Did you see something in the data sheets that might be a problem?
The Bud Toaster - (currently: Model 14, version 3)
- Thread starter Hippie Dickie
- Start date
well, i don't have the FDB8832_F085, so i can only assume it will work, but it looks the same to me.
Egzoset
Banned
Hello Hippie Dickie,
I wish to invite you as the author of the Bud Toaster project in this thread of mine to discuss similarities and differences, etc. Well, whatever that comes to your mind relatively to such matters:
This is a fantasy thread, please feel free to explore the full span of possibilities...
I wish to invite you as the author of the Bud Toaster project in this thread of mine to discuss similarities and differences, etc. Well, whatever that comes to your mind relatively to such matters:
This is a fantasy thread, please feel free to explore the full span of possibilities...
thank you for the invitation. Your design is so far removed from where my head is at that i'm afraid my views are just totally non-relevant.
And, right now (as i type) i am deep into the design of Model 14 version 2 (version upgrade). Production issues are driving me crazy so i'm hoping this iteration will be more manufacturable.
And, right now (as i type) i am deep into the design of Model 14 version 2 (version upgrade). Production issues are driving me crazy so i'm hoping this iteration will be more manufacturable.
Egzoset
Banned
Oh but i'm quite certain you still could have found relevant critics to make (and even more not so relevant ones) since this is only a fantasy! Manufacturing issues don't have to be taken into account, euh... But i can understand how preoccupied you've become with the Bud Toaster project, a real one. Perhaps later.
Status update ...
Last week was kind of interesting ... i absolutely, thoroughly convinced myself that i had to swap the MOSFET and the push buttons -- move the MOSFET to the bottom and the buttons to the top. Theoretically, that would let me switch back to the 90 angle buttons, so they would peek out the back face of the cube.
Last week was busy with real-world work, so all i could do was think. and think. and think. So after a week of tossing and turning ideas every which way, i finally realized the buttons can't be in that position due to a physical layout of the cube -- just won't work.
But i did work out how to connect the plug to the pcb, so i'm now working on the next pcb iteration. i also found some bugs in the code and the algorithm is much tighter and "better" -- still fucks up randomly very occasionally.
Then, last night, i broke this unit -- the green LED was OFF!!! and i was starting to panic ... but i calmed down and worked through the possibilities. Power was getting from the battery to the vReg and the vReg was pumping out 4.5vdc to the circuit.
Then i noticed the trace to the LED was fractured. Fixed that, reassembled, plugging it in several times during reassembly to be sure it was working. Then loaded and plugged in for a celebratory toke, and the heater turned on by itself and i had to yank the plug.
Totally stumped why this would happen, i opened it up again, and finally saw a broken ground trace. The lack of ground convinced the MOSFET to output full power. Patched the trace.
Reassembled, testing several times until the last screw turn -- plug it in and no blue LED. WTF!?
Opened it up and found the blue LED had been scraped off the pcb during assembly. Fixed it.
Reassembled, testing several times until the last screw turn -- plug it in and MOSFET goes to full power. WTF?!
Opened it up and found the patch to the broken ground trace had tore up. Fixed it "better" this time.
Reassembled, testing several times until the last screw turn -- plug it in and everything was back to normal operation.
And, it's been working okay today (except for one software error during one run. wtf???)
BUT, i did learn a very valuable lesson as a result of all the extending testing of this unit -- the traces are too fine. So this new iteration will have much wider traces -- at least as wide as the geometery allows.
This actually feels like progress.
Last week was kind of interesting ... i absolutely, thoroughly convinced myself that i had to swap the MOSFET and the push buttons -- move the MOSFET to the bottom and the buttons to the top. Theoretically, that would let me switch back to the 90 angle buttons, so they would peek out the back face of the cube.
Last week was busy with real-world work, so all i could do was think. and think. and think. So after a week of tossing and turning ideas every which way, i finally realized the buttons can't be in that position due to a physical layout of the cube -- just won't work.
But i did work out how to connect the plug to the pcb, so i'm now working on the next pcb iteration. i also found some bugs in the code and the algorithm is much tighter and "better" -- still fucks up randomly very occasionally.
Then, last night, i broke this unit -- the green LED was OFF!!! and i was starting to panic ... but i calmed down and worked through the possibilities. Power was getting from the battery to the vReg and the vReg was pumping out 4.5vdc to the circuit.
Then i noticed the trace to the LED was fractured. Fixed that, reassembled, plugging it in several times during reassembly to be sure it was working. Then loaded and plugged in for a celebratory toke, and the heater turned on by itself and i had to yank the plug.
Totally stumped why this would happen, i opened it up again, and finally saw a broken ground trace. The lack of ground convinced the MOSFET to output full power. Patched the trace.
Reassembled, testing several times until the last screw turn -- plug it in and no blue LED. WTF!?
