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

[Hippie Dickie]

Here's the current configuration:

i added a bracket to hold the thermometer in place ... i hate it, it's ugly, and once the duty cycle algorithm is fixed it is unnecessary ... but until then ... it's my security blanket.

Here's some big tube (16mm x 150mm) action:

Filled with vapor:

 

[Bubar]

Ah! Is your transistor is standing up! lol. I hadn't notice that before. The backs of my transistor pcb's are getting real hot and browning. That seems like a better idea. Also, my transistor, which unfortunately i have about 15 of seems to want 5.2V? I got a diff supply, so tomorrow I'll hack up the power connects. I may make a transistor set up with the transistor u used, I still have 2-3 that i haven't burned up.

 

[vtac]

Want!

It's sexy even with that bracket, so compact! I'm still impressed it's written in assembly.

 

[Hippie Dickie]

Bubar - yeah, it's a surface mount MOSFET, but i stand it up, solder a piece of 1/4" wide by 1/32" thick by 1" long piece of copper to that big drain connection on the back of the MOSFET, as a high-current bus bar to connect to the coil. Coil heats up a couple of degrees per second faster ... Keeps the heat off the pcb. There's plenty of room for that non-standard configuration with this design.

It seems ridiculous to me that the puny pins on the thru-hole MOSFET can carry 40 amps - far smaller than 12ga cable, but supposed to carry 40 amps???

vtac - i've been doing computers for 45 years, assembler is my native tongue.

i'm making progress on the pcb routing program -- just need one more footprint for the buttons and it will be done.

 

[Bubar]

Is it just me or did your method for attaching the heating elements change in these last set of picture?

 

[Hippie Dickie]

change from what? yeah, maybe -- but this current setup works well in this physical configuration.

By the way, i've been having a very interesting email exchange about problems with the MOSFET configuration.

Basically, the MOSFET is not being driven "hard" enough -- that is, the voltage to drive the gate should be way higher than 3.3v, ideally it should be 10v or more. Further, the red LED is clamping the voltage to the gate of the MOSFET at 2.7 volts (by measurement). This makes the Rds(on) absurdly large, so the MOSFET is wasting a lot of power and getting absurdly hot.

So, i'm deep into re-specifying the MOSFET and probably the Vreg, and i've got to do something different with the red LED. The redesign goal is to have the MOSFET run cool. The side benefit is the battery runtime will probably be doubled.

Perhaps this explains the problems you have been seeing with your setup.

 

[Bubar]

I think I said something about clamping before, the way to fix that put a resistor before the LED with it in parallel with the heating element. That is how I have duty cycle indicator LED's in other projects.

The stuff about the mosfet is interesting. My mosfet has a gate of 3-5v but is rated to 30v G-S. I think I need to find a better mosfet too. I am on my 3rd mosfet of the day now...

 

[Bubar]

Also I gotta say that I hate lead solder, tho that is all I have right now. The vape smokes with lead fumes the first few times it gets run..

 

[Hippie Dickie]

Yes, you did mention the LED clamping the gate ... i just didn't know what you meant. Today (!) i learned a fuck-ton more about how to read the data sheet for the MOSFET. And about LEDs. This project has been a mind blowing trip about learning a whole new field -- about learning, in general. So much fun it's hard for me to catch my breath. i thought my day job was a challenge ... this project is a challenge and a whole lot more.

re: LED in parallel with the coil ... i had that thought about 2 hours ago and need to test it. i wondered if the LED would cause a short. i need to go 10x the ohms because the LED will be seeing up to 30v on the coil, versus just 3.3v on the PIC pin.

i would really hate to give up the red LED ... it is so cool to see what the coil is doing.

when we get the MOSFET spec'd right there will be no heating.

Don't you have a Radio Shack around you? They have lead free solder. Flux makes all the difference in getting decent soldering joint with lead free.

 

[Bubar]

It shouldn't cause a short....

+ ==heating element ==Drain
+==330ohm resistor==LED==Drain

 

[Hippie Dickie]

thanks ... i'm going to try that tomorrow morning.

 

[Bubar]

I had lead free, but I used it up. I have an infinite supply of leaded solder...so that ends up being what i use. There is a radioshack though. I guess I will go buy some when I go buy a project box. I just snapped the shitty plastic box I made.

I finally ran the vape without killing the transistor, so I will probably take it home. I'll cross my finger for it workign at home. One thing I realized today was taht my transistor was turning on super slow.. like starting at 300mA and taking about 1min to reach 5A (max for the supply i was using). I thought that was due to the fancy lab bench supply I was using, but I think that is actually my retarded MOSFET, because when I give it 5.3-6v it instantly gives 5A. A lot of times I wish I had a more comprehensive EE background.

 

[Hippie Dickie]

i think i know just enough EE to be dangerous ... i go to Wikipedia to learn about electronics.

are you talking about giving the MOSFET 5.3-6v as the gate voltage? Definitely the gate voltage is a critical factor ... the curves go funky below 5 volts.

