


Good day, have fun!!

Hi again!

A couple days ago i posted this update after YouTube finally sent a critical echo to its foggy radar, it's about « a ghost in the machine » exactly:


And that's not all, here's a similar warning from another IH driver manufacturer:



We learn at 3m56s that this IH generator is rated 10 A @ 28 V, that's 280 W before losses, e.g. ~240 Watts if conversion efficiency is presumed to be around 85 %. In other words this is comparable to the Vestratto Vader/Forge.

Then there's another mention equally comparable, we get this clear warning at 7m46s:

« ...always fire the lighter before touching the metal... ... ...fire and then touch... ... ...don't do the opposite, don't touch and then fire... »

In this case no reference was made to some need for "dynamic coupling adjustment", not even a design including "every piece of protection". Actually the wording is vague at best except for a comment to the effect that failure to comply « might put more strain on the equipment »...



In all honesty i ain't surprized, years ago my evaluation of DynaVap's VapCap energy budget was 5~6 times the ~100 joules per puff range of my customized VaporGenie pipe. This late fall Vestratto announced 3250 Joules for their Tornado about which owners are now warned NOT to insert it in their Vader/Forge IH driver, while it's still OKay for the Anvil with 20 % less thermal mass. So what's happening i think is the testosterone-driven quest for more heat retention finally reached the limit of such tendency leading to monster clouds and consequently perhaps even a spiral of dosing abuse IMO! Lets keep in mind these manufacturers often dream of medical certification...

My pipe rations energy 1 puff at a time, these devices are meant to generate monster clouds and ideally seek multiple puffs per thermal charge. As a result it becomes too tempting to forget about auto-conscientization in a context of health-wise vaporism, there's so much heat accumulated it's starting to behave like a cigarette/"joint" that we typically wanted to finish once lit...

To each their own way, but my main concern at the moment is relative to now having 2 major players possibly exposing their clients to this same old Mazzilli « Royer ZVS » syndrome i heard about many years ago. In short this topology makes it necessary to implement protection mechanisms AROUND it instead of integrating it at the core of its twin power stage, and of course that sort of solution comes with trade-offs of its own, like to complexify a circuitry while the increased cost hardly reflects into additional convenience. On the contrary, it seems stoners are required to never mess up with the operation sequence or face some potential agravation!

Meanwhile, the vaporizers themselves continue to represent a risk of skin burns, or at least a threat for delicate furniture... Their cylinder structure still wastes half the radiative energy into empty space needlessly, on top of other improductive parasitic losses.

In any case i see no signs the Camouflet intends to work on correctives, as for the Vestratto Vader/Forge we can only hope for a rev. 3 knowing 10 month separated rev. 1 from rev. 2! All because at some point it finally seems justified to fear that the Mazzilli Royer ZVS became an element of endoctrination via "discussion" forums owned and controlled by a handful individuals who couldn't handle criticism and wanted to boss around being "king of the hill".

M'well, it's been 10 years since the failed Loto labs Evoke already. The bets are open, how many more for efficient correctives once the errors are recognized and their lessons properly assumed?

Today my old slogan never felt so true:


Good day, have fun!!

Think-tank epiphany?

Back in 2010 i suggested we should focus on what we're best at and rely on 3rd-party ready-made commercially-available products instead of getting involved with electronics. At the time i even considered it should be possible to hack an IH Cooker just by accessing its signals, because there was concern that it would auto-shutdown after detecting an absence of suitable workload. Since then i found out my customized pipe only requires torching for about 1 second, which is a tiny fraction of the time actually allowed.

My present thread illustrates some of the hair-pulling agravation that may result from an attempt to contemplate electronics design while being cursed with a pressing need to chase ghosts in the machine, then get rid of them forever, hopefully.

For a more accurate perspective have a look at what Tesla Coil enthousiasts must go through, for example:



The guy describes himself as an electronics engineer, a physicist doubled with a hobbyist...