Opened it up and found the blue LED had been scraped off the pcb during assembly. Fixed it.
Reassembled, testing several times until the last screw turn -- plug it in and MOSFET goes to full power. WTF?!
Opened it up and found the patch to the broken ground trace had tore up. Fixed it "better" this time.
Reassembled, testing several times until the last screw turn -- plug it in and everything was back to normal operation.
And, it's been working okay today (except for one software error during one run. wtf???)
BUT, i did learn a very valuable lesson as a result of all the extending testing of this unit -- the traces are too fine. So this new iteration will have much wider traces -- at least as wide as the geometery allows.
This actually feels like progress.
stinkmeaner
Well-Known Member
Sounds like a good place for the buttons from a user friendly/design perspective since after a while of use a user should get familiar enough to use the buttons without looking, buttons on the back would mean the user can just use one finger to adjust the temperature.
Egzoset
Banned
While reading about your problems with copper traces i was thinking perhaps instead of trying to control the trace's width why not just proceed in a complementary way and focus on the width between your copper traces instead?
This method comes with an advantage: less copper needs to be dissolved, which translates as using less etching liquid...
This method comes with an advantage: less copper needs to be dissolved, which translates as using less etching liquid...
@stink - i just can't physically put the oven tube assembly into the cube with buttons out the back -- the buttons don't have enough leeway to fit. the buttons are one of those issues that will take another iteration (or ten) to finally nail down. Buttons on the bottom work okay - just not as pretty as i want.
@egzo - i need to have wider traces to have more copper holding onto the pcb. i'm sure the pcb factory recycles all the chemicals, not to worry. When i finally get production quantities i'm going to have them do a solder mask and use 2.5 oz copper instead of 1 oz copper -- that should resolve any issues.
i don't etch my own boards anymore, but the etchant i use is dirt cheap -- muriatic acid and hydrogen peroxide (1 part to 2 parts), cheaper and faster than ferric chloride.
@egzo - i need to have wider traces to have more copper holding onto the pcb. i'm sure the pcb factory recycles all the chemicals, not to worry. When i finally get production quantities i'm going to have them do a solder mask and use 2.5 oz copper instead of 1 oz copper -- that should resolve any issues.
i don't etch my own boards anymore, but the etchant i use is dirt cheap -- muriatic acid and hydrogen peroxide (1 part to 2 parts), cheaper and faster than ferric chloride.
stinkmeaner
Well-Known Member
I can definitely see how that would be a hard one, especially since the Bud Toaster is the smallest PIC temp controlled vaporizer there is or will be. 
I heard this song and thought "Bud Toaster" Not that you would want to use something corny, it just seemed to fit in a funny sort of way. http://www.youtube.com/watch?v=uBPxMoEOmjo
Imagine>
Bud, Bud Toasters, We toast the bud when the others will not.
Bud, Bud Toasters, stay in the toaster when the toaster is hot.
You get the point,
I heard this song and thought "Bud Toaster" Not that you would want to use something corny, it just seemed to fit in a funny sort of way. http://www.youtube.com/watch?v=uBPxMoEOmjo
Imagine>
Bud, Bud Toasters, We toast the bud when the others will not.
Bud, Bud Toasters, stay in the toaster when the toaster is hot.
You get the point,
budballer
Well-Known Member
Hrm interesting progress overall, great job to you sir! I admire you're unrelenting persistance. But I wonder why all the complications with the pcb? I feel pike it'd be much easier to get a basic pwm controller like a motor speed controller and a dial. Then just use the thermocouple to measure what temperatures are maintained at 10 different degrees of the dial turn. All the way left. Low pwm or low heat, all the way right to 10 would be high pwm. Mark each section 1-10 with what temperature the heating coil produces almost like a classic volcano setup. I think i might use this format myself for its simplicity using rc products. It's just a suggestion for a simpler iteration. However I have full faith the bud toaster 14.2 will be superior, but it will also be harder to manufacture. So I'm just suggesting an easier alternative business wise.
not so ... all these iterations are to resolve fabrication issues, and each iteration removes a manual processing step and adds reliability.
The very first PWM module i got on ebay to try as a controller was like you suggest -- using a potentiometer to set the pwm output. i hated how it operated, how it was controlled -- compared to what i have now.
Also, temperature is affected by the amount of bud in the vial, and bud moisture content and trichome density, and ambient temp (to a minor degree)(lol) - so the dial would never be accurate.
i'm too close to perfection to change course now.
Most of my complications are due to my being a software guy and i'm just struggling through the hardware.
however, trace width is how you specify more copper on the pcb
yes ... i'm going to address trace thickness when i switch from 1 oz copper (per sq ft) pcb to 2.5 oz copper pcb for the production runs.
