Finally starting on this convection oven -> reflow oven conversion project

[mindcrime]

I've had this Black and Decker convection oven sitting on a shelf for a month or more now, so tonight I finally decided to get started on the project. All (or at least most) of the parts I need have arrived now: I have an SSR for the mains power side, a couple of thermocouples and a thermocouple amplifier board to detect the temperature, and a linear actuator to open/close the door. I have plenty of microcontroller boards laying around, so I just need to decide what I want to use to run the thing.



The sacrificial lamb before the hacking begins.



Opened up and ready for surgery.



That red wire looks important.



The existing controls. Looks like I can mostly leave this stuff alone and just insert the SSR "upstream" so to speak. Worst case, if that timer doesn't have an "always on" position, I can bypass that.

Note to self: I should have powered this thing up and frobbed the controls a little before starting the surgery. That said, I still could since I haven't touched any of the wiring yet.

Edit: on second thought, I am going to need one or two more parts that I may not have lying around here already. I'll need an h-bridge motor driver so I can reverse the linear actuator. I'll also need a transformer to give me 12V that I can rectify to DC to use for the actuator. I may have that in the parts box already, not sure offhand.

And the datasheet for my SSR suggests adding a fast blowing fuse on the load side, so I'll probably have to pick up a fuse / fuse holder. Still, the major elements are all here. 

[fourfathom]

I have what looks like a similar oven I use for reflow.  It has heating elements above and below the racks.  I have a single sheet of aluminum in the middle position to hold my circuit board being reflowed.  I use a thermocouple attached to a chunk of circuit board, screwed to the aluminum tray.  I don't have a door actuator, but rely on me being there to open it during the cool-down cycle.

My controller is an arduino-style "ItsyBitsy M0", and it drives an optoisolated SSR between the wall power and the oven power cable.

My oven is un-modified, and I found that the heating and cooling rates with the heater 100% on and off were pretty close to the desired profile.  I have an auto-learning program that adjusts the on and off timing to hit my targets after a few calibration runs.  There is no fancy PID loop or proportional heat control, it's just on and off.  Here's the sequence:
Preheat: turn on the heat, wait until it reaches about 100 C.
Soak: Turn off the heater. The heat continues to rise, but slowly, almost flattening out after 60 seconds.
Reflow: Turn on the heater.  Temperature climbs.  Keep heat on until temperature reaches about 205 C (this value changes as a result of calibration runs)
Peaking: Turn off the heater.  The temperature continues to rise, peaking at about 230 C (my target for leaded solder)
Cooling:  Once the temperature drops a few degrees, buzz the alarm to tell me it's time to open the oven door.

I have the electronics in a small aluminum chassis, with a couple of buttons and a display, and I can control it with those.  I also have a USB serial-port interface which lets me control the thing from a program running on a PC -- this gives me nice charting and logging.  I also have duty-cycle heater control capability, and sometimes use that if I want to use the oven for medium-temperature thermal testing, but for soldering the on/off heater control works well.  Duty-cycle control would only slow down the heating ramp, and it's already barely fast enough.

[mindcrime]

I have what looks like a similar oven I use for reflow.  It has heating elements above and below the racks.  I have a single sheet of aluminum in the middle position to hold my circuit board being reflowed.  I use a thermocouple attached to a chunk of circuit board, screwed to the aluminum tray.  I don't have a door actuator, but rely on me being there to open it during the cool-down cycle.

I'm never going to run this thing when I'm not physically present, but the door actuator seemed prudent based on the experience a buddy of mine had with his oven. He found that his oven doesn't cool fast enough during the "cool" parts of the cycle, to track the temperature profile well. He could only get the cooling to be fast enough by opening the door.

My controller is an arduino-style "ItsyBitsy M0", and it drives an optoisolated SSR between the wall power and the oven power cable.

Hmmm... opto-isolation, that's probably a good idea. I should do that. Glad you mentioned that.

My oven is un-modified, and I found that the heating and cooling rates with the heater 100% on and off were pretty close to the desired profile.

Gotcha. Other than the door actuator (and adding thermocouples), I don't plan to modify the core of the oven itself. It's all about switching the power on and off as needed through the SSR.

