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

[Bubar]

The load problem is annoying. In a future version I'd like to do 12 or 120v and rely on a huge heating element to make things easier. Other possible ideas: thermal grease layer in the gaps between the heating element and the oven tube and insulation layer over the coil. I think I might machine a body out of teflon for the next housing. I have access to a lot of scrap teflon pieces.

Also a screw hub for the connections might be a good idea. 4:2pwr,1thermocouple(the ground line would just connect to main ground), 1 gate of transistor. I with i put those connectors on the pcb. Then I could produce a prepackaged element, transistor, thermocouple that could be swapped at will.

 

[Bubar]

Grrr. I too am now having some trouble tuning. I switched out to a smaller heating element to get mroe amperage, but now it swings wildly.

Edit: I think I foudn the problem. The coil was shorting near the thermocouple, so the crazy swigns were the short getting super hot, then the heat spreading through the coil and cooling down. It would drop 10-20C in a few seconds which is waaay more than usual.

 

[Hippie Dickie]

What are your coefficients?

i have one battery pair that is giving me 1F accuracy ... it just hangs on a temperature for about 30 seconds, flicks up 1/2 F and holds a bit, goes back to SETPOINT, holds, and flicks down 1/2 F, holds, and comes back up.

The other pair (i switch between two sets) is giving me a 10F drift throughout the whole session with about a 20 second period ... wtf? i think i have noticed an intermittent when bumping this set, so maybe i've got a higher resistance connection on the plugs or battery tabs.

i can see how dynamic tuning of the coefficients might work. For example, when i take a hit, the temp drifts up, perhaps overcompensating, but it seems the P gain is outweighing the I gain, until it gets up +2F and then I takes over and regains SETPOINT. This has a lot of interesting possibilities.

For whatever reason, at the end of the last several sessions, after maintaining SETPOINT for 6 minutes, the temp jumps up 10F for the last minute? And of course, this has the effect of "squeezing" a little more THC from the vial. Damn considerate, i must say (although the last hit makes me cough -- needs some training), but wtf? Must be a software bug???

 

[Prankster]

Not being the brightest bulb.. But could the altered density of product in vial lead to awkward temp maintenance?

 

[Hippie Dickie]

Exactly!!! i must have read your mind as i was looking into the vial to stir and re-vape.

The vial had been filled about 3/4 full and a layer of kief sprinkled on top and the vaped material had formed into a crystaline-like structure that i broke apart with a wooden implement (toothpick).

and vaped. temp did not spike up 10F at the end this time.

7 minutes is too short a vape session for a vial more than a third full. stirred, not grindered.

 

[Bubar]

Would you mind sending the updated version of your code?

 

[Hippie Dickie]

Done. There are a couple of untested changes:

1) increased the delay before the green LED turns on to 60 seconds.

2) added a flag variable to keep track of how the Vape session ended.

i want to allow a BUTTON2 press to extend the session for another 2minutes in case the vial is stuffed too full and the vapor is still flowing ... not implemented yet.

Might be okay.

 

[Bubar]

Aha I was wondering why the green LED wasn't turning on. I didn't wait the 60s after reaching set point....EDIT: Seems that even after a minute the green LED isn't turning on. Perhaps during the minute it is swinging too much so it keeps reseting the ready timer. Could be due to differences in running volts or element resistance. This particular one runs at 5v and .66ohms -> approx 8A.

I have batteries now, but I went with the Blue Lipo which is 7.7 rated, so that might not work the greatest either. The batts are going on my design, but I gotta make another prototype because I hooked up the Vreg the wrong way and I think I fried the pic. I tried to program it after so it's also possible that is the problem. Do you think 8 or 12v would fry a pic? My guess is probably, but I haven't tried to kill a pic (these 18f's are $5 a piece!) to see if it does.

 

[Hippie Dickie]

re: PIC ... max voltage is listed at 6.5vdc, so, you probably fried it.

re:green LED ... it should turn on 60 sec after temperature exceeds 375F. Probably a bug in the count-down code. i usually get it wrong the first try. It doesn't reset it.

i'll load this version of the code and debug it tomorrow.

