Finally starting on this convection oven -> reflow oven conversion project

[fourfathom]

While I haven't added insulation to my oven, I do think that if you add it to the inside of your chamber this will accelerate the cool-down phase, and this is a good thing.  My oven is a bit slow on the cool-down, even with the door open (fanless), so using insulation to reduce the thermal mass inside the chamber can only help.

[coppercone2]

well I think its a bit of a trade off here, if you put it on the inside of the oven you loose oven space but you insulate the ovens thermal mass, but if you put it between the walls you maintain space but you still have the inside structure of the oven so its not as good (the air space in there is a somewhat decent insulator so it won't be magical.

What insulation is good for being thin, and clean? I thought about this but I am worried about dust contamination, particularly with the fan.

And the bottom of the oven typically will require the insulation. And I also wonder, what does it do for the weld joints that like to break in the oven? If they are insulated inside of the wall, won't they heat up more and distort more? The bottom and back wall IIRC are not doubled walled, but the side and top are. And if you insulate inside sidewall I think you mess with the thermostat, I don't think the temperature corrosponds well so if you want to maintain the 'dumb' features of your oven as a safety it will be bypassed.

I think insulating the interior of the oven will be the best. I thought graphene sheets but  I heard they are super messy, so you how to seal the seams? silicone? And if you insulate the inside not the wall, should you add some vents to the inner wall to allow the crimps etc to cool down a little? But if they run at a lower temp it might be better to leave it alone. lots of probing to test this IMO

i would test this stuff but taking the oven apart is such a hassle I don't want to lol, if I keep taking off that lid something is gonna break.

the temperature you achieved in the side chamber is a game changer, so long the crimps or welds are not under higher temperature

[mindcrime]

OK, just got done adding a bunch of insulation to the inner chamber, and doing another side-chamber test for 6 minutes.







Not a perfect job, but it makes a difference. I ran the 6-minute "full on" test again and now the side chamber temperature peaks at 64° C. That's down to within the specified range for both the SSR and the transformer (both list peak temperatures of 70° C). And I don't expect to see even the 64° during a typical soldering run. So right now I think I'm in pretty good shape with regards to that stuff.

Now I'll be really curious to see what the temperature curve inside the oven itself looks like. I'm probably going to kick off another test run of that here in a few minutes.
 

[mindcrime]

What insulation is good for being thin, and clean? I thought about this but I am worried about dust contamination, particularly with the fan.

Time will tell how "good" it is I guess, but here's the insulation I went with:

https://www.amazon.com/gp/product/B08GG22QHC/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1

the temperature you achieved in the side chamber is a game changer, so long the crimps or welds are not under higher temperature

It's even better now. On my last 6 minute run, the temperature in the side chamber peaked at a comparatively balmy 64° C.

[mindcrime]

OK, here are some more results. With the insulation inside the main chamber, I did one of my "run up to 250° and then shut off" tests. It took about 230 seconds for the oven to reach 250° this time, compared to about 181 seconds from the last run before adding the insulation.

"Wait", you're saying, "why did it take longer to get up to temp with the insulation as opposed to without it?"  Well, I can't be 100% sure, but the tests weren't identical. For one, the earlier run had the inner chamber at 32° when the run started, while the one just now had it at about 26°. I may also have had the thermocouple positioned slightly differently, making it not an "apples to apples" test. There are probably also other variables I'm not considering.

OTOH, there was noticeably less overshoot this time, which is good.

All in all, I'm fairly satisfied with this. I still need to write code to implement an actual reflow profile, but assuming I have the coding skill to do that (not a guarantee, but I'm a better coder than I am electronics engineer, or thermal engineer) then I think this is going to work out quite well.


Edit: looking at the data, on this last run it took ~30 seconds to go from 26° to 32° degrees. That alone represents the bulk of the delta between the two runs. I think that 181 second run was one I did where I didn't let the oven finish cooling down fully from a previous run, so it's a little bit misleading.

[mindcrime]

An early stab at prototyping what the control box might look like:

[mindcrime]

OK, latest results:

1. Added more insulation to the left hand side panel of the inner chamber. Did a "up to 250° C" ramp-up / ramp-down test from a cold start and found that it took ~230 seconds to reach 250°. That is consistent with my earlier observations when doing it with the oven cold. By and large, adding insulation to the inner walls of the cooking chamber did not seem to have a ton of impact on the "time to reach 250°" performance, but it did have a noticeable impact on keeping that side chamber cooler. So net-net, adding insulation there seems to be a win and a generally desirable thing.

