DIY Induction Heater Builds and References

[TommyDee]

LOL @PKOK - Welcome back buddy. Show us what you got.

Learned a few things, but in general, we have the basic module simplified and the new HalfPint half circuit.
Still no progress on charge regimes. However, I am qualifying some 18350 cells that are performing nicely. The 14500 were not quite tough enough. And the 18350 rate at ~1100mah - about a day's worth for me.

Looking forward to your build.

 

[Edgedamage]

My latest build based on the half pint, made to hide in plain sight "the router".
I salvaged the power jack, and LEDs one for power and the world logo lights when the IH is active. The fire button uses the original WPS switch.

 

[RedEyeFlightControl]

One other thing since this is a DIY section. The standard ZVS module is a fairly tall structure. If you want a shorter version of the ZVS, and a more robust one at that, I might suggest starting to look at this unit. It is basically the same unit with more robust fets to drive it harder - read; tougher circuit. You'll still only use 12V and it will still output about what the little module does. But since the DV craze has warped the pricing on modules, it is worth looking at these lower profile layouts. Perfect DIY boards!

I still can't attach images to this board for some reason. IDK why. But here's a visual of the other 2 popular ZVS boards compared to our little friend. I picked up one of each to screw around with. They're pretty much the same circuit, just bigger, and with a output for a center-tap work coil. The larger components look to be half-watt or watt, instead of the standard quarter watt SMD parts, presumably to handle higher voltage differentials without skewing performance significantly. Do note the single choke is significantly larger but uses the same 18g awg mag wire.

These have IRF250 fets on baord. Very rugged. Should be good up to 36V input, I would think. YMMV depending on your coil skills. If you're looking to make a custom coil set to drive something like a @phattpiggie triple out of the box, I'd start here with the big 36V guy. You could make a fairly boutique / artisan coil that should work pretty nicely, if your bendy skills are on point.

http://instagr.am/p/CCJiSRSHTts/

 

[TommyDee]

I am loving these flatter builds @Edgedamage . Great stealth implementation! Now if could only hide the click. Damn thing gives me away every time.

@RedEyeFlightControl - the FET rating for these circuits has a lot to do with the S-D breakdown voltage. The little guy uses 60V FETs. The MOSFET switches use 45V FETs. I am running the 45V FETs and they are holding up nicely.

These extra allowable 'volts' are needed when coils become bigger or the frequency needs to be reduced for electrical reference depth. The capacitors are the work within the circuit. The inductors are there to 1) pass the current to the circuit through the positive leg, where the wire must be sufficient, and 2) Generate the desired frequency in the LC network.

Yesterday I learned a neat trick that explained everything LC. Put the inductor across the cap and strike it with a single power pulse of whatever voltage will read on the scope across those two leads. The circuit goes into oscillation at the frequency based on the components. This is how you can tune your coils to electrical reference depth without all the fucking calculations. Now how do you mimic a scope if you don't have one?

The FET gate doesn't draw current but the frequency is generating hard edges. The zener and resistor are there only to protect the G-D differential voltage as a shunt. If you drive the gate at full S potential, the zener should be non-functional with the exception of some clipping of noise spikes.

Didn't dawn on me the inductor was bigger but great observation, thanks. How did they do the center-tap thing?

 

[Edgedamage]

Just took the router for a test drive, works nicely. I have the vapcap sitting ever so slightly higher than my normal builds. My normal depth has a spacer three wraps from the bottom. I am about 3 1/2 wraps from the bottom.
First click 6sec, no vapour.
Second click nice taste light vapour.
Third click lots of flavour and good vapour.
Fourth click taste dropping off good vapour.
Also noticed it runs just fine using a 12v 4A power supply. Before this same unit would just flash the LED and quit on the same 4A supply.

 

[TommyDee]

I suspect your power supply is putting out a bit more than 4 amps... maybe 5-ish. I have a 65 watt brick that puts out 77.5 watts. That just makes sure they are always in spec.

Love the build. Perfect office accessory. Now I can see why one would want about, oh, a half dozen modules to 'disguise' around the house.

