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

[Hippie Dickie]

well, as usual, the way it works doesn't fit my a priori imaginings. yes, the small probe fits completely in a smaller diameter polyimide tube, but ... first try had the probe on top of the heater ribbon, and while that was interesting to see the instantaneous temperature changes of the coil - getting 20°F swings to maintain the glass tube temperature - it is not a user friendly display. so that was yesterday's annoyingly disappointing test result.

but i thought about it last night and remembered a different configuration that forces the probe directly against the glass surface of the heater, was finally able to get that method installed, and it is now showing a more averaged temperature ... all the actual temperature jitter is gone - smoothed out, as it were. so, this is fine for presenting the heater temperature - glass has got some lag for heat up and cool down.

and, it appears the display temp change with orientation of the cube is gone. more testing required to be sure.

i need to go back to the BBQ thermometer people and see if i can get some program changes: (1) no timed cutoff of display - not sure how many minutes it allows, but it is a lot less than the first BBQ thermometer i got, (2) power on as soon as the power is applied. i'm not too hopeful about either being possible ... i know if it were me, i wouldn't do it - not for less than 100K units - certainly not for 100 units or 1000 units. and even then it would be a pain in the ass. maybe there's a "pain in the ass" fee?

but at least i have a working cube with the new BBQ thermometer ... so i need to retrofit the other cube that is in service to work some more on assembly procedure ... and then get on to testing the new pcb.

 

[lazylathe]

well, as usual, the way it works doesn't fit my a priori imaginings. yes, the small probe fits completely in a smaller diameter polyimide tube, but ... first try had the probe on top of the heater ribbon, and while that was interesting to see the instantaneous temperature changes of the coil - getting 20°F swings to maintain the glass tube temperature - it is not a user friendly display. so that was yesterday's annoyingly disappointing test result.

but i thought about it last night and remembered a different configuration that forces the probe directly against the glass surface of the heater, was finally able to get that method installed, and it is now showing a more averaged temperature ... all the actual temperature jitter is gone - smoothed out, as it were. so, this is fine for presenting the heater temperature - glass has got some lag for heat up and cool down.

and, it appears the display temp change with orientation of the cube is gone. more testing required to be sure.

i need to go back to the BBQ thermometer people and see if i can get some program changes: (1) no timed cutoff of display - not sure how many minutes it allows, but it is a lot less than the first BBQ thermometer i got, (2) power on as soon as the power is applied. i'm not too hopeful about either being possible ... i know if it were me, i wouldn't do it - not for less than 100K units - certainly not for 100 units or 1000 units. and even then it would be a pain in the ass. maybe there's a "pain in the ass" fee?

but at least i have a working cube with the new BBQ thermometer ... so i need to retrofit the other cube that is in service to work some more on assembly procedure ... and then get on to testing the new pcb.

Another milestone in the Toaster's history!
Are we going to see a video of it in action again?
I still watch the others ones every now and then and am still amazed at what you have done!

 

[Mr. Whitewall]

 

[Hippie Dickie]

omg … so embarrassing to be so slow to execute, but fuck it, that's just the way i roll, so ...

IT'S ALIVE!!! first vapor from M15V1 was today at 4:20pm (or thereabouts). i ordered this latest pcb on 4/4/2018 and it has taken me until TODAY! to finally get the old WinXP HP TC1100 tablet running (after just sitting in the corner looking neglected and forlorn for several years!!!) - i thought for sure the battery would have ruptured and leaked, but it is still intact, even if it can't hold a charge anymore.

i wish i could have had steady time to devote to this project (i wish i may, i wish i might, have the wish i wish tonight, etc.)

but, a definite accomplishment … a way back … i can once again download a program into the 12F683. Now i have to update my tools to work off my Surface Pro 4.

and evidently i didn't screw up any of the component routing when i re-did the pcb layout. i moved everything to the other side of the pcb, opposite from the heater coil, except the three LEDs.

and the mounting of the BBQ thermometer probe … that used to be a solid pain in the ass to place against the glass oven tube … so embarrassing (again), but a solved problem - BY ACCIDENT - when the new BBQ thermometer vendor used a smaller RT probe (that stands for resistor thermometer, and yea, that is actually a thing - not a K-type or N-type, but an RT - go figure). it is fucking hard to do a project when i am so fucking ignorant of the technology … however, at age 71, i think i am just too old to learn this lesson. so, yes! the RT probe is sitting right next to the K-type probe that the PID uses. so, reliability, manufacturing consistency. very exciting!

okay, but things aren't perfect (yet) … the BT only ran for a minute or two and then shut off … i suspect the over-temperature condition was triggered … the ABV was much darker than normal ... this pcb has much thicker copper (3 oz) so the PID parameters are way too large … i think … easy to adjust and retry ... but i was seeing wild swings in the temperature and the heater was turning OFF for seconds at a time - which the control device NEVER does - so mysteries galore.

i haven't been into the PID assembly code for several years … this should be fun.

