Building a Digital Vaporizer (Brain Dump)

[skippymcware]

Hippie,

4 amps but how many volts? At the moment, I am constrained to industrial heaters which use a bunch of power and have high voltages. For this reason and several others, I am starting to think nichrome wire is a very smart move. Anyway, I have been looking for ac/dc converters that can put out around 150 watts. Like so:

http://www.trcelectronics.com/IPD/nxt-175-1006.shtml

This unit may be a little pricey compared to what I have been seeing, but I have never found one under $100. Really, I think we are just coming up against fundamental design differences in our units. I don't mind having AC power because it is a whip vape. 120 Vac also lets me use smaller currents and, thus, smaller everything else and less heat generated by the electronics. Also, without a controller that can correct the voltage sent to the heater as often as a pulse width modulation based controller or a SCR based controller, I think I have little chance of being successful with a low thermal mass and low wattage heater. If I tried to use one with such a slow controller, I would have to either submit to a long heatup time with a large thermal mass or have very little thermal mass in the system. Well, like you said, I am not into long heatup times. Avoiding that is a big goal of mine. On the other hand, if I used a low wattage heater with a low thermal mass system and a slow controller, the thing would be a piece of crap because you would start vaping with it and it would suck all the heat out of the system before the controller even knew what was going on. Then, the heater wouldn't have the power to correct it. So, even though there seems to be an infinite number of solutions, there are only a few reasonable ones. Right now, I am just trying to learn a little more before I jump into the more complicated project.

I have no idea if that makes sense. With any luck, it does. I know it made sense in my head.

-skippy

 

[Hippie Dickie]

i understand what you're saying.

i started with a nichrome ribbon from a car cigarette lighter. that was hard to source so i switched to nichrome wire. i found a chart that claimed that 3.2 amps would heat the 16ga to 400F. So i determined (via Ohms law) that the resistance of 72" of wire with 5vdc would draw 3.2 amps.

so a 5vdc, 4 amp power supply was selected. this required 15 minutes to reach vape temp. of course, there are reality interactions which changed the theory. The 3.2 amp was based on a straight wire in a vacuum (of course!). i found a coiled wire would heat more - i attributed it to "thermal feedback between the loops of wire". so i started chopping the wire until i ended up with about 48" of wire.

this took 15 minutes to reach vape temp. the first hit would drop the temp 50F. then i had to wait for the vape temp to recover. eventually, over about 10 minutes the temp would continue to drift up. this despite the fact the resistance is supposed to increase with temp and should have limited the temp.

so there was some kind of "heat saturation" effect going on.

then i started playing with several pwms that i picked up online.

then i found the MAX6675 and my whole world changed.

by the way, the MAX66675 can only sample the thermocouple and output the temp to the PIC processor 3 times per second. that is more than adequate, because with the instantaneous current output of the batteries, the temp can be maintained. Quite nicely.

 

[skippymcware]

Hippie,

[EDIT: This is wrong. A magnetic field is only generated when you have a change in current. Induction also only works on ferrous metals.]
I would attribute the increase in efficiency in the coil to electro magnetic effects. I think this is also called inductance in a coil? Putting current through a coil generates a magnetic field which then generates more current. It will also heat anything in the coil. There are tons of videos on youtube, but I picked this one because his work bench made me laugh out loud:

http://www.youtube.com/watch?v=ctPtOONtGGg&feature=related

Here is another one:
http://www.youtube.com/watch?v=LT5VvBQfOoo&feature=related

Anyway, that is very interesting that your system only samples 3 times per second. Gives me quite a bit of hope. I found this unit last night at 2 in the morning:

http://www.nuwaveproducts.com/ssrman-1p.htm

The fact that it clips on the top of a normal SSR is totally awesome. It looks like a very small company, so I will proceed cautiously. The unit can accept an analog input or a PWM input from a controller. I still have a lot to learn about this whole thing, but I think I am starting to close in on a really cool solution.

-skippy

 

[Hippie Dickie]

hmmmm ... i thought electro magnetic effects like induction are due to alternating current.

i don't understand that power supply you linked: 165 watts? 24v? why so much power?

a computer power supply is cheaper.

a nichrome coil heater -- which also is basically what a ceramic heater is -- just needs the 3.2 amps to maintain 400F, and at 5 vdc, that's just 15 watts. [Edit: that's for the 22ga nichrome-60 wire - i may have said it wrong in a past post, sorry.]