Egzoset
Banned
There you go, the amount of ounce(s) per square feet is how one specifies "more copper on the PC Board". This parameter affects trace thickness directly on a blank PC Board while trace width is not how one specifies "more copper on the PC Board". My suggestion stands, if you need larger/stronger traces which won't lift then why not simply reduce the void between them? Instead of drawing traces of controlled width i strongly recommend you draw (air) insulation space of controlled width between those tracks. This will improve "adherance" (e.g. the force which bonds copper traces to a PC Board's insulation substrate).
when i mentioned more copper on the board i was referring to the size of the footprint, i.e. the trace width. The wider the trace, the larger the footprint of copper to adhere to the board. Hence, more copper on the board.
i guess i don't understand what you're saying. if the trace width is now 0.025", it doesn't matter how much space there is between this trace and the trace next to it. The traces have, for the most part, been widened to 0.062".
The pcb layout software works with trace width and trace placement. The voids are where the traces don't run. Yes, i am reducing the void areas because they are being filled with a wider trace. It is not possible to draw air insulation spaces directly.
i guess i don't understand what you're saying. if the trace width is now 0.025", it doesn't matter how much space there is between this trace and the trace next to it. The traces have, for the most part, been widened to 0.062".
The pcb layout software works with trace width and trace placement. The voids are where the traces don't run. Yes, i am reducing the void areas because they are being filled with a wider trace. It is not possible to draw air insulation spaces directly.
Isn't enough for what?
Why do you think drawing voids on the pcb is better than drawing traces? Seems to me the pcb software is very logical and actually a nice program ... now that i've mastered it (more or less).
Egzoset
Banned
Lets see...
Here's a visual exercise, just for you: keep focussing on the central area of this graphic animation while all of the slides are being displayed...
Now, is that sufficicient to get through to you (read "to be understood")?
Dark brown represents the amount of copper to be etched away. The more brown there is the more excess copper gets washed out, leaving relatively weak conductors behind which are easier to lift (e.g. seperate from the PC Board's substrate because of their little surface), not to mention that unsufficiently large copper traces provide less current capacity, a most undesirable feature in power applications! Briefly put, more brown is synonymous of wasteful/innefficient design in this case.
Golden represents the copper traces. Three layout geometries are illustrated, it's better to avoid sharp edges so i favour a trade-off between the usual fancy (round) solder pads versus huge copper traces with straight corners: why not have large conductors with rounded ends instead? By the way, varnish is a suitable feature in power applications so there's no need to count on the surface tension of conventional round solder pads to keep solder where it's required since varnish accomplishes this task very well anyway...
Finally but not least, please take note that the minimum distance between my copper tracks is the same in all of those three layouts (this parameter depends on etching resolution).
Here's a visual exercise, just for you: keep focussing on the central area of this graphic animation while all of the slides are being displayed...
Now, is that sufficicient to get through to you (read "to be understood")?
Dark brown represents the amount of copper to be etched away. The more brown there is the more excess copper gets washed out, leaving relatively weak conductors behind which are easier to lift (e.g. seperate from the PC Board's substrate because of their little surface), not to mention that unsufficiently large copper traces provide less current capacity, a most undesirable feature in power applications! Briefly put, more brown is synonymous of wasteful/innefficient design in this case.
Golden represents the copper traces. Three layout geometries are illustrated, it's better to avoid sharp edges so i favour a trade-off between the usual fancy (round) solder pads versus huge copper traces with straight corners: why not have large conductors with rounded ends instead? By the way, varnish is a suitable feature in power applications so there's no need to count on the surface tension of conventional round solder pads to keep solder where it's required since varnish accomplishes this task very well anyway...
Finally but not least, please take note that the minimum distance between my copper tracks is the same in all of those three layouts (this parameter depends on etching resolution).
how is what you just explained different from what i already stated? That is, i want to make the traces wider, too. Which is what i did when i increased the width of the traces.
what position are you "arguing" for? -- that is, what am i supposed to do different, in your mind, than what i have already explained as that which i am doing?
What exactly am i supposed to be learning from your animation that i have not already explained?
Current carrying capability is not an issue in my circuit layout. All the electronics are running about 1 ma each, except the LEDs which are hitting about 10 ma each.
The copper path to the coil is cut from a "bus bar" made of 1/32" thick copper plate, about 1/4" wide -- enough to handle the anticipated 12 amps current flow to the coil.
what position are you "arguing" for? -- that is, what am i supposed to do different, in your mind, than what i have already explained as that which i am doing?
What exactly am i supposed to be learning from your animation that i have not already explained?
Current carrying capability is not an issue in my circuit layout. All the electronics are running about 1 ma each, except the LEDs which are hitting about 10 ma each.
The copper path to the coil is cut from a "bus bar" made of 1/32" thick copper plate, about 1/4" wide -- enough to handle the anticipated 12 amps current flow to the coil.