2-slice PBC toaster?

[KreAture]

I picked up a cheap 2-slice toaster today :)
The thought was to modify it to toast PCB's.

I'll be making a small EVR controlled circuit that will be used to monitor temp and regulate the power applied to the heating elements. It should be possible to make a quite ok ramp for SMT reflow work.

The toaster I got can handle up to two 100x100mm boards which is ample for my needs and the first tests with manually regulating the power worked great!

I will be posting schematics, pictures and code later as I progress, but the PCB layout will only be availabe upon request as this is a mains powered project. I do not want those with no experience/knowledge about such matters to attempt to build a regulation unit as the risk of injury/death is not to be taken lightly.

Modifying the unit for final use will involve removing the pringloaded slice-supports as well as fitting feet so it can lie on the side.

Attached is a picture of the toaster as well as a shot of a 0805 LED on a scrap piece of pcb I had. As you can see, the solder has flowed nicely (albeit a bit oddly due to lack of pads/mask) and the LED is not been singed...
Good enough for first tests :)

[rmteo]

You may aware of this... Thought you might find it interesting:
http://www.sparkfun.com/commerce/present.php?p=Reflow%20Toaster

[KreAture]

Thanks rmteo, I am aware of that.
It uses a relay, I'll use a triac. It also has some extra components I won't use.
Also, those ovens are big. My toaster is just a dinky little 2-slice. It only takes up as much space as a 10-pack of jewel-case CD's and it handles a 2-3 minute full reflow cycle using 600 watts.

My design will be a 8-pin ATtiny chip, (One with an adc for the thermocouple), and a simple power-by-mains circuit to supply power from the same feed as goes to the heater. It will use a 1-4 Hz pwm to avoid having to synch to ac mains and will use a simple temp-profile suitable for most solder paste.
I'll probably add more profiles as I go, to allow for quick small boards and large slower ones. Using longer preheat times to evenly heat the larger boards.

The controllerboard will replace the original board in the device, using the original mounting and I think I can even reuse the button. :D

[rmteo]

You're welcome. I only hope that your toaster will not pop the PCB clear across the room when its done. :D :D :D

[KreAture]

Hehe! That was the first part I disabled...

[Pete P]

After some range tests, of course?

[Acetronics]

Hi,Kreature

Elektor magazine modified a micro wave oven for that purpose ...

http://www.elektor.com/default.lynkx?pointer=1-28-16120-16381-16382-27272

Deluxe ...

Alain

[KreAture]

Alain, your link is to a toaster oven, not a microwave.

Anyway, I found my thermocouple!
Now I can start testing the temps :)

[Acetronics]

Hi, Kreature

You're right ... a Pizza oven !!! , looks like the microwave one in my kitchen !!!

Apologies, then ...

Alain

[kentuckyglider]

Pardon my ignorance, but could you explain in layman's terms what exactly a PCB is, and why you're toasting them? I have very little experience with electronics, but this caught my eye...

KG

[jeffs555]

KG,
PCB is printed circuit board, ie the things in your electronic equipment with all the little parts all over them. The rest is probably more than you wanted to know, but here goes.

In the old days, most boards were assembled with thru hole parts. The parts had wires which were inserted through holes in the PCB and then soldered. Most commercial boards were wave soldered. A wave solder machine has a vat of molten solder over which the boards are moved. A wave is induced in the solder to cause it to rise up and solder the bottom of the board.

Most boards now are built with surface mount components. These parts just sit on top of the board. To assemble them, the copper pads on the board are first coated with a solder paste, then the components are placed on the board, then the board is heated to melt the solder and attach the components. Commercial PCB ovens are rather expensive. Many hobbyists are using toaster ovens to heat the boards to melt the solder paste, but this is the first I have seen using a pop-up toaster.

Jeff

PS Has anyone here used a griddle as suggested at Sparkfun? http://www.sparkfun.com/commerce/present.php?p=Reflow%20Skillet#Skillet
I just bought one, but have not tried it yet. Hopefully I will be able to try some boards in a few days.

[Peter Seddon]

How do you ensure all the parts will stay on the board when the heat is applied?

Do you go through the full preheat etc. cycle.

Just curious as I've only just satrted to used SMD devices and hand soldering is painful!

Peter

[Acetronics]

Hi, peter

Never heard of cyanoacrylate glue ???

Alain

[jeffs555]

How do you ensure all the parts will stay on the board when the heat is applied?

The solder paste is somewhat tacky so the parts will stick, but you do have to be careful not to bump the board. When the solder melts, the surface tension tends to align the components with the pads, so their placement doesn't have to be perfect.

[andrewm1973]

Kreature,

How are you doing the thermocouple with an 8 pin AVR ? None of them have built in temp sensors for the ice point do they. Only one I know of with a temp built in is the Tiny44 that is a 14 soic. Or are you using external ice-point compensation.