The choice of voltage is completely arbitrary -- as long as the 3.2 amps are there. i picked 5vdc cause that was easy to get.

i've just switched to 3.3vdc cause it is easier to get LEDs than at 5vdc. Radio Shack has some nifty blue LEDs, but they're pricey compared to, say, Newark electronics distributor for more common 3.3v LEDs.

Edit: this 3.3v is just for the electronics. the coil is seeing the full 2 x 3.3vdc from the two battery pack. my coil is 24" of 16ga nichrome-80 (medical grade, hah!) which is 0.5 ohms.

Edit: i switched to 16ga several versions ago in a failed attempt to get more thermal mass into the heater. but the larger gauge provides more physical structure for the coil. and with current control, the length of the coil is completely arbitrary, too. except i want to draw a high current at full current, so that keeps the length down.

i've been thinking about two coils around the borosilicate glass oven tube -- for "zone" heating. since i'm now pulsing the current, there may be some beneficial mutual inductances??? i'll have to figure this out over a vial of bud and with a deep dive into wikipedia. if this could work, i could put many zones along the oven tube with a different PIC with more PWMs onboard. all for no change in parts cost.

 

[niall]

Fascinating stuff guys I've been thinking about this and keeping an eye on Hippie Dickie's posts over the years, been planning to try some of this myself but haven't had time for it until now.

Has anyone tried using the iolite as a base unit providing heat, and replacing the bimetallic with a thermocouple, maybe with a modified brass crucible to give it improved thermal mass and reduce the frequency of gas firing?

 

[skippymcware]

Hippie,

Ya, after I posed that I thought about it a little more... totally wrong on my part. First of all, it can only heat htings that are ferrous. Secondly, you have to have a change in current in order to generate a magnetic field. So, you are exactly right when you say it has to be AC power. Ooops. I do know that I saw a chart on sizing nichrome wire for certain temperatures, and it recommended half of the straight wire lenths. You must be right about the proximity of the wires to each other making the system heat faster.

The only reason I am seeking so much power is just because heater manufacturers and suppliers keep telling me I need a minimum of 175 watts. Now, clearly, you have proved this wrong. But, as I mentioned earlier, I can't achieve what you have done without a much better control system. At the moment, I am going to use the auber controller and a large heater to try and get a large thermal mass system to work. As I am starting to learn, the industrial heater design world really isn't designed for addressing a system with a very low thermal mass that requires a fast response control method. The only fast response system they ever deal with are air heating systems, but, even then, it is always constant flow which dumbs the whole thing down to slow response again. They just don't deal with small fast response systems very often. As such, I think I am asking them to help me design something that they don't really know how to design.

All of that being said, I am starting to see your point about the necessity of AC power once you get down to a low thermal mass system. Once I don't need a bunch of current from the wall, I might as well switch to DC power... I think. I am going to have to think about this more. There must be some other benefits to AC power.

Why did you decide to pulse the power? Is it just more stable? Zone heating would be pretty cool, but do you think it would actually improve your level of control?

As always, thanks for the input


niall,

I haven't explored butane heaters much. I have very little experience fooling with that type of thing, so I decided to go with electric heaters very early in the design process. It sounds like you might be on to something there. Only way to know for sure is to tear one apart and start changing stuff.


I think my next step after doing some testing on this simple system will be to buy a watlow controller that has a 10hz sampling frequency and an 0-20ma analog output and couple it with a phase angle fired power switching control for a SSR from that nuwave company. The guy that owns nuwave told me that he used to work for Athena, and they have the exact same product.

http://www.athenacontrols.com/pages/products/powercontrollers.tpl

they call it a pwm series converter. All of this is assuming I still want to go with AC power.

-skippy

 

[Hippie Dickie]

Why did you decide to pulse the power?

That's fundamentally what the pwm actually does. it only passes the 12amp current during the duty cycle portion of the period. i have the timer set for 1kHz period to get maximum resolution in the pwm output. and fortunately there is no audible tone at the selected period. one of my previous vapes had a pwm that bothered the cats.

i think zones could be very useful. i forgot each zone would need its own thermocouple and MAX6675 for temp feedback. that adds about $5 per zone - mostly for the MAX6675 chip.

a 5-pack of type-k thermocouples is $25. but if i get a special welder ($1.5k) to create the thermocouple joint from the two wire types, the cost drops to $0.15 each.

