1. What does SSTB mean? See our glossary of acronyms.
    Dismiss Notice

The auto-magic super rZToaster, or something like that..

Discussion in 'DIY' started by rz, Apr 3, 2017.

  1. rz

    rz Well-Known Member

    Driven by my love for the @VapCap and inspired by @Pipes awesome idea to use an induction heater for it with his VapCap Induction Heater for Desktop and in Car Use project, I set out to build an induction heater of my own.

    Since I'm a hardware guy in a software company, I crave for the chance to build things when I can, and this project really gave me so many cool things to play with.

    I started out building a barebones royer oscillator - the driving circuit that creates an immense oscillating magnetic field in a coil. When a conducting object is inserted into this coil, such as a vapcap :brow:, large currents of electricity are produced in the item, causing it to heat up.

    Now since I enjoy playing with microcontrollers, and have played with many little components, the possibilities suddenly seemed endless. Before I knew it, I had a small OLED display, and a contactless infrared temperature sensor hooked up. The plan was to have the IR sensor look at the end of the vapcap, and turn off the induction heater when it reaches the right temperature. It was soon obvious that the microcontroller was overkill for just that, and I needed to give it some other things to do :D


    Now, I'm not the first 'mad scientist' on FC :science:. Egged along by other buildings and innovators here like @hardboiledfrog, @Andreaerdna and others :tup::tup::tup:, and inspired by their respective temperature controlled projects, I implemented a PID algorithm. This is a smart way to get an object up to temperature quickly, without going PAST the temperature, and keeping it there - SOU VIDE FOR THE SOUL :D:D:D

    Since then, and MANY blown transistors along the way, my project has evolved.. just a little :brow:



    It's still very much a work in progress, and I'm currently waiting for parts to fix the oscillator that's built on my prototype toaster, but in the meantime, here are some of the features:

    • Adjustable temperature, maintained by a PID system.
    • insertion detection
      • This is achieved by the awesome property of the induction heater, that when an object is placed inside it, the frequency of the oscillations changes! So when the device is ready, it consumes very little power unless an item is INSERTED. A baseline frequency is set when activated, and a threshold change needs to happen before properly heating up begins.
    • Battery monitor - Both voltage and current. By using the microcontroller's built in differential analog inputs and programable gain amplifier, power consumption can be monitored. System will shut down if running out of spec.
    • Multiple control / safety points. Will have at least two internal fail-safe systems that need activation before powering up. One will be a transistor that supplies the full battery voltage to the oscillator, and another transistor that supplies actually controls the circuit. Additionally, everything will go through a fuse.
    • OLED display! :) It even has a graph showing a line for the target temperature, and the actual tip temperature.
    • Powered by 3x 18650 batteries for hours and hours of fun :D
    • watchdog timer will quickly reset the system if it gets stuck in any state for some reason, which will promptly power down all circuits.
    • Low power mode (not implemented yet), hopefully for weeks of standby. I haven't found a nice switch that I want to mount on a case yet, so I may skip the main switch all together.
    • TBD

    For now, I haven't had EVERYTHING working all at the same time yet, but each component and feature is coming along nicely, and I hope to be done (enough) soon.

    Not sure how much time I'll have to push this project further with my third kid about to pop out any day now, though I'd LOOOOVE to turn this into more of a product.. get it printed on a proper PCB, etc, and will gladly work with some of the community here to move it along if possible.

    In the meantime, here's the latest video showing insertion detection, with glimpses of the baseline frequency detection and threshold checks, the huge mess left in the wake of building such a project on a small desk, specially after hours of testing this thing out :lol::lol::lol:

    So, who wants one ? :worms::rockon::science: (don't hold your breath, its many many months away from that, at best. Go grab a portside in the meantime! :) )

    Click to play YouTube Video

    For more info about the start of the project and the journey it's been on, check out @Pipes's VapCap Induction Heater thread. Thanks to a bunch of you guys over there for the help and support! :)
    Last edited: Apr 3, 2017
    SavageCore, szai, Cl4ud3 and 55 others like this.
  2. Quetzalcoatl


    SDCA, 3rd Planet
    Fucking awesome!!
  3. KeroZen

    KeroZen Chronic vapaholic

    On Air
  4. Hogni

    Hogni Honi soit qui mal y pense

    Berlin, Germany
    Great project I'm already watching for a while :tup:
    Good idea to create an own thread on this project.
    It's a overkill for sure tu use it just for an induction VapCap heater.
    But might be a base for an induction enail-heater/-controller or an induction vape in the future?
    Will follow with much interest...
    asdf420, Kermitt, Prolusio and 10 others like this.
  5. CheeseSandwich

