Building a Reflow Oven | Hacker News Reader

Building a Reflow Oven

whizoo.com

72 pointsmindcrime38 comments

This is cool! I can buy a $30 Black & Decker toaster oven and turn it into a reflow oven.

There are a few worrisome steps, like this section:

> Ceramic fibers are very fine and sharp. They don't penetrate though thick skin, but can irritate thin tissue like eyes, nostrils and lungs. If you exercise precautions you won't have any problems. The blanket has amazing insulating properties, and the top of your oven will be barely warm during a reflow.

> Things to keep in mind:

• Unwrap the ceramic fiber blanket just before you use it - not earlier.

• Clean up your workspace immediately after applying the blanket and closing the oven

• If possible, wear a face mask (we recommend using N95 masks)

• Wear shoes. Our preference is flip-flops that we can easily wash afterwards

• Take care not to touch your eyes, lips or ears

• Keep the area free of drinks, and use the restroom (toilet) before starting

• Work in a draft-free environment. Don't use fans

• Make slow movements to minimize dust

• Make sure you have all your tools and parts close by (you need to close up the oven too)

• Make sure pets, kids and neglected spouses don't come into your work area

• Immediately after working with the blanket you should take a shower and wash your clothing

> Scared? Don't be. Ceramic fiber is easy to work with as long as you take a few precautions.

OK! I am not scared at all.

I do have to wonder about the cost effectiveness, at least for myself. It sounds like a solid couple of days work to make this thing.

Suppose I were looking for a ready-to-go reflow oven at some kind of reasonable price. Where would I start? Any recommendations?

I don't think the ceramic fiber blanket is necessary. I made my own reflow oven (PID temperature controlled) and was surprised to find that my roommate was able to reflow a PCB with his completely unmodified toaster, just popping the PCB in and bringing it up to the right temperature. Obviously it didn't follow the temperature curve like mine did, but the end result wasn't really different, at least for our hobbyist uses.

Suppose I were looking for a ready-to-go reflow oven at some kind of reasonable price. Where would I start? Any recommendations?

Here's the rub... there are relatively inexpensive, COTS reflow ovens out there[1], and they do work - more or less. The problem is that many people find them to be more on the "less" side of that spectrum, without putting considerable work into upgrades and mods anyway. And once you take a cheap T962 off the shelf and start modding it, you might start to ask "why not just build my own from a toaster oven?"

The toaster oven conversion winds up being cheaper, or - at worst - a wash, from a price standpoint. And many people find that they get better results from the converted ovens than they do from a T962.

Unfortunately, the next "hop" up in quality in terms of commercial ovens winds up being a pretty big jump price wise.

I haven't tried this myself yet, but I have a friend who has both a T962 and a conversion oven (made from a Black and Decker convection oven, so a little pricier than the model cited in TFA), and he only actually uses his converted oven. I trust his opinion a lot, so if/when I decide to get a reflow oven, I'm leaning hard towards doing this.

[1]: https://www.amazon.com/Happybuy-Soldering-Machine-Infrared-A...

I've also had less screwups using the regular toaster oven instead of T962. I suspect that's because T962 uses IR heaters instead of convection.

I've regularly had slightly charred or even molten connectors on boards reflown with T962. The most extreme case was broken inner layer traces (on a board with black soldermask, which apparently had a bit too much moisture inside- the board layers visibly popped).

An interesting side note- large LEDs don't like being reflown in IR ovens. The lens on large LEDs also works in reverse- it happily damages the bond wires by focusing incoming IR radiation to a point.

I've also had less screwups using the regular toaster oven instead of T962. I suspect that's because T962 uses IR heaters instead of convection.

I suspect so. Especially if one follows the advice of adding an extra heating element to their converted oven. It's not just about having more raw heating power, it's about having the heat coming from multiple places, so you get more even heating.

I have one of those (T-962) at home, and I am pretty happy with it.

I keep thinking about buying one, but for now I've convinced myself that the amount of SMD work I do is small enough that I can rely on either hand soldering under magnification, or using my hot air rework station.

Most of what I do prototype wise is still through-hole, or the larger SMT packages (SOIC, etc.)... the only reason I'd use one of the smaller packages would be if the component I need is only available in that package. Unfortunately, that's becoming more and more the case.

Anyway, if I need a component that's a real fine pitch QFP or something, I'd just hot air solder it to a breakout board, and then jumper it in... unless it's something where that little bit of extra capacitance would screw something up, or where EMI would be an issue. But I'm a hobbyist, so that stuff isn't usually an issue here.

It's extremely irreponsible of this author to not list a respirator as a requirement for handling ceramic fiber, given that the audience is likely to be unfamiliar with the materials.

