Another DIY Reflow oven project - Help needed!

[huseyinozsut]

Hi, this is my first post on this forum.

For a few weeks I am researching for small scale reflow ovens. My budget is tight. T-762 is around 200$ and the max. money I can give is around 200$. But here is the thing: I live in Turkey!  Our customs office will consider this product as "commercial" and I will need to hire a customs counselor and pay warehouse expenses also. Which will double the cost of T-762.

So, I am planning to make my own reflow oven. But I have another problem... In Turkey we don't have toaster ovens!  The smallest ovens here are around 30 liters, with metal heating elements on them. My plan is to buy one of these, find a non flamabble insulation thing (like rockwool) and reduce the inside volume, then use custom length heaters on both bottom and top. Then I want to use raspberry pi to pid control the oven. I don't need any screens, buttons, etc. Just emergy stop probably, which is easy. The rest can be controlled through a pc or mobile phone with a browser. For Raspberry, I found this:
https://github.com/apollo-ng/picoReflow

I found a infrared heater repair shop. They can cut the infrared glass in any given length and with resistant inside, it costs 7.5 Turkish Liras, which makes 1.5 dollars!  Thin IR heaters are two type here. 400w and 600w. My plan is to use 3 or 4 400w heater at bottom, 3 or 4 400w heater at top. So my plan is to have an oven, which is 2400w or 3200w. I want to reduce cold/hot spots as much as I can and increasing the number of heaters would help I guess... Bu also, increased heater number means increased thermal mass, which can cause more overshoot. Can overshoot be reduced with fine tuning? What do you say?

Also I will probably make my own reflector for heaters (by salvaging ir heaters) and L shaped aluminium covers to reduce the IR light directly hitting the components. That is a must, isn't it? Also, I want to use very thin aluminum sheet both top and bottom of the pcb area. They will be smaller than the tray itself, but bigger than the cards I want to populate.

What do you say? If I don't electrocute myself (220V would hurt alot!?), do I have flaws with my project? What should I consider? Do you have any recommendations?

[Peabody]

I've never done a reflow oven, but I am currently working on a reflow hotplate for smaller PC boards.  My experience with the hotplate is that the more thermal mass, the slower the response, and while you can prevent overshoot by cutting power early, nothing can give you the rapid response needed to follow the reflow temperature protocol.  In the case of my hotplate, I run it at full power for about 90 seconds, but then have to wait for two minutes with the power off while the temperature of the board continues to rise into the soak phase, but not overshoot.  Then I give it another 30-40 seconds of power to begin the reflow stage, but actual reflow only begins later after the power is off, but the temp continues to rise.  Then of course cool-down is far too slow, so I have to remove the board from the hotplate.  So there's no way I can follow the factory profile.  Either I take too long, or I overshoot.  I suspect taking too long is less harmful.

I think the key to your project may be using quartz heating elements.  They heat up and cool down rapidly. As for PID control, I've asked about that here and have been assured that PID controllers can be tuned to work with even the very slow response of my hotplate setup.  But so far I've just relied on trial and error.

You didn't mention anything about a thermocouple, but you will need one for the PID controller.  For my hotplate, so far I just run the fixed power timing sequence with no feedback, so no thermocouple is needed.  And since I'm just switching the full power on and off twice, I don't need a solid state relay.

If you have any interest in the hotplate idea, I can show you what I've done.  Of course it is a completely different way of heating the board, but appears to be very forgiving.  But iI don't think it will do large boards because of the center hole.

[IconicPCB]

PID will not work satisfactorily... You will find PD may be the only elements in play.

A better more stable algorithm is the so called TAKE BACK  HALF  approach.

[huseyinozsut]

Thank you for your responses. I found several quartz heater (medium wave) producers here in Turkey. But also, some of them also produces short wave heaters. As I understand, short wave heaters heats up much faster than quartz heaters (they are also called "halogen heaters" here). According to the information in their sites, halogen heaters heat up in 1 sec! And quartz heaters heat up in 30 sec. So, it can be even better alternative, if it is not way too expensive.

But the thing I don't know is, what would be the surface heating difference? I will be using some black plastic header pins for my production and they need to withstand the heat. And heating with light emitting heater means higher temperatures on darker parts...  As I understand, short wave heaters emits much more light... I found also a quartz heater that is called "fast quartz heater". I will learn the difference of this product tomorrow...

I also found ready made heater group, 1000w at total. One on top, one on bottom would solve my problem (if they are heating/cooling fast enough). It has 12 quartz heater element in one case. This would definitely reduce hot/cold spots:
https://www.e-bym.com/quartz_kaset_rezistans_245120mm_1000w

---------------------------------

For the electronics part, I will try to use ready made picoReflow software for Raspberry Pi. The hardware is standart. I will use SSR (with heatsink), MAX31855K module and a high heat resistant, fast heating/cooling termocoupler. If the software works ok, than I should be fine. I'm not good at programming.  All I can do is to use ready made softwares and tweak them. So, I won't write the PID (or in this case PD) code.

Do you have any suggestions for me to look at?

[jmelson]

I modified a GE toaster oven over a decade ago to reflow PC boards.  I got a "ramp and soak" temperature controller on eBay for about $75, they usually are pretty expensive new.  The ramp and soak feature adds programmable steps in the temperature, such as "start at current temp, ramp to 180 C in 3 minutes", "hold at 180 C for one minute", etc.
I hooked this to a solid state relay wired in series with the oven's thermostat.  I got some miniature thermocouple wire on eBay for a steal.  I poke the thermocouple into a plated-through hole in the board, close the oven door and start the controller.  Because the thermocouple is reading actual board temperature, the reflow profile is pretty well-behaved.  (I did try it with the thermocouple in the air, and it was a disaster - the board got really singed!)

This oven uses metal-clad heating elements.  With the on/off cycles of the controller, they just become visibly red at the peak reflow temp.  I have no idea whether quartz heaters will be better or worse.  But, I do know that the black chips seem to absorb heat more than the green PC board.

I've done close to 2000 boards with this rig so far, it works great!

Jon