I was thinking of doing my own oven controller too. I just thought the MAX6675 sounded a whole lot easier than wrtiting my own software for the TC :D

[KreAture]

Andrew,
actually I am not going to be that accurate.
It appears to me that most k-type sensors have measurement errors consisting of 2 things: offset and response. Offset is usually small, and response seems to be important when you need high accuracy.
Polynomials of 9'th order seem to be normal for high accuracy and 5-6 orders for lower accuracy. The base sensitivity of a k-sensor seem to be around 41 µV/°C. Staying below the Curie point of 354°C helps keep the world a simpler place too.

I will start by just using 41µV/°C as my formulae and the first profile I am implementing will be:

Time (s) Temp (°C) Probe (mV)
0 25 1,025
50 120 4,92
120 170 6,97
130 230 9,43
140 230 9,43
160 170 6,97
250 25 1,025

The targeted temps between each setpoint will be interpolated.

If someone has a good formula for a generic K-type probe they are welcome to find my temp errors based on the abowe voltages. Would be interesting to see how far off I'd be without offset compensation and response-adjustment.

Thanks to Jeffs for the nice explenation. Saved me the typing :)

To those still wondering why I am attacking a 2-slice bread toaster the ansver is simply it takes up much less space than a big toaster oven.

Also, as for the pin-issue on avr's...
Since I do not want to bother with the ice-point references and such I only need 1 or 2 ADC inputs to handle the one or two sensors I will be using.
I also need only 1 software pwm output for the triac as it will be running at a very low frequency.
So, that's: power, ground, reset, temp 1 in, temp 2 in, pwm out, start button, stop button.
Total of 8 pinns. Just perfect for a ATtiny13 that has an adc.

[Dan Baldwin]

The problem with using thermocouples without a cold junction (ice-point reference) is that thermocouples are not absolute temperature devices, rather they are temperature differential devices. The output voltage will be relative to the temperature differential between the leads, and the junction. If you have 15 deg drift in ambient temperature, you will have a 15 degree drift in the temperature you are trying to regulate. With very high temperatures, a 15 degree drift might not be significant, but at the temperatures you are using, it could be a problem.

Dan

[kentuckyglider]

Thanks for the answer, Jeff; it was just what I was looking for.:)
I'm majoring in computer engineering, but haven't gotten to the fun stuff yet.:D

KG

[KreAture]

Thanks Dan, that was what I was assuming.
For such simple DIY stuff I can live with 15 deg offsets maby even more.
However, using just 1 sensor I free a pin and could add a NTC to compensate for it easily. Not a bad idea really, considering the minimal cost it adds.

Anyway, I did a test today with probe a tad closer to heat than the board and checked for the reflow point of my paste. The paste is R276LF from Kester and it has a melting intervall of 179-189°C. Reflow would be higher and in my test it reflowed with probe showing 9.8mV meaning 239°C or, according to Dan, only 214°C due to my 25°C room temp that would offset the reading due to it being my interface-junction.
Since this probe wasn't in the right plane relative to the PCB (tad closer to lower heatsource), I assume the pcb did not reach this temp but only slightly abowe the melting point of my paste causing it to reflow.

As soon as I get the first prototype of the controller done I can start fiddling with profiling.

Like to point out the goals again:
1. SMALL reflow unit.
- Basically taking up as little space as possible while still doing the job.
2. 100x100 mm boards max
- Simply but, this is for small projects and not bulk processing. More like one-offs, and up to maby 10-offs.
3. Cheap!
- For me the major cost is the probe. Seeing how I happened to have one lying around it would be a shame not to use it...
- The toaster is very cheap but still looks good :)
4. Accurate enough
- Not a production unit
- As long as it can do most stuff without ruening the components or shortening the lifespan to just weeks, it will do.

So far, this is looking promesing!
Oh, and did I mention I am having barrels of fun? :D

[PlaneCrazee]

To find your offset you could use one of these (http://www.analog.com/UploadedFiles/Data_Sheets/630448615ad22103.pdf) to determine your ambient couldn't you?

[andrewm1973]

I reckon as soon as you need to add anything external to the CPU then you may as well just pop in the Maxim part. Sure its $10 but how much effort does it save especially for a one off.

If you are going to use the Tiny44 with the temp sensor inside then by all means do the TC stuff yourself. Adding a thermistor here or a 3 wire temp senesor there take more time than the nice little 8 pin IC that does it all :D

Equation might change a bit of you want to monitor multiple sensors - but still do the time v's money maths.

[KreAture]

I have done the ntc code many times so it will be copy/paste.
The NTC is $0.15 so I'll go for that if anything.

I am still looking for a practical triac... Too large triac and the base current becomes impractical. Too small and the current rating won't do.

[xtal]

Why not use a surplus solid state relay....