 

[skippymcware]

Yes, pwm's pulse the power, but I thought you meant pulses in the heat output. At 1 kHz I wouldn't think the heat output is pulsed. At that frequency, in my understanding, you really are generating an effective voltage equal to duty cycle x input voltage.

 

[Hippie Dickie]

it's really an effective current - a current chopper, because the voltage is fixed and the current draw is fixed (by the coil resistance), but the current is interrupted. so by Ohm's law, there is an effective voltage, but i think about it from the current side.

Yes in practical terms you are right that the heat output is probably not going to be pulsed -- the inertial of the nichrome will dampen out the pulses.

And there is a snubber diode to try to balance the current switching. i only have it in there because the theory says it is useful, but is it needed? i don't know.

 

[skippymcware]

Hippie,

I am thinking about using some insulated nichrome wire. My one concern is that my wire temp will exceed the insulation temperature rating while it is heating up the system. I noticed in one of your pictures that you have your wire wrapped very tightly. I am assuming this is insulated nichrome. What type are you using (other than 80), and what max wire temp do you see?

 

[Hippie Dickie]

oh ... that's a misleading image -- that's right after forming the coil. when the coil is actually mounted in the vape, the coils are stretched out and do not touch. i used teflon tubing on earlier versions/models, and it never melted, but it was a fabrication hassle. so i've kept the nichrome bare since then. from my experience (with this coil design: 24" in a 16mm (ID) coil) insulation is not needed.

the max temp seen in the wire is around 450F. but that is rare, and the vape algorithm adjusts the "trip point" so the peak of the full-power-on phase should be right at the setpoint (about 420F, i kid you not). the coil temp takes about 3 seconds to hit a peak (overshoot) after the pwm kills the power during the initial max power startup phase.

 

[vap999]

the max temp seen in the wire is around 450F. but that is rare, and the vape algorithm adjusts the "trip point" so the peak of the full-power-on phase should be right at the setpoint (about 420F, i kid you not). the coil temp takes about 3 seconds to hit a peak (overshoot) after the pwm kills the power during the initial max power startup phase.

With the temperature of the resistance element at most reaching 420?F, and wanting heater air output temperature to be about no more than 50-60?F less than this (which considering inevitable heat loss from convection and radiation is not that much less than the heating element), your heat sink/exchanger must be rather efficient. What basic design are you using to heat the air? [Or did you cover this earlier?]

 

[Hippie Dickie]

vap999 -- check out my thread "the bud bomb" for more info about my design and pictures.

also, i'm measuring the temp of the coil on the outside of the glass tube that holds the glass vial that holds the bud. temp at the bud surface is a bit less than 420F, i'm guessing.

 

[vap999]

Skippy and Hippie:

With all of this discussion of voltage, power levels and heating elements, I will add some information about the prototype vaporizer I am slowly building. I am not yet ready to fully discuss or answer questions about this vaporizer (it's too early in R&D; I've only just started to use my first prototype), other than about heater component-related experiments and some general observations. Someday my vaporizer may deserve its own thread, but for now the FC members most interested in what I have to communicate are likely following this thread.

THIS VAPORIZER IS VERY MUCH UNLIKE YOUR DESIGNS! I am pursuing simplicity in design using the best possible materials. Heating time is irrelevant to me. Like with the PD and other Eterra class vaporizers, who cares about heating time if it is consistently hot when you go to use it? Recall, I am experimenting with a simply-designed Eterra/PD class vaporizer with a massive heat sink/exchanger (currently ~230 grams of pure copper); using a custom industrial square cartridge rated at 25 watts, 14 volts; using the best insulation available (silica aerogel and microporous insulation); and so far (for test bed use only) using only quartz glass (liquid colloidal silica ceramic binder; hardens to pure silica/quartz) to hold everything together (i.e., no screws, solder, wires, pins, etc.). There are no organic materials, such as wood, in this vaporizer.

