Author Topic: Reflow Controller Design - Comments, Critiques, Suggestions, Corrections  (Read 1388 times)

0 Members and 1 Guest are viewing this topic.

Offline nuktuk

  • Newbie
  • Posts: 3
  • Country: us
So this is my first post to the eevblog forum and I'm hoping I can get some feedback on my reflow oven controller design. It is a standard DIY toaster oven reflow oven using SSRs, a microcontroller, and a thermocouple.
I decided to attempt to make my own board for the fun of it, and because I think it will be good hardware design experience.

Please don't hold back on your suggestions! I want to learn new/better ways of doing things so any feedback is appreciated.

Also, if you have any suggestions for making the PCB please include those too!

Thanks!

PS. Sorry about the zip file. I didn't know how else to skirt around the 1MB upload limit, and still preserve the ability to zoom in on the circuit.
 

Offline Ian.M

  • Super Contributor
  • ***
  • Posts: 8067
To crunch the filesize for schematics,  you need to reduce the colour depth to 16 (4 bit) and save as PNG.  Irfanview saves a 3261x2059 greyscale 4bpp PNG of the schematic copied from your PDF with a filesize under 250K

You'll need level shifters between the 3.3V MAX31855 Thermocouple interface and the 5V ATmega328P.  Don't forget to make the MISO one tristatable, controlled by the MAX31855 /CS  so it doesn't conflict with the display.  Check whether the display uses 5V logic levels - Raspberry PIs use 3.3V levels so your direct connection to the ATmega328P is suspicious.

It would probably be worth adding additional SSR control outputs so you can support control of an oven fan for more even temperatures during rampup and soak and extra boost elements for faster peak reflow.

Consider supporting multiple thermocouples with additional MAX31855 chips for better temperature control. There's no need to populate all the MAX31855 and Type K connector footprints until you need them, but I'd put four footprints on the board with a 74HC138 3 to 8 line decoder to drive the /CS pins from only three I/Os to select the SPI device to access + master /CS to a '138 enable pin.   

Also add a header with all the spare ATmega I/Os on it, the SPI bus, all spare '138 /CS outputs, 5V, raw DC in, and Ground for future expansion.
 
The following users thanked this post: nuktuk

Offline nuktuk

  • Newbie
  • Posts: 3
  • Country: us
You'll need level shifters

Oh man! I completely forgot about the different logic levels! Thanks for the headsup!

I am already adding an extra SSR control because the oven I'm looking at only has two main elements, but its a good idea to add more just in case I decide on a nicer oven.

I had thought about adding a slot for a second thermocouple to get a better ambient temperature reading but have been debating the idea. I think the suggestion to add one probably settles it, and besides, as you said, I could always decide not to populate it.

I like the idea to add headers to unused pins. I might want to use them one day.

make the MISO one tristable

Didn't know about this. Thanks for the tip, I'll need to readup about it a bit more and make sure I get that right.

Thanks again for the feedback, I really appreciate it!
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 6338
  • Country: de
The way the SSR are driven does not work this way.

With more thermocouples i would consider separate amplifier, cold junction sensor and ADC instead of the special TC chips.

A second TC might be useful to get separate readings for the furnace surface and directly at the board.
For just environment temperature a diode as sensor might be enough.
 

Offline Ian.M

  • Super Contributor
  • ***
  • Posts: 8067
The way the SSR are driven does not work this way.
Whoops!
I didn't notice they were wired to crowbar the supply!
The SSR 'coil' should be in series with the MOSFET drain.  I'd feed the coils from the raw DC input voltage, after a polyfuse in case some idiot shorts something.

I'd also stick a 0.1uF ceramic across C1 to improve HF decoupling and add local decoupling for each MAX31855.
 

Offline nuktuk

  • Newbie
  • Posts: 3
  • Country: us
Holy crap, I can't believe I didn't put the SSRs in series with the drain  :palm:

That's why I posted though, I was sure I made some dumb mistakes and yall have been a big help.

I'll have to add the local decoupling for the max31855s. I didn't already have it because I was originally using a different 3.3V regulator and it used a small ceramic on the output anyway.

I will look into individual amps and sensors for the TCs instead of a bunch of max31855s. It would certainly be more cost effective
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf