Reflow oven idea, need a second opinion (or more)

[jeremy]

I've got a 2kw toaster oven which heats up too slowly for decent reflowing, and I don't think adding insulation will improve it enough. Is it reasonable to buy a smaller oven (say, 800w-1kw) and swap out the electronics/heaters with the 2kw ones? Of course, I intend to put a PID controller on the heaters.

[jeremy]

Yes, it is too slow. I get around 1C per second best case, but the oven has a lot of volume so that makes sense.

[Falcon69]

I bought a $34 1500W B&D one from walmart, home of the late night weirdo shoppers.

It has 2 elements on top, 2 elements on bottom.

Using an SSR and a PID controller from Auburn Instruments, it works fairly well.

your 2k watt should work, how many elements does it have?  What PID are you using?

[jeremy]

3roomlab: for some jobs yes, but I'd like to fit the profile a lot better as I want to start playing with bgas. 2-3C per sec is what I am aiming for. Since I won't be getting an X-ray, I'd like to give the soldering the best possible chance 

Falcon69: unfortunately, I think mine is something like 25 or 29L, so I think that is the problem. It has 4 elements; 2 top and 2 bottom. It works quite well at low temps, but it is very slow above 150C.

I don't have a pid controller on it yet, this was just tested by turning it to full blast. I was planning on using a 40A ssr I have lying around and a custom controller based on an arduino.

[mjkuwp]

I used a relatively small oven that came with about 1200W of power from 4 elements.  I disassembled the elements and shortened the heating wire.  Just had to cut it off shorter and stretch, then crimp a new connector.  It has 1700W now.. I was aiming higher but this is how it turned out.  got that idea here:

http://www.buildlog.net/forum/viewtopic.php?f=2&t=1013

How are you getting feedback of temperature?  I think this is the hardest part and I am still working to improve mine.  For my first reflows I soldered the sensor directly to the target board.  Worked perfectly of course but that is unreasonable for the long haul.  Second I tried using a bare board next to my target board with the sensor on a bare board.  That works but the bare board gets cooked over and over and over again.  I am not sure that is ok and it also just doesn't look elegant.

One feature I have is that I blackened a sheet of copper and use this to help gather the energy and also to distribute the heat.

http://themzlab.tumblr.com/tarnishing

This can only work if you do totally surface mount parts and also only one side.

I have not taken a photo yet but the most recent change is that I removed the wire shelf and installed a sheet of glass as the shelf.  The local hardware store (USA) will cut glass to the size you want.  The heat is just light so it goes right through the glass and now i don't have to worry about the metal grounding the thermocouple.

My next task is to find a good way to solder the thermocouple directly to the copper sheet.  I have been using extra-high temp solder (obviously so reflowing doesn't melt it).  of course.. it is very hard to solder to a giant copper sheet

Mine will bring the board up at about 1.3C per second under control and I think that is fast enough.

[wazzokk]

Would it be practical to raise the lower element to reduce the volume of the oven.
How much heat escapes around door etc. Some ventilation is probably desirable.
Just ideas from experience building glass kilns.
Regards Dave

[mjkuwp]

....Some ventilation is probably desirable.
...
Regards Dave

oh, speaking of ventilation.  At the peak temp desired, I open the oven door and aim a pretty big fan at the board.  after it reaches solid - just a few moments I remove the board and leave the fan blowing on the oven to cool it down.

I learned the hard way not to leave bare test boards unattended in an oven cycle.  Even though the oven cycles off (and I have a redundant off switch in the mechanical timer) the thermal inertia of the elements once over-cooked a bare board.  This made the most nasty smell come out of the board because who knows how hot it got but... way too high.  That smell is still in my basement even though I ventilated for a long time with two window fans. 

Now I have a vent tube to the outside and a fan to push air.  I am going to  put fans in the other windows to push air in, making the basement positive pressure to help my vent fan and make sure fresh air comes in from the outside.

For small boards on rare occasions this process does not seem very bad but please proceed with caution in this area.  best to do this outside or in a garage that can be opened up fully.  Closed basements are not good for this kind of work.

[Wim_L]

I have not taken a photo yet but the most recent change is that I removed the wire shelf and installed a sheet of glass as the shelf.  The local hardware store (USA) will cut glass to the size you want.  The heat is just light so it goes right through the glass and now i don't have to worry about the metal grounding the thermocouple.

Glass is not transparent to mid-infrared. What you're seeing is likely the glass getting heated as it absorbs the infrared, and then acting as a heatspreader, similar to what a blackened copper sheet would do.

[mjkuwp]

...
Glass is not transparent to mid-infrared. ...

oh! ?  This is a pretty new addition so I have not done much...maybe just one board this way.   Any idea where to get data on this?  I also have a copper sheet in there and there are also top and bottom heater elements.  Maybe I need to do an experiment with this.  for sure the copper spreads the energy around really well 

There is also this project - they use some reflective tape but I don't know if there is evidence of what it does for the oven.

http://whizoo.com/reflowoven

[jeremy]

...
Glass is not transparent to mid-infrared. ...

oh! ?  This is a pretty new addition so I have not done much...maybe just one board this way.   Any idea where to get data on this?

Here you go: http://www.crystran.co.uk/optical-materials/silica-glass-sio2

Thanks all for the suggestions so far. I think I will try to modify this oven by moving the elements first.

[bench_knob]

Quote from: jeremy on Yesterday at 09:15:02 AM

I've got a 2kw toaster oven which heats up too slowly for decent reflowing, and I don't think adding insulation will improve it enough. Is it reasonable to buy a smaller oven (say, 800w-1kw) and swap out the electronics/heaters with the 2kw ones? Of course, I intend to put a PID controller on the heaters.

I constructed a successfully operating reflow oven. I looked carefully around for a unit with low thermal mass heaters. Sounds to mee that you bought a unit that uses iron wire heaters, which exhibit considerable thermal latency. The oven I bought was a Black & Decker InfraWave 1200 Watt oven with three 'crystal' heaters. These exhibit much less thermal latency.

I removed all the electronics from the oven and rewired the three heaters in parallel. I brought out the heater wiring to the power-plug which I plug into a 30Amp duplex AC socket that is wired to an 120VAC (rated at 240VAC 50 Amps) SSR relay. I modified the duplex socket by breaking the brass colored dual socket parallel-link bridge. One socket has uninterrupted AC on it all the time, while the other socket is isolated (via the broken link) and the AC to it is switched by the SSR. The duplex socket, the SSR relay, a fuse holder and the PIC PID controller are all mounted inside a dual duplex metal square junction-box (J-Box). I mounted a miniature toggle switch, a power-on LED and a thermo-couple connector on a small piece of single-sided PCB board affixed to the other duplex socket cutouts on the dual J-Box.  The reflow-oven power cord plugs into the switched AC socket, while the PID controller wall-wart plugs into the unswitched socket. The J-Box power-cord plugs into an AC mains wall-socket and the associated wiring to the fuse box is rated for 50 Amp operation.

The PID controller was a commercial product which worked exactly three times and then failed. When I attempted to contact the company for warranty repair, I discovered that they had disappeared. In fact the company had been OOB for several years. I reverse engineered the controller and drew a schematic and in so doing discovered that one of the USB data lines had opened in an interior layer. Bridging the open-trace fixed the controller and to boot, I have full schematics, along with the controller firmware, controller source code (provided by another fellow) and the installable Windows XP PID reflow-oven control software.  All of this information was posted on the AVR Freaks website in 2012. The controller company was named "Silicon Horizon". You can search on that name over on the AVR Freaks website to locate those files, should you desire to make your own PID controller that was designed specifically to control home-brew reflow ovens. The PID controller, in my opinion, is fairly simple to fabricate and should be easy to program using an el-cheapo Olimex programmer (~$25). 

My oven, I discovered leaked too much heat causing it to not reach full heat and-or did not accellerate fast enough. To remedy this I acquired a few sheets of high-temperature non-flammable insulating material which I stuffed inside the interior cavities between inner and outer walls of the oven. Hah! After which, the oven didn't leak enough heat to enable PID control or even oven cool-down, so I punched a hole on the left side of the oven and mounted a 12 Vdc blower-fan which is also controlled by one of the PID controller outputs. I routed the thermocouple cable through the rear walls of the oven, using a ceramic sensor tube pushed through the walls. The tube was glommed from a BBQ igniter kit. The oven works perfect now. 

bench_knob
ps, these are just some tips and a few recommendations, use at your own risk.