Well, perhaps there's a significantly more affordable alternative, similar to my initial intuition of 2010, except i like it better:



Then over 14 years later the thought hit me like a truck, while standing in the middle of a highway:

Why would i want to even open my IH Cooktop unit when it might suffice to simply place a similar refurbished coil on it and drive some vaporizer selenoïd with it?! In other words this would turn my IH Cooker into the isolated (safe) primary of a transformer, while i might also take care to put that flat coil inside a box with its IH ~20 mm (dia.) or so aperture on its top side...

The Salton iD-1081 was rated 1.4 Kilo-Watts as i recall, so i'll be damned if i can't expect to extract at least a couple hundred watts from it before it times out.



A ferrite flux-concentration core inserted into that IH aperture is probably going to allow power injection into a workload such as my customized VG pipe - once it's been featured with a 17-holes SS430 disc, that is.

There are persons already used to smoke under the stove hood of their kitchen, so what's more stealthy than an IH Cooker nearby?...

170 mm measures about 6 11/16 inches and the coil's center hole seems to be a perfect location where to install its vertical selenoïd, as required to drive the full pen-shaped flame-compatible vaporizer suite (DynaVap, Vestratto Anvil/Tornado, Mad Scientist Tempest, etc.)! Besides, that adapter box wouldn't look too obvious, specially when conveniently stored among books on a shelf.

Not to mention those IH Cookers ain't too expensive in comparison to Ispire's Wand, Camouflet's Inductor or Vestratto's Vader/Forge (or else) - and relatively easy to find should replacement become necessary...

Merry Christmass!!

Check that one:


...and remember, my customized VG pipe only needs 100 Joules to be injected in 1 second. It would still work even if i missed by a factor of 2!

Update:


There's been a sale in a local superstore offering some IH cooktop for only 60 $ Cdn - but there's a catch: unlike the 2010 model this one won't generate induction power unless i put a real IH casserole on it. Nothing happens if try it with my refurbished IH workcoil as imagined in the December 21st post, no matter how loaded or not. In addition, the sort of 10 cm flat disc salvaged from a peanut bottle 14 years ago will no longer suffice as it seems i now need its diameter to be ~40 % larger, while the manual states a 10 to 26 cm casserole should do fine - though casseroles have vertical edges (none is flat after all)!

Anyway it appears i won't be able to drive the old IH workcoil at all, so here's the situation: my vintage 2010 unit used to start driving immediately for about 10 seconds and then it would abort, this means i'd have to salvage the workcoil of my fresh-new purchase to benefit from such odd feature - which still defeats the idea of buying 2 identical units knowing now they all come with some unavoidable casserole detection mechanism triggered in a split-second if no suitable workload is put on. Moreover, temperature/power control is NOT proportional and works similar to a 2-states thermostat controlling the duty-cycle times between full-On and full-Off...



Is it possible the YouTube video(s) was(were) fake? My next step i guess will be to experiment with alternate workcoil layouts.

Good day, have fun!!

Welcome to 2025!

Sometimes a need to mention technical details such as component/model numbers hardly becomes more evident, most certainly this would be the case here except i must be cautious not to make it look like SPAM, a somewhat moot argument considering i'd actually advise against purchasing the right-side upgrade product while my other device is 14 years old... So lets put it all in pictures anyway:

Late in 2010 my experiments with a 4 inches (dia.) disc went quite nicely, as long as it wasn't seperated from the glass surface by more than 1 toothpick's thickness (and a flat one at that, or vibrations would cause displacement and result in some premature shut-off). This same experiment failed with my fresh-new upgrade, the reasons being an insufficient size and also a less permissive cookware detection strategy - which in principle should still fit the overall margin of ¾" allowed by a CD caddy, though in practice i sense it's expecting something around 5½" or 140 mm (dia.). Go figure, perhaps a hack can be performed.

One way or another bye bye stealthy CD/DVD caddy i'm afraid! One alternative perhaps could be to find less restrictive brands/models if that's possible (which i suspect being the case with my 2 previous YouTube linked videos), but it would involve too much shopping around for my own taste so that's no real option in the end.