 

[RedEyeFlightControl]

I am loving these flatter builds @Edgedamage . Great stealth implementation! Now if could only hide the click. Damn thing gives me away every time.

@RedEyeFlightControl - the FET rating for these circuits has a lot to do with the S-D breakdown voltage. The little guy uses 60V FETs. The MOSFET switches use 45V FETs. I am running the 45V FETs and they are holding up nicely.

These extra allowable 'volts' are needed when coils become bigger or the frequency needs to be reduced for electrical reference depth. The capacitors are the work within the circuit. The inductors are there to 1) pass the current to the circuit through the positive leg, where the wire must be sufficient, and 2) Generate the desired frequency in the LC network.

Yesterday I learned a neat trick that explained everything LC. Put the inductor across the cap and strike it with a single power pulse of whatever voltage will read on the scope across those two leads. The circuit goes into oscillation at the frequency based on the components. This is how you can tune your coils to electrical reference depth without all the fucking calculations. Now how do you mimic a scope if you don't have one?

The FET gate doesn't draw current but the frequency is generating hard edges. The zener and resistor are there only to protect the G-D differential voltage as a shunt. If you drive the gate at full S potential, the zener should be non-functional with the exception of some clipping of noise spikes.

Didn't dawn on me the inductor was bigger but great observation, thanks. How did they do the center-tap thing?

@TommyDee I suggested up to 36v as max value as the board shows this, but I'm not recommending it. The IRF530 vDS is ~100v, and accounting for resonant voltage multiplication, this gives us a nice round target of 25 volts to maintain a somewhat healthy resonance with those fets, at max power for the fets to be within "tolerable" operating ranges for any length of time. Though ideally you'd want to run this one around 18V which is just shy of ~80% max of vDS/4 based on our assumed values and I think you'd find it ran nice and cool.

---
Edit - I misread these fets. They are both IRFP250N's, not 530s (smaller). These are rated for a vDS of ~200v and a max current of up to 120a, pulsed, or 30a, continuous. BEEFY transistors. They're logic level, still, with a vGS of ~2-4v. So these can potentially run much harder with minimal exertion at the gate. The limiting factor here is probably about 50V max working before you grenade something, making 36V just about 75%, a nice comfy value for these modules to run at without exceeding thermal thresholds leading to positive heat accumulation at the gates.
---

If your current fets are rated for ~45 vDS and you're hitting ~48 with the resonant factor of 12v working, they would probably see some minor to moderate heat dissipation, I would suspect. Have they gotten warm on you? My little standard 12v miniZVS doesn't break a sweat throwing ~70 watts through the 30N06's that it came with, the vDS on those is 60V IIRC, and that's running off of 12v / 6a with spikes to 7 as demanded by changes in workpiece. I would guess ~45vDS fets would be underrated for a full 120w rig, but probably sufficient with the lower thresholds of something like the HP, at least. The currents generated by the primaries should be half if my understanding is correct, so even at 12v I'd suspect impact to the fets to be significantly lower.

The toroids on these two boards don't appear to be center-tapped. That means the work coil will need to be. - That is, if my assumption is correct. There are a few sneaky ways you can do this. I state this based on the fact that there is only one winding on the primary toroid on these inductors. The Mini makes use of two individual primaries rather than a center-tapped primary, or center-tapped WC. Since there is only one winding there, my first thought is that the work coil needs to be commoned-out in the center (also supported by the evidence of a three-port work-coil blocks on the output side). One of these two kits is actually a jacob's ladder driver (came with a BIGASS coil I'm going to use to zap stuff with, hehehehe). The function is the same, I believe. They are strikingly similar despite having different labels.

That said, they will probably run for short periods of time upwards of 30-36v provided inrush and sustained current is stable, but the parts will get hot, fast, the closer you approach and exceed that 25v threshold. Likely just before it. Like any thing else, the less you beat it up, the longer it will last.

There are a bunch of inexpensive arduino based scopes/kits if you want to go cheap. I'm actually looking at getting a Pokkit to go with my cheapo (yet novel) 200khz DIY scope. I'm waiting for the Pokkit Pro, I think. My toy is probably sufficient for this kind of work for now.