[richard.cs]

A cheap and cheerful power boost for mains heating elements (up to double) is to feed them from a bridge rectifier and capacitor. With a sufficiently large capacitor you get sqrt(2) more rms volts and therefore double power, with a smaller capacitor you get a smaller boost. I've used this to halve the boiling time of an electric kettle* but for air-cooled elements you might want to be a bit more cautious.

The input current becomes a bit (well, very) peaky so be careful with control circuitry but it should be ok on a sufficiently rated SSR doing cycle-by-cycle control. Do not use with a controller that does trailling edge phase angle control! Also beware capacitor ripple current ratings.



*It's good but wiring between phases for tripple-power is better 

[Jeroen3]

The input current becomes a bit (well, very) peaky...

Are the emissions, apart from heat, from elements of any concern to the components?
A colleague once mentioned to only get ceramic heating elements because of their low magnetic field emissions. Unsure about more details, never put effort in getting one.
However, here it looks like all kinds of elements are used, and there is little mention of harmful EMI.

[richard.cs]

For anything other than induction heating (not used in reflow to my knowledge) I seriously doubt there will be any problems with emissions. I can't see element choice affecting that and would be inclined to ignore it, and perhaps instead go for elements with a short thermal time constant. For the capacitor boost circuit the current is spiky in that it contains lots of harmonics of 50 Hz but there will be very little energy above a kilohertz.

[richard.cs]

It's the same power envelope until you add the capacitor, then with a sufficiently large capacitor you get d.c. equal to the a.c. peak voltage and your power doubles. Smaller capacitors give you a lot of ripple but with an rms in between the two. The extra power comes from drawing the input current in big gulps.

[SeanB]

The heaters will work with 300VDC, as they are required to withstand 500VAC during testing. Best heaters probably would be to use wound elements in a quartz glass sleeve, as those are low thermal mass, and will heat up fast. You probably could buy spares and add them into the over as well.

[mrpackethead]

I used a kambrook oven that had two sets of elements,  a grill element + a bake set.    By themselves, they were only 1500W, but together, i get 3000W, and i get around 4C/ up to about 180C, and then it slows down a little.    It was wired up so you could use the bake or the grill but never the two together.

I ripped everything out the of the oven, apart from the elements, and rewired. it, and put all the elements in parallel.

About to upgrade to a new fang-dangled oven, but i'll never chuck out 'toasty'.. She's been a good work-horse.

[mjkuwp]

It's the same power envelope until you add the capacitor, then with a sufficiently large capacitor you get d.c. equal to the a.c. peak voltage and your power doubles. Smaller capacitors give you a lot of ripple but with an rms in between the two. The extra power comes from drawing the input current in big gulps.

This is really cool but the cost for this much capacitor is pretty huge, right?  That is on top of the cost of the bridge rectifier sufficient to pass this much power.  It does seem like a nice system design because then you can also control with simple mosfet drive instead of SSR.  Maybe this is still a 'win' when one considers the labor of doing other solutions such as trimming the elements for power or adding additional elements to the oven.

[SeanB]

A half dozen old scrap pick up PC power supplies and you get more than enough capacitor for essentially free. As well the input socket, some wire to wire them up, a decent bridge rectifier ( or at least the space for one), some RFI filtering and a case to place them in.

[richard.cs]

I've always considered it as "cheap, ugly but effective" and used it for things like faster boiling kettles or a "boost mode" for using a soldering iron outside in the wind. Keeping the bridge in permanantly and switching the capacitor gives a choice between normal and boost. By having several sizes of capacitors there is scope for multiple power settings. An auto transformer is more elegant but heavier and more expensive.

Some common methods for mains pwm:
1) rectify, don't smooth, chop with 1 mosfet. Load gets pulating d.c. with pwm superimposed.
2) use two mosfets connected source-source, floating gate driver.  Load gets a.c. with pwm.
3) use a single mosfet inside a bridge rectifier. Works the same as 2) but has higher voltage drop when on. Was more common before mosfets got so cheap.