I have an auto-learning program that adjusts the on and off timing to hit my targets after a few calibration runs.  There is no fancy PID loop or proportional heat control, it's just on and off. 

I thought about going that route as well, but this is as much a didactic exercise for me as it is just a tool building exercise, and I've always wanted to learn more about control theory, so I think I'm going to go the PID route. Although once all the "pieces" are in place (thermocouples, door actuator, SSR), then I can tweak everything else through software, so I may experiment with different algorithms and approaches as I learn more about control theory.

I have the electronics in a small aluminum chassis, with a couple of buttons and a display, and I can control it with those.  I also have a USB serial-port interface which lets me control the thing from a program running on a PC -- this gives me nice charting and logging.  I also have duty-cycle heater control capability, and sometimes use that if I want to use the oven for medium-temperature thermal testing, but for soldering the on/off heater control works well.  Duty-cycle control would only slow down the heating ramp, and it's already barely fast enough.

Sounds pretty close to what I'm planning. I definitely want to support data logging / control via a PC. I might use wifi for the connection instead of a USB cable though. Heck, I could even experiment with using Bluetooth, I suppose.

That said, I like the idea of having a display and some sort of input mechanism right on the oven... at the very least I'd like  a way to select which temperature profile to run, a "start" and "stop" button, and a display of the instantaneous temperature in the oven.

That's an interesting thought about using the oven for thermal testing, and possibly other uses, besides just for soldering.  I hadn't even considered it having any application besides soldering (and learning control theory stuff).

[rstofer]

I used this controller.  I remounted the display in a metal box mounted above the oven.  Clean installation.

https://www.rocketscream.com/blog/product/tiny-reflow-controller/

That version is retired, the new version is:

https://www.rocketscream.com/blog/product/tiny-reflow-controller-v2/

I installed the Solid State Relay in the right hand sidewall.  I also added high temperature mat insulation in the various cavities.  Without the insulation, I couldn't get up to temperature fast enough.  McMaster-Carr carries it but I don't remember which variant I used:

https://www.mcmaster.com/insulation/insulation-for-equipment-ovens-and-furnaces/

My oven is based on the Black & Decker InfraWave toaster oven.

[fourfathom]

Yes, if you leave the oven door closed after the end of the reflow stage you will never get close to a recommended cool-down rate.  I might eventually add a door-opener.  Another option is to add an air duct and a blower, but that probably adds other thermal complications.  With my beeper alarm the system works well enough for now.

I've been asked to write up my toaster reflow, and when I've collected the data will post more here.  I've also hacked into a "Reptile Incubator" Thermal Chamber (see eevblog video: ), modifying it and adding an external controller (much like my toaster-reflow project).  This does used a PID control loop, and I'm writing that one up as well.

[mindcrime]

https://www.rocketscream.com/blog/product/tiny-reflow-controller-v2/

Nice, I'll have to check that out.

I installed the Solid State Relay in the right hand sidewall.

Yeah, that's pretty much what I'm planning on as well.

I also added high temperature mat insulation in the various cavities.  Without the insulation, I couldn't get up to temperature fast enough.  McMaster-Carr carries it but I don't remember which variant I used:

There's some firewall insulation that race-car builders use that I'm familiar with, I was thinking of using that if I need extra insulation. Not sure if it's more or less the same stuff or not. I may even be able to get some for free, if my dad (who builds race cars) has some scraps laying around his shop. I'll have to ask next time I talk to him.

[Miti]

From my experience with toaster ovens, they never cool down fast enough, even when the door is fully open. The reflow profile is at the board level, not the air inside the oven. If you attach a thermocouple to the board, you’ll see that it stays at dangerous high temperatures for way too long if you don’t blow air on it.

[mindcrime]

From my experience with toaster ovens, they never cool down fast enough, even when the door is fully open. The reflow profile is at the board level, not the air inside the oven. If you attach a thermocouple to the board, you’ll see that it stays at dangerous high temperatures for way too long if you don’t blow air on it.

Blimey. Well, worst case, I can always bodge in a blower of some sort of push in cool air from outside the oven.