Worse, once i got one in my own hands i quickly realized such new "upgrade" has a few turns so close they're actually touching together, which is a very bad prospect that worries me. By chance i didn't care to buy a pair of identical units since i already had that old thing awaiting for attention in my storage room, somehow preserved in good working order.

Yet evolutionary differences don't stop there, here's my 2nd IH workcoil - probably manufactured in September:

Not so bad for a 60 $ Cdn bargain. Clearly it wasn't going to be real "Litz" stuff though its 2-layers stack came as a surprise. I believe its sole purpose was about making room to augment wire gauge, using a same number of turn, just more crowded in the center.

Most unfortunately this manufacturer was negligent and chose to sacrifice an essential feature despite a critical context: they thought saving on insulation was clever when in fact the wires touch! E.G. given enough time i smell a fire coming...



That said lets finally proceed with a 3rd picture illustrating one of the multiple configurations i was able to evaluate:

All this could manage to accomplish was to induce strong clicks in my dynamic earplug phones connected to the "Mic" (audio input) of a cassette Recording Walkman - more than twice as old as the cooker by the way...

No matter what, both IH workcoils just won't replace a compatible casserole that's suitable for completion of their power injection cycle.

What's next?

M'well, i keep thinking it's no definitive defeat. My guess is no gadget sold 60 canuck bucks would be sophisticated enough to tell the difference between real IH cookware and a virtual echo emulating some similar signature (made of frequency, phase & amplitude range)... In other words this might be a stimulating challenge for ham radio hobbyists who also happen to be vaporists.

Good day, have fun!

Salutations,

4 years ago i illustrated my reflexions on workcoil geometries, starting with a 2-D tri-phase 9-poles alternative to the traditional 1-D single-phase 2-poles air-core selenoïd:

The month after this was followed by a shift to 2-D variations including some dual-phase cardan:

Then i graphically illustrated the 2-D stacking of an inter-penetrating dual-phase layout where the flux concentrator might also support a set of R.F. Chokes:

The idea of such R.F. Chokes felt too unconventional but i continued to wonder about novel geometries rendered possible by flux concentration cores, then i imagined this:

My reasoning was that 1 of the 2 poles get wasted in air-core selenoïds unless a susceptor will fit inside; actually, years later we can observe that even when heated from the tip of Camouflet's Inducor (which does have flux concentration) its other pole happens to terminate toward empty space, so why not dare sacrifice windings density as a trade-off to bringing the 2 poles in close opposition?! After all, my weaving style of windings cause the workcoil assembly to become mechanically rigid, which i viewed as an asset in the perspective of exploiting microphonics with haptic feedback...

The aerial interstices which such frankenstein workcoil left behind made me want to add a secondary circuit dedicated to "closing" a feedback control loop, so i thought i'd keep that in mind while evaluating a few layout strategies, until the prospect of fitting it all inside the volume of a King Size cigarette pack struck the ex-smoker in me. Too bad the 14500 power cells only generate burst currents below 900 mAh @ 3.6 Volts or so, hence it would become neccessary to choose between having 3 cells or a storage compartment for its consumption tool.

Almost a year passed before i switched to drawing full bifilar workcoils using the same wire diameter, after realizing the Hall-Effect chips made the tiny wire secondaries redundant.

Here's one of my artworks which got published after i performed a bit of touch-up:

This prove most useful in trying to understand the consequences of such pins layout and assessments, previously confirmed in those exploratory layout attempts starting with the basics and aiming for additional complexity:

The main features put forward were inclusion of a PC Board with USB-C connectivity on it, plus continued emphasis on an opportunity to scan the cannabic vapour on its way out. Mention was also made about my intention to include a micro-switch for mechanical insertion/removal detection relative to its associated consumption tool.

