The Bud Toaster - (currently: Model 14, version 3)

[zenmasterofzinfandel]

Hey Norb,

while I know most people in this short life we have (believe me I know some hyper people) have a limited amount of patience, therefore the faster the unit can heat and start delivering the goods, all the better.

however for ultimate portability, perhaps a compromise model would work well, also?

I'm thinking the much higher capacity 18650 cells you can get with relatively safe (as good or better than a cell phone burning in your pants ) protection via circuits and cell chemistry and design over typical Li-Ion or Li-poly we know now.

Size is another factor, as well as weight for portability, you (whoever is designing the ultimate compact portable) might obtain by using 18650 cells.

These are built with Panasonic NNP series LiNiCoO2 chemistry. 2900ma..but only 5.8A max discharge rate. means 2x or more slower heating, but still acceptable for positive qualities over the other cells. Would be easier to source chargers for these.

*http://www.cpfmarketplace.com/mp/showthread.php?t=230876* don't know if hotlinks are allowed so I post just a static html addy.

Sorry, did not have the time to read the whole thread on my slow connection.

If it were up to me, I'd only want a single sealed unit, batteries and all, in the smallest form factor possible. Whether or not the batteries can be replaced by user is up to you, but I'm not looking for a desktop unit as a 'portable'. needs to be built well enough, in a single, easy to hold unit.

 

[Hippie Dickie]

If it were up to me, I'd only want a single sealed unit, batteries and all, in the smallest form factor possible.

i thought so too... for a while... until i made one... (about a year and a half ago -- picture in post #104). it was just too heavy with the two 26650 batteries -- one of many problems. A123Systems has a 18650 size also, but it is not that much smaller. Battery technology just isn't there yet for an all-in-one unit, imho.

As i see it, the cube and the cylinder are just 2 of many possible configurations. Exercise left to the reader, eh? The circuit diagram and software are all there to help you get what you want.

Actually vaping unit is fairly small, but the pix here and in many parts of the thread don't show the base unit 'tethered' to the vap

well, the base unit is just 2 cells and a cable. Here is the complete vape ( cylinder version):

 

[Hippie Dickie]

I'm not very well versed on the technical details of the cube but I must say aesthetically it's very appealing. It's really awesome how the design is so small and minimal.

thank-you. two chips is all it takes for accurate temperature control. we live in amazing times.

 

[Hippie Dickie]

Okay, so that picture two posts up is from several months ago. Here is today's picture of the "base unit" (battery pack) tethered to the vape, a caddy to hold the tweezers and glassware (stem and vial, when not in use), and my most favorite version of the Bud Toaster in 9 years -- the Perfect Cube:

The two cell battery back is 2 1/4" wide x 2 5/8" long x 1 1/4" high. The cable is 36" long, so the battery pack can stay in one place (on my desk) and i have a very light, small, vape body to hold and toke.

i still haven't worked out the design for a battery sled that holds the two cells, incorporates the caddy, and winds up the cable. It's eluding me, but really, i've been focused on other issues.

The Bud Toaster is definitely NOT hands free -- my philopsophy is that the glass should be able to fall away -- fall clear -- rather than be knocked over and break into glass shards (which ALWAYS happened with my hands-free designs). That means i must hold the stem against the top of the cube to take a hit. This works for me quite well. On the other hand, this may not be the vape for everyone.

People like to joke about all the testing required (well, at least i do) but i take my vaping very seriously. If my bToaster isn't working everything else stops until it is up and running again. Fortunately, i now have 3 in-house bToasters with the same software version -- so i'm covered.

 

[Egzoset]

Hi,

I'm curious, with some more code would your software be able to plot temperature vs energy curves?



The reason for this question is that i guess such data could be exploited, eventually. For example:

http://www.fuckcombustion.com/viewtopic.php?pid=92661#p92661

Evidently, i don't have all of the answers but i sure do have a few questions in mind...

 

[Hippie Dickie]

you mean, plot on a screen? but no anyway ... i don't have a sensor input to read current and voltage which i would need to calculate energy -- ran out of input pins on the 8-pin 12F683 processor i'm using.

Perhaps i could infer the energy used from the temperature rise for a fixed duty cycle, but i haven't found the temp/energy relationship to be needed. i'm getting 1F temperature stability with my integer implementation of classic PID control theory -- that's 0.25% error ... that's good enough.

 

[Egzoset]

Hi again!



In this particular case the on/off ratio controls how much energy gets delivered to the heater, which means the load goes 100% ON or 100% OFF during periods of variable duration; compared to using a linear solution your method reduces heat dissipation in the MOSFET transistor and it leads to a more elegant design as a by-product - which i find suitable myself!