Also, do you get the 3 green flashes after you power up, but before you hold BTN3 to start it running? ... that would mean the PIC is running okay.

The 12f683 are only $1.10 each in 10 quantity.

 

[Hippie Dickie]

Yes, of course, there was a bug in the countdown timer code. i just love to make simple things more complicated. Of all the predilections i could have, why that one? ....

Anyway, new code package has been sent.

Also increased the run time to 7 minutes and 30 seconds to allow the 30 seconds to get to vape temp, and increased the "Toke It" delay to 60 seconds after reaching vape temp.

So, 1 minute and 30 seconds after power on, the green LED comes on to signal time to start toking.

 

[tinkeringtoker]

hello all,
it's inspiring to read this thread. after looking at a friends digital vap I thought to myself "I could make this..".. shortly after I did some research, bought some parts, and have been doing a little tinkering. Eventually I stumbled upon this thread and was thrilled that I wasn't the only one who had that thought. Like you I learned about PID and have been working on implementing it satisfactorily with little success, but instead of a thermocouple I'm using a thermistor. It's a glass bead encapsulated thermistor rated to 250C, similar to the one used in the RepRap project to monitor the temperature of molten plastic (http://reprap.org/bin/view/Main/Thermistor). Also, unlike the two leds or serial LCD module, I opted for 4 7 segment displays and two LED indicators. The setpoint & current temperature are displayed on the 7 segments and the leds indicate status. In addition, I am using an Atmega328p from AVR as opposed to a PIC; higher cost per chip but I like them better besides, not having to pay for the expensive max thermocouple chip makes up for it. I got real thick 12awg nichrome, but I might drop down to 14 or 16 because the resistance demands such a low voltage. I'm using a bench power supply now but I can't exactly buy one for every intended vap... I kinda worked in reverse as compared to you; I have yet to taste any vapor but I have all the uC software written and wrote software for my mac to graph & log temperature & PID coefficients, which are controlled by 3 knobs. world's first USB vap I can't wait till I have fairly stable hardware and can work on the software more, but for now I gotta focus on finalizing the first prototype's hardware and designing PCBs. I also need to find a good source for glassware, do you know of any? I have small quantities that have worked for prototyping thus far but they keep having unfortunate run ins with gravity... Thanks for the sweet thread!

 

[Hippie Dickie]

thanks! Sounds like you're making good progress. Interesting about the thermistor -- i didn't think they could read that high. Thanks for that link -- lots of new stuff to learn.

The MAX6675 is only $3.50 in 1000s. And a K-Type thermocouple can be fabricated for $0.17 with a $1500 welding machine.

If you use a MOSFET to control the current you don't need to worry about resistance of the coil, but 12 awg must be difficult to work with, being so thick.

re: glassware -- these are not single quantity purchase websites, but:


Lab Depot - Glass Test Tubes - Culture Tubes, Disposable, Borosilicate Glass (47729-566)
http://www.labdepotinc.com/product_details~id~33~pid~12627.aspx

Kimble-Kontes - Vials
http://www.kimble-kontes.com/html/pg-60931.html

Fisher Scientific
http://www.fishersci.com


Single quantity can cost a couple of $ each. Do a google on "borosilicate" or "kimax" or N-51 -- stay away from "pyrex" as that brandname has been sold to china and the formula is now a soda-lime type glass.

 

[tinkeringtoker]

Hmm what do you mean by I don't need to worry about the resistance? I have some STP20NF06's, n-channel MOSFETs, 60v, 20a continuous, 80a pulse. One of my coils has a resistance of 1/10th of an ohm meaning it wants 50a at 5v; wouldn't the extremely high (ideally 100%) duty cycle during early stages of warm up cause the tranny's to fry? I'm using the bench supply to limit current but maybe I'll try stacking 3 of the fets for a theoretical max continuous of 60a... not quite sure.

and yes, the 12ga is a bitch to bend, but the thought of hitting the vap in 30s is too tempting XD I just found a similarly sized cylinder and did my best, it's far from perfect...

