The auto-magic super rZToaster, or something like that..

[GreenHopper]

Looking good man.

I don't know if @Ratchett would be a good guy to ask about enclosure design. I know he's been quite busy lately with Delta3D studios but he might have time for a chat?

He's a 3D printing whiz and might be able to advise on what the best way to design an enclosure would be. Maybe he could advise on the feasibility of 3D printing them or possibly just 3D printing some prototypes?

@Pipes was using off the shelf aluminium cases sourced from one of the chinese online markets. Not sure how easy it would be to adapt one of those considering the screen and buttons, etc...

@phattpiggie has made a wooden enclosure for his Induction Heater, although I appreciate your board has a number of significant differences such as the screen and the buttons.

Just throwing out ideas really.

 

[phattpiggie]

I've been keeping eye on this thread but most it goes over my head, especially when @rz starts talking Swabloodyhili.

Perhaps source an off the shelf box, box mod suppliers or as @GreenHopper says an ali case.

Off the shelf or custom enclosure getting the buttons in the right place would need some sort of template so everyone turned out the same. One hole out of place and it's bin it and start again.

To be honest my head is all over the place with this one, every time I read thru the posts I end up with the same thoughts running over and over.

This could be a more than a VC accessory.

This could be a vaporizer.

 

[jimt0r]

Love that Display

if ya do a nekid kit, i would recommend mountable buttons and a wire leads, ribbon wire for the display. to allow for a little flexibility for custom boxes

i kinda see where phattpiggie might be going with this, glass tube with a inductive load wedged in the tube to perform heat transfer in airpath

 

[toofluff]

Teensy is awesome and a great choice. Personally I would recommend an ARM Cortex-M chip now anyway over Atmel AVR's. They're usually much cheaper, more powerful, and use less power. Arduino is a very fine choice too, even for production stuff, and they're supporting more and more ARM/non-AVR cores so it's not a big deal to keep using it and just compile for other chips.

For a DIY kit, I like what you did with your board, to just lay out some through-holes to BYOT(eensy) right onto the board. That's probably best for a kit in my opinion.

As cool as the screen/display is, some RGB LEDs to indicate temperature, like a Pax vape, would be a lot cheaper and take up much less space. The RHoS thing, needing to add buttons for interacting with it, etc can be a pain.
An alternative is to use something like a Nordic nRF5x Bluetooth chip (Arduino supports it too) to run the code both for the heater PWM, temp measurement and all that; but also act as tor he "display" interactive part using BLE with a phone app. Used to be that would cost a lot having to write apps for both Android and iOS but nowadays there are some pretty decent frameworks that can do that for both platforms to let users adjust temp parameters and all that. Hopefully it's not something people would need to be adjusting each time they use it anyway. You'd save a bunch in BOM cost and space; and with using pre-certified modules you wouldn't have to worry about FCC/CE stuff for the antenna/RF parts.

I wish I had enclosure experience...but that's a whole 'nother ball of wax. Like GreenHopper said, 3D printing could work well for smaller batches. But after that, even when you do get a nice one designed, the tooling cost for injection molds is still quite high and needs pretty hefty volume/margin to pay off. Like GreenHopper said and Pipes did, maybe look for some off-the-shelf enclosures first, since they can usually mod them for you to cut out holes and stuff for pretty cheap, without needing to make your own molds.

If you do ever decide to open source it, I'd love to have a look and try to build my own or buy your kit!

 

[rz]

OK, so my logistics setback should be solved soon. I should finally have my parts to build up the first prototype PCBs in about a week! However, my kids surgery is in two weeks, so.. not sure when I'll get to build it.

A step in the right direction at least. All will be well in good time.

@toofluff yeah I thought of a bunch on LEDs at some stage, but the OLEDs from china are just a few $'s and they're really cool I think I'd prefer to over-kill this thing than to super-minimize it due to costs. I hope most ppl would prefer to keep some cool factor overkill for a few extra $.

I also like ARMs. I will keep the Teensy LC as the target MCU platform for now. The second iteration of the board will be without the MCU on the main board. It'll be a module with just driving and monitoring circuitry along with the IH unit on one board, with breakout to a second board that'll have controller+display+buttons. This will also be pretty easy to DIY mod your own controller.