    CheeseSandwich Well-Known Member

    Wow! Very cool.:tup:
  6. JCat

    JCat Well-Known Member Accessory Maker

    Ontario, Canada
    If I could like this more than once I would! Super cool!
    Prolusio, GreenHopper, muunch and 9 others like this.
  7. The Beagle

    The Beagle Bubbles & Bags

    Copacetic, Quetzalcoatl, rz and 2 others like this.
  8. VaporSipper

    VaporSipper Well-Known Member

  9. rz

    rz Well-Known Member

    Worked on sleep mode a bit. Consumption is down to 3.8mA, and that's probably mostly peripherals (IR sensor and OLED). I'm pretty sure I can cut this down significantly, probably by cutting their voltage off completely while sleeping. Even at 3.8mA, the battery would take almost a month to drain in sleep. I'll be happier when it's under 1mA in sleep.

    As for waking up, it requires a long press. If not pressed for long enough, backwards (literally) to sleep. I added some graphics for that :)

    Click to play YouTube Video
  10. Joaon

    Joaon Well-Known Member

    That wake up/back to sleep graphics are awesome!
    Great job ! I 'll keep a eye on your thread :cool:
  11. rz

    rz Well-Known Member

    YAY. So, the IR sensor that I have doesn't have a sleep mode, and it consumed about 1.5ma - right on spec per it's data sheet. The OLED driver, even while off, also consumed about 1.5mA.

    After putting them both under a PNP switch, I can turn them off completely :) Unfortunately, the re-init sequence took me a few hours to get right, but it now works perfectly, and I'm measuring ~0.9ma, which is good enough for now :)

    geez though.. for instance, if I tried to init the IR sensor too soon after powering it back up, it started consuming about 50ma.. wtf.. :bang:

    I also cleaned up the board a bit since I'll have my mouser order in hand soon. Going to rebuild the whole oscillator, so I just removed that whole section. I'll also have my precision resistors for more accurate voltage detection and current sense. Note to self: Make sure to hook the ground back up to the bottom of your voltage dividers, or you'll get the full 12 volts at the input of your analog converted. Yep. I almost lost a microcontroller, but now the ADC on those pins just doesn't work :bang::bang::bang:

    Good thing I have a new microcontroller in my parts oder :cool:. I'll keep using this one though while I finish building things, and will pop the new one in when I determine it's all safe.

    At the end of the day, all's good. OLED and IR sensor now work as expected through sleep cycles and I'm under my 1ma target. I expected it to go lower than this, but good enough for now :rockon::D
  12. phattpiggie

    phattpiggie Well-Known Member Accessory Maker

    All this electronics stuff goes straight over my head so it's great seeing yourself and @Pipes exploring new ideas.

    The creativity this little device is spurring on is unheard of in the vaporizer world.

    Looking forward to seeing how the final product looks and performs.
    MoltenTiger, Squiby, V-apE and 12 others like this.
  13. whammy_time

    whammy_time Member

    So bad ass! I'm an arduino hacker myself, so I immediately took a liking to this setup. Reminds me of looking at my own "work bench" (dining room table...).

    I'm curious how you do temp regulation? From that video, the microcontroller is in one place, your coil is in another, and you contact-less insert the vapcap into the coil, right? Is there a known frequency relationship with the desired temperature, or do you just have to eyeball it? Sorry if I missed that explained somewhere.

    I haven't tried do do induction yet, but this makes me intrigued for sure!
  14. rz

    rz Well-Known Member

    MiniMe 3.0 has finally be released! :D :D :D

    Under the circumstances, baba and mama are doing GREAT, and may even come home today :)

    In the meantime, I received my package last week, managed to put the oscillator back on the board and get it rocking. still need to hook up control circuits, no idea when I'll get a chance.. though I'm bound to neglect other more important things I need to do and get more done sooner rather than later :science:

    @whammy_time I'm using a contactless IR temperature sensor for temp sense. I've had some prototyping challenges so at times I've tested and worked on different sections independently, hence the decoupled mcu and coil. If you look at the second picture on this page, you'll see a little board (one of these) where the IR sensor sits behind the work coil. In the video that you saw without any sensor hooked up, I was testing out a way to sense the frequency of the oscillations. Since the frequency changes when I insert the vapcap, I can detect the insertion and start heating when it's in, and stop heating when it's removed :)
  15. Copacetic

    Copacetic Somewhere North of The Wall

    Congratulations to all three of you! (or does 3.0 mean that there are now 5 of you?!)