[fourfathom]

I just ran some tests with my un-modified Black & Decker toaster oven:

Reflow cycle, door opened after peaking


Door remaining closed

I'm fairly comfortable with the cool-down when I open the door.  This is in a garage, fairly cool but no particular airflow.  I also think that the pre-heat, soak, and reflow are reasonable.  I'm using a leaded solder paste and have a reflow target of 230 deg C.  Perhaps with added insulation the cool-down might not be as rapid?

(is image in-lining still not working?)

[coppercone2]

if you open the door it won't be so bad.

[fourfathom]

if you open the door it won't be so bad.

If you're referring to my charts above, in the first one I do open the door after temperature peaking, and indeed, it's not so bad.  The other chart shows the temperature profile with the door kept shut, and I agree, it stays too hot for too long.  BTW, The temperatures shown are measured by a thermocouple that is attached to a small circuit board located adjacent to the board being soldered.  This is the sensor that my program is monitoring.

[coppercone2]

I did similar experiments with my PID oven, mine is a multi zone controller or whatever that can turn on a cooling system, but I have not done it. I thought to put a blower fan on top and channel it in, but I think its good enough when you open the door. you could just put a dryer flap on it so when you turn on the fan it opens the flap and cools the oven quickly, you need 2, one to protect the fan, and the other to open the oven.

but its a fair bit of sheet metal work because I don't want hot air shooting out of the side of the thing, I would need to make a diffusor system so when I leave something on top of it, the damage is not catastrophic

if you want a 'nice' mod, you need to do this on top of the oven, and figure out the correct springs, and install duct on double wall thin sheet metal.. its a fair bit of work.

I used a expensive SSR for mine ($60) because the parts I put in there can easily be a couple of hundred dollars, so its pointless to go with china reliability here, since I am not drying painted door handles.. personally to me a ruined pizza bagel is enough of a problem I don't want to go to china. Seriously when you pop something complicated into the toaster oven chances are you are tired as hell and going to be mega pissed off if it burns food lol..


They are useful also for doing things with varnish, I varnished my transformer in it from a HP supply to stop buzzing, after vacuum impregnation, since you can set the temp precisely to datasheet parameters.

I am super interested in high quality pizza bagels though, I wonder how much customers they lose because of bad thermal control. I have no idea how such a good food got a reputation along the lines of a pop tart.

Also getting a steady flow inside of the thing to cool it down nicely without thermal gradients is gonna be hard, because if you spot cool it with air flow you can break parts. I think basically opening the door is the safest option. They sell a nice D-cell powered fan for 25$ you can set next to the oven, open the door and turn the fan on to blow into it, it lasts a long time on a set of D's, does not have a cable and you can put it far away enough to cool the board gently.

What might be best is another chamber called a cooling chamber, so you take the board out when its solid and put it in there with something designed for cooling rather then heating!


I think I will use an AUX on the pid controller to sound a buzzer and turn on a light when it detects the solder is solid and then make a cooling box some how.

Keep in mind too those elements will last shorter if you cool them fast, so its probobly better not to even do cooling in the oven, and close the door after you take the board out to cool off.

[Miti]

I just ran some tests with my un-modified Black & Decker toaster oven:

(Attachment Link)
Reflow cycle, door opened after peaking

(Attachment Link)
Door remaining closed

I'm fairly comfortable with the cool-down when I open the door.  This is in a garage, fairly cool but no particular airflow.  I also think that the pre-heat, soak, and reflow are reasonable.  I'm using a leaded solder paste and have a reflow target of 230 deg C.  Perhaps with added insulation the cool-down might not be as rapid?

(is image in-lining still not working?)

More than 2 minutes above 200 C?  Good luck!

[fourfathom]

More than 2 minutes above 200 C?  Good luck!

When I open the door after peaking I'm only above 200 C for 90 seconds.  Perhaps you're looking at the "door closed" profile, where I delay opening the door until long after cool-down in order to show why you need to open the door?

[johofz]

I just recently converted a similar oven into a reflow oven. It has a hole on the top originally used to grill stuff on a plate you can put there. This is very convenient, since you can open the door aswell as the hole on top to cool down very fast. The only thing is, that it needs some thermal isolation to increase dT/dt for the soldering phase.

