I've been using a halogen oven with a home made controller as a simple reflow oven. The controller uses a thermocouple to measure and control the temperature profile.

The problem I have is that I'm getting inconsistent results depending on the component. As a test I made a simple board with 8 resisters and 8 LEDS. In this case I've put a little solder on the pad using a soldering iron, positioned the components on the pads and then re-heated in the oven. I cleaned the board with liquid flux before applying the solder and again after applying solder (In one case - I had similar results when I didn't do this second clean)

The resistors have all come out fine. They are positioned well and the solder looks good. The LEDS however all come out terrible. They all just about have proper electrical contact so the board "works" but they are very poor indeed.

The components are all about the same size and have similar contacts. There was no difference in the preparation. The LEDS are very slightly thicker I suppose. They are within 1cm on the board and I don't believe would have significantly different temperatures.

I'm wondering if anyone has any explanation or suggestions about how to improve? Maybe I need to increase the temperature? Or hold the temperature a little longer?

enter image description here enter image description here

  • 2
    \$\begingroup\$ I think it is most likely a temperature effect. I don't think the LED's are fully reflowing. Maybe due to being so close to board edge. Maybe try more heat or more time at max heat to see if the LED's self-center a little bit better. You could also try using solder paste instead of pre-soldering the pads. It will then be more evident whether proper reflow has occurred, because the solder paste will not look shiny unless it reflows. \$\endgroup\$
    – user57037
    Commented Nov 16, 2016 at 23:25
  • 3
    \$\begingroup\$ The problem is called "component shift". It looks like your pads are possibly too far apart for the LEDs. Notice how they've tried to sit so that one end of each part is in the center of one of the pads? \$\endgroup\$
    – Polynomial
    Commented Nov 16, 2016 at 23:26
  • \$\begingroup\$ I think the centers of pads are too far apart, you maybe need slightly less solder, because that much seems to lift the components, and it is obvious that solder does not stick well to the components because of lack of flux - big deal. \$\endgroup\$
    – Nazar
    Commented Nov 17, 2016 at 0:01
  • 1
    \$\begingroup\$ How about the idea: your parts have different ROHS palting/coating, and you solder is not well compatible with LED plating? As you can see, the resistors usually have have some solder build-up, while LEDs don't have it. \$\endgroup\$ Commented Nov 17, 2016 at 2:19
  • \$\begingroup\$ so.. more flux, less solder and better pad spacing. maybe a bit hotter. thanks. I'll try all of those. \$\endgroup\$ Commented Nov 17, 2016 at 7:00

1 Answer 1


If you must use solder instead of solder paste then you really need to ensure that there is plenty of flux on and around the pads when you reflow. Without the flux the solder won't flow into the components properly.

Solder contains flux in the core, but that burns off as you melt it onto the pad leaving just (almost) pure solder behind. You need flux for the solder to flow onto the metal of the component.

Solder paste is made from millions of little tiny beads of solder suspended in flux. If you're going to reflow you really should be using paste. It's not expensive (well, it is, but you use such small amounts a little goes a long way). Personally I use a nice low-temperature one (138°C melting point) with my home made "toaster oven" reflow oven (ChipQuik SMDLTLFP) in a syringe.

Here's an LED I reflowed earlier today using my homemade oven:

enter image description here

BTW, the pad layout for the LED above is exactly as specified in the datasheet for the LED. Check with the datasheet for your LEDs that the pads you are using are suitable. If they are spaced wrongly (and they look like they may be) then reflowing may be problematical at best.

It's even good for fine-pitched QFN:

enter image description here

I wouldn't even think about doing those with normal solder without having the board swimming in flux when I reflowed.

  • \$\begingroup\$ thanks. ill try solder paste again but i found it really hard to work with before. I'll also use more flux. ill admit i didn't use much on that one. \$\endgroup\$ Commented Nov 17, 2016 at 7:03
  • \$\begingroup\$ In what way do you find paste hard to work with? \$\endgroup\$
    – Majenko
    Commented Nov 17, 2016 at 9:26
  • \$\begingroup\$ I've bought several types. I found it was either too thick and I couldn't spread it properly. Or it was too runny and didn't really work. I found it often preferred to stick to the spreader than to the pad. It was also expensive in small quantities and messy. Perhaps I had bad paste or need to experiment more with the temperature of it. Perhaps i need to clean/flux the pads better. But I just found it a complete disaster every time. \$\endgroup\$ Commented Nov 17, 2016 at 9:32
  • 1
    \$\begingroup\$ If the paste is clumpy then it has gone off and needs replacing. If it is too runny then it is too warm. Runny is better than clumpy. Unless you are using a stencil then you want your paste in a syringe with a fine needle like (not a real needle, blunt) applicator. When you reflow the preheat phase will make the paste go runny and it will spread and pool anyway. Don't worry too much about it being runny. \$\endgroup\$
    – Majenko
    Commented Nov 17, 2016 at 9:35
  • \$\begingroup\$ I will certainly try solder paste again but I'm now getting good results with what I'm doing so I'm happy for now :) \$\endgroup\$ Commented Nov 18, 2016 at 13:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.