The cartridge heater is working better than expected. Tightly inserted within the central heat sink/exchanger (surrounded by ~140 grams of copper) with this wrapped in insulation, and measuring the temperature at the hottest point (thermocouple separated by 1/8-inch copper from the outer surface of the heater), I have measured consistent low 800s?F. This is when supplied the rated 14 volts (at 1.85 amps, which by Ohm's Law says its operating at 25.9 watts, 7.6 ohms, with measured resistance very close to this theoretical value). This indicates more than enough raw heating power, allowing lower voltage for actual use. Powered by 12.0 volts DC, 1.60 amps (by Ohm's Law, 19.2 watts, 7.5 ohms) and fully installed (a working vaporizer), it performs very well -- running at consistent temperature; delivering consistent vaporizer-temperature air (at the hot end of the spectrum, resulting in darker brown ABV material than to most peoples taste; indicating use at even lower power would be OK); easy draw with no real resistance, with the smallest constrictions being the 1/16" x 1" air flow areas around the surface of the main heat sink and a 3/8" diameter output pipe (compare to the few small holes in the Bud Toaster; allowing large quick hits or whatever kind of draw one wants, not just the slow draws seemingly the norm with the PD and many other convection vaporizers); no noticeable decrease in performance between hits when unplugged and up to 8 deep hits taken in rapid succession (heat sink is doing its job); hits are as potent as one would expect (much like most every 100+ gram heat sink Eterra class vaporizer I've owned, repaired or built for myself in the past 14 years; although I've never tried a PD); no noticeable off tastes or odors (as expected, copper is no problem in this respect); and even when run at full 14 volts, 1.85 amps, mounted/insulated in a standard 14 oz., 4.5" tall, 2-7/8" wide food can, weighing about 14 ounces total, the outside only gets slightly warm. [So far, everything is working better than expected].

Running at 12 volts, inside the fully-assembled vape (measured at the same hottest area on the outside of the main heat sink), the temperature consistently measures in the low 500?F (but this is the hottest part of the rather large heat sink/exhanger). Whether supplied with 14, 12 or 11 volts, the vaporizer consistently takes 30 +/- 1 or 2 minutes to heat up to steady-state temperature (apparent no worse than the PD), averaging +14?F/minute no matter what voltage is supplied. And so far, all of this is with worse case construction decreasing efficiency, including long wimpy power supply leads; no insulation at all on the top of the fully-exposed heat sink/exchanger (much heat lost by convection and radiation); no screens or bends in air input and output passages (hot air likely being lost to chimney effect); and straight-through air flow, with no measures yet taken to increase internal heater air turbulence (which should further increase efficiency), such as perhaps inserting twisted copper strip helices in the main heating air passages. Conceivably, this vaporizer could eventually be operated to most users' tastes at 10 volts or less.

Other than heating time, this vape may well be meeting and beating most or all of your performance goals, and with no complex electronics and at relatively low power! I have instant, on-demand, always-there, consistent temperature output hot air controlled by simple adjustment of voltage/power supply. I anticipate users either using a standard regulated 14 volt, 2 amp fully-variable power supply (such as used worldwide to power tattoo machines) or costing much less, a standard 12 volt wall power adapter with a variable DC power control, such as the 12 volt, 8 amp DC dimmer switch commonly on Ebay for less than $15; e.g. see http://cgi.ebay.com/LED-Dimmer-Adju...66:2|39:1|72:1205|240:1318|301:1|293:1|294:50. I'm awaiting delivery of one of these. And all of this costs just a fraction of your high-power, complex electronics-loaded designs, and with very few parts by comparison. I might eventually add a thermocouple with digital display, but would never bother with any near real-time voltage/power controls (irrelevant with a large heat sink).

As I would expect, others are apparently pursuing similar vaporizer development paths. For example, Bob, the original Eterra (inspired the PD) developer and patent holder (also developed the Flash Evaporator, Pnuema, Omega and other classic convection vaporizers), appears to be developing a new Pneuma model similarly using a large heat sink and 12 volt operation (see http://www.flashevap.com/pneuma.htm).

 

[Hippie Dickie]

I've only just started to use my first prototype

Congratulations on getting your vape running! i find this phase of development to be very rewarding.

Yes, very different design goals from mine.