[KreAture]

Relays are a good idea but I don't have any 2-phased ones.
The 1 phased ones I got wouldn't fit in the toaster.
Besides, I really like the solid state stuff as it's small, reliable and won't wear down the contacts if I decide to do something synched to the phase of the supplied power.
As of now I don't know how accurate I will need this to be, now how accurate control I need to use to achieve it. I can get a 600v, 25A triac for only $5. I think it's a good bet. BTA140B-600 by Philips.
A nice thing about solid state: No clicking...

[kfong]

Great idea on the toaster, I'm interested in the final results. If you need accuraccy you can use this board.

http://www.embeddedtronics.com/signal.html

Thermocouple Amplifier
Based an Analog Devices AD595 cold junction thermocouple amplifier chip. You can generally get samples from Analog Devices.

Kin
www.embeddedtronics.com

[Dan Baldwin]

I have done the ntc code many times so it will be copy/paste.
The NTC is $0.15 so I'll go for that if anything.

I am still looking for a practical triac... Too large triac and the base current becomes impractical. Too small and the current rating won't do.

Mouser electronics shows a sensitive gate triac that is rated at 8 amps and 600 volts (Mouser PN 863-MAC228A8G). Max gate current for turn on looks like 10 ma. Would that work for you?

Dan

[KreAture]

Dan,
That would be perfect!
The triac I found with the 25A capacity has a very high max gate current of 35mA.
10mA would be much better. I will look into that one. My biggest issue is what I can get ahold of easily through my local EE suppliers though. I may design the pcb with optional component footprints so one can use devices available locally.

[xtal]

You could use a moc3041 or similiar [opto isolated triac] to turn on larger triac

RS may still carry them

http://www.rs-components.com.au/electronic-components-au/86418-triacs-opto-isolator-opto-isolators-low-holding-current-triac-moc-3041-opto-isolator.html


The solid state relay I was referring to is an encased opto isolateded triac..
of course single phase

[Comatose]

I don't want to discourage experimentation, but for prototypes I've had very good success for years with a $25 Milwaulkee heat gun from Home Depot. Put it on "low" and preheat the board for maybe a minute from 2" away, then turn it to "high" and reflow. Takes a couple minutes.

Of course, we have a real reflow oven, but sometimes its not on and it takes the better part of an hour to heat up.

[KreAture]

I have a heatgun. It works great for everything except some of my mlf parts that want to fly away and the 0603 LED's that tend to roast gently and turn a yellowish brown.

[KreAture]

Got my triacs today. Will have some fun tonight :)

[KreAture]

I have to confess I am a DC/digital guy and really hate working with AC.
I am however trying to make this somewhat safe so I wish to switch both the neutral and the live wire for the toasters heater. (Being that the neutral wire here tends to be anything but neutral so you can easily find 100v AC on it. I think it's because we get 2 of the 3 phases distributed so none of the phases are "silent".)

Anyway, I have drawn up an idea I am wondering if would work for a resistive load. Or atleast have a chance of working with some modifications.

See attached schematic.

Basically the avr is fed via a simple transformerless supply with a resistor/diode limiting and rectifying the initial voltage before a resistor/zener/cap is used to deliver 5.1v to AVR. The AVR will be working with long on/off periods covering multiple AC cycles each for the power control. T1 is controloled directly while T2 is controlled through T3. Heater is connected between H1 and H2. The idea being that both H1 and H2 will be dead when AVR is not appklying power to allow stupid people with metal objects being poked inside the toaster to keep their life and sanity. S1 will be used for starting the reflow cycle as well as aborting it at any time during the cycle. I will add ISP header later but for now I am focusing on the actual functionality and basic design idea. The actual Triacs I got ahold of were MAC16M and it doesn't have specs for Q4 so I may be out on the ice already.

I'd apprechiate any feedback from people fermiliar with triacs and AC switching as this is NOT my field. Help me get this right the first time...

[KreAture]

Nobody ? Where's the experts when you need em :O

[Dan Baldwin]

I've never used triacs, but I think you'll need a current limiting resistor between T3 and the gate of T2.

Why do you have the 1 ohm resistor after the diode in the 5 volt power supply?

Dan

[andrewm1973]

Sorry - I have always been scared of 240V AC so have always used the old motorola optocoupled triac drivers. So can't help you there.

On the input side are you connecting directly to Active and Neutral or do you have a capacitor in-line ?

[Comatose]

Why use triacs? Rectify it and use a MOSFET. Its a toaster, it doesn't care if its being fed AC or DC.

[KreAture]

Comatose, you have a point.
Hmm. It would change the power to the heating element though.
Also, from my understanding high voltage DC is more dangerous than AC due to how it doesn't allow your muschles to let go. I might be wrong but that's always been my understanding.

[KreAture]

I've been pondering...
Parhaps using opto-isolated triacs will simplyfy matters? They do not have polarity on gate and have the LED portion seperate from the nodes. They would allow me to run them both from TTL levels.