 

[Hippie Dickie]

Day 2 with the new model: had a chance to try a second session today … definitely hitting the high temperature cutoff of 500°F, even saw 500 on the BBQ thermometer for an instant … had to keep restarting the vape and after about 6 tries the oven was heat saturated enough that the temp swings were only about +/- 20°F … on the second restart, i tried to cool the oven heater by inhaling more forcefully, and only succeeded in having a slight combustion incident (damn, does that hit hard!) - that PID algorithm and high battery amps can NOT be overcome.

looks like i need to do some debugging … the PID writes the last 20 seconds of the calculated PWM dutycycle values to EEPROM, which i can read out with the PICkit connector … but i looked at the PIC on the pcb, and it isn't soldered square on the pads - which makes connecting the SOIC-8 connector an iffy thing - i need to remove and reinstall. tomorrow. first thing. well, after my exercise, of course.

 

[Hippie Dickie]

last two days have been frustrating … the RT probe is just not working as i thought it would - reading 50°F low when sitting right next to the K-type probe - it needs to be right against the glass - so, not a solved problem.

i did get the PIC fixed up - removed, pins adjusted, resoldered - and also discovered the resister feeding the green LED had a bad solder joint and fixed. i really need to have larger pads for the components - so at least another iteration on the pcb.

working with WinXP is a drag - must retire that tablet … so i ordered the latest PICkit iteration, and downloaded the MPLAB X IDE … even tried the online version, but that was throwing too many errors, and i ought to have it on my local PC anyway, so, time to put the childish things aside and man up to the latest technology, eh?

i did manage to adjust the PID parameters and get that downloaded into the PIC and it is behaving a bit better.

So i have a day before the PICkit gets here to get the assembler IDE up and running and accepting my code.

thrill a minute.

 

[KeroZen]

time to put the childish things aside and man up to the latest technology, eh?

You mean porting the Bud Toaster to ARM?!

(just nagging you, don't take it seriously)

 

[lazylathe]

@Hippie Dickie

I love updates in this thread! Lately there has been a flurry of activity!!
Every time I see an update , I get ready to sell off a kidney to afford a bud toaster.
Safe for another while...LOL!

And yes, updates are good. Sometimes expensive but at least give you access to the latest tech.
I have a Mac desktop stuck on Lion and that is as far as it goes...

 

[GetLeft]

so, time to put the childish things aside and man up to the latest technology, eh?

And yes, updates are good.

Looks like the Bud Toaster project has been a ploy to justify upgrading to a decent laptop.

 

[Hippie Dickie]

to justify upgrading to a decent laptop.

no no no, i have a Surface Pro 4, i have waited 50 years for this computer ... ever since Alan Kay showed his Dynabook prototype (cardboard mockup, actually, since the tech took 45 years to evolve to a Chinese assembly line). i may have to get the Surface Pro 6 because they are so cheap at $800.

So i now have a complete development system - pcb design, PIC programming, 3D design, website programming - in a clipboard size PC.

no, it's the old dog - new tricks thing ... i really don't mind getting into a new digital infrastructure, takes a mental shift that feels grueling to get into.

However! installed the MPLAB X IDE (free software) on the Pro last night , and lo and behold, there was an "Import - Legacy" option. omfg. it gobbled up the files and the build worked first time. Feels like i'm carrying Thor's hammer when i carry this pc.

So, did some more testing this morning ... found a workable probe placement under the heater ribbon and against the glass ... performance matched what the thermometer was saying ... but +/- 20°F swings ... should receive the new PICkit tomorrow to start tuning the PID parameters ... i guess it's time to put it in a cube.