That "if possible" is terrible advice. If you're working with a ceramic fiber blanket, always wear a respirator. And the fibers will continue to hang in the air afterward, just like asbestos.

High temperature turns the fibers from irritants into carcinogens like asbestos. The suggestion to build one of these things with an uncoated ceramic fiber blanket is horribly negligent.

High temperature turns the fibers from irritants into carcinogens like asbestos.

Well, high temperature in this case being above 1800 F/1000 C, over 4 times the operating temperature of these ovens.

I built one of these about a year ago. I spent about 5 hours on it, over a little more than 2 days (some things need to dry or set up before going to the next step).

To reflow a small board (or 3) on an extreme budget, use a frying pan and an infrared thermometer. Not super accurate, but if your board is on a piece of foil you can pull it out before it overheats so your components don't cook.

I handled the ceramic fiber part by working in a room with a powerful vent fan that exits far from people, taping my cuffs to my gloves, and staying upwind of the insulation. I cleaned the room as soon as I was done. It only took a few minutes and the top cover goes on as soon as you get the insulation in place.

You can buy pre-built Controleo3 ovens, but you can probably get a used industrial oven for about the same price as the prebuilt kit. I guess it depends on what you consider a reasonable price, and whether you know people who like building things from a box of parts.

I've worked with fiberglass insulation before, is this stuff significantly worse?

I thought the worst thing that could happen to you with fiberglass would be to inhale the fibers. Isn't a mask and gloves enough?

> This is cool! I can buy a $30 Black & Decker toaster oven and turn it into a reflow oven

$20 at Bed Bath & Beyond [1].

[1] https://www.bedbathandbeyond.com/store/product/black-decker-...

Maybe $16 with one of my lifetime supply of 20% off coupons!

This gives me an excuse to share the Three Secrets of Bed Bath & Beyond in-store coupons:

1. Even though the coupons have an expiration date printed on them, they do not expire. Keep your old ones and use them when you want.

2. You can use more than one coupon in a single visit. You can't stack them on one item, but if you have N coupons you can use them on N items.

3. If you forget to bring a coupon to the store, you can come back within 30 days with your receipt and a coupon and they will refund the discount amount.

None of this applies to online purchases, only in-store.

Seems like it would be physically quite similar to asbestos. I wonder if it's comparable.

AliExpress expect to pay maybe US$140ish plus shipping

Finding a good reflow oven is frustrating. The Puhui T962 series units, available since 2014, were around $150, but they had wildly nonuniform temperature inside. I've had boards scorched at the center and failure to solder properly at the edges. There's an open source mod to those which makes them reasonably usable. New firmware that runs the cooling fan a bit during heating to distribute the hot air, and a proper cold end reference for the thermocouples.[1] With those mods, those things are usable. These things have been cloned in China; you'll see "T962" units that don't have a Puhui logo on Amazon and eBay. None of the cloners seem to have made these well-known improvements and delivered a good out of the box experience.

The next step up costs much more. A nice little prototyping convection oven from Germany costs about $3500. There's not much in between the $200 units and the $2000 units.

The article is really an ad for Whizoo's Controleo3 board, which is $249 with all the parts you need. So this isn't that great a deal.

Most of the cheap solid-state relays on eBay and Amazon are fake. The current ratings on the unit are far beyond what the electronics inside can handle. Google "Counterfeit Fotek SSR" for details.[2] Since the failure modes are 1) fail in the ON state and 2) catch fire, this is not good for anything that draws significant current. The real ones are a lot more expensive than the fakes, and cost goes up with current rating.

[1] https://hackaday.com/2014/11/27/improving-the-t-962-reflow-o... [2] https://www.ul.com/node/81246

https://www.estechnical.co.uk/ has a very nice modification kit for the T962 that makes it quite usable, as well as controllable via USB.

Unfortunately it's about as expensive as the oven itself, but the end result works very well and no longer torches boards.

An alternative controller is the Reflow Master by the Unexpected Maker:

https://www.tindie.com/products/seonr/reflow-master/

US$89 on Tindie (just add Toaster+SSR!)

It's been used to reflow _thousands_ of boards including the majority of the TinyPICO's:

https://www.crowdsupply.com/unexpected-maker/tinypico/

I also designed and built a reflow oven! My goal was to make it simpler / smaller: https://nathan.vertile.com/blog/2019/04/24/fluxlamp-the-sold...

Email me if anyone wants control boards and I can try to make some more.

I like the light housing based design, pretty clever!

Hrrmmm... I like SMT. It's cheap, easy, and performs well. I haven't tried the ceramic paper, but I've only tried single sided reflow, and it worked fine without. I'd also worry a bit about fumes during reflow so doing it under a vent or hood (they recommend no drafts?)