I guess I didn't think about 1000s quantity haha that's a little outside of my intended scope; and the double-digit small quantity price really threw me off (though I suppose I should have busted out the free samples like you guys haha). Thermocouples seemed like they would be much cleaner to use but I like to do as much on-chip as possible... I'm driving the whole display without any external chips so it seemed like I might as well use a simple voltage divider circuit rather than a whole new i2c chip if I could. I mean they flew rocket ships with simpler uC's, I feel required to take advantage of what's available plus I can read it faster than 4hz and considering the thing is barely smaller than a pinhead it reacts fairly quickly.

For a pretty good break down on math behind a nice voltage<=>temperature curve, check this out http://hydraraptor.blogspot.com/2007/10/measuring-temperature-easy-way.html

thanks for the glassware sites I'll check em out, and yeah I hate how "pyrex" no longer means "borosillicate" but merely "brand name glass". what a load of horse shit...

best of luck

 

[Bubar]

You already don't have to pay for the maxx6675. You can sample 5 of them for free from Maxim's website.
Part of the reason for the thermocouple is the absoluteness of thermocouples and the lower error. I'm not sure what the resistance and beta tolerances on that, but it is probably less tight than for using the maxx6675. Of course, the maxx6675 does have an abysmal sampling rate.

 

[Hippie Dickie]

tt - i find that the bud needs to sit in the oven for 60 seconds before i can take a decent size hit. Maybe i could speed that up by setting the temp much higher at first, but i am very pleased with the current 90 sec startup.

And, yeah, the MOSFET has to be selected to match the current flow - so that puts a lower limit on the coil resistance. what voltage are you using to drive the gate?

 

[Bubar]

tt- be very careful if you're going to use that many amps. Bad choices of insufficiently rated connectors or mosfets could result in very bad stuff. If you look back at the pics I posted of melted glass due to a short on the mosfet, which was in a system only using 8A! Imagine 50A. I don't think I will even try fucking around with anything over 20ish amps, since 8-12A will heat ~30ish seconds.

Also, welcome to the party. It's good to have another contributor. What compiler/langauge do you use?

Once I get some free time I'm gonna start on auto-tuning PID algorithms. I started looking at the stuff a year ago when I didn't know enough, but I may be competent enough now. That would solve all of our problems.

 

[Hippie Dickie]

i had a thought last night about the gains ...

first a sidebar: the PIC is really cute little processor, but it is kind of simplistic in its math abilities. i find the best way handle numbers is in "sign-magnitude" format. That is, when i subtract, if the answer is negative, i immediately do a 2's complement to get the absolute value and set a flag bit.

So, considering the gains for the PID, this gives me two temperature ranges:

vBelow, where the coil temperature is below SETPOINT
and
vAbove, where the coil temperature is above SETPOINT

So the thought i had last night, is the gains should be different in each region, because the coil heats up much faster than it cools off, by about a factor of 5, but there is instantaneous delay in response, so, say, a factor of 3 is enough.

i just loaded and tested some new gains ... (bump!) ... and this works much better. Now, of course, this is just the first test, and i'm not anal if i'm not thorough, and gf says i really am "detail oriented" ...

So here are today's gains:

vAbove: (-30 * P), (-.25 * I), (either -20(+) or +10(-) * D)
vBelow: (+10 * P), (+.125 * I), (either -10(+) or +5(-) * D)

D, the derivative term, is also a signed value, with a positive slope(+) or a negative(-) slope.

Also there are limits on changes to the P, I and D values, which seems to dampen out jitter in the MAX6675 readings (if there is any).

These gains keep a tight limit on the upward error. That is, the green LED bounces off every now and then as the temp dips below SETPOINT before being corrected. i didn't see the annoying drift up on the high side.