 

[Hippie Dickie]

goodness gracious … that was so easy … i soldered a 5 conductor cable to the SOIC8 clip to connect to the PICkit4 and, damn, it worked first time … so i am totally on the new system (Thor's Hammer swings some strong kung fu).

so i just tried my first take on turning down the PID parameters … taking too long to settle but kind of hits a +/-5°F range … seems to be running hotter than the 410°F on the display - almost pyrolysis (nice buzz).
...
been adjusting PID params and testing all afternoon … looks like the MAX6675 gets a spurious low reading every so often, which throws off the algorithm for several seconds … need to do some filtering of the readings.

but, all in all, a very productive day on this project.

 

[Xclerk]

Just decided toncheck in the progress any finished version. I really lov3d this and still looking to play w one. Keep it up

 

[Hippie Dickie]

@Xclerk - thank-you for your interest in the Bud Toaster.

not much progress to report, but i do keep coming back to the punch list ... and perhaps made a little progress there, but i need to test the ideas i came up with.

i'm getting closer to committing to this full time, particularly since my real world business has simply imploded as a direct result of the great recession of '08 (and, well, millennials) ... so i've been looking around for successful business models: the vig - but if i knew that much about finance i wouldn't be working everyday for a living at 71; selling to an addicted market seems to be a winner - but i don't want to sell sugar; or, selling to connoisseurs in a reliable, strong, niche market - i.e. cannabis users, which is what i have great personal interest in, and feel like i understand somewhat after 50 years of cannabis use.

yesterday i did a quick search on alibaba for a ready-made custom leather case 6" x 3 1/2" x 2 1/8" - expecting to see usable stuff in the $2 - $5 range, but no, it was $22.50 - perverse as it may seem, i take that to be an indication that my case design is worth doing. i think i have resolved the aesthetic and manufacturing issues ... finally ... at least enough to move forward with another iteration. happy about that.

i need another iteration on the pcb for manufacturing issues - the major debate raging in my head about this: should i change from a circular pcb to a square pcb (this affects how the cherry cube is CNC'd) ... i don't know why things like this become so important ... and shuffle a few components around in the layout to de-clutter a bit.

and i keep second-guessing some of the foundational design concepts: use a cable or plug batteries directly into the cube; material choices; anything, really, that i can fret about ... but whenever i do this, i usually find the first design choice is the best, even if it needs gentle tweaking or manual practice to get right.

to motivate myself i keep coming back to the facts: that this design is complete, it has worked reliably for years, and it vapes flower really well - at least it satisfies me: truly all glass vapor path, 2 minutes from power on to first hit, thick vapor on first hit, and temperature control.

meanwhile, in other news, i need to completely redo my digital infrastructure (before 2020) and made a little progress there - moving from Windows 7 to Windows 10 and getting a monitor that can almost fully support the resolution of my Surface Pro, which makes a big difference for the PIC programming tool and PCB layout software.

tl;dr - not much progress ...

 

[lazylathe]

tl;dr - not much progress ...

Any forward movement is progress!!
Love the small or large updates and the info you provide.

Really like the circular circuit board as it would fit perfectly into the hole CNC'd for it.

Super professional!!
Still very interested in owning one of these cubes if the stars and moons align!

 

[madhockeydad]

What does the prototype look like now?

 

[Xclerk]

Awesome progress. I cant wait to see a for sale model or beta for testing. I would be totally down and ok to pay for beta

 

[Hippie Dickie]

What does the prototype look like now?

the cube looks the same, and really the size and shape of all the components is the same … i am totally satisfied with the overall design and functionality.

the main difference now is the final resolution of how to hold the corners of the case together … this has been bugging me for literally years … and i finished testing the latest concept today and i think it will work … kind of hard to explain without pictures (coming soon), but the seams of the leather case will be perfect, and something i will be willing to offer for sale.

this design will never be a pocketable vape - for out and about vaping - just a desktop device, but easy to move around - on the balcony overlooking the river (my favorite vape spot) or in my old man chair in the tv room or at my computer.

beta for testing

i actually feel like i can do this now … still kind of tedious to make the whole assembly of wood, glass, leather and electronics until i get some production tooling in place … but moving forward.

