I just tried to reflow my first boards this weekend and I had mixed success. I used a hotplate, varister temperature probe and a cast iron skillet.
I did a test setup with a piece of pcb with a blob of solder paste in the middle and the temp probe touching the copper layer and cranked up the heat.
The reflow profile I used is:
Preheat: 25C to 100C(212F) as fast as possible
Soak: 100C(212F) to 180C(356F) in 1-3min
Reflow: 180C(356F) to 220C(428F) as fast as possible but for no longer than 1min
I was going to do the smart thing and log when my paste began to melt but ran into problems.
I turned the skillet up to max and temperature rose to 100C in 3min and at 130C my solder melted very nicely then at 150C the board started to melt and reek! These temps are nowhere near the reccomended!
I use [this solder paste and maybe its special to melt lower (leaded melts lower I know) but if my pcb board is melting 70C before every reflow profile I have seen whats wrong?
The pcb material I use is MG Chemicals copper clad I have 1/16 and 1/32 0.5 and 1oz copper layers. All my tests where on the 1/16" 0.5oz boards as this is what O use most commonly. I figured as FR4 is an insulator the thicker boards would take longer to heat up the top copper but would burn at the same temperature…
I am putting together a SCR based controller and don’t know if I should mod this hotplate or go get a toaster oven instead (overhead elements to heat copper directly)
I think your temperature measurement is WAY off, or the temperature probe you are using has a very long thermal time-constant. Solder does not melt at 130 deg C (more like 180 deg C minimum). All fiberglass boards I’ve ever used could tolerate much higher temperatures than what solder melts at.
A suggest you get a cheap non-contact thermometer - Ebay has lots of cheap units that would do the job.
I agree with Sparkfun that a skillet is better than a toaster oven in many cases (heating from the bottom has advantages, especially when you have large components like connectors that are difficult to solder in a toaster oven):
I have no experience with these things, but why would you go that much hotter than the melting point of the solder. Seems like once it’s melted it should be ready to ramp down?
I’ve been using hot air reflow on MG Chem boards with no issue; I know that doesn’t help much.
MichaelN:
I think your temperature measurement is WAY off, or the temperature probe you are using has a very long thermal time-constant. Solder does not melt at 130 deg C (more like 180 deg C minimum).
My temp probe is from a digital kitchen thermometer with the casing removed to expose the little bead that does the sensing. Its seems pretty active, if I turn up the heat it changes rapidly and when I take the pan off the heat the temp only goes up one or two degrees before heading back down so I figured it was pretty accurate.
I don’t have a non contact thermometer as I didn’t know how I would incoperate one into a reflow ovens control setup
Most people don’t bother with automation when it comes to skillet reflowing. There’s NO way you should be seeing those affects on the boards with the temperature profiles you mention; the fact that solder is melting when you’re measuring 130 deg C suggests your measurements are not accurate.
MichaelN:
I think your temperature measurement is WAY off, or the temperature probe you are using has a very long thermal time-constant. Solder does not melt at 130 deg C (more like 180 deg C minimum).
My temp probe is from a digital kitchen thermometer with the casing removed to expose the little bead that does the sensing. Its seems pretty active, if I turn up the heat it changes rapidly and when I take the pan off the heat the temp only goes up one or two degrees before heading back down so I figured it was pretty accurate.
I don’t have a non contact thermometer as I didn’t know how I would incoperate one into a reflow ovens control setup
That doesn’t mean that it is accurate! Have you checked the calibration?
I have no means of checking it because im still in college and on EI, I wont have much of any space change for a $50 IR Thermometer. Ill check ebay but wouldn’t I have the same temperature calibration issues on an uber cheap non contact?
I think my boards are burning because on the top end the burners get very hot very quickly, I mapped the temperature rise of the pan by itself and it was under 1C/sec and rose to slightly above 1C/sec on the top end. Maybe the temp differential between the pan and the top copper layer is greater than the FR4’s rating.
buzzard192:
Stick the cup in the freezer with the probe in it. When the water is frozen, what does it say the temperature is?
Freezers typically reach as low as -20 deg C, so this could be inaccurate. I suggest putting a bunch of ice cubes in a glass of water - that way, the ice will be melting when you’re taking the measurement (ie at 0 deg C).
MichaelN:
Most people don’t bother with automation when it comes to skillet reflowing.
I second this comment. I’ve done many personal PCBs and in my old job the assembly people did a TONNE of boards with hot-plate reflow for low quantity prototype work, including working with some fine pitch devices, tonnes of QFN, etc… We never worried about the profile - As long as the hot plate is regulated to the appropriate temperature, it worked with a VERY good success rate to put the PCB(s) on the hot plate, wait for it to re-flow, and once all the solder had re-flowed, take it off to cool down…
MichaelN:
Most people don’t bother with automation when it comes to skillet reflowing.