Also, hitting BUTTON1 to raise the temp by 10F was amazingly accurate and fast.

 

[Hippie Dickie]

i took some time today to push ahead with production unit #2, the second unit with the factory-made printed circuit board. i still need to install three LEDs, the k-type thermocouple, and the outboard thermometer probe.

And the pink skin.

Here's the back end view. i decided to try a plug instead of attaching the cable:

Here's the front side showing the copper strips for mounting the coil, and the three downward facing buttons:

 

[Bubar]

Looks sweet. Did you post your cutting method for the tubes. I just ordered a glass tube cutter from the net, so I should have some nice looking oven tubes. I got tired of scoring with a file and then snapping them. Sometimes it work ok, mostly was jagged though.

 

[Hippie Dickie]

i haven't used a photo before ... here is my glass cutting jig:

my oven tube cutting technique:

1) BIC the tube at the break line area for about 30 seconds to make it too hot to hold

2) rotate the tube against the glass cutter wheel to score the break line - i go a dozen or so times around so the score is very visible.

i took the cutting wheel part off a standard glass cutter (for cutting a sheet of glass into window panes) and mounted it 1.5" from the block where the bottom of the tube sits,

3) BIC the score for about 20 seconds

4) take a previously dampened cotton rag and wipe the score

5) wrap the tube with the rag and snap at the score.

This usually makes a clean break.

Then i file the edges of the cut so they are dull. And BIC the end to anneal the micro-fissures.

This is my least favorite fabrication step. i checked a glassware supplier and i need to buy about 5 tons of glass to get it custom made. i know glass is heavy, but that is a bit more than i need.

 

[Progress]

It could be worthwhile to see if a glassblower would be willing to make them for you, and send them a sample to use as a template (as it could be much cheaper, easier, and quicker for them to make them due to tools and know-how).

Once you are ready for mass production (assuming that you are interested in going that route), you may even find one by posting what you are looking for here at FC (or some other, more-related, site)

 

[Hippie Dickie]

Thanks for the idea ... but i don't know if it can be as cheap as the 16mm x 150mm glassware i'm using now, which is $0.25 each in 100s. i can cut a hundred in a couple of hours.

If i need to cut more than 100, i'll get a CO2 laser to do the scoring. And the laser will also do most of the wood milling.

 

[Purple-Days]

Hey there Hippie, thought I would pass on a little tip on glass cutting.

Get some kerosene and apply it to the cutting wheel (or the workpiece in this case since the wheel is below the work.

When cutting glass you get a much better score with the kero. This was taught to me by an old glass man and I have used it ever since.

 

[Hippie Dickie]

Thanks, Tom, i'll give that a try ... i should be able to "borrow" a cup of kerosene from a neighbor.

On another note, the second production Bud Toaster had its first test run tonight and ran the 7 1/2 minute session just fine. No vapor yet, just a test run. Maybe tomorrow i can put the skin on it and it will be done.

So far (one test run), i like the plug on the Bud Toaster better than soldering the cable to the printed circuit board. The copper bus bar connecting the plug to the pcb makes the unit solid as a rock. It was easy to extend the 12awg copper wire downward into the wood to function as strain-relief for the plug, so plugging/unplugging will never move the plug even a little.

It amazes me how design decisions for prior implementations continue to propogate through later designs, and cause little fuckups. For example, one of the copper wire bus bars is just a smidgeon too close to the PIC chip, so the programming clip almost doesn't grab the SOIC chip. Fortunately the clip has a lot of unneeded plastic that could be melted out of the way.

So, now i need to move the PIC up about 1/8" on the next iteration of the pcb. i've got a list of 8 changes for the next run of pcbs .... so far.

And the other copper bus wire blocks where the k-type thermocouple connects, so i need to move the solder pads over 1/4".

So, two done, 998 more to do.

 

[Purple-Days]

Your welcome, funny you mention a 'cup of kero'. I used to see the clear liquid in a cup, the glass men, dipped the cutting wheels into. I knew it wasn't water and asked.