My next drawing suggested how to implement dual-frequency support while also starting to envision where the high-current copper tracks would fit:

This exercise allowed me to realize that i could shorten those copper tracks somewhat, by reversing the winding's direction in one of my workcoils (identified as a "mirror-image" of the two others):

At that point i managed to figured out one scenario about how to enslave Mazzilli's « Royer ZVS », by resorting to the addition of a pair of Hall-Effect chips, e.g. 1 per branch:

This helped a lot because i could begin to foresee a probable combination of components offering these features: a) Over-Current protection; b) Optical scanning of the fumet; c) Audible 2-channels feedback providing a hint on workcoils activity; d) Detection/Encoding as a function of the type of consumption tool being inserted, with dual-frequency control via some Solid State Relay (SSR); e) A Smart-Slave USB-C option accomodating external advanced functions performed by a computing device, like sessions recording for later pairing with physical cotton samples analysis (both for study & entertainment purposes).

Then another windings compromise imposed itself, confirming my belief that « function defines form »:

So far i had resisted the urge to return to classic selenoïd windings with a flux concentration core inside. It still brought came with a magnetic bridge turning it into a "U" shape inductor but all pins were gather near the center, which actually implied my copper tracks would get even shorter at last!



Past this step i felt quite comforted this was ready for further 2-sided "U-shape" cores supperposition with the PC Board between two sets.

Little tweeking remained necessary, my copper tracks had moved into the central zone and the number of permissible coil turns was about to be known now that it all seemed to fall in place:

This last illustration is nearly finished, i knew which wire sizes would fit in and what capacitor values were possible. The next step could have been to draw copper tracks again and decide how/where to match them to a "daughter" PC Board supporting the Master-Slave & battery recharge USB-C connector:

Meanwhile non-rechargeable 14500 cells have appeared which can generate 3.6 Volts @ 2600 mAh, those could be useful in prototyping but they sell 40 $Cdn for 2! Consequently, i don't expect this particular « HR » compatible vision of mine to get evaluated in real form until battery technology has improved sufficiently. So, i sure like having a bit of similar day-dreaming - with a grain of nostagia since « It's already been done before, just not yet! »...

Good day, have fun!!

Egzoset said:

Salutations,

4 years ago i illustrated my reflexions on workcoil geometries, starting with a 2-D tri-phase 9-poles alternative to the traditional 1-D single-phase 2-poles air-core selenoïd:

​

(2020-Aug-10)​

The month after this was followed by a shift to 2-D variations including some dual-phase cardan:

​

(2020-Sep-24)​

Then i graphically illustrated the 2-D stacking of an inter-penetrating dual-phase layout where the flux concentrator might also support a set of R.F. Chokes:

​

(2020-Oct-22)​

The idea of such R.F. Chokes felt too unconventional but i continued to wonder about novel geometries rendered possible by flux concentration cores, then i imagined this:

​

(2021-Feb-20)​

My reasoning was that 1 of the 2 poles get wasted in air-core selenoïds unless a susceptor will fit inside; actually, years later we can observe that even when heated from the tip of Camouflet's Inducor (which does have flux concentration) its other pole happens to terminate toward empty space, so why not dare sacrifice windings density as a trade-off to bringing the 2 poles in close opposition?! After all, my weaving style of windings cause the workcoil assembly to become mechanically rigid, which i viewed as an asset in the perspective of exploiting microphonics with haptic feedback...

The aerial interstices which such frankenstein workcoil left behind made me want to add a secondary circuit dedicated to "closing" a feedback control loop, so i thought i'd keep that in mind while evaluating a few layout strategies, until the prospect of fitting it all inside the volume of a King Size cigarette pack struck the ex-smoker in me. Too bad the 14500 power cells only generate burst currents below 900 mAh @ 3.6 Volts or so, hence it would become neccessary to choose between having 3 cells or a storage compartment for its consumption tool.

Almost a year passed before i switched to drawing full bifilar workcoils using the same wire diameter, after realizing the Hall-Effect chips made the tiny wire secondaries redundant.