Also glad that my other two goons went back to kindergarden this week, so I should have some DIY time again soon. Been a busy summer.. Meanwhile, prototype v0.01 keeps toasting away

 

[rz]

My kids surgery went well!

And.... My parts arrived!

(almost, minus an item I mixed up during my component selection. ponied up to get it via local supplier, should be here soon.. I have enough to start building in the meantime.)

We still have a busy few weeks ahead but that won't stop me. Slowly but surely, gonna get this thing going

Stay tuned..

 

[GreenHopper]

My kids surgery went well!

And.... My parts arrived!

(almost, minus an item I mixed up during my component selection. ponied up to get it via local supplier, should be here soon.. I have enough to start building in the meantime.)

We still have a busy few weeks ahead but that won't stop me. Slowly but surely, gonna get this thing going

Stay tuned..

Awesome news bud, glad he's through the worst of it.

Looking forward to seeing this heater in action.

 

[rz]

sooo..... I managed to play a little here and there over the last few days

(post may be a bit a little technical..)

When I first got a chance, I continued building up on the board I had started on already. The board with the OLED, buttons, and IR sensor. I added voltage regulators, gate control transistor pair, some other bits and pieces.. and.. DOH! I forgot a few things, and got something in the wrong size

The digital current sensor chip I ordered is too big - need to order smaller ones.
Forgot to order bts50060 high side smart switch
Forgot to order LM27313 - Boost converter

Still plenty to work with, so..

Then I got to the induction heater bits, and couldn't get it to oscillate! I even zapped a transistor along the way..

I have five boards, and bought about four bunches of different transistors to test different ones out. So I built another board, starting from the heater bits this time using an external LC tank, and it din't F'n oscillate!

HMMM.. went over everything, parts, specs, etc.. dunno.. built another board with a different set of transistors, and VOILA! WE ARE IN BUSINESS!

The transistors got really hot, and I still wanted to experiment with the boost circuit. The boost circuit is designed to drive the transistors gates with higher voltage so they work in more efficient way. It will also allow the thing to work on just two batteries! So I found a cheap little external boost circuit laying around (obviously..), and it works a charm! The transistors didn't get nearly as hot! But their gate resistors.. GOT HOT AS F DAMN! MFKR..

I'm not done playing yet, but I've learnt a few things so far..

I probably over-killed some of the copper traces width. It was sometimes difficult to solder as the board took a lot of heat. It didn't heat up too much when in use. Will consider thermal relief solder points.

Transistor's gate resistors get hot as F when using the boost circuit. Will try absorb more into PCB in next build.

Circuit works a charm at 6V (~2 batteries) consuming 5A for 30Watts. It is slower than at 12v of cource, but it'll be controlled, so not a huge issue. Even at 5v it ran at 20 Watts. Helped keep transistors cool enough they may not need heatsinks, but HOT DAMN OUCH those resistors may if I don't suck their heat into the PCB.

Fortunately I still have the analog current meters. I'll continue building this working board up with everything else. I can short circuit what would have been the high side switch for now, since I'm using transistors to control gate voltage for PWM anyway.

For now, looks like I may really go down the two battery path since the boost circuit works well.Waiting for some 2S BMS boards already anyway. I'll try up the power with more capacitors.. so still lots to play with!

Here's some work in action. Excuse my mess

Any questions? glad to discuss..

 

[KeroZen]

Any questions? glad to discuss..

Yes! What is the pink blob right to the copper wire spool that looks like a silicone tit inside a dish?!

 

[rz]

Yes! What is the pink blob right to the copper wire spool that looks like a silicone tit inside a dish?!

Ah! That's important! It's a glass sphere. It works really well as a magnifying glass. I mostly use it to check solder joints and my buddies

(Mod note: I was playing with img embed on my post above and the images are not doubled. too late to edit now. plz fix tnx tnx)

 

[Hippie Dickie]

i don't know if this relates: my heater current control MOSFET got too hot until someone contacted me and taught me how to read the spec sheet ... then i realized the gate voltage was clamped by a LED at about 2.7v - way, way too low. so i moved the LED to a different part of the circuit and the gate operated at 4.5v power supply voltage, and got a MOSFET that had much lower R-on, and now it is always room temp when running.