    Great work on the vape too :science:
  16. rz

    rz Well-Known Member

    He is our third boy. So now we're 4 monkeys and wonder woman :)

    And I've even managed to make some progress :D

    OK.. A LOT of progress :rockon:

    To cut to the chase this this is ROCKING :D My favorite new feature is incremental mode!
    I like to start it at tasty 98°C, and every time I remove the vapcap from the coil, the setpoint jumps up by 2 degrees. The system then waits in low power READY mode, only firing up the induction circuit 1% of the time to measure frequency (optocoupler schmitt trigger + extra diode and resistor work perfectly connected straight to the work coil). When it detects a frequency change indicative of vapcap re-insertion, it fires it up to two degrees hotter than last time, and so on. I just place the vapcap in the toaster between every puff, and it takes care of ALL the rest for me. :brow:

    Now the click is a tricky thing with induction heating. If I set the thing to say 115°C, the cap will click pretty quickly, around when I measure 110°C, and I've even slightly charred a bowl in a race to just 120°C.
    However, if I heat up much slower (incrementally between puffs), it only clicks when IR sensor reads 125°C. Either way, I've been finishing some bowls without even hitting a click at a measured 120°C.

    I say measured a lot since I'm not claiming that that's the actual tip temperature where it touches the bud. This is obviously some science involving a bunch of coefficients I don't really want to start thinking about yet, but hey, it works :D

    Dealing with the noise generated by the oscillator was a challenge. A bunch of bypass capacitors and I was over an annoying OLED image disruption almost every time the thing fired up. This will really benefit from a properly designed PCB.

    The first set of MOSFETs that I tried worked but had some noisy switching issues. Either way, when I finally got the case I ordered, I was trying to squish everything in. I bent the mosfet heatsinks back and their pins broke :o When I changed them to a different set, oscillations were MUCH smoother :tup::tup:

    Alas, the case is about 1mm too small inside. The clearance of the battery, the board, and the coil is a tad too tight for my comfort. I've come up with some ways to reduce coil thickness when I wonder about coil design..

    Now the batteries.. Up till recently the work tank on this thing would consume about 7Amps while heating a vapcap. This is a problem since the voltage drop on the batteries internal resistance increases, and the BMS would shut off due to low voltage cutoff, as it should. I've since removed one of the two 0.22uF caps I was using, and it is now consuming a more modest 4 Amps. This should reduce BMS issues. She now oscillates at 430Khz while idle, and uround 465Khz while toasting a vapcap, compared to circa 340Khz with 2x0.22uF (@Pipes you should experiment with this too, remove one of those 0.33s and check consumption and voltages under load). I'll be happy if it can consume even less.. Heat up speed does not seem to a problem, and if it goes a little slower battery life (BMS shutoff wise) will be significant longer.

    So time to first setpoint is now about 10~15 seconds from cold. Re-heat after a puff or two even to the next incremental setpoint is very quick though, about 2~4 seconds, and since temp is maintained it's perfect to just place VC there between puffs. This is reeeally convenient and I'm loving experimenting. I just can't experiment as much as I'd love to since.. well.. life :lol: but ya, I'm happy it's finally all working.

    There are still so many little things to play with, so this will keep me busy for a while. I'd love to share it, though it really needs to be on a properly designed board and built properly to be safe in anyones hands, and that'll take a while at best ( Anyone wanna help? :) )

    OKK... it's late but I'll go take some pics quickly. One day maybe even a video :lol:


    P.S This is the closest vape experience to smoking a spliff I've ever had.

    Last edited: Apr 13, 2017
  17. hardboiledfrog

    hardboiledfrog tinkerer

    nice progress, very cool with the frequency detection! :cool: good idea to reduce the cap, this application doesn't require much power. you could also drop down to smaller mosfets in your ZVS to gain space, TO220s aren't really needed. I would recommend going to surface mount components when you do your board. they really speed up assembly and save a lot of space. :tup:
    so, are you finding the IR sensor gives pretty consistent readings? does the VC rest against it or is there some kind of spacer in there?
    is that a Ti VC? does the body get very warm if you finish off a load in one session?

    for an induction vape I was thinking of using one of these as a hot air engine to feed a glass stem with a basket like the shitfucker uses. not sure if that is enough Ti mass though. :hmm:

    I'll have to get one someday but I already have too many irons in the fire ATM that have a very low WAF. :rolleyes::uhoh::D
    MoltenTiger, GreenHopper, b0 and 4 others like this.
  18. rz

    rz Well-Known Member

    @hardboiledfrog yep def. could do with smaller transistors. These puppies hardly get warm. + The capacitor gets a little hot, specially multiple vapcaps are being toasted.

