17
\$\begingroup\$

I have read many website and forums about DIY reflow ovens for SMT soldering. I have also seen many solder profiles as specified by the solder manufacturers, component manufacturers, and other self-proclaimed experts.

I am having trouble understanding what is the best way to control the temperature. Unless I am mistaken, all of the recommended profiles that I have seen indicate the profile that the solder should undergo. But you cannot monitor the temperature of easily unless you have an infra-red camera which is difficult to obtain on a budget. All of the DIY projects that I have read about, including the nice pre-made controller from sparkfun, utilize a simple thermocouple to monitor the air temperature.

In my own reflow oven I soldered a thermocouple to a board and compared the board temp to a second thermocouple monitoring the air temperature. The two profiles were very different, as expected. The temperature of the board and the solder is going to vary based on many factors, including board size, which change the heat capacity of the board. Everyone is trying so hard to follow a specific profile as close as possible but if your feedback temperature is bogus then your controller is useless, right?

I have thought about putting a small piece of glass inside my reflow oven and attaching a thermocouple to the glass and using that to monitor the temperature because glass has a specific heat capacity very similar to that of FR4. But it would still not be perfect for every board of varying sizes. So what is the best approach to monitor the temperature?

\$\endgroup\$
10
\$\begingroup\$

If you are a manufacturer who assembles lots of PCBs for a paying customer, it is good economics to get the solder temperature profile exactly right, to reduce the incidence of tombstoning and other soldering defects.

On the other hand, if you are a hobbyist cooking boards one at a time in a toaster oven, then achieving the perfect solder profile is a waste of time. The best hobbyist profile is:

  • Heat on high until the solder melts
  • Shut off the oven and crack the door open until the solder freezes
  • One or two passives will have tombstoned due to uneven heating. Use a soldering iron to rework them.

I realize that some hobbyists build elaborate temperature controllers, but this is because hobbyists enjoy building things, not because it is necessary for a reflow process.

\$\endgroup\$
6
\$\begingroup\$

As long as the rate of rise is reasonable (1-2 degC/sec) , the temperature of the board will be fairly uniform, and is what will determine solder melting, so if you measure temp, board temp is better than air temp. Board temp will depend on a mix of air temp and radiated heat absorbtion - you don't want the latter to be excessive as this can cause uneven heating or scorching depending on component IR absorbtion characteristics - I usually run heaters at the point just below when they start to visibly glow, which seems to work OK. You should of course avoid using leadfree in a toaster as there is less headroom between soldering and burning.

\$\endgroup\$
4
\$\begingroup\$

Component size, density, ground planes and placement in the oven will all greatly affect the temperature at a particular point, so putting a thermocouple on glass won't tell you much more than putting one in air. However, for hobbyist applications, you may not need the precision. Although an airborne thermocouple won't be reading the exact temperature of a particular joint, it should still allow your controller to recreate a consistent profile on a run of boards.

If you want more assurance that the profile is correct, you should monitor several points with thermocouples that are attached with thermally conductive paste or epoxy. (Omega sells this).

Typically, larger components will reflow last, as well as ones closer to the front glass of the oven where it is colder. Try to monitor the hottest and coldest points. Manufacturers often use multiple thermocouples while they're getting the process figured out.

I agree with other posters that the simple approach of turning on the oven, waiting for the last joint to reflow, and then opening the door is effective. Many ovens can't warm up or cool down fast enough to exceed the maximum temperature rates specified on component profiles, so the feedback loop supplied by the thermocouple may only be telling the controller to turn on and then off anyway.

\$\endgroup\$
3
\$\begingroup\$

I've used a rework station which had an under-board and above-board preheater, and a spring-loaded thermocouple to measure the PCB's temperature. When the board was properly preheated it would raise the temperature for the actual reflow phase. The difference with an oven is that heating was through radiation instead of convection. The convection heating of an oven is too slow to even approximate the prescribed profile. You need radiation to change temperature as fast as reflow prescribes. The only thing I can think of that comes close is use two ovens, one for the preheating and a second one for the soldering. And you would have to move fast from one oven to the next.
Air temperature is totally irrelevant, IMO.

\$\endgroup\$
0
\$\begingroup\$

I also played with a toaster reflow oven and I realised that the airtemp sits at a cosy 150 deg Celsius and then the silkscreen burns (it turns brown) and my capacitors are not that effective anymore after the reflow so I stuck about 5 K-type thermocouples in the oven and started playing to get better control in the end i found that the thermocouples had a big profile on the measuring end and that when i turn on the IR it takes about a minute to go from 40 to 80 deg Celsius but the element had already evaporated a small bit of water that I placed in the oven while testing that means the Temp was already over 100 degrees and I am still moving up to 80 degrees on my thermocouple

i bought other thermocouples that has the bare wires and it picks up faster now also Itake my readings in the air and directly on the biggest component as well as underneath the board but right against the board then I work out a kind of an average where the temp at the component is a bigger factor than the rest and this works for me

my silkscreen doesn't turn brown anymore and a hardly get tombstones (it still happens)

One more thing because I am using an Infra red heater I placed a steel mesh in the oven to avoid having direct heat on to the pcb this slowed it down a bit but I got better results

\$\endgroup\$
-1
\$\begingroup\$

I have an old Sears 4-element toaster oven that I use for reflow, with a "Extech EX330 Autoranging Mini Multimeter with NCV and Type K Temperature" that I use for temperature monitoring. The probe wire is held in place in the door so that the end of the probe is near a board surface. The boards being reflowed sit on a metal AL tray. My tray will hold 3 10cmx10cm boards or the equivalent in smaller boards.

This sequence works well for Kester EP256 leaded solder:

Temperature full on, dial it back to hold around 150C for 90 seconds

Temperature full on again dial it back to hold ~190-195 for 90 seconds. DO NOT EXCEED 205C. That's not good for the parts.

Turn off the heat, crack open the door a little, let the temperature cool down, once below 180C you can open the door full. Solder solidifies at 183C.

Will have to turn the heat up & down a little at the hold temperatures as the oven cools, and then heats up again. Temperature monitoring is the key.

\$\endgroup\$
  • \$\begingroup\$ Why the downvote? Is using a thermocouple too complex a situation for you? \$\endgroup\$ – CrossRoads Jan 7 at 13:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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