Ultimate DIY Reflow Oven, Solder Surface Mount PCB Components With Ease 

At 8:30, don't worry about having a slower ramp time, too fast and thermal dynamics (thermal shock) can damage the components and/or PWB. Most profiles are as follows: 1. Room temp. 2. Ramp to hot and stabilize that is safe for components but NOT hot enough to reflow the solder paste (usually 40 F below the reflow point). 3. Hot soak (wait) to be sure every component and the PWB are all at the same NEAR REFLOW temperature (usually about 10 minutes). 4. Ramp up to 10 degrees past the Reflow temperature. The faster the better, but slow is ok as long as you stop heating as soon as all solder points reflow to the solder specs. This is most critical, you don't want to be at the reflow temperature (or much above it) for very long, just enough time to get the solder to reflow correctly. 5. Cool off from the Reflow temperature back down to the safe Soak temperature as soon as possible (shut off hearers and open door for a few seconds). 6. Ramp down to Room temperature (heat off with door closed), slower is actually better here - you are not a high rate production line, so time is not a big factor.

Great job💯🤛🏻

I designed a controller about 3 years ago for a friends powder-coat oven. That oven is big enough to park your car inside - Pretty big. The heat source is two kerosene fuelled central heating units - Big power. Insulation is generous on this unit - it is used as part of a business, so it has to be efficient. The doors are opened often (high tunover of product) and this dumps the hot air from the oven into the room (wasteful and generates a high workload for the heating system). I used similar thermocouple (TC) amplifiers to you, and they feed back to an Atmel running code generated in Arduino IDE. I averaged the TC readings like you're planning to do, so that I could get ONE number to make on/off decisions. --------------------------------------------------------------------------------------------------------------------------------------- Usually, these amplifiers will generate an error message if a TC is grounded or open/short circuit - Use this in the code to generate a number variable: Variable = 0 Read thermocouple1: If result == error; Variable = Variable + 1 (variable now equals 1) Read thermocouple2: If result == error; Variable = Variable + 1 (variable may now equal 2) Later, when you do the average, you add up the result of all the TCs, and then divide it by the number of TCs but you subtract the above variable from the number of TCs first. This way, a TC can fail and the averaging still works with the remaining TCs. By the way, they WILL fail - this is just a facet of TCs. --------------------------------------------------------------------------------------------------------------------------------------- Temperature differences between TCs: They're shit. TC's are apalling sensors. Daft little LM35s walk all over even the most expensive and exotic TC's for accuracy and repeatability. What silicone sensors can't do is high temperatures, and this is where TC's shine. Don't rely on the TC to provide Seebeck voltage uncalibrated - and here is your problem - you need a calibration source. I have a Druck DBC650 temperature bath which is specifically designed to calibrate temperature probes. Its an expensive bit of gear to buy new - about $3000 (I got mine 2nd hand and very cheap). You're already using polynomials from spreadsheets to represent traces so thank God I don't have to describe that (I'm crap at math). You need to calibrate each TC in a known temperature source (as above), and record the reading at the microcontroller end. Plot that in a spreadsheet and then step the temperature up. Repeat. Repeat. Repeat. Once you have enough readings (every 10deg C?) generate a polynomial and then use it as a calibration in your code to display a REAL temperature. Sound a bit much? I was disgusted with just how far out TC's can be, and I would never have known if it wasn't for the fancy calibration source. You can generate a reliable 100degC from boiling water, and 0degC from ice water, but those two points are 100degC apart and only give you a linear response (all of the K-type TC's I've worked with were all over the place and totally not linear). Here we get even deeper - If the temperature at the measurement end of a TC changes - so will the reading you get, even if the temperature at the "hot" end stays the same. The only way to counteract this is to either control the temperature (submersion of TC amp in artificially cooled/heated liquid) or to KNOW the temperature of the amplifier end. Mount the amps to a strip of aluminium and use an off the shelf LM35 to measure its temp - again - this is where the calibration in situ plays a part. Also, you have some of the thermocouple lead inside the oven, and the lengths are not controlled. This is where TC's can get to be a pain in the arse. The bulky looking stainless bodied units you have look good with a clear line between what is the sensor and what is not. Install them properly and your readings should stabilize. Be prepared to have to calibrate in-situ because the amplifier end of the TC wire IS sensitive to temperature (Seebeck effect). Actually, don't bother with the big stainless bodies sensors - They're bulky, so have a lot of thermal mass, so will lag behind reality. This will be a total pain to code out the overshoot. Instead, buy a reel of TC wire and just crimp one end with something like a ring-terminal - instant TC for cheap! It has the advantage of being low thermal mass, so will be agile as hell. If the oven changes, so will the sensor. Ughh.. anything else..... Insulate the oven for efficiency, but don't pack it tight because the insulation will start to act as a thermal conductor. That internal fan will generate more than enough air movement to equalize the temperatures. If you're seeing dirfferences, it will be a mixture of the TC's needing calibrated, the TC's needing to be installed rigidly and direct view to the heating elements (radiant heat). On radiant heat, you're stuck between a rock and a hard place. You ideally want to block radiant heat to prevent hot-spots, but to do you have to add a shield which then adds thermal mass which resists temperature change. Try tin foil. I need to lay off the coffee. Good luck! Subscribed just for the oven project.

At 7:04 you just need to put in PID code and adjust the coefficients by trial and error until you get very stable temperature set points. It's not very complicated code, there are plenty of examples of PIDs online.

Awesome! Instead of using a solid state relay, a TRIAC would give you a smoother gradient control by triggering it with an Arduino. It also dissipates less heat and is significantly less bulky. You don't have to create a PCB for it, since you can salvage one from a typical dimmable light switch. It a matter of replacing the potentiometer with a simple opamp circuit and connecting the hot and neutrals to the heating elements.