It may be tricky to tell what's the energy level in absolute numbers but the ON/OFF ratio is all that should be needed to detect transitions in the delivery of energy. Of course, having all of your microcontroller pins busy makes it impossible to share such information with the outside world so it may be difficult to determine what's really going on.

Re-assignment of the GP0/GP1 lines to a serial-port may enable communications with a computer but then chances are you just might want to get rid of the microcontroller completely as well and this would probably interfere with your primary objective anyway. Well too bad, lets say i'm short of luck!



Nonetheless... Do you believe it's reasonable to assume that matter absorbs energy differently when there's a transition as in the graph from my link? In other words, what happens to the energy output once all of the THC has been vaporized?... My guess is there would be a momentary surplus of energy and that free energy would cause the temperature to rise because a part of that heat isn't being absorbed anymore (relatively speaking). If that is true, this might provide the means to optimize battery use, among other things. Though, i suppose it's also possible no transition would ever be monitored due to the tiny amount of matter involved in a typical vaporizer...

I started to value your opinion a great deal after reading this most creative thread...

Any comments?

 

[DeepFried]

People like to joke about all the testing required (well, at least i do) but i take my vaping very seriously. If my bToaster isn't working everything else stops until it is up and running again. Fortunately, i now have 3 in-house bToasters with the same software version -- so i'm covered.

That is called "passion" and it is your passion about your bud toasters that makes this one of the best threads on FC! With your dedication and passion you will most certainly succeed with your vision. Your honesty and ability to listen to other peoples ideas instead of just dismissing them is a rare trait. Keep up the great work and thread, I am sure there are a lot more lurkers like myself that really enjoy your posts but don't comment much.

 

[Hippie Dickie]

@egzoset - not quite sure what you're asking, but ... the 12F683 has an on-board PWM, so, yes, it is 100% on or 100% off, but it is running 250 PWM cycles per second -- i am re-calculating the duty cycle of the PWM 3 times per second. The nominal duty cycle value for 420F is 144. (out of 1024.)

3 of the 8 pins are used to activate and clock data from the MAX6675 (and 1 for PWM output, 1 for green LED, 1 for 3 button input (resistor divider array into analog-to-digital converter), 1 for power, 1 for ground). This duty cycle value is highly dependent on the physical configuration of the vape -- the cube contains the heat much much better than the cylinder where the heat shield was 1/4" farther away. Can't do without the PIC 12F683, but i think this is a minimal design for temperature control.

Once i got the MOSFET properly specified (thanks to advice from a reader), i am seeing no heat dissipation in the MOSFET.

And you're right that toward the end of the vape session, when most of trichomes have spilled their guts, the temp does rise a couple of degrees -- this is where i need to be more agressive about dropping the duty cycle value. Heating the empty trichomes doesn't absorb as much energy as vaporizing the THC goo. The transition can be monitored, but there are several confounding influences -- such as voltage drop as the batteries discharge over multiple sessions.

In terms of outputing operating data, i am still recording in eePROM (memory that retains contents after power is disconnected) the most recent 20 seconds of temperature readings and duty cycle settings, which i can unload using the PIC programming board. This has been an incredibly valuable debugging aid.

i still have 1/3 of the code space unused, so i can crank in some more AI. But i'm thinking about some user interface mods -- such as selecting temperature profiles, i.e. varying the temperature over the duration of the session. Or selecting a "bud drying" mode to purge excess moisture from the bud before initiating the vape session. Only three buttons makes this a challenge -- i could probably add another button before i run out of space on the pcb.

@DeepFried - i think passion says it all. i always ask any adolescent that will stand still long enough to chat with a gray head what is their passion. Passion is the energy that powers life -- at least a life worth living -- imho, of course.

 

[Egzoset]

So, if i understand you correctly, you have collected 20s records of the thermocouple (temperature) readings from the MAX6675 chip combined to the PWM duty-cycle ratio (energy) settings from the PIC 12F683 microcontroller...

 

[Hippie Dickie]

yes ... there are two ring buffers that overwrite after ~18 seconds (112 bytes, 2 bytes per reading, 3 readings per second). i reserve 16 bytes of eePROM for operational settings, e.g. SETPOINT and maximum duty cycle (adjusted to accommodate higher voltage power supply up to 20vDC). And another 16 bytes for "factory settings" so the user can reset these values if something stoner happens.

 

[stinkmeaner]

People like to joke about all the testing required (well, at least i do) but i take my vaping very seriously. If my bToaster isn't working everything else stops until it is up and running again. Fortunately, i now have 3 in-house bToasters with the same software version -- so i'm covered.

That is called "passion" and it is your passion about your bud toasters that makes this one of the best threads on FC! With your dedication and passion you will most certainly succeed with your vision. Your honesty and ability to listen to other peoples ideas instead of just dismissing them is a rare trait. Keep up the great work and thread, I am sure there are a lot more lurkers like myself that really enjoy your posts but don't comment much.