 

[rz]

I added a bit more capacitance. now at ~1uF. I get 5A with 5V supply with an inserted Vapcap, which heats to click in about 20 seconds~30 seconds from cold. This would be with a (worst case) flat battery pack. I'm going to up capacitance to 1.3uF, which will improve it even more. With a nominal 7.4 volts, the two battery set up would probably work twice as fast. Looking good

i don't know if this relates: my heater current control MOSFET got too hot until someone contacted me and taught me how to read the spec sheet ... then i realized the gate voltage was clamped by a LED at about 2.7v - way, way too low. so i moved the LED to a different part of the circuit and the gate operated at 4.5v power supply voltage, and got a MOSFET that had much lower R-on, and now it is always room temp when running.

It does (kind of) relate.

IIRC, you are using DC current through your MOSFET, and it is indeed important to turn it on properly to get down to Rds(on) but once on, it's on.

The induction heater resonates at tens of Khz, so there is a lot of switching on and off and transitions in between. The problem during these transitions is that Rds is much higher than Rds(on), and a lot of heat is generated in the transistor. If the transition is much faster, the transistor heats up less.

The gate acts kind of like a capacitor, or battery.. it needs to be charged. It doesn't go straight from 0 to the driving voltage, it ramps up. The harder you push it, the faster it ramps up. We just need to make sure it doesn't go to high. Using 12v zener diodes is common to clamp the gate at a max voltage.

The MOSFET I'm using now has a Vgs(on) threshold below 2v but only get to Rds(on) at about 10v (figure 6).

When using a 3 battery design, with voltage around 11v, the gate doesn't turn on quickly enough, and only gets to about 6~7 volts before it starts turning off - this still works, but does generate some manageable heat.

When using a 2 battery design (~7v) , the gates only get to about 4~5 volts before turning back off, which is way too inefficient, and get really hot.

Using a voltage boost circuit, we can bump that 7v supply voltage from the 2 batteries up to, say, 20~30v. This will make the transition much faster, and the gates actually fully open during the cycle, which this keeps them MUCH cooler so they won't need a heatsink. However, pushing the gates this hard causes the gate resistors to heat up quiet a lot. Not too much of a problem, it just needs to be handled. for now I'll use a little heatsink. I'll try design some more copper around them on the next PCB to help, but a tiny heatsink is not a showstopper

 

[pxl_jockey]

@rz I've been lurking here, popping in to check your progress and that of Mini-rz3.0 since I found this shortly after the release party. Glad to know all is well and as settled as a home with 3 boys can be, even more impressed that you find the time/energy to press on with your passion-project! Sending you the best vibes so that you might continue trouble-shooting and refining, although much of your posts make my brain hurt. Keep up all your innovation, I would love to have a kit someday! As always, please let us know how the monkeys are doing!

 

[jimt0r]

Ah! That's important! It's a glass sphere.

thats classic mate I gots to get me one

 

[rz]

Thanks for those vibes @pxl_jockey, every bit helps Also thanks to ALL you other followers!! If it weren't for you, I'd still just be rocking my V0.001 over the next few years!

My three monkeys are doing well. Little one is recovering like a champ

It's been a bit hard to find the time, so I only get a little bit done here and there.... Enough for an update though

So, after my first parts order which was missing bits for my boost circuit and the digital volt/current meter, I got the rest of the board built up. I used an external boost circuit which works really well. I found some transistors that work perfectly, I was missing a pull up resister here, a pull down there, and needed to modify a bit for the frequency measurement system to work at a two-battery voltage (the whole insertion detection is based on measuring the frequency), and I finally got the whole franken-thing working stand-alone with just two 18650 batteries!

I have had a cold over the past few days, so I haven't extensively tested heat-up times etc, but will get back to that soon. I was testing with an external power supply at the minimum two-battery voltage (5v), and it was pretty slow to heat up, but now with a fully charged 18650 pack it's definitely quick enough. I'll fix up lower voltage performance when my new parts get here.

I've also spent some time narrowing the board design down by 30%, to the width of about two 18650 cells (40mm currently), incorporated fixes for issues I found on first PCB and removed the microcontroller (which will be on a separate module).

I had a logistics opportunity for parts, so I just placed an order for a whole bunch of new bits to get my current prototype fully up and running with plenty leftover for next design iteration.