I found wire gauge can be a limiting factor for electrical performance. The best wire (incredibly low resistance) is the multi-strand (1660 strands!!!) 12ga Deans Wet Noodle -- very flexible but kind of big and bulky. Available from Radio Control hobby shops. Those R/C dudes know about large current flow.

 

[vap999]

I found wire gauge can be a limiting factor for electrical performance. The best wire (incredibly low resistance) is the multi-strand (1660 strands!!!) 12ga Deans Wet Noodle -- very flexible but kind of big and bulky. Available from Radio Control hobby shops. Those R/C dudes know about large current flow.

For any wiring, I use high-purity oxygen-free copper -- standard good grade multi-stranded copper speaker wire, also designed for high current flow. I use naked wired, stripped of plastic, for any use within the vaporizer. At low power (<12 volt DC, <2 amps), I don't see the need for such extreme high-performance wire, but it's good to know what's available.

How do or will you attach wires and components? Silver solder?

 

[Hippie Dickie]

Yes, silver solder - SMT paste. And all the components are isolated from the air path.

 

[skippymcware]

vap,

Congrats on the prototype. It sounds like it is going to work great. Once you get a little closer I would love to see pictures of the heater. How long is your air path? If you have measured, what kind of temperature drops do you get at the bud when you take a big hit?

As you say, this has and could be done with a simpler approach than what some of us are working towards, but I am in this for the challenge and fun of a good design project.

Thanks for sharing!
-Jake

 

[vap999]

vap,
How long is your air path?
-Jake

I have a rectangular central heat sink boxed in with air flowing around it on both sides (a very simple design). I am using a 3/8" x 2" square cartridge heater (3 sq. inches exposed surface) inserted in a milled slot in a 2" x 1" x 1/2" pure copper bar, with this covered by a 1/8" inch copper plate, forming a central heat sink of around 140 grams (with 2 sides of 2 sq. inches area, 4 sq. inches, in the air path). This is covered on both sides, with 1/16" separation (air flow passage), by a 1/8" copper plate on each side, with these acting as secondary heat sinks (not as hot as the central one, but surely holding and contributing much heat), with these adding another 4 square inches of air contact. Air input and output is at the top.

Currently, there are 8 square inches of pure copper heating area in the air path. This is working very well as it is. From here there are many ways to easily boost hot surface area and air turbulence (and associated efficiency), such as more air passages and filling the passages with shaped copper pieces, tubes, ball bearings, etc. (also adding heat sink mass).

I am just starting to think about what I could do with a more powerful AC model, particularly with a much larger heat sink.

Coming back to a recurring theme in this thread, voltage/power control, if I want a variable voltage control for say a 100 watt cartridge heater, can I simply use a standard solid-state (digital?; voltage on-time-slicing?) light dimmer switch, such as standard radial dial models from Lutron. If these can commonly handle many more 100s of watts of white-hot heaters (light bulbs), shouldn't it be no problem to use these to control a much cooler vaporizer heater?

 

[skippymcware]

Vap,

When you are looking at controllers, you just have to make sure that they are designed for the voltage and current that your heater requires. If you are using a cartridge heater, it will most likely be 120 or 240 volts. Given that it is 100 watts, this would give you a current of less than an amp. Interestingly, a dimmer switch is actually just a silicon controlled rectifier (SCR) device. My guess is that it would be a "zero cross" version. That is, it turns the voltage on and off when the sine wave of the AC voltage is zero.

From wikipedia:
"Modern dimmers are built from silicon-controlled rectifiers (SCR) instead of potentiometers or variable resistors because they have higher efficiency. A variable resistor would dissipate power by heat (efficiency as low as 0.5). By switching on and off, theoretically a silicon-controlled rectifier dimmer does not heat up (efficiency close to 1.0)."

So, in summary, a dimmer switch would function quite well for what you need as long as it is the proper size and is rated for the current and voltage you plan to use it for. I think there may be more similarity between your vaporizer and what is going on in this thread than you might imagine.

Good luck,
-skippy

 

[skippymcware]

This is the first prototype of many. Some terminal blocks would have made this much cleaner, and will be added soon. I just wanted to get to testing. It seems to work pretty well. We are debating whether or not to put this version in an enclosure or go straight to the next iteration. We will see after I get a chance to play with this one for a while. Anyway, here are the pictures:






I will post more details when I get a chance. I have lots of other stuff to mention, as well.