 

[Bubar]

Wow. Ok. I was retarded and misread threshold and gate as being the same thing without consulting wikipedia first.

 

[Bubar]

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=NDB6060LCT-ND
This might work. It's listed under Logic Gate MOSFET's which seems appropriate.

 

[Hippie Dickie]

That MOSFET is basically like the FDB6030BL i've been using.

Take a look at Fairchild FDB8832, 30v, 80A, 2.1 mOhm -- this is an order of magnitude lower Rds(on) -- this is the resistance between Drain and Source when the MOSFET is on, and a pure waste of energy. Look at Figure 9 in the data sheet. A gate voltage of 4.5 volts is below the knee of the curves.

Edit: And, the 2.7 volts that my circuit is feeding the gate would drive the Rds(on) to infinity -- basically, the MOSFET isn't really turning on.

Since i'm using batteries, i'm now looking for a Low Drop Out voltage regulator that provides 4.5 v output with a .1v drop out -- meaning the batteries have to provide at least 4.6v to allow the regulator to provide 4.5v to the circuitry (most vregs are 1.5v drop out) . my batteries are discharged at 2.5v, or 5.0v for the 2 cell pack. This should work fine to minimize the Rds(on).

The downside of using LDO vreg is the max voltage input allowed peaks out at about 20 volts. Still, that's well above a car 12v battery or a DeWalt 18v power pack.

 

[Bubar]

So what is your problem then? Your transistor is fine, but your Vreg is not providing enough voltage to you circuit?

 

[Hippie Dickie]

Basically, yes. The Vreg is putting out 3.3v which is what i wanted, which is what the PIC provides on the PWM pin, but that's really too low to drive the MOSFET gate hard enough.

So, since i need to change the Vreg anyway, i might as well get a more optimal -- i.e. lower Rds(on) --MOSFET.

Also, the 1K resistor from the PIC to the Gate is too much current limit. It should be about 10 ohms - the MOSFET switching rate depends on how fast the capacitance in the MOSFET is charged, and 1k is a limiting factor. The slower the switching rate the more wasted power in the MOSFET and the hotter it gets.

By the way, i moved the red LED to be across the coil and it works just fine.

So, i just need to get some new parts and i can finally get the PCB fabricated. i've about mastered the layout sw and almost have the finished design. The LDO Vreg is a different footprint.

 

[vapordank]

Hippie Dickie now thats a name i recognize!!. You helped me through the construction of my own bud toaster about 4 years ago!@ So cool to see that your still building vaporizers. I got really into blowing glass, and have been using that to my advantage with some of my newer vapes. Im going to go look through this thread and see what youve been up to!

 

[Hippie Dickie]

Hello vapordank ... good to see you again.

Here's a synopsis of what i'm up to these days:

computer temperature control - accurate to 1F,
power on to first toke in 60 seconds,
battery powered -- one charge lasts for over 30 minutes of vaping,
red LED that lights when the coil is getting power,
green LED that lights when at vape temp - the "take a toke" signal.

Cost of the computer controller is about $20 in parts.

The Bud Toaster looks pretty much the same on the outside - my avatar is the current prototype.

Note: the first post of this thread has been updated with the latest (1/10/10) circuit diagram.

 

[Bubar]

I have a temporary solution before I get new transistors ordered. I will use another transistor to give the mosfet 12v from the PSU.

 

[Hippie Dickie]

yeah, i considered doing something like that myself -- instead i just ordered a low drop out voltage regulator that puts out 4.5v. This looks suitable for the Fairchild FDB8832 MOSFET.

Max input voltage will be 20v - okay for car (and 18v DeWalt battery pack) in "emergencies".

However, my red LED blew -- i guess it can't tolerate the 6 volts across the coil. i used a 1k ohm resistor -- should be drawing only 6 ma - only a third of the 25 ma max - and it is definitely dimmer, but it sure is dead.

yet another puzzler. maybe i can use a transistor to isolate the red LED from the MOSFET gate circuit???

 

[IAmKrazy2]

Really some awesome ideas. I will have to go back and look in more detail but i am very impressed.

 

[Bubar]

Ok. Got some of those transistors (fdb8832) in the mail. This is gonna be exciting when the transistors stop melting.

For you LED try a larger resistance maybe? I forgot to take into account in my other project the heater line is only seeing 3.3v. I've never had an led die with that set up.

Here is my implementation of the nichrome mounting.

 

[Hippie Dickie]

Bubar - looking good. i find that mounting technique to be the best one i've used.

larger resistor won't solve the problem since i'm allowing such a large range of input voltage ... and the 1K makes it too dim at 6.6v as it is.

i have an EE question: if i want to put the red LED back on the PIC pin that controls the MOSFET, can't i just add another MOSFET to control the LED? That extra MOSFET won't clamp the voltage like the LED diode itself does, will it? i'm considering this one: NEC MOSFET 2SK1958 ($0.12 ea). It shouldn't clamp the voltage on the gate because it's not a diode, right?