 

[Hippie Dickie]

been a busy several months ... but productive!

i worked out all the design issues with the leather case, except ... i inadvertently included a part made of unobtainium - seemingly simple, just a 6" U-channel formed from a 1/2" wide by 6" long piece of stainless steel strip, 0.012" thickness. i did get a quote to have it fabricated, and although the price is several dollars each (and i need 10,000 for this initial run), that wouldn't be so bad, except the quoted tolerance is (+/-) 1/32" - which is totally unacceptable for that price.

but, "i've got a plan" ... more, later.

so i did a pivot back to the heater control pcb and after starting totally from scratch, the "right" way, starting with the parts list, the schematic and then the pcb layout, i have model 15 version 3 about ready to ship off to the pcb fab house. The boards will be about $20 each in 20 quantity.

this is Bud Toaster model 15 version 3 ...

here is the schematic - same as it always was:

and here is the top of the new board:

and the bottom of the new board:

the big cutout on the bottom edge of the pcb is to flush mount (and solder) the Deans connector to the pcb.

hell, yeah! things are really rolling now.

 

[Ivan van Ogre]

Very Nice, its neat to see your work in progress. Keep on keepin on!

We are witnessing genius unfold.

i don't know if a separate section is needed, but whatever management decides is fine with me ... as i learned from my time in CA in the early 70s, i just go with the flow (and i ain't no dead fish, either).

We face reality in the one and dream our dreams in the other.
There is definitely room for both to grow.

 

[Hippie Dickie]

okay, that was a stressful two weeks … finally emerging from a hell of my own making … so, i get the new boards … and they just don't work. i populated 3 just in case i was doing it physically wrong … nope! as it turns out, i had a software bug that i introduced when trying to tidy up some code in the READCOILTEMP subroutine. i sure did spend a lot of time trying to debug the hardware that just wasn't buggy. i never seem to learn. maybe 71 1/2 is getting too old - but i was the same at 20.

it has been about 6 years since i dug into the code … sometimes i wonder who the Hippie Dickie was that wrote this code so long ago … at least he left a LOT of comments - that often helps. And the MP Lab X assembler is a really nice system for developing code - with a very useful DIFF utility to compare two versions of a file.

there are three main aspects to this project: the pulse-width-modulator software that provides the temperature control, the pcb layout tool for defining the bill of materials, creating the circuit schematic and positioning the components and running traces between them, and the case design. i have spent literally years on each aspect, with years between iterations of each aspect. Not a recommended way to bring a product to market, but that's the way i roll.

but persistence has paid off. the first session on a new version of the vape (this is officially "The Herbal Cube, version 1) is so satisfying.

also i found an Alibaba supplier for the LiFePO4 battery charger (2S) - with the DEANS male plug already installed.

so, onward to making the first new cube in 5 years (or so) - milling the new cube from Cherry this time.

 

[Hippie Dickie]

well, this is definitely the fun part of the development cycle … i.e. testing the new pcb and tuning the software. turns out the clock is running slow - a session that should take 7 min 30 sec is now running 15 seconds longer. that's a puzzle because i haven't changed the clock frequency in 5 years or so.

and this new board is definitely running hotter than the BBQ thermometer is showing - when it says it is running at 400°F, the ABV is coming out charred and the high is more intense - more like a joint. so i need to calibrate with a second thermometer (i got like a dozen from Alibaba last time i ordered).

the minor iteration just before this one had a design flaw - the K(-) lead was floating - not grounded as recommended - so it was getting occasional zero readings … at least i think that was happening. Anyway, tied pins 1 & 2 together on the MAX6675 and that problem is gone. next iteration on the pcb will fix it. god only knows why i missed that detail last year - it was okay the iteration before that.

it was while trying to program a fix for an occasional zero reading that caused me to mung the code and spend 2 anxious weeks chasing a non-existent hardware issue. not being particularly hardware savvy i wondered if the step up from 1 oz. copper to 3 oz. copper might have caused an issue if my circuit layout was marginal. nope! appears to be as easy as it should be.

my belief is that 3 oz. copper will be more reliable. i don't know it's true, but it just seems like it would be a good thing when the internal temperature is 400°F - although the 1 oz. version is running many times every day for years with no problem - still, not expensive and i am seeking ultimate reliability - if my vape breaks i deal with it immediately, top priority, everything else waits … my security teddy bear - i like that it has a heartbeat @100 msec.

so many more steps - i must get this version in a cube - i keep burning myself on the exposed heater coil.

first a little bit of desk work, then i want to sink into The Herbal Cube v1 for the long holiday weekend.