Here is my repository for the Hardware and Firmware:
https://github.com/johofz/Reflow_Oven_V2

I did a V2 because I had some issues with power dissipation and wanted a bigger touchscreen. I choose a rather beefy STM32f415 in LQFP-100 package, since they have FSMC and I wanted to use a 16-bit parallel 3.2" TFT with a touchscreen.

I can share some pictures and graphs if someones Interested.

It can also be used to heat things to a controlled temperature. It can keep temperatures stable as low as 10°C above room temperature.

I didn't had the time to test the reflow part yet, but I will shortly recive some PCBs and will share the results once I'm done.

[mindcrime]

Project update 05-01-2021 (repost due to database issues blowing away previous posting)

I spent some time a couple of days ago working with the thermocouple amplifier I am planning to use, and the thermocouple, making sure I could get useful temperature readings with the Arduino. That part is handled now, so the next major task is to sort out switching the power on and off with the SSR. I didn't want to start messing with the wiring inside the oven itself yet, so in order to start experimenting with the SSR (and leave myself with a possibly useful gadget) I took a sacrificial extension cord and one of the extra SSR's I ordered, and built this:





So far this works a treat, although the only load I've used is my desk lamp. That's fine, the principle is the same. The issue here is that I've realized that I'm going to need a couple more parts to make all of this work. The minimum rated voltage for the control signal on this SSR is 4VDC, and my arduino is a 3.3V unit. So I'm going to need a separate power rail (also needed for the linear actuator) and a transistor to control the SSR.

My current (no pun intended, hahaha) thinking is to add a transformer to step the 120VAC mains current down to 12VAC, then a bridge rectifier to give me 12VDC. That should serve just fine for the linear actuator, and I can probably just use 12V for the SSR control as well. It's rated from 4-30V so that should work. Or I can always add a linear regulator or a buck converter to give me 5V or whatever.  Actually come to think of it, I think I'll need a low voltage rail to power the Arduino anyway.

I also recently realized I'll need an h-bridge for the linear actuator. And the production version of this will use opto-isolation on the SSR control, and a fast blowing fuse on the load side of the SSR.  So despite my laughable assertion earlier that "all the parts I need are here", I now find myself waiting for a box from Jameco with the transformer, h-bridge, opto-isolators, fuse holder, rectifiers, etc. Not a ton I can do until then. I may also add a manual toggle switch upstream of everything else on the mains input.

[mindcrime]

Project Update 05-09-2021

OK, making more good progress on this. I have the arduino now controlling the SSR, by using a little SN2222 as a driver for the SSR, and an independent 12V power supply for the relay. I didn't do all the stuff with the transformer and rectifier yet, rather I just used my bench power supply for that for now. Nonetheless, now the 3.3V digital pin on the Arduino can turn the SSR on and off with no problems. So now I've combined that and the thermocouple stuff, and today for the first time I plugged the actual oven into the SSR instead of my desk lamp "dummy load".

I wrote code to turn the oven on, monitor the temp, and then turn it off once it hits a certain temp. This is not any actual soldering profile, but just a proof of concept to convince myself that I can detect the temperature in the oven and turn it on and off in response.

Here's a sample of the output from the serial monitor with the thing running:

Code: [Select]

MAX31856 thermocouple test
Thermocouple type: K Type
C = 22.0
F = 71.7
Turning OVEN on
C = 44.1
F = 111.4
monitoring OVEN
C = 44.1
F = 111.4
monitoring OVEN

<snip>

C = 89.9
F = 193.9
monitoring OVEN
C = 89.9
F = 193.9
monitoring OVEN
C = 90.0
F = 194.0
Turning OVEN off

Next steps now:

1. Build the power-supply that will live inside the oven to power the SSR, power the Arduino, and power the door actuator.
2. Wire up the door actuator controls to the arduino and get code written that can open and close the door.
One open issue I have here is figuring out the door position at a moment in time, so I don't try to drive it beyond it's range of motion. Since the linear actuator works at fixed speed, I could just "do the math" and not worry about measuring the door position empirically. Or... dunno. Haven't given this part a ton of thought yet, TBH.
3. Physically attach the actuator to the oven and couple it to the door.
4. Provide a way to store and identify particular temperature profiles.
5. Implement PID in software (probably using a library of some sort, not necessarily going to try to write PID from scratch) to try and match the selected temperature profile.
6. Test it by baking some boards.
7. Add a little LCD display or something to show the selected profile, instantaneous temperature, elapsed time, etc.
8. Implement a keypad control or something to select a profile, as well as "go" and "stop" buttons.
9. Package everything up neatly and close the case up.