Here's one of my artworks which got published after i performed a bit of touch-up:

​

View attachment 2365172​

[ https:// i.postimg.cc/ZRDZXwN3/Egzoset-s-artwork-symetry-issues-trying-to-illustrate-2-sided-PCBoard-stuffing-2022-Jan-24-580x40.png ]​

This prove most useful in trying to understand the consequences of such pins layout and assessments, previously confirmed in those exploratory layout attempts starting with the basics and aiming for additional complexity:

​

View attachment 2365173​

[ https:// i.postimg.cc/fb8h08BF/Egzoset-s-novel-IH-Prototyping-Platform-concept-General-Purpose-Template-for-LAVACapsule-2022-Jan.png ]​

The main features put forward were inclusion of a PC Board with USB-C connectivity on it, plus continued emphasis on an opportunity to scan the cannabic vapour on its way out. Mention was also made about my intention to include a micro-switch for mechanical insertion/removal detection relative to its associated consumption tool.

My next drawing suggested how to implement dual-frequency support while also starting to envision where the high-current copper tracks would fit:

​

View attachment 2365174​

[ https:// i.postimg.cc/W4Y1PpLh/Egzoset-s-Half-Mazzilli-Royer-Center-Tap-Routing-Dual-Frequency-Switch-2022-Jan-12-800x800.png ]​

This exercise allowed me to realize that i could shorten those copper tracks somewhat, by reversing the winding's direction in one of my workcoils (identified as a "mirror-image" of the two others):

​

View attachment 2365175​

[ https:// i.postimg.cc/Y2WqJwXC/Egzoset-s-2nd-Half-Neo-Mazzilli-Royer-Optimized-Mirror-windings-U-Cores-layout-2022-Jan-15-800x80.png ]​

At that point i managed to figured out one scenario about how to enslave Mazzilli's « Royer ZVS », by resorting to the addition of a pair of Hall-Effect chips, e.g. 1 per branch:

​

View attachment 2365179​

[ https:// i.postimg.cc/8PbDSkLN/Egzoset-s-early-Cube-of-Borg-Panic-On-Off-dual-Freq-even-Audio-2022-Jan-20-800x800.png ]​

This helped a lot because i could begin to foresee a probable combination of components offering these features: a) Over-Current protection; b) Optical scanning of the fumet; c) Audible 2-channels feedback providing a hint on workcoils activity; d) Detection/Encoding as a function of the type of consumption tool being inserted, with dual-frequency control via some Solid State Relay (SSR); e) A Smart-Slave USB-C option accomodating external advanced functions performed by a computing device, like sessions recording for later pairing with physical cotton samples analysis (both for study & entertainment purposes).

Then another windings compromise imposed itself, confirming my belief that « function defines form »:

​

View attachment 2365180​

[ https:// i.postimg.cc/Vk6yPqYW/Egzoset-goes-further-into-Cube-of-Borg-suite-2-Sides-PCBoard-conflicts-2022-Jan-26-720x600.png ]​

So far i had resisted the urge to return to classic selenoïd windings with a flux concentration core inside. It still brought came with a magnetic bridge turning it into a "U" shape inductor but all pins were gather near the center, which actually implied my copper tracks would get even shorter at last!



Past this step i felt quite comforted this was ready for further 2-sided "U-shape" cores supperposition with the PC Board between two sets.

Little tweeking remained necessary, my copper tracks had moved into the central zone and the number of permissible coil turns was about to be known now that it all seemed to fall in place:

​

View attachment 2365181​

[ https:// i.postimg.cc/FzZ62R86/Egzoset-s-intermediary-IH-Artwork-for-Cig-Pak-Cube-Winding-Terminals-2022-Feb-5-840x700.png ]​

This last illustration is nearly finished, i knew which wire sizes would fit in and what capacitor values were possible. The next step could have been to draw copper tracks again and decide how/where to match them to a "daughter" PC Board supporting the Master-Slave & battery recharge USB-C connector:

​

View attachment 2365184​

[ https:// i.postimg.cc/FsPYFtqL/Egzoset-s-Primordial-2-Sided-General-Purpose-Backbone-for-Cig-Pak-Cube-2022-Feb-4-960x800.png ]​

Meanwhile non-rechargeable 14500 cells have appeared which can generate 3.6 Volts @ 2600 mAh, those could be useful in prototyping but they sell 40 $Cdn for 2! Consequently, i don't expect this particular « HR » compatible vision of mine to get evaluated in real form until battery technology has improved sufficiently. So, i sure like having a bit of similar day-dreaming - with a grain of nostagia since « It's already been done before, just not yet! »...