There's still quiet a way to go, but I'm really glad that I'm past the proof of concept stage with a mostly functional PCB design. My next step is to verify the boost circuit when my new parts arrive, finish up the new narrow PCB, do the microcontroller+display+buttons module, order those PCBs, test, modify, and then hopefully be at a stage that I could get some initial 'production' PCBs made.

I haven't even gotten to enclosure design yet, so first gen will probably be naked, but there seems to be plenty of good DIY case makers around for that
So, to good health and progress!(+beer. mmmm.. I love MJ&Beer)

 

[GreenHopper]

Awesome update bud, glad the young ones are well, hope the cold passes quick.

Seems like you are making good progress. It's really unique to be part of a community where you can see people actually engrossed in the creation and realisation of ideas. There are a few sites like instructables but those are more geared towards offering a how to guide after the fact. Although half the stuff you write goes completely over my head it's still fascinating reading your posts.

Regarding the enclosure, I figure it's not worth worrying about until you have settled on a base design. If the first few are naked/skeletors (as @Pipes would put it) I think you'll get a bunch of folk coming up with interesting enclosure ideas.

Thanks for checking in.

 

[Mr. Me2]

Regarding the enclosure, I figure it's not worth worrying about until you have settled on a base design. If the first few are naked/skeletors (as @Pipes would put it) I think you'll get a bunch of folk coming up with interesting enclosure ideas.

I still like my lego idea...

 

[stark1]

Latecomer to your work in progress, @rz. Looks like you may well be putting out some
for the FC Community soon. To marvel.

Can I please get on your growing list for one?

 

[Dustydurban]

Latecomer to your work in progress, @rz. Looks like you may well be putting out some
for the FC Community soon. To marvel.

Can I please get on your growing list for one?

Me Too Please

 

[rz]

So.. still (too?)busy with life/work, but I've managed to make some slow progress, an hour or two here and there.

I've been fiddling a lot with the design. I managed to find an orientation that I like and ended up putting the microcontroller back on the board, so not so modularized, but keeping it compact. I put the bulky coil and capacitors on one side of the board and all short components on the other side of the board. The two batteries will go next to the coil/caps, so it'll be about the size of 3x18650 + some added thickness for PCB and components. Still haven't found the right case.

I got my boost circuit working on board I then promptly killed it with my antique oscilloscopes grounding behaviour..I forgot GND in DC/GND/AC selection actually grounds the probes. WTF. I then botched up some PCB traces while fixing. I did re-build a boost circuit on a spare PCB so I could verify behaviour and do some further testing. I've since dived deeper into boost circuit behaviour and optimized component selection and layout a bit.

I've created the little boards, one for the switches on the front panel, another for the IR sensor with a whole bunch of red/green LEDs directly under the glass tube.

I also really want to attach the battery holder directly to the PCB, and I've been diving into charge management (USB charging?), cell balancing, overcurrent protection and other related designs. I keep moving things around on the PCB to make space but I think i'll go bonkers if I keep adding complexity right now. I need to clean up current design and rely on external BMS board for now, maybe add those features in future design... I'm still waiting for my BMS boards and 8.4v charger from china..

So plan for next few weeks: No more features on board. Stop. Clean up current design. Print PCB and make another component order. build. verify, then order parts for like 10~20 boards to build up. This may take a month or two.. then I'll still need to hand-build them, so that'll take time... but moving along.

Prototype 2 is working really well I get about hmm... 10~15 stuffed caps (1~2 sessions per cap) out of 2x18650 batteries. Enough for a good few days of (average?) use.

 

[rz]

Some technical rant...

I've had a bit of a development bump around battery connection / protection but think I'm honing in on a solution.

I apparently don't want to burn my house down, nor anyone else's, so battery protection is paramount and I've been thinking about it a lot.

At first I was going to rely on an assembled BMS, but the one I ordered ~two months ago STILL hasn't arrived.

Fortunately I'm not prototyping without protection. I'm using a ~16A hold current resettable fuse for now (I'll be drawing about 10A for a few seconds here and there so I can probably reduce to a 10A poly)

Anyway, I also want something like this two-battery holder soldered directly to the board. I could probably piggy-back the Chinese built BMS to it without adding much bulk.. but well.. I haven't received it yet...

