Reflow oven first tests with questions...

[mjkuwp]

thanks for sharing your project in so much detail! : )

i recommend having a look at the recommended maximum TAL (time above liquid) for the solder and also heat tolerance of some of your parts.  I feel like you should be getting the boards cooled down faster.  What I typically do is open the oven door at 215C to 225C and aim a fan into the oven to force the temp down.  Even just opening the door does not bring the temp down fast enough.

so far as the controller goes...

If you have a look at the first plot in my write-up you can see that sometimes this PID loop boils down to... full on - light duty - full on.  OFF.    however, i still like having the PID loop.  The controller is helping to compensate for the lag in response of the elements.

http://mjkuwp94.tumblr.com/ReflowOven


I have a moderate sized Black & Decker convection oven purchased at Amazon or Wal-Mart in USA.  It is 1200W out of the box but I shortened the heating elements and boosted this to about 1700W.  I also insulated the oven and even with these two efforts it will barely do 2 deg C/ second.  I wonder what rate you are achieving when you target 3C per second?

I get bad readings from my thermocouple and board as well so I just made sure to write code to help ignore them.  I think this was faster than trying to chase down the hardware problem - at least in my case.  I have done more than  a dozen boards and they have all gone well.    Most recently I have been holding the thermocouple to a scrap pcb using a combination of aluminum tape and kapton tape.  One time I used wrapping wire to tie one in place.   I definitely got the best results in my first few boards on which I mounted the thermocouple directly using extra-high temperature solder.  however it is far too tedious to do that every time.

I would be curious to know your oven wattage and interior size.  It would be interesting as a comparison to my own oven.

[alank2]

mjkuwp - I have the same black and decker oven you do.  How hard was it to shorten the elements?  What did you use to insulate it?

[mjkuwp]

9379K93 from McMaster-Carr

Ultra-High Temperature Foil-Faced Insulation Strip

and also a chunk of stovepipe wire from the local hardware store to knit it down and into place.

I should have mentioned that if I had to do this again I would try to make the oven work without doing the insulation and I would test it also without boosting the elements.  It was a lot of labor to do those things and I'm not sure it is worth the time.  this of course depends how much time you have.

I got the idea for shortening the heater wires from here:

http://www.buildlog.net/forum/viewtopic.php?f=2&t=1013

It was fairly easy to do the shortening.  Just involves disassembling things carefully.  The wires have to be terminated by crimping so you have to have some good crimp connectors and tools.

[alank2]

How important is it to measure the actual boards being baked?  Is it workable to just pick an old board roughly the same size and attach the thermocouple to it permanently so you don't have to clip it on and off other boards?

[mjkuwp]

How important is it to measure the actual boards being baked?  Is it workable to just pick an old board roughly the same size and attach the thermocouple to it permanently so you don't have to clip it on and off other boards?

yes, this can work.  It is what I do now.  Since I started using a 'donor' or 'sacrificial' board I noticed that the temperature would be off by about 10C.   I keep on eye on my temperature reading and note when the solder actually melts.  In my head I do the math and make an adjustment to when I open the door on the oven.  I know this is not perfect but it has worked well so far. 

[alank2]

Well, I baked my first board, and it works perfectly!

I only had some really old solder paste to use (probably >2yrs old), but I tried to rejuvenate it a few months ago with some liquid flux when I got my pneumatic dispenser.  My only complaint is that it puddled out even before I could place components on it - I couldn't see the soic pads any more, but it still lined up.

I'm going to try my hand at some stenciling with some newer paste soon.

[StrangeloveMD]

Awesome, alank2.  You can probably dry your current paste out a bit, and carry on using it.  Still, it would be nice to have a fresh tube around.  (You know, just to eat up its shelf life )

mjkuwp, your efforts were probably worthwhile.   If nothing else the insulation will effectively help to even the heat distribution in the oven body.

It is nice to be able to jump to your peak (air) temp right when you hit liquidus, but heating is generally less of a concern in hobby ovens, compared to cooling.  Cooling rate, meaning everything after the "peak dwell" period down to liquidus point, is going to determine the internal structure of the solder joint.  Cooling too quickly/slowly may lead to cracked joints down the road, especially on devices "gettin through dem thermal cycles".  (rapid and/or extensive thermal cycling.)  Of course, this is all assuming you've met a profile that doesn't cook anything to begin with.

Also, I made a post a bit further up with some links that may warrant a second glance.  I've had to go back over them a few times in working on my oven (in slow progress), but there have been a few questions asked here since (board vs air temp, &c.) that are at least partially covered.