That's probably missing a few things. I may need to fiddle with adding insulation, sealing air leaks, etc., if my  one buddy's experience is any indicator. And who knows, maybe I'll need to add a fan/blower of some sort to get the cooling to happen fast enough. Time will tell on that part, I suppose.

[Miti]

Take a look at this project:

https://apollo.open-resource.org/mission:resources:picoreflow

I tried it, it works well but my oven cannot follow the profile, I did not bother insulating it. In the end, I’m controlling it manually, with a low mass thermocouple connected to the board. I turn it fully on up to 150 C,
about 25% duty cycle to 180 C, then fully on to 215 C, this is for leaded, I don’t use lead free with my projects. After about 10 second I open the door and use a 120mm fan to blow air from about half a meter or more so I don’t cool it down too fast. This method never failed me for years. I thought about coding this in an Arduino but I couldn’t find the motivation. I bought recently a T962 for 120CAD but it’s a piece of crap. The temperature difference between left and right is about 10 C.

[mindcrime]

I bought recently a T962 for 120CAD but it’s a piece of crap. The temperature difference between left and right is about 10 C.

Funny you mention that... one reason I decided to go the DIY route - aside from it being a learning experience in and of itself - is that a buddy of mine had a T962, and then did a convection oven conversion, and ultimately found that his converted convection oven did a better job than the T962. In the end, I think he donated the T962 to the hackerspace where we were both members and kept his DIY oven in his lab for his own use. I haven't heard a lot of good things about those cheap Chinese made reflow ovens, and the conversion project looked fun and not too expensive, so here we go...

[phil from seattle]

Wire up the door actuator controls to the arduino and get code written that can open and close the door.
One open issue I have here is figuring out the door position at a moment in time, so I don't try to drive it beyond it's range of motion. Since the linear actuator works at fixed speed, I could just "do the math" and not worry about measuring the door position empirically. Or... dunno. Haven't given this part a ton of thought yet, TBH.

One possibility is to use a hobby servo.  Because you command it to a specific position, you will never over drive it or position it in the wrong spot. Plus, you can use the PWM hardware of the arduino so there is no software overhead to maintaining a position.  They make all sorts different kinds of servos and they are farily inexpensive.

[mindcrime]

Project Update 05/12/2021

Made some solid progress tonight. I have the low voltage power supply pretty much finished. It's nothing fancy, just a transformer, a bridge rectifier, a couple of LM78xx series regulators (one for 5V, one for 12V) and a few capacitors. But it should serve my needs.

[beanflying]

You might get some ideas from this Coffee Roasting firmware/hardware including PCB shield for SSR and fan controls. Coffee Roasting and PCB ovens share a similar temperature controlled ramp soak and rapid cooling process. In particular look at the slow PWM part of it to drive the SSR. https://github.com/greencardigan/TC4-shield People have used small turbo ovens with a Rotisserie and drum for roasting but the standard mod was always Insulate the body for better stability and control and generally adding some airflow from adding a fan and exhaust.

greencardigan is a member here too.

[mindcrime]

You might get some ideas from this Coffee Roasting firmware/hardware including PCB shield for SSR and fan controls. Coffee Roasting and PCB ovens share a similar temperature controlled ramp soak and rapid cooling process. In particular look at the slow PWM part of it to drive the SSR. https://github.com/greencardigan/TC4-shield People have used small turbo ovens with a Rotisserie and drum for roasting but the standard mod was always Insulate the body for better stability and control and generally adding some airflow from adding a fan and exhaust.

greencardigan is a member here too.

Radical! I had no idea that roasting coffee beans was at all similar to reflow soldering. Maybe this oven *will* turn out to have multiple uses in the end.  Of course this one will probably wind up contaminated by soldering byproducts, so if I wanted to roast my own beans, I'd probably build another one. Just to be on the safe side.