I want to reduce relying on other add-in circuits too, and have been curious about BMS design.. so.. I wanted to try incorporate protection on the board, with all them features of course .. so I spent a(n obsessively) long time looking at available power monitor and protection chips..

I'd LIKE to build what looks like a good combination, something based on S-8209 as primary protection which has over/under-voltage protection and cell-balance, and is scarcely available(in small quantities) at-least. It is missing over-current protection, which can be added by using S-8253 but this is NOT available at all in smaller quantities. hmpf. I think this would offer great all-round protection and feed my BMS obsession, would be scalable to many cells, etc, but would be a little project in and of itself, so I'm putting it aside for now.

I've been charging my 2S 18650 pack with an external power supply (bench supply with CC/CV modes that I manually set/monitor). I charge every 4~5 days, so this 2S pack has been charged a handful of times already, and the cells still seem to be pretty well balanced (with my crappy multimeter at least.. I really need some new tools...)

SO.. for now.. for the sake of progress.. I'm going to try work with S-8252, which does not have cell-balance, but DOES incorporate overcurrent protection in a much simpler to implement 2-cell targeted chip. I think this will fit well on the board and be much quicker to implement. I believe if two good quality cells are used from a good source, 2S cell-balance will not be much of an issue, and at least for the next few iterations of my board, they at worst may just need an external charge cycle or two once every few months. In a future iteration I may add a bq29209 as secondary protection and cell-balancing. I believe this will be sufficient protection for this project.

But wait.. there's more

I haven't only relied on BMS. In my current design already, incoming power is split directly (short connection), and only, into two paths:
1. ~120mA polyfuse for the digital path, to the 3.3v regulators.
2. smart high side switch which has integrated short-circuit protection, thermal protection, and undervoltage protection at a perfect bottom line for 2S lipo 4.1v typical (4.4v max) (I'll prevent getting this far down in software anyway, but it's a good minimum cutoff). The output of this switch goes to the boost circuit and induction unit's chokes.

I think the 10A main polyfuse, smart high side switch, 120ma polyfuse for digital, and S-8252 OV/UV/OC is enough for prototype v0.03. cell balancing and integrated charger will have to wait for now (think boost circuit charger ala USB WOOHOO MP2639AGR-Z looks perfect but isn't really available yet anyway..)

Any thoughts? Any engineers with relevant feedback? A lot of new territory for me.. I may need to open a thread on a different kind forum sometime.. However, I'm really appreciating the enthusiasm and interest I've seen on this thread and my inbox lately. Thanks guys! Without it, I'd have moved on to something else by now.. like sleep.. oh yeah.. sleeeep...

 

[Hippie Dickie]

i just serial charge my 2-cell power pack. after the full charge one cell is approx 0.5v higher than the other, but within several minutes they have equalized - automagically. my cells are soldered to a pcb, so not possible to charge them individually - and my 2S charger doesn't have a balancing cable anyway.

i think i'm pulling too much Amps to use a protection circuit (18A max) - thinking about a thermal cutoff if the temp outside the heat shield is >150F or so.

nice that you're making progress. it is rewarding.

 

[KeroZen]

That MP2639A doesn't look like it's doing any balancing, at least they don't list it in the features summary. If you are willing to go this far, at least do it fully! Also the discharge part would be a bit useless in your application... unless you use it somehow to implement the balancing yourself (many chargers work this way for the balancing part, normally during the CV phase: the charge voltage is always applied but periodically all cells that are at or above the cutoff get discharged, faster than they charge obviously. So they are emptied a bit then re-toped in a loop, giving the lagging cells enough time to catch up. This also allows to keep enough current flowing as to not prematurely reach the termination cutoff)

But there is a "mid" pin and a "batt mid" tap on the (very partial) schematics though, so there is hope maybe?

The Infineon smart high side switch chip is interesting. Seems to already offer a lot of what you want.

Adding the TI chip in a later iteration on top of S-8252 sounds like a good plan to me. The S-8253 just to add over-current protection is a bit overkill since you already have a 60A short circuit protection in the Infineon smart switch.