    The IR sensor is working very well, very smooth readings. What it's actually measuring, and how accurate, is another issue. It measures the cap centre, for sure, but most of the heating occurs in the circumference of the tip, not the centre. This reveals all sorts of interesting things about click temp vs bowl temp. Like i've mentioned, when heating quickly (like a triple torch would?), the click can happen when centre temp is 100 even, or at random temps depending on the gradient. When I heat slowly, I consistently get my clicks at 148 and about 152 for the second click - and is a little too toasty at this point - P.S: This could be why some slower single jet lighting could result in an over toasty bowl.

    The coil is wrapped around a glass tube, kind of like a flat bottomed test tube with a large hole in the bottom, so it sits in nicely but it stopped by a glass lip. I then used jb-weld to make a little mount on the glass so it fits nicely on the sensor. The sensor itself is about 3~4 mm away from the tip of the cap.

    Body doesn't get warm at all. Heating seems very concentrated.

    My new mode is slow mode - I heat it rapidly till 80 degrees, then I heat it at a rate of 1 degree a second by slowly increasing the setpoint up to a pre-set limit. This gives very consistent behaviour since the heat sinks in nicely. I like to toke it at a reading of 120~130.

    So so far 3 decent modes:
    SET - gets there quickly.
    INC - starts at predefined start point, increments 5 degrees each toke.
    SLOW - insert, and it gets warm quickly, then slowly creeps to set max. remove when you want or settle at max.
    + Found myself incrementing the SLOW max by 5 degrees, so SLOW+INC mode ? :D:brow::science::worms:

    Also, since I don't know the exact emissivity of the cap, these temps may be off by a bit - however, the behaviour is consistent.

    I forgot to ask @VapCap to throw in some titanium shavings for me in my last order - I think those would be interesting to work with for convection testing..
    Last edited: Apr 15, 2017
    MoltenTiger, Kermitt, marduk and 11 others like this.
  19. rz

    rz Well-Known Member

    So, it seems that my IR sensor needs the tip to be VERY close to it to get a good reading, due to the (too) wide field of view of the sensor. I thought it was close enough, but the temps I was getting didn't make sense, and I couldn't figure out why I was getting SUCH different cap readings depending on heat speed. It's just that some average between actual tip temp and the surrounding view (the induction coil, once it warmed up) was throwing it all off. :doh: There was still a large cap temp element in the reading, so it managed to work before, just with other variables involved

    After thinking that it's worth investigating, I bent the coil out (and the glass tube + adaptor) and away from the IR sensor, heated the cap separately, and put it up against the sensor - EUREKA! Now I', getting about 210c (~410F) clicks up and about 155c (~310F) down. This much MUUUUCH more sense with the camabinoid evaporation temps.

    So I filed my adaptor down a bit, and got the cap to rest ON the IR sensor, and it's working as it should now. I need to re-tune the PID a bit, and make a better coil.

    Sorry if I caused any head scratching about my temps (specially to my own balding head), glad I'm back on track.

    ..Now I need to go experiment from scratch :brow::science::leaf::D
    Last edited: Apr 27, 2017
    MoltenTiger, Kermitt, marduk and 14 others like this.
  20. rz

    rz Well-Known Member

    I finally started playing with board design software, Here's a sneak peek at what I've been up to.

    Still very much WIP. I decided to modularize things. This is the high voltage / induction board. It will house the induction circuit, the optocoupler, current sense circuit, the 5v regulator for the digital side (I think I'll actually remove it, and stick it on the digital board), and the IR sensor (not yet on the board). It also has control circuitry for the choke supply and the gate resistor supply.

    I'm yet to design a daughter board for the digital side.

    In the meanwhile, since I re-oriented the IR sensor, it is working really well as long as the vapcap is sitting nicely. I've ordered some IR sensors with narrower field of view which should make it even more stable.

    Last edited: May 4, 2017
  21. GreenHopper

    GreenHopper 20 going on 60

    This is all looking mighty awesome to me!

    Love that 3D render.

    If there is a wait list I want on it, if there isn't a wait list I still want on it.

    Hows your new kiddie doing, everyone well?
  22. rz

    rz Well-Known Member

    @GreenHopper I'm on that list too! I can't wait to get this thing fabricated.

    Kiddie is doing well, all things considered. He's loving mama's boobies :brow: exactly what the doc wants before surgery, for him to chubby up. His bloody oxygen is saturated so thats great, though they can see deterioration already.. We hope to be able to wait a few months to get it all done in one go.
  23. GreenHopper

    GreenHopper 20 going on 60

    I'm delighted to hear he's growing stronger, I wish you guys all the best.

