Important safety update for Toast-R-Reflow controllers

Everyone who has a Toast-R-Reflow model I or model II controller should obtain the latest firmware from GitHub and install it. The latest version for model I is 0.5, and for model II it's 1.2.

The change adds support for the AVR watchdog. The Toast-R-Reflow installed at The Hacker Dojo had some unknown event happen that wedged the controller with one of the elements switched on. This is obviously an undesirable outcome. It's unknown whether the watchdog would have been effective or not, but it certainly isn't effective if it isn't used.

If you are unsure how to update the firmware or have any questions, please feel free to contact me and I'll be happy to help. If you have no other way to apply this update, you can mail your controller back to me and I will apply it for you.

Next steps for Toast-R-Reflow

Toast-R-Reflow was a success, but I think I could take it to the next level. I think there's a market for a hobbyist tabletop reflow oven that would cost $200 or less, be able to handle up to a square foot of PCB, and be able to successfully keep up with a RoHS paste profile. 

The trouble is that the idea isn't quite to a KickStarter project level. There are two important parts of the project that I don't know how to scope: custom heating elements and custom sheet metal enclosures. 

Everything else is doable. But without those two, there's no product. 

Who does one even talk to about those sorts of things?

Toast-R-Reflow and PowerSwitchTail

I just discovered the PowerSwitchTail on AdaFruit. For what it's worth, it appears to be an almost "consumer friendly" solution to the problem of controlling reflow ovens or hot plates with a micro controller.

If you wanted to build a reflow oven without actually taking the oven apart, you could connect its power cord through one of these, and then connect one of the two Toast-R-Reflow outputs to the control input, turn the oven's own controls on as high as they will go, and go.

The only reason it might not work is if the oven's own temperature controls do not allow the oven to heat as hot as the reflow controller desires. Toast-R-Reflow's SnPb profile goes up to 225°C, which is 437°F. That's close, but not quite 450°F, so if the oven's controls go up to 500, and they're even in the ballpark, it should be ok.

The one gotcha is that Toast-R-Reflow's controller comes with the LED series resistors on the controller, while PowerSwitchTail has them as well. If you want to use a Toast-R-Reflow controller with PowerSwitchTail, you need to replace the output series resistors with either zero ohm resistors (yes, those exist) or wire jumpers. As a special option (and at no extra cost), if you want a Toast-R-Reflow controller with no output series resistors, just mention it in your order and I will do that (if you do, DO NOT use the controller with unmodified Toast-R-reflow power boards - you will destroy the opto-isolator!).

Woah! Toaster oven sale at Amazon!

If you've been procrastinating trying to build a reflow oven, this may help.

Amazon is selling the Hamilton Beach 31138 for only $23.99! With the Toast-R-Reflow backpack controller and power board at $50 for the pair (as kits), that's a reflow oven for less than $100!

Giddy-up!

New TRR backpack controller

It was a big OSH Park day today. Among the arrivals were the latest design for the TRR controller+display backpack. This design replaces the ATTiny85 and the MCP23017 with an ATTiny84.

The initial cut at the firmware didn't work perfectly, but it wasn't hard to sort out. Unfortunately, unlike the ATTiny85, the 84 doesn't have a 2.56 volt analog reference built in, so far as I can tell. So instead, it must use Vcc (5 volts), which means that the full 10 bit range covers 1000 °C, which is way, way more than a reflow oven needs. The downside to that is a loss of resolution, but even 1 degree is fine for this application. In principle, this design could also be used to control lower temperature processes, like a sous-vide heater. If you do, then using the 1.1 volt analog reference would yield a maximum temperature of 220 °C (just short of what's necessary for a reflow oven, unfortunately), but with quarter degree resolution.

The alternative part is the MAX31855, but that part is twice as costly and requires a 3.3 volt power supply, as well as level shifting on its input pins. It does, however, offer 14 bit resolution - .25 °C - over an incredibly wide range, which even includes negative Celsius temperatures.

Anyway, I built two of them this evening - and the first built the second. One of them for me, the other for the Hacker Dojo oven. I'll install it sometime tomorrow evening.

Price reduction!