 

[rz]

@KeroZen correct. The MP2639A does not do any balancing. In fact, I couldn't find ANY 2S chargers that incorporate balancing, and IIRC, none of the multi-cell charger ICs that I encountered even integrate balancing. It seems that balancing is left to the monitoring and protection ICs. However, some chargers DO monitor individual cells for OV.

The MP2639A looks pretty cool cause it has boost/buck for charge/discharge, as in, it can boost 5v from USB to charge 2S up to 8.4V, and it can buck 2S down to 5V to charge your other devices This would turn the toaster into a proper external USB battery pack, and you could charge your vape from your phone or your phone from your vape Anyway, not there yet.. but on my mind.

There are a few different balancing mechanisms not too far from what you describe though I don't think it's too critical with 2S if using a good matched pair of cells. I find they stay more or less in balance so even a very small bypass current around the alpha cell during charge would allow the weaker cell to catch up.


While balancing is important, I think the more important part is to limit EACH cell individually rather than the pack as a whole. That's why a battery monitor IC is important as it will prevent individual cells from going out of safe bounds. The Infineon short circuit protection is just another fallback, as is it's under-voltage protection as it looks at whole pack voltage.

@Hippie Dickie, I'm glad your cells are still behaving (more or less). How long have you had them in use, and how many charge/discharge cycles have they gone though ? I'm glad you're still getting decent capacity out of them, but it seems they may need a little balance correction. I think that 0.5v is pretty significant TBH. The difference seems to vanish while you're at the batteries nominal value around 3.7v, about where both cells will spend most of their time anyway. I assume you're charging to about 8.4V? with 0.5V difference does that mean one cell is 4.45 and the other 3.95? In which case, one is over-charged which will impact it's life. the lower charged cell isn't utilizing it's capacity either then, and will hit the bottom ~2.5v before the stronger cell has exhausted capacity. How are you limiting over-discharge? are you stopping at 5V? how is the balance then? What I suggest you do is try manually get them back into balance, and try keep them balanced every few weeks. I think this could be done pretty easily in two ways, even with soldered in cells. You could charge your pack until the stronger cell reaches ~4.2V, then apply a CC/CV on the weaker cell alone with an external lipo charger or bench power supply set to 4.2V with current limit @ say 1 amp. Alternatively, you can charge up your pack as you usually do (try avoid OV), then just slap something like a 50ohm 1/2W resistor on the stronger cell and monitor it till the voltage drops to the same as the weaker cell. Then remove resistor and re-apply your charger. It may take two or three cycles to get them synced up, and they should stay balanced more or less, depending on their health difference. 18A is pretty steep. Ss it 2S? I haven't seen 2S 18A hold current BMS on ebay either, though I'll be glad to help you design one after I burn enough transistors to know a bit more what I'm talking about In the meantime, consider popping in a poly fuse. Check out figure R19 on www.mouser.com/ds/2/240/Littelfuse_PTC_Rline_Catalog_Datasheet.pdf-1021735.pdf - looks like RHEF900 would give you 20Amps for ~35 seconds, 30A for about 10s, 70A for 1 second. Enough for your use case, not ideal response time, but at least some limit on how long it can keep burning up for.
edit: hmm I'm not sure how suitable a poly is. There isn't enough margin between your working current and cell max current, and poly's have quote a bit of R... maybe a 30A fast blow at least. I'm poop scared to hook up lipos without any protection specially if I'm leaving the room.

 

[Hippie Dickie]

okay, i misremembered the difference after charging - i just charged and measured and it is only 0.05v difference - 3.67v and 3.62v. and these are LiFePO4 cells, so max voltage is 3.2v, 2300 mAh. kind of small capacity, but i get twelve 7 minute sessions per charge. charger goes to 7.3v at 3A. (CCCV)

the recommended cell cutoff is 1.6v. i use a LDO 4.5 voltage regulator, with a .2v drop out. so the pack only goes down to 4.7 v, or 2.35v each cell before the power is too low to drive the regulator. i'd use a lower voltage vreg, but the MOSFET needs at least 4.7v to get low on resistance. i need to re-spec the MOSFET before i think about production.

i just measured after a session and the cells were identical at 3.25v each. (this was before the recharge and the measurements above).

I've been using the same pack for several hundred recharges. the A123Systems cell is rated for >1000 recharges with capacity still at 95%. not your typical lipo.