

I pause the heating at 160C for about 20 seconds to activate the No-Clean flux, then let 'er rip up to a peak of 225C before turning off the heating elements and slightly cracking the door open while the fan still runs to cool the board down at the recommended rate of 1-2C/second.

Smd oven reflow full#
Now I just run the heater at full blast on Toast mode while monitoring the actual board temperature with a Type-K thermocouple (usually at a via NOT going to a ground plane) through a hole I drilled in the top of the oven. I used to have a PID temperature controller with the recommended ramp/soak profile for Kester EP-256 programmed in, but I needed the PID controller for another project one day and it turns out that it isn't really necessary. Just don't ever use it to cook food again. For prototyping and modest volumes (say, 1-4 boards per day), it is really and truly hard to beat a table-top *convection* toaster oven, at least if using leaded solder with a typical RMA or No-Clean flux (lead-free, however, is indeed much more difficult to solder with this get-up). In my experience, the only kind of reflow oven worth a crap for production manufacturing is a conveyor belt type with at least 5 heating and 2 cooling zones that has been well profiled. I'm sure someone can get the same result with a T962, but you will have to modify it Drawback: that oven costs around 4000 USD, you can build it much cheaper if you have the time. The firmware always waits until a certain temperature is reached and starts counting. The temperature is very well sealed inside, during the reflow process you will not notice any heat leaking out of the oven.

Just putting the PCB at the bottom of the reflow oven will not give the same result as if the PCB will be put on some spacers (the same happens with every small reflow ovens). However the PCBs should be monitored at the beginning and a profile for the particular PCB can/should be created. After some time I bought the second one RO250BF, it's really just putting the PCBs in and waiting no more unexpected results. At the beginning I used a T962 which even burned some of my PCBs.