I second this comment. I’ve done many personal PCBs and in my old job the assembly people did a TONNE of boards with hot-plate reflow for low quantity prototype work, including working with some fine pitch devices, tonnes of QFN, etc… We never worried about the profile - As long as the hot plate is regulated to the appropriate temperature, it worked with a VERY good success rate to put the PCB(s) on the hot plate, wait for it to re-flow, and once all the solder had re-flowed, take it off to cool down…
Yep, I have to admit, when I am building with the hot skillet, I use eutectic solder paste, I put the board in the skillet and I crank it to the top. I just make sure to watch it closely and look for everything to flow. Once it flows, I cut the skillet off and let it cool.
On occasion, when I have somthing like a metal connector that won’t flow, I slide the board around a bit to get that part over the heat coil and direct the heat a little more. That always gets the “stragglers” to flow.
From the sounds of everyone this is a pretty simple thing to do… Has anyone used a cast iron skillet as the heated surface? My hotplate has electric stove elements and I chose the skillet as they give very even heat distribution but I’m wondering if somehow that my be my issue. I’m going to test the calibration of my meter tonight and mock up a smd board and try to reflow with less aggressive heat and a standard surface.
How long can FR4 stand the heat? Components can not handle an extreme increase in heat and I imagine the FR4 could only handle say 15min at high temperatures.
EDIT:
All right I tested my temp probe and it is plenty accurate. 99C(211F) to a rolling boil and 1C(33F) when frozen inside an ice cube. Ill give my board a try in an aluminum pan tomorrow.
Should I try again with my 1/32" 0.5oz board or up the size to 1/16" 1oz?
MicroMinded:
From the sounds of everyone this is a pretty simple thing to do… Has anyone used a cast iron skillet as the heated surface? My hotplate has electric stove elements and I chose the skillet as they give very even heat distribution but I’m wondering if somehow that my be my issue. I’m going to test the calibration of my meter tonight and mock up a smd board and try to reflow with less aggressive heat and a standard surface.
How long can FR4 stand the heat? Components can not handle an extreme increase in heat and I imagine the FR4 could only handle say 15min at high temperatures.
I suspect that cast iron is not the best material, as it will take too long to cool down. I use an alumium skillet (similar to the one Sparkfun use in their turorial). Even that doesn’t cool quick enough for my liking, and I often put a fan or similar to cool it down quicker. Remember that you are trying to approximate the soldering temperature profiles that semiconductor manufaturers recommend.
Well over this week I have tried different materials with varied success.
I have tried aluminum and cast iron on full spiral electric elements, a wrap around electric skillet and in an old toaster oven. Sofar it seems that the cast iron actually gains an extra 10deg after turning the power off where as the aluminum only gains ~2deg so I am using aluminum now. The full spiral element worked much faster than the skillet as it has much more heating surface area and I did not run into cold spots.
I think im going to goto wallmart and pick up a $30 1300w toaster oven and move the bottom element to the top and put my SCR controller on it for better temperature control. By putting both heating elements on the top I should eliminate my problems of fickle burning.
Out of 7 boards I got two that did not burn and one the solder did not get hot enough to properly flow around an SOIC part so only one that actually work.
Hopefully I will have luck with the toaster method but I go back to school next monday so I probably wont have much for updates for a while.
A couple of people have suggested using a toaster oven. I’d strongly suggest that as well. Toaster ovens with convection (which you want for even heat distribution) cost $40 and (in my experience) work quite well. I have not tried lead-free, as many lead free solders require higher temps (like 260C) than standard toaster ovens offer. I do prototypes with lead, then do volume production lead-free. At some point I’ll get a hotter over for lead-free prototypes, but frankly there’s not a very compelling reason to do so (IMHO) for prototypes.
I use a temperature probe with my multimeter to keep an eye on temperature, but anything that can read over 200C is fine. You stick your boards with components in the oven, turn it up to 450F (usually as high as they’ll go), and wait until the temperature passes 200C, which most ovens will do in a fashion that’s compatible with a standard process profile. I don’t think you need a closed-loop controller. It’s not that finicky. Then, after reaching 200C in a couple of minutes, you turn the oven off. You let it dwell at 200-205 for just 10 seconds or so, then crack the door to begin the ramp down. Once it’s back below 160C, you can cool it more rapidly. The entire process takes just a few minutes and I have yet to botch a batch. You can put the board back in upside down, and do parts on the other side, too (the parts you already put on the other side don’t fall off, believe it or not). This works with very small parts like 0402 and fine pitch, lead-less packages like QFNs quite well.