Not many of us can comment much due to the specific technical aspect but it is enjoyable to read nonetheless. What ever happened to the guy that was building his own Bud Toaster type version? Bubar.

I wonder if we will soon see an offering come on the market from him.

 

[Hippie Dickie]

i often wonder if we will see a product come on the market from me.

 

[Egzoset]

Well, do you think it would be possible to use a visual signal to warn about energy level transitions?

 

[pakalolo]

Well, do you think it would be possible to use a visual signal to warn about energy level transitions?

The Extreme does that. It might be the thing people complain most about. People hate that the little red light flashes.

 

[Hippie Dickie]

Well, do you think it would be possible to use a visual signal to warn about energy level transitions?

yes, the Bud Toaster does that, too -- the red LED under the oven tube is on when the heater is on, so it dims and brightens as the duty cycle is lower or higher. It's the red ring in the photos:

 

[Egzoset]

No, my question wasn't about the heater. I guess i should have been more specific by asking about the energy absorbtion level transitions in the blend...

 

[Hippie Dickie]

can't do that in this minimal version with the 8 pin processor ... determining the transition would require more sensor inputs and more LEDs or a display -- probably step up to a 14-pin or 28-pin PIC.

 

[Egzoset]

I'm surprized. Isn't it possible to get green, yellow and red out of a same twin back-to-back LED using only one control line? As for sensory inputs, i thought the transitions were already being recorded via the thermocouple and PWM paths... Actually, i was expecting an objection relative to a lack of space for the additional code this would require instead.

 

[Hippie Dickie]

Isn't it possible to get green, yellow and red out of a same twin back-to-back LED using only one control line?

i would have thought it takes two control lines to select one of 3 or 4 choices, but i haven't looked into using multi-color LEDs so i don't know -- even so, i don't have any unused pins to use for control. Maybe voltage level selects color? that would require a voltage divider and that's not available in this minimalist design.

Actually, i was expecting an objection relative to a lack of space for the additional code this would require instead.

probably plenty of code space -- 1/3 is unused, and all the hard stuff (start, stop, error detection, PWM control, PID calculations) is done.

And why do you want to signal this transition? How does the toker use this info?

 

[andrewburgess]

Well, do you think it would be possible to use a visual signal to warn about energy level transitions?

i've been thinking about a way to measure the vapor density with perhaps several (IR, UV, various colors) LEDs shining through the vapor stream onto a photodetector. maybe with an led bar graph for output.

while filling a bag one could detect declining vapor levels and automatically slow down the fan to keep the density on target.

plus it would be major eye candy for the hi tech vape crowd

does this help what you are trying to accomplish?

PS this probably needs its own thread unless we call it bud toaster 2011 design ideas...

 

[Egzoset]

Hi,

Here's the type of Multi-Colour LED i had in mind:



Only one line is required. Yes, it would require a Vcc/2 supply though...

:/

Anyway, if it's true that detection of the transitions can be used to warn when vaporisation begins/ends then i'd find this information interesting but in your context things may be very different. In any case, please rest assured i don't want to bother you with such matters if it's of no practical use.

 

[Hippie Dickie]

it's not a bother -- i enjoy hearing other people's ideas. But at this point, for the Perfect Cube, the feature set is fixed -- the only design changes going forward on this model are for production issues. It's very tempting to keep "optimizing", but it's time to kick this bToaster out of the lab and into the street.

 

[Egzoset]

Hummm...

In that case may i suggest that you check if the addition of a jumper (J1) as shown here would be possible?

I didn't familiarize myself with the data sheet but lets suppose GP0 can sink as much current as it can source...

Euh... Or vice-versa(?)...

I wonder, perhaps a multi-colour LED would be able to convey two different pieces of information with a minimum of modifications to your circuitry and software:

1) Colour = Duty-Cycle (Heating)
2) On/Off and/or Flashing State = Wait/Ready (Inhale!), whatever...

 

[Hippie Dickie]

i don't get what J1 is doing ... if J1 connects to MOSFET(D), then it is the same as the red LED already there??? And how do i throw J1 to one side or the other?

The green LED is already doing what you want ... when the green LED is ON, it means the temperature is at or above SETPOINT temperature, so take a toke ... it is delayed by 60 seconds after reaching SETPOINT, and after that 60 seconds, the bToaster is always ready to toke. Come to think about it, the green LED is really redundant to the digital thermometer when running, but it also provides a way to signal the toker before and after a vape session (7.5 minutes).

Edit ... also, i find it is very easy to know when the bud is cashed out ... the vapor goes from thick to thin. it is possible to continue toking vapor, but it is of a different quality and no longer worth the effort ... and meanwhile, the bud in the vial has turned to a very uniform darker brown.