Good day, have fun!!

Click to expand...

Hi, you seem quite knowledgeable on induction heaters, what do you think of the circuit this guy has made?
He claims it could be used for induction heater, it seems compatible with my 20v 10a power supply that I'd like to get to work.
My goal is to get an induction heater up to around 180w without it self destructing within a year or overheating too much. I found this schematic and some pictures of his latest builds.

I sent him a message but he seems inactive on reddit now, the design uses a gate driver and a feedback loop to oscillate around the natural frequency of the circuit leaving the mosfets cool to the touch; He is able to play music through his arcs.

gomaruanna said:

Hi, you seem quite knowledgeable on induction heaters...

Click to expand...

Actually it's no major difficulty to search through the available literature and eventually realize that the « clever » element in Mazzilli's "Royer ZVS" resides in a core's saturation response, which implies the popular "Tesla modules" being sold without their Flyback Transformer have been a conceptual fraud since over a decade ago without anyone complaining...

In addition, this subject fueled scholar papers which i refered to on reddit before ending in a war zone where it all got rejected/ignored in block on the basis of identity alone. Quite possibly that's even impossible to trace back, but i remember the conclusions suggested this was an unfinished (prototype) topology plagued with some serious tendency to self-destroy in situations of improper/deficient power-up sequencing.

Much more recently the manufacturer of Vestratto's Vader/Forge announced there would be recalls followed by a delayed release of these previous versions:


This occured 3 months ago and i ain't been made aware of any updates so far. But at least instructions were provided on how to avoid catastrophic failure, a communications strategy also adopted by Camouflet about its Inductor... So to be honest i ain't gullible enough to think YouTube's access restriction that was later imposed on the December 15th video only translates as a coïncidence.

gomaruanna said:

...what do you think of the circuit this guy has made?

Click to expand...

Another skill that's shared by quite a lot of people is the ability to "ready" electronic schematics and it turns out the sample which was published here only describes how to implement an On/Off function by controlling a simple relay; to be more accurate its double feedback path remains fixed (e.g. dumb/blind) while the associated power twin-stage gets controlled by brute force instead! In other words it should appear quite noticeable there's no hint of a « zero-crossing » circuitry for the audio modulation...

The gate driver has no basic Rss/Rg(on)/Rg(off) structure dealing with the "Gate Charge" (Qg) of Infineon's medium voltage 5th-gen silicon N-ch HexFET, much less a version equally featured with 1 external capacitor plus diode. Maybe i'm day-dreaming but a transition to GaN technology would reduce the gate driver's parasitic capacitance workload 10-fold, which then would translate into less energy being wasted as heat.

gomaruanna said:

...seems compatible with my 20v 10a power supply...

Click to expand...

A ready-made commercially available IH cooker offers ~15 times the power, even with a 5-yr warranty if careful when shopping around. In such case the challenge would be to rewind its workcoil and make sure the workload gets detected or accommodated at least long enough to inject sufficient energy into your favourite consumption tool.

gomaruanna said:

My goal is to get an induction heater up to around 180w without it self destructing within a year or overheating too much.

Click to expand...

Yes, that's the idea of course.

gomaruanna said:

He is able to play music through his arcs.

Click to expand...

Percussions perhaps, but no music and certainly nothing with voices until the schematic is updated!

I'm sorry but those people who insist over use of the Royer ZVS have failed at the test of time and what i observe through their ~240 Watts flagships still isn't comforting unless they manage to compensate through more cheating like some "snubber" arrays which waste a little more energy while persisting over a feedback loop where magnetic saturation played an important role.

To top it all the matter of an energy budget needs to be addressed, for example Vestratto's 3250 Joules (for the Tornado) seems like moving far away from any form of rationalization in such arena!

That doesn't imply you can't have any fun at all, just that without integration of self-protecting mechanisms there's always a risk it will be short lived, as the prototype it always been.