 

[Hippie Dickie]

my, what in "interesting" last several months … i have officially switched from my tried and true, ole reliable m14v3 to this current generation m15v3 … but not without some struggles along the way (seems to be my way). this one, and the next dozen or so, are cherry, rather than the mahogany i was developing with.

on the software front, i tuned down the PID gains and the temperature control is rock steady - kind of unbelievable (given the last 10 year history), but the temp doesn't budge for minutes at a time, whether puffing or not. now, for sure, part of this is an indicator effect, since (i have recently learned) the RT thermometer, which the BBQ thermometer is, has a much slower response rate (2 to 3 seconds) than the k-type (3 times per second) that the PID algorithm uses … the green LED ("toke me i'm yours" signal) is slowly blinking on and off, indicated the heater temperature is gently above and then gently below the SETPOINT.

and, instead of a dramatic overshoot on startup, it slowly approaches SETPOINT … i want to see if i can speed this up without the old overshoot.

also! the problem of the temperature changing with orientation of the cube has completely disappeared … a different method for holding the k-type thermocouple and the RT thermocouple - where they are spooning under a loop of the heater coil - has solved that issue.

i guess i am joining the temperature stepping gang - i like to bump the temp up by 5°F every couple of tokes. When it gets to 430°F the vapor is harsher, but more intense. However, at 450°F, the herb goes to pyrolysis - based on how dark the ABV turns - like i mentioned in a previous post, much more like toking a joint - sometimes that's just what i need.

so, after many weeks of thrashing, i have finally convinced myself i didn't screw-up this iteration of the pcb and it is working EXACTLY as intended. well, one minor issue: i had hoped to separate the RT thermocouple from the BBQ thermometer pcb - use some type of plug in connector - however, as i just recently learned, the RT probe's copper lead length affects the accuracy of the result - must be accounted for in the BBQ software algorithm. But, my goal to feed the 1.5v to the BBQ pcb with a 1/3 voltage divider from the BT pcb's 4.5v line works perfectly. So, no need to ever change the LR44 battery in the BBQ thermometer (nor buy nor install).

at this point, i am in a feasibility study to see if i can actually make The Herbal Cube for sale. hand populating the pcb is INCREDIBLY tedious and prone to poor solder connections because the components are so fucking small. i am this close to just ordering a bunch of factory-stuffed boards (all USA made, of course).

and i finally figured out what to do with the unused button press during the session:

Hold BTN1 > 2 seconds during the session

- this will reset the session SETPOINT to the starting SETPOINT … this will allow correcting a bunch of +5 or -4 adjustments during the session … trust me, if you temperature step, this will be useful.

that's it for now, must go populate another pcb.

 

[PoopMachine]

Really have enjoyed following your progress over the years.

 

[Hippie Dickie]

taking the plunge into the deep end …

so, after 10 and a half years of working out the details of this design - and relentless testing (this ain't no brand new, never tried, design) i have taken the next step in my manufacturing feasibility study - can the pcb design standup to automated assembly … i have definitely proven to myself that i can't hand assemble the pcb - only 22 components but debugging a soldering problem can take hours - and definitely a skill i don't want to have to get good at, for 1000 pcbs.

so, new learning curve … i only hope it ain't toooo expensive a lesson.

so, two weeks to mill some cherry cubes and cut some leather … and for sure a couple of tweaks to the software - it's been too long since i've looked at the code.

here is the new pcb for The Herbal Cube version 1 (it is Bud Toaster model 15 v4 in a long line of Bud Toasters):

Here are the first 2 of 3 cherry wood cubes using this layout (except without the 5-position header at the top of the pcb to allow easier connection to the PIC for loading the control software - finally getting rid of the finicky SOIC-8 clip!!!):

and unlike Henry Ford who said, "any color as long as it's black", these may be the leather case colors:

let the waiting begin! will it work??? details to follow as news develops …

 

[Hippie Dickie]

o … m … g … they're here!!!!

here is the top of the board:

and the bottom:

and the header fits the holes at the top perfectly:

so, i soldered a DEANS plug and a k-type thermocouple to the board, connected the header to the PICkit4 programmer, downloaded the program (download worked PERFECTLY the first try) and started the vape ... it appeared to work - won't know for sure until i install the heater coil and BBQ thermometer to display the temperature.

a game changer. (but non-vegan) (lol)