P.S.: I am a bit jealous of your controller UI, mjkuwp.

[alank2]

My measurements so far show it will heat around 1 deg C per second and it cools with the door open at 0.35 deg C per second.  Do I need to consider doing some things to improve these?

[StrangeloveMD]

I'm not sure if you need to, but my understanding is it will give you a much better joint, especially on the long term.  (Ahhhh!  Thermal cycling!)  Of course, the more control you have over either of these rates, the more capable your oven will be, especially in the future.  I haven't looked into the window you have to hit on some of the "high pin count" bump packages, but I can't imagine they enjoy being thrown in a blast furnace, as it were.

Cooling rates are usually recommended in low single digit Celsius. 2*C/s to 4*C/s, with "more than 6ish" seeming to get fairly dangerous.

But what is cooling down at that rate?  Air, PCB, joints?  Is that a window you must get through or a range in which the value you must "hit" lies?  I can probably clarify if you have some questions on the above links, as I am just barely moving on from them to some slightly more in depth stuff.

[hamster_nz]

High cooling rates cause issues with the grain in the solder, making the joint brittle, and more likely to fail due to mechanical shock.

http://www.ami.ac.uk/courses/topics/0244_tsm/ has quite a lot of interesting info, and on a side note http://www.ami.ac.uk/courses/topics/0132_rs/ is a quite interesting read about the reflow process in general..

[alank2]

What about low cooling rates - mine has been averaging 0.35 deg C per second, is that too slow?

[StrangeloveMD]

What about low cooling rates - mine has been averaging 0.35 deg C per second, is that too slow?

hamster_nz has provided some great direct links to information on over/under cooling.  I specifically recommend the "Self Assessment Questions".

Also, the welders have been going on about this sort of thing for ages: http://www.aws.org/wj/supplement/WJ_1972_11_s536.pdf

For anyone looking for a simple answer:
Undercooling (providing nothing gets cooked) is probably not as dangerous as cooling too quickly (>6C/s seems an absolute max, but I'm still reading) in terms of likelihood of damaging a board.  I would say leave it in the oven a lot longer, though.  The "slowly cooled" joints will want to be nice and stationary,  asforto grow whatever structure they can, prior to encountering loads.

At the end of the day, one simply reflows.  If the board works, one has succeeded. 

[alank2]

Hi Everyone,

How does this look?

I'm trying to get it to match this type of solder / profile A:

http://amtechsolder.com/pdf/4300_5Fspec.pdf

135s in <140
140s in 140-180 range
70s in >183



What do you guys think?  Any recommended changes?

[hamster_nz]

I think that is looking pretty good - did you reflow any components?

The one thing I am not sure about is the climb to the peak temp. IIRC you want to take it to peak temp pretty quickly (at up to 3 degrees per sec IIRC), and then drop off once everything should have reflowed. Looking at your controller output it is aiming for a soft arrival at the target peak temperature.

A soft arrival at the initial heat/ramp phase is what you want, but for the peak it keeps your components hotter for longer than they need to be.

Maybe when you jump your set point to 210, instead jump it to 250 for 50 seconds, then back to 210. That should steepen up the climb, rather than it backing off  after 25 seconds (between 275 and 300).

Also, your peak temperature is looking a little low. Shouldn't it be around 220? (as per the PDF).

[alank2]

I think that is looking pretty good - did you reflow any components?

Not yet with this profile, but I did do a board the other day and was very pleased that is worked!

The one thing I am not sure about is the climb to the peak temp. IIRC you want to take it to peak temp pretty quickly (at up to 3 degrees per sec IIRC), and then drop off once everything should have reflowed. Looking at your controller output it is aiming for a soft arrival at the target peak temperature.

I'm using my own controller code.  The final stage "reflow" changes the SP to 205 and the stage ends when it hits 205.  I thought about the quick jump, but the CO does bump to maximum so I figured I was pushing it hard enough to try to get there and then let it cool.

Also, your peak temperature is looking a little low. Shouldn't it be around 220? (as per the PDF).

I'm being a bit conservative, don't want to burn or pop anything but still get the job done.  Do you think I really should get it up to the peak temp?

Thanks,

Alan

[mjkuwp]

yes, you can go ahead and let your board get hotter.  Caveat being that you make sure to consider that the feedback is not directly on your pcb.   

 Many or most components have been designed to handle lead-free solder so you have a pretty big window.