Egzoset said:

Actually it's no major difficulty to search through the available literature and eventually realize that the « clever » element in Mazzilli's "Royer ZVS" resides in a core's saturation response, which implies the popular "Tesla modules" being sold without their Flyback Transformer have been a conceptual fraud since over a decade ago without anyone complaining...

In addition, this subject fueled scholar papers which i refered to on reddit before ending in a war zone where it all got rejected/ignored in block on the basis of identity alone. Quite possibly that's even impossible to trace back, but i remember the conclusions suggested this was an unfinished (prototype) topology plagued with some serious tendency to self-destroy in situations of improper/deficient power-up sequencing.

Much more recently the manufacturer of Vestratto's Vader/Forge announced there would be recalls followed by a delayed release of these previous versions:

View attachment 2393898​

[ https:// i.postimg.cc/fW365vKy/Vestratto-Forge-rev-1-2-2024-Dec-8-600x360.png ]​

This occured 3 months ago and i ain't been made aware of any updates so far. But at least instructions were provided on how to avoid catastrophic failure, a communications strategy also adopted by Camouflet about its Inductor... So to be honest i ain't gullible enough to think YouTube's access restriction that was later imposed on the December 15th video only translates as a coïncidence.



Another skill that's shared by quite a lot of people is the ability to "ready" electronic schematics and it turns out the sample which was published here only describes how to implement an On/Off function by controlling a simple relay; to be more accurate its double feedback path remains fixed (e.g. dumb/blind) while the associated power twin-stage gets controlled by brute force instead! In other words it should appear quite noticeable there's no hint of a « zero-crossing » circuitry for the audio modulation...

The gate driver has no basic Rss/Rg(on)/Rg(off) structure dealing with the "Gate Charge" (Qg) of Infineon's medium voltage 5th-gen silicon N-ch HexFET, much less a version equally featured with 1 external capacitor plus diode. Maybe i'm day-dreaming but a transition to GaN technology would reduce the gate driver's parasitic capacitance workload 10-fold, which then would translate into less energy being wasted as heat.



A ready-made commercially available IH cooker offers ~15 times the power, even with a 5-yr warranty if careful when shopping around. In such case the challenge would be to rewind its workcoil and make sure the workload gets detected or accommodated at least long enough to inject sufficient energy into your favourite consumption tool.



Yes, that's the idea of course.



Percussions perhaps, but no music and certainly nothing with voices until the schematic is updated!

I'm sorry but those people who insist over use of the Royer ZVS have failed at the test of time and what i observe through their ~240 Watts flagships still isn't comforting unless they manage to compensate through more cheating like some "snubber" arrays which waste a little more energy while persisting over a feedback loop where magnetic saturation played an important role.

To top it all the matter of an energy budget needs to be addressed, for example Vestratto's 3250 Joules (for the Tornado) seems like moving far away from any form of rationalization in such arena!

That doesn't imply you can't have any fun at all, just that without integration of self-protecting mechanisms there's always a risk it will be short lived, as the prototype it always been.

Click to expand...

Thanks for your detailed explanation. From your insights, I understand the common pitfalls in designing a robust ~200W induction heater PCB include:

• Relying on basic Royer ZVS topology without proper resonance tuning and zero-crossing detection.
• Inadequate MOSFET gate-driver circuitry (missing essential Rg(on)/Rg(off), Rss resistors, diode-capacitor structures).
• Not considering newer transistor technologies like GaN, which significantly reduce gate losses and heat generation.

Could you confirm if these indeed represent the main issues for creating a reliable, commercial-grade induction PCB at this power level?


Additionally, since I'll soon be hiring an electronics engineer through Fiverr, could you briefly suggest any essential design points or instructions I should clearly communicate to ensure the best possible results?


Finally, what topology would you personally recommend as ideal? I'm particularly curious about your opinion on the RM Cybernetics induction heater (link), which appears stable and reliable. Does their robustness stem primarily from using an advanced microcontroller (e.g., STM32) for precise control, feedback, and resonance tuning, or is there another key factor at play here?


I'll gladly share the final PCB design results here once completed—your insights would be much appreciated!

gomaruanna said:

Could you confirm if these indeed represent the main issues for creating a reliable, commercial-grade induction PCB at this power level?