 

[TommyDee]

I've already stress tested the 45v FETs. This is on a HalfPint build so it already derated to half. I made the mistake of upping the windings on the inductor and the FET were boiling. I could feel the heat through the board. I went back to the center-tapped single coil and the FETs remain cool even with the most aggressive heating, at under 80 watts. I fully appreciate that I now have a 60 watt heater as the original is indeed 120 watts. So 80 watt performance with under-rated lesser FETs is a an absolute Godsend. Those FET are are a buck a board. I got a shitload of em and don't need a one, except for spare FETs.

However, thanks for the expected resonant voltage. I am fully up on your analysis but I just don't have a scope. My VAS is screaming at me to toss out $200. Then I could see all these things... but I'm not going to tweak them anyway. This HalfPint, even hobbled with derated FETs, does the perfect bake.

I wasn't suggesting pushing input voltage; I'm still on 3S operation looking for 2S optimization. I'm done with the "12V" ZVS solutions, and there are 3
A) ZVS 120W 12V Brick at ~96 watts peek with a solid power source with regulated DC for consistent in home performance. Hot, fast, reliable, good form-factor. 3S capable with very tolerant voltage sag performance.
B) ZVS 100W 11.1V 3S high output power source. Typically runs at 60-70 watts throughout the voltage range of the cells. More sensitive user experience to voltage sag. As a corded device, it makes for a compact wand for on-demand heating.
C) HalfPint as an integrated device, it makes for a nice tabletop cordless that you can walk off with when needed. Not too clumbsy and it will drop into a coat pocket in a Flatpack configuration.

Flix and Fluxer have B) and C) covered with much more sophistication and the power level of A). Until @mr_cfromcali clones himself, people will try to build the experience before their order will arrive. I did make a nice unit while waiting for my Alpine 2020 preorder and that was only an extra month. I'm happy as PigPen using this little black box in front of me. It now has a Flowering Quince activation button.

Yes, as to the positive input, I suspected those need a center-tapped coil. I definitely don't want to do that for the IH.

 

[PKOK]

LOL @PKOK - Welcome back buddy. Show us what you got.

Learned a few things, but in general, we have the basic module simplified and the new HalfPint half circuit.
Still no progress on charge regimes. However, I am qualifying some 18350 cells that are performing nicely. The 14500 were not quite tough enough. And the 18350 rate at ~1100mah - about a day's worth for me.

Looking forward to your build.

Huh I've got some 18350's and 18500's kicking around somewhere. Well TD I'm not sure what to build now that some time has passed. Some of the minor parts I was waiting on never came, need to reorder from North American supplier.
The PCB's, BMS and Balancing Board those did come, two of each. The project box, I knew would be big, is a beast. Well now that I'm done with the house, maybe just maybe I'll
have the chance to finish one of my projects!

 

[TommyDee]

Special 10 amps 18350's. If you need MOSFET switches I have extras.

 

[TommyDee]

Yuck Had a connector failure today. One of those green-w/blk-overmold screw on type. Female side on the meter. Pushed up to eighty watts and suddenly went for run-away smoke from the 3S pack. Mind you, the pack has a 20A BMS set to disconnect @30A. I've done it twice before and it didn't hesitate. It requires cells to be removed to be reset. This time, it delivered current and stink! Lucky it was just a connector but shit, it was going for broke!

 

[TommyDee]

I decided that plain black boxes, or clear boxes are not something that has any class. I have designs for HalfPint's case but while I was playing with "The Standard" I decided to venture out on dual-color printing on the 3D filament printer. It is a messy process and the quality usually goes down as a result. Dual color is much better managed as 2 parts rather than mixed in the same print. Problem being, as with all color printers, bleed.