I decided to go ahead and operate on the assumption that extra insulation will ultimately be useful, so I ordered some of this stuff from Amazon. So hopefully it will do the job as far as further insulating the internal cooking chamber.

[wizard69]

This is an interesting thread that I just dropped into.   I have not built a reflow oven yet as I'm thinking more along the lines of a multiple use oven.   However some comments from working on industrial ovens of various types over the years:

• First insulation is your friend.    One of the gotchas with conversions is making sure the oven is easy to insulate.   It makes a huge difference in the temperatures you can achieve and the power you will use to get there.
• If you are going to use SSR you might as well buy an optically isolated one.   More so I believe that most on the market are isolated in some manner.
• Air circulation can make a huge difference in the ultimate oven temperature profile.   It can also be integrated into the cooling profile.
• If you don't have air circulation consider the aluminum plate somebody mentioned above.
• Make sure your wires are temperature rated for the location they are in.
• Be careful not to disable any over temp safeties built into the oven or provide your own.   This is extremely important for safety.   These ovens can get very hot in some cases even over the course of a trip to the bathroom.   The alternative is to implement your own, which in some cases is a better idea.
• A door switch can offer up additional safety, if designed to shut off heater power on door opening.   This may be an issue with fan based cool downs though.
• Don't assume that the unit will never be run unattended, thus design in safety in case something of higher priority comes up and you get distracted.   Your best intentions can be trumped by a kid crashing a bike and bleeding all over or the wife saying her water broke and you need to go to the hospital.   Or if you are like me you can some times just focus on something to the exclusion of the entire world and forget what is going on.   This goes back to over temp protection and frankly a limit on cycle time.
• Kinda related to the above add a buzzer or bell to alert you to the end of cycle or other program steps.   You might not need it all the time but it can really help when developing recipes and doing things like manually opening the door.
• Data logging can be a great diagnostic tool.   Even if it is a file on secondary storage that get over written with each run, it can be a bit help.   So consider that when writing your oven control software.   The only trick here is to find the right sample time to catch glitches yet not blow up file size.
  
  In the industrial world this has made what might have been a difficult to solve issue a simple task of reading a file and seeing what happened.   At the very least you want time and temperature data but the more the better though "more" gets into a lot more control hardware.

Hope this helps some.

[mindcrime]

• First insulation is your friend.    One of the gotchas with conversions is making sure the oven is easy to insulate.   It makes a huge difference in the temperatures you can achieve and the power you will use to get there.

That definitely jibes with my one buddy's experience, as well as what I've heard from others here. I just got  a couple of sheets of that aluminized fiberglass insulation that they use for race car firewalls, so we'll see how well that works out.

• If you are going to use SSR you might as well buy an optically isolated one.   More so I believe that most on the market are isolated in some manner.

Ah. Had I thought about that earlier, I would have done so. But I already have the SSR I plan to use, and it's not explicitly opto-isolated AFAIK. I do however, have standalone opto-isolators that I plan to incorporate.

• Make sure your wires are temperature rated for the location they are in.

Definitely. In addition I got some high temperature insulated wire sheath material to use anywhere I have to route wires where they might be exposed to a lot of heat.

• Don't assume that the unit will never be run unattended, thus design in safety in case something of higher priority comes up and you get distracted.   Your best intentions can be trumped by a kid crashing a bike and bleeding all over or the wife saying her water broke and you need to go to the hospital.   Or if you are like me you can some times just focus on something to the exclusion of the entire world and forget what is going on.   This goes back to over temp protection and frankly a limit on cycle time.

Agreed. I expect I'll incorporate both an over-temp limit and a maximum cycle time.

• Kinda related to the above add a buzzer or bell to alert you to the end of cycle or other program steps.   You might not need it all the time but it can really help when developing recipes and doing things like manually opening the door.

Agreed. A nice, loud bell or chime of some sort should be there. I hadn't planned on that initially so I didn't buy a part for that yet, but it should be easy enough to find something suitable.

• Data logging can be a great diagnostic tool.   Even if it is a file on secondary storage that get over written with each run, it can be a bit help.   So consider that when writing your oven control software.   The only trick here is to find the right sample time to catch glitches yet not blow up file size.
  

The tentative plan as of right now is to push logging / telemetry data to a server on my network.