-Jake

 

[vap999]

This is the first prototype of many.

The combined heater/output steel pipe looks to be very sturdily bolted down. But all of that high heat-conducting copper being used for this is probably acting as a major heat sink. With your goal being rapid, near real-time temperature control, you probably should insert some insulation around the pipe where now contacting its mounting hardware, or at least use less massive steel mounts, with steel being a much poorer heat conductor than copper.

What heater wire/coil unit did you end up using? It appears to be pre-made commercial product wrapped around the pipe, rather than something hand-made. Did you somewhere purchase nichrome wire with heat-conducting electrical insulation? [I presume one can't simply wrap bare nichrome wire around steel or any other electrical conductor, with this shorting the circuit, while this is commonly done with glass (non-electricity conducting) vaporizer heater pipes].

Where is the thermocouple sensor? I only see one lead wire going to the pipe. Is it built into your heater coil?

This is a whip type vaporizer, traditionally requiring a large diameter air output. To increase heating efficiency and reduce heating response time, have your considered using a smaller diameter, perhaps a little longer and thnner-walled, heating pipe and maybe one made of high-purity copper (better and faster heat conductor than steel), and then simply attaching a larger diameter steel compression fitting or other adapter on the end?

 

[voltaic]

Damn fine work y'all done here!!!

Just stumbled upon this site today and have definitely found what I've been missing.

I've been tossing around the idea of a DIY vape for a year (given their ABSURD costs) or so and finally began penciling down some ideas a few weeks ago.

I just began throwing together v.1 of my vaporizer (still PLENTY to go) but I'll dump a little of what I've done here to stir the pot a little more for ya skippy

The only design point I am really centered on is accurate, consistant air flow (well I guess that's kinda the goal of all) and being able to make this thing for around $100. On that point, I was riding the PWM train will y'all on that one. However, I went about it slightly different. My PWM will be supplied from a 555 timer that will have duty cycles ranging from 5% to 90% (at least the prototype version until I figure out the sweet spot) with a natural frequency range of about 3-30 Hz. The PWM will feed into a solid state relay (SSR) to vary an AC supply across the heating element. I kept the frequency of the PWM low to compensate for the natural switching delay of my solid state relay which switches on (worst case) at 100 Hz and switches off (worst case) at 25 Hz. I will have feedback control provided by a thermistor array (for more accurate readings) placed directly below the screened bowl, in between that and the heating element. The thermistors will feedback into the 555 timer to infinitely vary the PWM supplied to the SSR.

That's the just of the control details with mine, hope it helps. But I'll end up starting a thread to get some user feedback on mine as well.

I just started soldering some components on the perf board last night and should have the circuit completed sometime this week. So if all goes well, I should have a crude, but functional, prototype by the end of next week.

skippy- if ya feel like bouncing some electrical issues off of me feel free (I am an EE) as long as I can pick your ME brain for some goodies- like some CFD on Solidworks- so that by the end of this thing Volcano will be calling us for ideas

Cheers all and happy 4th,

-DP

 

[Hippie Dickie]

voltaic - i'm with you on the PWM as the best current control to a heater, but the 555 seems like a lot of extra effort. i use the PWM function of a 12F683 PIC processor, with a MAX6675 (cold junction compensated type-k thermocouple converter) to do the temperature feedback. Parts cost is about $12. And the duty cycle is of 0% to 100%. i still need some work on the duty cycle algorithm, but i'm pretty confident i can achieve +/- 0.5F.

i'm an EE, too, but i turned to the soft side - hardware hurts my head, but when i get a problem into the digital domain, i fucking rule.

Did you see my thread on the Bud Toaster?

 

[voltaic]

Extra effort?? I thought you said your code was some 40 pages long! I'd say we both still have our work cut out for us- while you'll be refiguring code to fine tune, I'll be swapping resistors and maybe a cap or two. But, there's always more than one way to skin a cat!!

Someday I'll probably take a look at your code and see what kind of other options it provides. I believe you stated it was written in assembly (SWEET!!), the one language I don't mind poking around in- I feel that the high level languages disconnect ya to much from that beautiful black art of analog design.

I missed your thread on Bud Toaster, did ya have a link?

Well now it's off to the best hangover cure out there- a tasty pull of some fresh herb.