Click to expand...

My comments were in reaction to Mazzilli's Royer ZVS topology which failed to evolve since 2008~2009, still avoiding the matter of core saturation being initially exploited in some unusual manner while the insertion of gate driving within its feedback loop still won't address such odd characteristic that made it "clever" in days when it was built around thermo-ionic tubes.

gomaruanna said:

...could you briefly suggest any essential design points or instructions I should clearly communicate to ensure the best possible results?

Click to expand...

Although i've practically abandoned my hopes to see that old riddle fully resolved i keep a trinket on my desk as one fair reminder of what i once envisioned as the avenues to explore. It dates back to the early pay-TV era, it's a pair of daisy-chained "dual monostable multivibrators", more exactly the 74LS221 with Schmitt triggers plus the 74LS123 retriggerable version. This goes back to over 40 years ago but i recall they formed a free-wheeling loop able to regenerate the suppressed synchronization signal in baseband video... For me the main difference today is that i'd go full digital and delegate a maximum of functions to software for the candy part, similarily to this:



Admire the simplicity in a Tayloe detector, that's all it takes to implement quadrature IQ demodulation using software-defined radio!

gomaruanna said:

...what topology would you personally recommend as ideal?

Click to expand...

My reasoning in terms of "pitfalls" is this: dealing with induction heat it should be clear that D.C. ZERO Hz is a forbiden state, hence the idea to rely on pulse timers. But those are actually hybrid digital/analog chips so i wonder how noise from the high-power stage would affect them. Years ago i found a chinese datasheet where the timing elements were made of down-counter logic, and i think its supply voltage was around 12 ~ 15 Volts, which suggest high noise immunity.

Yet i believe some aspects of the Royer ZVS remain desirable if one cares to have protections built-in. That's why i considered the Allegro ACS733 Hall-effect current sensor featured with evolved integrated A/D functions, though these days i might prefer Infineon's TLI4971 as this other device raises the current detection ranges to 25, 50, 75 or 120 Amperes:


The trade-off here being a lesser bandwidth and phase error margin, but it's not going to operate above 100~150 KHz anyway.

So much for the "robust" part, insert 1 chip in each branch to simplify the detection/ban of D.C. & over-current states while trying to stick to a K.I.S.S. design orientation.

After that i'd feel tempted to experiment with electric stress distribution both to improve sturdiness and fall-back redundancy via cooperative magnetic addition, which then brings the matter of spreading that energy through multi-poles XY (Cardan) fields also applied to back-EMF detection, ideally even in a 3-D XYZ configuration. Another trick for dealing with radio parasitic emissions might reside in floating/flying switched capacitors as in the MERU Class-D HiFi Audio (96 KHz) amplifiers already mentioned in February '21.

I wish i could relocate hint notes about new hybrid power devices combining the excellent Qg of GaN with other qualities of the more conventional iGBT transistors, all in one single package. On another hand it's quite probably overkill on top of getting obsolete fast as technology improves, which in the end would require searching through up-to-date catalogs - a demanding task.

Honestly, this is all starting to feel overwhelming, not to mention it remains to be evaluated via prototyping. Well it felt good to put it in words again, too bad i can't really offer specific advice to the attention of your Fiverr correspondant. Since you have microcontrollers in mind i can only suggest the "ideal" topology would be a synthetic master/slave Royer ZVS built around a minimalist number of parts as illustrated previously, keeping in mind there's room for modifications like to include a classic air-core helix-shaped workcoil.

gomaruanna said:

...I'm particularly curious about your opinion on the RM Cybernetics induction heater, which appears stable and reliable.

Click to expand...

I yet have to see its schematic diagrams. To me robustness needs to be demonstrated with a long/patient evaluation of catastrophic failure modes, e.g. attempting to make it burn... So far i generally get a vague impression they all gamble on over-sizing, but i'd be glad to get proven wrong!

gomaruanna said:

I'll gladly share the final PCB design results here once completed...

Click to expand...

It will be with great pleasure, every step forward counts!!