What is "The Standard" today you ask? Slightly modified I'm afraid. The coil is hot, 9-turn at the original ID making it ~20mm tall with closely spaced coils (unlike original). I also left the lead wires long-ish trying to maximize wire length. The reason for the 9-turn coil rather than the standard 10-turns and long lead-wires is that 9 turns is the right stack-height. A 1/10th turn has never proven useful and only helped char the vape in the tip past the CCD. Having the larger ID matching the standard coil works well for making robust cups out of dowel or other hardwood. They char on digger but the shape of the cup can minimize that. I just don't have a select glass piece that has a stop in the bottom and is 16mm OD. I prefer a 16mm ID coil. The ID of the standard coil is just over 18mm. Very forgiving in depth settings when you increase the ID. Very mellow warming pattern with a good heat soaking due to the more reduced focus of the magnetic field. The field becomes more 'homogeneous' so to speak. These more lenient builds don't target the heat so much as just 'provide heat more broadly'. Unless it is stupid-aggressive, you can cook anything DV can throw at it. It is a safe bet for an IH.

The second modification to The Standard is the gate-circuit isolation allowing for the little bobble-switch to control the full circuit. I now know I have a rapid disconnect from the power source by yanking wires as that safety aspect was tested yesterday. I needed a 10-15 amps fuse on the battery pack! The BMS was insufficient to protect lower power circuit elements. It delivered way to much current for my liking. I am still of the opinion that a power switch on a corded shouldn't be necessary. I would have jumped for the switch first if it exited and the fault would only have gotten worse in the meantime. Simple is good in some respects. Safety-wise, if there is a switch, it can be left on as the single-fault failure. A fuse must function is a FET fails closed in a "hard-wired" configuration.

This is "The Standard" performance metric with an '18 tip and cap. This is the hottest among '18, '19', '20, and Ti. The huge voltage drops has been a standard feature in my circuits since day 1. I live with it and understand the implications once the circuit is fully hardwired. This represents a hotter IH than the standard 120W mini-ZVS module due to the shorter coil wire.

Free-running current with a fresh charge will exceed 20 watts. This is not bad for this circuit. It is one of the opportunities in HalfPint as free-running current is reduced. It is what it is. Notice the classic "blue-dot effect" with the translucent red 3D print patches. That's gonna happened somewhere for sure! It even has a faint purple halo at the wooden cup.

If I had to recommend a heater to someone on a budget, I'd point them in this direction. A bit of artistry but in the end, cheap and functional. A case for this is also simple enough just about any skill someone may throw at it. And as this unit runs cool to begin with, even a hot-glue cardboard box would be functional.


Again, I am happy with this build. I don't need to 'finish' it for any particular purpose but I will be keeping an eye out for a good box or box fodder. The lack of the magnet bugged the shit out of me after having one on HalfPint. Spent almost a day getting the printer dialed as best I could. Now I know how to deal with it in the future.

The Standard is only marginally larger than its little brother, HalfPint. Both corder still makes them desktop/tabletop units. I like the freedom of a battery pack.

I have limited experience with the HalfPint circuit to know if it falls short. I look at it as a more "conservative" approach. 6-8 second heat-up times are calm where a second or two more are a matter of choice rather than a worry. A mellow targeted coil with a tuned depth for the VapCap is highly repeatable, and adjustable. I've chosen to freeze the insertion depth in a fixed position on both units. The Standard will now be the backup. It is complete until a real case comes along for it. The compactness and the sharp looks of the HalfPint make this the solid contender for my coffee table for a corded IH device. All else being equal, of course.

The Standard and The HalfPint

 

[Edgedamage]

Thought it would be fitting to have a computer power the "router".

 

[TommyDee]

Oh man, I hope the drives can handle the brown-outs!

Did you do some power readings on your unit? Something in the 70 watt range? @Edgedamage

 

[Edgedamage]

No power readings taken yet. The system I used is a back-up music server. Only one drive in it and and it runs headless, the 850 watt power suppy has enough overhead to run the IH. And the drive only acesses data when a song is playing.

 

[Edgedamage]

Ok did the test. 37 ish watts? Excuse the nasty looking thumb nail, I run wire forming and wire drawing equipment.