    I can only imagine how awesome this thing is going to be. You are essentially heading towards having an induction based vape with the potential for accurate temperature control as well as custom session routines, now that's taking things to another level. :tup:
  24. rz

    rz Well-Known Member

    A little update.. cause well, I've been busy :D

    Since this is my first board design, this is very much an experimentation board. Though now that I can build much more complicated circuits, I'm... experimenting :)

    Here's what it should look like now:


    Why yes, what a sharp eye! That is indeed a DC/DC boost circuit right next to the gate resistors! :science::science: (LM27313/-Q1 1.6-MHz Boost Converter With 30-V Internal FET Switch in SOT-23 ). I followed the design rules (I hope), and have it set up to output 18V for the gate resistors. This should help keep the main oscillator transistors cool, and allow me to experiment running this thing with just 2x18650 batteries.

    Gate resistor voltage is switched by a nice little complimentary N+P channel set in a single package. The actual voltage source can be selected between Battery voltage (~12v) or the boosted voltage (~18v) by a jumper, just in case the boost circuit doesn't work (yet).

    Power to the chokes is also switched, though I'm trying out a "smart switch", basically a transistor with all the bells and whistles around it - BTS50060.

    Current is measured with 2 separate solutions cause well.. :science:.
    I'm using 2x INA210. One for charge current detection and one for drain detection. These little puppies basically output an analog signal that the MCU can read with it's analog to digital converters.

    I'm also doing the same with 1x INA219, which outputs a digital signal with battery voltage and current. This covers both charge/discharge. This is a much nicer solution than 2x INA210 (I could have used one, but didn't want to deal with bias offset in a noisy environment).

    These little current monitors have pretty high gain (I'm using ones with 200x amplification) so they enable the use of very small shunt resistors. I'm using 1mOhm for now, which will hardly impact the system while running.

    Currently, I think the BMS board probably has the highest unwanted resistance in the circuit (~100mO). If I go for 2x18650s, I may skip the balancing part since boards without them have much lower resistance (~10mO), critical if i'll be drawing higher current to compensate for the lower voltage.

    The BTS50060 smart switch for the choke also measures current, along with delivering a bunch of diagnostics info.. so, lots of info to process :)

    I've ordered some IR sensors with a 35 degree FOV rather than 90 degree that I had now. This will improve the reliability of my readings since now I have to make sure my vapcap sits in 100% flat or stray IR gets into that huge FOV.

    I have 2x 3.3v regulators with very low quiescent current. One for MCU and current measurement circuits - things that I want to be on all the time. The other for all other peripherals (frequency counting optoisolater, IR sensor, OLED) - basically anything that may use too much power while in sleep mode. I previously used some transistors to switch these things on, though a whole regulator with an enable pin costs just as much (<$1), and is much simpler :D

    I'm currently at the refinement stage. Going over datasheets again for proper pin assignment and component value selection. I hope to be done in a few days and get this board manufactured somewhere. I don't think I'll get all my parts in and built in the next month or two though, but hopefully not much longer than that.

    I'll build a few of these boards with little variations to refine my selection and have my local vapcap buddies test them out. Once I know the circuits themselves work, and I have final component selection, I'll work on removing the unused bits and redesigning the whole board layout for a smaller form factor.

    For now, my franken-prototype toaster is working like a king. It needs an occasional fixup every now and then since I haven't put it in a proper case yet, and I need to hold the VC straight due to wide FOW, but it's reliably giving reproducible results..

    So, don't hold your breath, but it's coming along :science::D :rockon::rockon::rockon:

    Otherwise, little baba is doing well and keeping us busy and tired as expected. He manages to keep his oxygen levels mostly saturated most of the time, and we hope he manages that for another few months so they can do full repair in a single op.

    P.S Not bad for my 100'th post :D
    Last edited: May 17, 2017
  25. GreenHopper

    GreenHopper 20 going on 60

    Been on vacation recently so only just read this update.

    I'm amazed that you've gone from general chat about potential fine tuning of a VapCap session based on induction charging to prototyping and alpha testing. Truly blown away fella.

    It's all looking awesome to me, I mean I don't know the in's and out's but I'm following the general gist.

    I'm liking the idea of dropping down to 2x 18650's making the device more compact.

    Not sure losing the BMS balancing is going to make too much of a difference, one can always chuck the batts onto an external charger once in a while to compensate. :shrug:

    I'm curious to see how things go with your alpha testers.

    Defo sign me up for one of these :D

    Glad to hear baba's doing well :tup:
    MoltenTiger, Sketch420, Joaon and 3 others like this.

Support FC, visit our trusted friends and sponsors