Well, despite the fact that no one has yet actually bought one, I've decided to get a bulk order of Toast-R-Reflow parts. This bulk ordering has allowed me to drop the price 20% on the power board. So if you don't mind doing a little soldering (none of it surface-mount), you can get a complete TRR kit (consisting of a power board kit and a backpack controller "quick kit") for just $50. All you need to add is a 7-12 volt @ 250 mA power supply, a K type thermocouple and a toaster oven.

Step right up, don't be shy!

The Toast-R-Reflow boards and kits are available if you want to convert your own toaster oven into a reflow oven. It's not that hard. I've even written a nice instruction manual you can peruse before you take the plunge. Assembling the power board is just about the easiest through-hole soldering I've ever done, and the controller comes with all surface mount parts pre-assembled and the controller pre-programed. All you need to do to finish the "quick kit" is attach the through-hole connectors and mount the LCD. And all you need after that is a K type thermocouple and a toaster oven.

Who'll be the first?

Toast-R-Reflow schematic round-up

So here are the schematics for the Toast-R-Reflow project so far. For the Eagle .brd files, you can go to my OSH Park profile.

First, the power board:

This has remained, more or less, constant throughout the whole project. The design is quite simple, and taken straight from one of the data sheets (for either the BTA-20 triac or the MOC-3020 optoisolator. I forget which). The only possibly controversial aspect is the inclusion of a snubber (the .01 uF cap and 39 ohm resistor on each output). The loads are not particularly inductive, and the triacs are billed as being "snubberless," but I included it anyway. It doesn't do any harm, and it's not like the components are pricy.

But once that's done, to what do you connect it?

That's the first controller design. It was intended to connect to an OpenEVSE style RGB LCD backpack or to an AdaFruit RGB LCD shield. It works perfectly, but I imagined folks railing against the need to buy two pieces of kit. So I came up with...

This design combines the backpack's circuit with the controller. That worked too, but it occurred to me shortly after I finished it that the price of an ATTiny84 was going to be less than the price of the 85 and the MCP13027 GPIO chip. And that led to...

That's what's in at OSH Park right now being fabbed as we speak.

The Hacker Dojo reflow oven

The oven is complete and operational. In case folks want more information, I'll put up this page as a clearing house.

While I was there, I ran through a cycle while someone took a video. When that video gets posted somewhere, I'll link to it here.

The gist of it is that you need to supply the oven itself with 120vac (just like its makers intended), and the controller board (with the LCD) needs 7-12 VDC @ 200 mA, 2.1mm center positive.

The normal waiting state will show "Waiting" and the thermocouple temperature on the left side of the display.

The thermocouple is the brownish wire-looking thing snaking up from the bottom of the oven. You want to put your boards as close to the end of that wire as practicable, but you want the thermocouple in air, not touching your board or any other part of the oven.

There is a button on the back left (when viewed from the front) side of the controller. Press it to start a cycle. The display will show four pieces of information. Top left is the name of the current step in the cycle. The steps are "Pre-Heat," "Soak," "Reflow," and "Cool." The bottom left will show the current temperature and the bottom right will show the current set point (target temperature) of the oven. The top right will show the elapsed time from the start of the cycle.

The way the firmware works is best illustrated with the concept of a graph. The axes of the graph are time and temperature. The beginning and end of each step represent points on that graph. The end of one step is the same as the beginning of the next. The set point of the oven will gradually move along the graph described by line segments joining each of the points.

The points are:

1. The beginning of pre-heat starts at 0 seconds, 25 ºC. Pre-heat lasts 90 seconds and ends at 150 ºC.
2. Soak lasts 60 seconds, and ends at 180 ºC.
3. At the end of Soak, the line has a discontinuity. It breaks and immediately snaps to a set point of 225 ºC.
4. The Reflow step stays steady at 225 ºC for the duration, which is 75 seconds.
5. The Cool step goes downwards from 225 ºC to 150 ºC for 60 seconds. In reality, the oven will cool slower than this left on its own. Impatient users may wish to open the door a crack to speed cooling, but there is a danger of thermal shock if the cooling proceeds too quickly. The cool stage will attempt to prevent this by heating the oven further if cooling is proceeding too quickly.

This profile is appropriate for Sn-Pb solder paste. I have not attempted RoHS soldering with this oven. I am skeptical that it can make it up to 260 ºC in a reasonable amount of time, but I'd be happy to be proven wrong.