The challenge I saw was with silk screening the solder paste with the right alignment and thickness. Placement with tweezers and a binocular microscope (or young eyes) went fine once you do a couple. Post reflow rework substantially increased yield from ~80-90% to 99%.

Make sure you have a test for those parts before you put then all together wrong!

I recently did a smd reflow using a cast iron pan and a few handfuls of sand. Worked like a charm.

Got the idea here: https://hackaday.com/2014/02/15/smd-soldering-on-hot-sand/

Granted that was for fairly large, hearty components, but for many hobbyists it seems to me an oven is a bit overkill.

Does your method work for QFN and DFN? I currently find them to be the biggest headache in prototyping. If you coat some solder on the chip and board using an iron and then reflow with a hot air station, it's easy to miss a few pins underneath due to uneven solder distribution; on the other hand, if you use solder paste and reflow with hot air, again, due to uneven solder distribution, you'll short some pins and miss others, and there is no straightforward way to fix it, the thing gets even worse when solder on the ground pad suddenly melts and causes unwanted movements (during rework) that misaligns the chip. I got my board working only by pure luck. I'd like to hear your recommendation.

QFN/DFN are pain, but I've had decent luck with them using the toaster oven method. Of course that's using a solder mask from JLC PCB, as without the mask you just end up with the incorrect amount of solder. Still you get the occasional missed pin you need to re-solder but it works.

> Still you get the occasional missed pin you need to re-solder but it works.

How? Did you mean to remove all the existing solder, reapply the paste, and reflow again (can that even be done on an already populated board, where do you place the mask/stencil?), or it's something else?

More of just putting a bead of solder on a soldering iron tip and fixing that pin (best with a microscope setup). If you get the technique just so, you can do a form of "drag soldering" to fix a missing pin or two. Otherwise you can do a quick dab with a bead of solder since with most QFN's you can still see the side of the pin but you may need to rotate the board. I think there are a bunch of great youtube videos on doing that.

Getting a proper solders for QFN's really does best with a proper metal stencil. If you use JCL PCB, use their "electropolishing" option otherwise on fine pins the paste sticks more.

Removing and reflowing the part is tricky and often damages the part. Also to not ruin other parts of the board, a heat gun is often easier. We've never really tried reflowing parts unless a board didn't "bake" properly.

Confirm you meant to say stencil rather than solder mask there? (Solder mask is the colored material over the board, while a stencil is a separate piece of metal with holes cut to allow solder paste application with a squeegee.)

> Confirm you meant to say stencil rather than solder mask there?

Yes, stencil is what I meant to say. The "mask" part of solder mask seems to throw me everytime I mean to say stencil, as in I want to mask the parts of the board _not_ to solder. Someways I wish English were more like Swiss/German or Hebrew.

All the ic’s on the board I used this technique on had leads, so I can’t comment on how dfn/qfn parts would work. Pan+sand would definitely be the first thing I try tho

Thanks for the confirmation anyway!

I would not follow the tip about connecting multiple thermocouples in parallel. The voltages involved are very small and the Maxim chips are very sensitive to noise.

Source: myself, as I had to put in a diff 10nF capacitor between T+/T- of the 31855 to get good consistent readings. Maxim confirmed this is a known solution and they were supposed to put the recommendation in the datasheet.

So, be careful with those thermocouples.

When I was a student, pretty much every student organization that did PCBs had one of these (FSAE Electric, Robotics clubs, etc.)

This + boards from OSHPark or Advanced Circuits' 33/66 each deal were a great way to cheaply get started with PCB design.

I've got a Zallus reflow controller off an old kickstarter - finally installed it this weekend into an old oven but haven't had a chance to try it properly. They seem quite a bit cheaper than the whizoo controller although it's much more limited. Mine doesn't control the door or heating elements independently.

Apparently a popular use case for these is roasting your own coffee beans.

http://www.zallus.com/

Instead of that nasty ceramic blanket and that reflective tape why not just use ceramic tile? Its cheap, easy to find locally, and will work better. True, it will be heavier, but it would be easier to maintain a consistent temperature. Another alternative would be the ceramic cement products like FiberFrax. It dries solid, so no floating fibers to be concerned with.

Puhui T-962A @ $250 does the job if you don't want to DIY.

I have both the Puhui T-962A and DIY conversion oven. The Puhui T962A will consistently burn your board. You solder joint will come out black as the temperature goes way beyond proper lead-free temp. You will never be able to solder any film capacitor. Any film cap you put into T962A will come out melted/exploded. It actually save your time to build you own oven then by the off-the-shelf T962.

Works fine for us! Haven't had any issues to date, must have done over 100 different boards. There is third party firmware[0] for better control including an active fan if you run in to issues. Thanks for the tip though.

[0] https://github.com/UnifiedEngineering/T-962-improvements

Reply on news.ycombinator.com