 

[TommyDee]

LOL I guess. Yea, you could run a bar with that PS. That's a pretty tame heater at around 40 watts. maybe 50 watts with the umph behind it at 12 volts? 50 watts is a very nice soaking rate. At 36 watts I start wondering if my cells need recharging, which they do at this point [9V].

 

[Edgedamage]

LOL I guess. Yea, you could run a bar with that PS. That's a pretty tame heater at around 40 watts. maybe 50 watts with the umph behind it at 12 volts. 50 watts is a very nice soaking rate. At 36 watts I start wondering if my cells need recharging, which they do at this point [9V].

Just tested the pack. Standing voltage=12.07v IH firing=11.07v

 

[TommyDee]

Are you happy with its performance as it is? Certainly adds to the safety factor.

 

[Edgedamage]

Are you happy with its performance as it is? Certainly adds to the safety factor.

Yes works fine, even lowered the vapcap insertion depth to my standard. Which is three loops from the bottom of the coil. And amp draw rose to 3.8 amps.

 

[TommyDee]

Nice. The whole idea of DIY is to dial exactly what you want. Thank you again for taking those readings.

This is the tuning of my HalfPint...

It might look innocent but it's as hot as all the rest.

The original Mini ZVS on the bench was running 10V/6A|60watts consistently. That is the baseline. The 9-T standard is hotter than original. HalfPint matches the original circuit in power draw with the '18 VC.

 

[Edgedamage]

The BMS you see in my pics is the reasion I pulled a sale posting on reddit. I did a couple of tests and as a result I will never sell any build with one of them in it. I took three matching cells all posting 4.02V I drained one to 3.85V. Put them in the holder and charged them using a 3s charger without balance leads. What I found was the two 4.02V cells where allowed to over charge while the one low cell kept charging. A real BMS stops charging cells as 4.20V and allows a laging cell to charge alone. I am thinking all future builds using batteries will be for personal use only. And they will have a 10A fuse with voltage display and no BMS. I will use a 4 bay external charger.

 

[RustyOldNail]

The BMS you see in my pics is the reasion I pulled a sale posting on reddit. I did a couple of tests and as a result I will never sell any build with one of them in it. I took three matching cells all posting 4.02V I drained one to 3.85V. Put them in the holder and charged them using a 3s charger without balance leads. What I found was the two 4.02V cells where allowed to over charge while the one low cell kept charging. A real BMS stops charging cells as 4.20V and allows a laging cell to charge alone. I am thinking all future builds using batteries will be for personal use only. And they will have a 10A fuse with voltage display and no BMS. I will use a 4 bay external charger.

The BMS board in my Fluxer seems to charge the 3- 18650’s properly, though I did not test using any lower voltage cells in the mix, the Sony VTC 5’s were all bought at the same time, kept at 80% charge in frig. Tested their IR, and matched the listed spec.

 

[Edgedamage]

The BMS board in my Fluxer seems to charge the 3- 18650’s properly, though I did not test using any lower voltage cells in the mix, the Sony VTC 5’s were all bought at the same time, kept at 80% charge in frig. Tested their IR, and matched the listed spec.

Married cells should be just fine, however almost every 18650 battery pack I salvaged cells from. Always had one under 2.0v cell in it. I can't speak for the fluxer BMS but the no name ones I got off ebay will not balance charge a low cell properly.

 

[RustyOldNail]

Married cells should be just fine, however almost every 18650 battery pack I salvaged cells from. Always had one under 2.0v cell in it. I can't speak for the fluxer BMS but the no name ones I got off ebay will not balance charge a low cell properly.

I’ve believe the Fluxer BMS is similar to:

Amazon.com: 2Pcs 3S 11.1V 12.6V 25A W/Balance 18650 Li ion Lithium Battery PCB Protection Board: Home Audio & Theater

Buy 2Pcs 3S 11.1V 12.6V 25A W/Balance 18650 Li ion Lithium Battery PCB Protection Board: Batteries, Chargers & Accessories - Amazon.com ✓ FREE DELIVERY possible on eligible purchases

www.amazon.com

There are various types, seem pretty inexpensive, I know the included AC charger is about $6.