The low voltage cable coming out of the oven to the controller is directly connected to a pair of optoisolators with a common cathode (the green wire). There are no series resistors - you must add them yourself. The target is 20 mA of current. Do not exceed 50 mA or you'll damage the optoisolators. If you want to experiment with alternate control systems, the optoisolators are MOC-3020s. The thermocouple is a K type. Note that the K type thermocouple lead color code is yellow for positive and (perversely enough) red for negative.

The controller board that's there is a prototype for the backpack controller design. There is a wiring error in the LCD wiring that was worked-around in software. You must modify the LiquidTWI2 library to reverse the order of the LCD data bus lines for this particular board. I have an order into OSH park for corrected boards and will bring one in when they arrive.

See you Saturday

I will be at Hacker Dojo this Saturday at noon. I'll be there to do the Toast-R-Reflow conversion operation in person. I'll leave the completed oven there for folks to use too. I'm not a member, so I can't reserve a room or create an event or anything like that. But I visited (for the first time) this evening and met some folks there who dug it and they said that Saturday at noon-ish was a good time, so that's the plan.

I've already set aside most of the parts. I'll be using the standalone controller and an LCD backpack, but I will eventually replace that with the ATTiny84 backpack design as soon as the boards arrive from OSH park.

Yet another TRR controller design

To some extent, I sort of wonder why I keep doing this. So far as I know, I'm the only one who's actually used any of my designs to build a Toast-R-Reflow oven. But still.

There are three basic ideas I have for the TRR controller:

1. A standalone controller, intended to plug into a separate i2c "backpack" display. This one is based on the ATTiny85 and has an AD8495 thermocouple amplifier. That design is basically a done deal.
2. A display "backpack" controller. The first design still uses an ATTiny85 and the MCP23017 i2c GPIO chip to connect up to the LCD. Although my first prototype had a wiring error, it did prove the concept.
3. A "second generation" backpack controller, which uses an ATMega rather than an ATTiny to give it more flash space. This one is still TBD.

What I've done is come up with a new design for controller number 2: Instead of the ATTiny85 and the MCP23017, I've gone with an ATTiny84. It will use the traditional LiquidCrystal Arduino library intended to talk to a directly connected LCD display. The LCD data bus will be shared with a MAX31855 digital thermocouple amplifier as well as the ISP bus for programming, so that the pins can sort of serve double duty. I've stuck with the same analog thermocouple amplifier for now. It's much cheaper, and seems to work more than well enough.

The swap of the Tiny84 for the 85 + MCP23017 saves about a dollar off the design.

Toast-R-Reflow variants

I've already now designed three different Toast-R-Reflow controller boards... First, there was the ATTiny85 board with an AD8495 and an i2c header for an OpenEVSE LCD backpack. The second design was the same, but done as an LCD backpack by itself. The third is going to be an ATMega328 based LCD backpack.

But it occurred to me today that the problem really breaks down into two hardware pieces:

1. Power control for the oven
2. Reading the thermocouple

The nice thing about the Toast-R-Reflow design is that the high voltage switching and control apparatus is a very static design. Since there are optoisolators driving the triacs, it really just amounts to treating the oven elements as if they were LEDs, and that's fairly trivial for Arduino/ATMel based systems.

There are a number of ways to read the thermocouple. There is, of course, the AD595AQ that I started with. Nowadays I'm using AD8495s instead, but in principle, you probably can get more accurate results with the MAX31855. It's a particularly good choice because the A/D converter is built right into the device. You just need to connect it up to 3 digital pins to make an SPI interface. SPI seems at first to be unfortunate, since most of my projects use i2c, but in this case, the interface is fairly easy to bit-bang. The other downside is it's a $8.19 part - double the price of the AD8495. But for your trouble, you get quarter-degree resolution and a signed 14 bit range - meaning you can read negative temperatures properly without resorting to supplying a reference voltage like for the AD8495.

I originally went with the AD8495 because the ATTiny85 only had a single pin to spare, so using it as an analog input was the direction I chose. One way to get more pins is to shift up to the ATTiny84. I was contemplating such a design when it further occurred to me that the i2c LCD GPIO chip has lots of pins to spare!

Of course, that means a fully customized version of the LiquidTWI2 library. And I'm already - more or less - heading down the road of of an ATMega328 based design with the LCD directly connected. In that design, I left the ISP pins unused so that I wouldn't have to protect them during programming, but as long as the CS line of the MAX31855 is not asserted (meaning there's a pull-up on its !CS pin), it can share SCK and MISO with ISP.