I think your profile is looking really good.  : )

[alank2]

Last night went pretty well.  My first stenciling was great, a friend of mine gave me some fresher paste.  Wipe the paste across, scrape the excess, next pcb.  Surprising how fast that is to paste a board.  The tough part for me was placing parts without smearing the paste, hopefully practice will improve that.  I baked 10 boards and tested one so far and it works great.  All in less than 2 hours.

[alank2]

Here is the latest profile I have been testing.  I finally added graphing to my pid control application:

My oven can only heat at a maximum of 1.1 C/s - will this be a problem and do I need do think about tweaking it for more heating power?

I'm trying to follow profile A fro this:
http://amtechsolder.com/pdf/4300_5Fspec.pdf

My oven spends 160s in 0-140 deg, 105s in 140-180 deg, and 95s in 183-218 deg...



My profile:

Code: [Select]

  //{SET_AUTO,        "Optional",                         SP,           MaxChange,    Fan           },
  //{DELAY_THRESHOLD, "Optional",                         0(<=) 1(>=),  PV       ,    Seconds       },

  {
    {SET_AUTO,        "Stage 1 - Preheat",                140,          0,            1             }, //sp=140
    {DELAY_THRESHOLD, "",                                 1,            140,                        }, //wait for pv>=140

    {SET_AUTO,        "Stage 2 - Soak",                   170,          0,            1             }, //sp=170
    {DELAY_THRESHOLD, "",                                 1,            140,          90            }, //wait for pv>=140 for 90s

    {SET_AUTO,        "Stage 3 - Reflow",                 230,          0,            1             }, //sp=210
    {DELAY_THRESHOLD, "",                                 1,            210,          10            }, //wait for pv>=210 for 10s

    {SET_AUTO,        "Stage 4 - Cool Down (Open Door)",  0,            0,            0             }, //sp=0
    {DELAY_THRESHOLD, "",                                 0,            70                          }, //wait for pv<=70

    {END,             "Completed",                                                                  }, //boards can be touched
  };


[mjkuwp]

In declaring that your oven can do 1.1 C/s do you mean that the board heats up at this rate?  Not a thermocouple inside the air of the oven?

I will assume you have attached to the board.  My oven has been boosted and has insulation inside it and in the walls.  It can heat up a board at about 1.5C per second in the first stage (as it approaches the soak).  I put foil-backed insulation inside the oven in an attempt to reduce the volume.  The foil on the insulation hopefully reflects energy back to the board. I also think it is important that the foil has less mass than the oven walls so it tends to have the affect of reducing the thermal mass of the oven.  I also have insulation between the walls of the oven and the decorative outside sheet metal but I am not sure what that really does.

All that work... and I only got 1.5C versus 1.1C.

My suggestion is that you change your profile to support 1.1C rise and see where that gets you.   

or.. Maybe go full on power until 80C and then try to have a gradual rise from 100C to 150C by the end of your soak period.  At the end of soak or at 150C go full-on power again.   {Between 80C and 100C there is thermal inertia of the elements as I am sure you have seen}

I think the above is about what my PI controller ends up doing.

[alank2]

In declaring that your oven can do 1.1 C/s do you mean that the board heats up at this rate?  Not a thermocouple inside the air of the oven?  I will assume you have attached to the board. 

Yes, it is a thermocouple attached to a board.

My oven has been boosted and has insulation inside it and in the walls.  It can heat up a board at about 1.5C per second in the first stage (as it approaches the soak).  I put foil-backed insulation inside the oven in an attempt to reduce the volume.  The foil on the insulation hopefully reflects energy back to the board. I also think it is important that the foil has less mass than the oven walls so it tends to have the affect of reducing the thermal mass of the oven.  I also have insulation between the walls of the oven and the decorative outside sheet metal but I am not sure what that really does.
All that work... and I only got 1.5C versus 1.1C.

Were you the one who shortened the elements so they would run hotter too?  That did look like a lot of work.

My suggestion is that you change your profile to support 1.1C rise and see where that gets you.   

I may play around with it some more, but it somewhat adapts itself to what the oven is capable of.  I set it up so that it will stay in a step until the step is satisfied (for example the soak from 140-180 is 60-90s and I set it to stay until 90s above 140 is obtains, so while I can't get the oven hot enough fast enough to do it as quickly as they do, I do make sure the step has been satisfied with a delay that says we need to be above 140 for 90 seconds consistently before it moves on.

Thanks,

Alan