I was contemplating adding a 1.8 volt analog reference to that design to make better use of the available A/D resolution, but using the MAX31855 would obviate that. The only other downside is that the cost goes up, as it's a more expensive chip.

But I think, in the end, that that's the direction I'll go for the "next gen" controller design. It will eventually have a parameter and reflow curve editor UI built-in (since the ATMega has enough flash space). It'll have 3 buttons instead of 1 - up, down and select.

TRR backpack controller pictures

Despite the error in the LCD wiring, I did get the backpack controller working. I put a hack in the library to rewire the interface in software to match the error, but once it's up and running, it's good to go.

The button is the 'start' button for the oven.

It functions identically to the older controller design. While it's running, the temperature on the left is the current actual temperature, and on the right is the current target temperature just below the time within the current reflow session (generally the set point drifts gradually across a phase from one point to another, with the different phases effectively representing different line slopes).

Design choices

I have a quandary.

The Toast-R-Reflow controller works really well... so long as you don't want to change any of the parameters of the reflow cycle without uploading new firmware. About all the ATTiny85 can handle is letting you push "start" and displaying what it's up to. If Atmel (seriously, you have GOT to go check out their website right now. It's got pictures of SOIC and TQFP packages and calls them "stocking stuffers." LOL) were to make an ATTiny165 - that is, an 85 with 16 kB of flash instead of 8, then that would solve an awful lot of problems.

But, in principle, I don't have to use the ATTiny85. I could swap it out for a TQFP ATMega328. The price is $2.80 instead of $1.20 (in single unit quantities. Not that you'd buy them that way, but it makes for a valid comparison) - which doesn't break the bank, but it is more.

What do we get for our trouble?

Well, we get a whole bunch more pins, most of which we don't need.

However, we may be able to do away with the MCP23017 i2c GPIO chip to drive the display. The LCD would take most of the digital pins, but so what? That would make up for most of the cost difference (that chip is $1.44).

What we really get out of the deal is quadruple the flash. That should be more than enough for a parameter / curve editor UI. Perhaps even one that could handle multiple time/temp profiles.

I might have to consider that...

Toast-R-Reflow power board pics

I built one of the v0.2 TRR power boards today. It all went together pretty well, although I do think that the QD terminals for the hot and out lines are uncomfortably close to each other and the triac heat sinks. The v0.3 boards will have a bit more room around them. 

These boards definitely are an upgrade from the prototype board. I'll be installing this one in my oven in exchange for the prototype tomorrow. 

Basic Toast-R-Reflow controller design validated

Well, the Toast-R-Reflow project is now officially "ready for prime time."

The design of the power and controller boards and functionality of the firmware have been validated in the field, and that means that if you want to play with reflow paste, I've got a pretty good starting point.

Here's what you need:

1. A toaster oven. I used the Hamilton-Beach 31138.
2. A K type thermocouple. SparkFun sells them.
3. An OpenEVSE LCD backpack. You can get those from the OpenEVSE store.
4. A Toast-R-Reflow power board. You can get one from my SquareUp store as a kit.
5. A Toast-R-Reflow controller board. You can one from my SquareUp store as well. Those come assembled and programmed because they're made from surface mount components, and if you're building a reflow oven that's a chicken-and-egg problem.

If you buy either board from my SquareUp store, it includes an instruction pamphlet as well. Of course, it's open hardware, so you're more than welcome to do it all yourself, if you prefer.

Now that the basic controller design has been validated, the next step is the all-in-one controller/backpack. Stay tuned…

Today is a very "meta" day

The board for the first version of the reflow oven controller that actually had a chance of working arrived. And so the reflow oven controller on the breadboard toasted its last board - it created its own replacement.

From what I can tell, the AD8495 is doing its job perfectly. The software may need some tweaking, though, because the temperature being displayed doesn't really jive with the voltage being presented on the V.temp test point.

The latest firmware has some code in it that attempts to self-calibrate the D/A converter. That may be getting in the way. It seemed to help on the breadboard version of the circuit, though.

The board is going to get a couple of tiny tweaks, but I think the really big deal is going to be the LCD "backpack" version of the controller - the whole thing will be all-in-one!

Note that the thermocouple wires are just sort of stuck in the holes. The connectors for that haven't arrived yet.