2. Suspecting that the glass panel on the front door was a significant source of heat loss, but not wanting to permanently block my view into the oven, I made two slip-in heat shields out of two of those pieces of aluminum flashing and some of the insulation. The flashing pieces coincidentally turned out to be exactly the right height, so no cutting was necessary. And there's a little ridge near the bottom of the door that the heat shield will catch on, which helps keep it in place while shutting the door. Of course I could bodge in some Kapton tape to hold them in place if I wanted to.



These pieces just lie inside the door frame and are held in place by the door itself.



Reflecty stuff, cool.



The way they lean up against the frame. When the door is shut, they're locked into place well enough.



View from outside with the shields in place, and the door closed.

Did another 250° ramp-up / ramp-down test from cold start, and found that adding those brought the "time to 250°" performance down to 196 seconds. So yeah, it appears that the glass front is a major source of heat loss. Closing down that channel even partially helps by a decent amount.

Now to finish up that control box and get some proper software written for this thing.

On the software side, here's what I think I'm going to try for a first, "strawman" stab at this. Instead of dealing with PID and complicated algorithms, I'm going to take the data from one of my test runs, run linear regression on the side of the curve up to 250°, and use the slope as an approximation that's good enough to use heuristically. Then I'll do something similar to what user @fourfathom was doing, where I'll just rely on timing for the various states, based on the expected change in temp over time. I'm just going to hard-code the timings into my control program for a trial run, and test that approach and see how close I can get to matching a published "official" soldering profile.

I also just ordered a few more of those SOIC-14 to DIP adapter boards and 30 more NE556 timer chips in SOIC-14 format, along with a fresh tube of Kester 63/37 solder paste, so I can start experimenting on actual boards soon. 

[fourfathom]

Then I'll do something similar to what user @fourfathom was doing, where I'll just rely on timing for the various states, based on the expected change in temp over time. I'm just going to hard-code the timings into my control program for a trial run, and test that approach and see how close I can get to matching a published "official" soldering profile.

Looks good.  Have you done any door-open cool-down measurements with the new insulation?

You probably know this, but my method is temperature threshold-triggered timing, not just blind timing.  The ambient and starting chamber temperature still have some effect on the heating curves, but with the temperature thresholds the results are reasonably accurate.

[mindcrime]

Then I'll do something similar to what user @fourfathom was doing, where I'll just rely on timing for the various states, based on the expected change in temp over time. I'm just going to hard-code the timings into my control program for a trial run, and test that approach and see how close I can get to matching a published "official" soldering profile.

Looks good.  Have you done any door-open cool-down measurements with the new insulation?

You probably know this, but my method is temperature threshold-triggered timing, not just blind timing.  The ambient and starting chamber temperature still have some effect on the heating curves, but with the temperature thresholds the results are reasonably accurate.

Not sure if I realized that earlier or not, but that's good to know. I plan to try the dumbest possible approach first just to see what happens, and then gradually add complexity as needed to get a good track against the desired profile. If we're talking about the same thing, then "temperature threshold-triggered timing" sounds like what I'm guessing with be the best combination of "simplest thing that could possibly work" AND "achieves reasonably good compliance with the desired profile". 

[fourfathom]

Not sure if I realized that earlier or not, but that's good to know. I plan to try the dumbest possible approach first just to see what happens, and then gradually add complexity as needed to get a good track against the desired profile. If we're talking about the same thing, then "temperature threshold-triggered timing" sounds like what I'm guessing with be the best combination of "simplest thing that could possibly work" AND "achieves reasonably good compliance with the desired profile". 

Yeah, I mentioned this a couple of times, but probably not clearly.  I have three  temperature trigger points:
* First at 100 deg (start of soak).  When the oven hits this I turn the heater off for 60 seconds, then back on.
* Second  at about 215 degrees (set during calibration runs).  I turn the heat off at this temperature, and if it's going as planned the chamber temperature continues to rise until it hits my target of 235 deg.
* Third when the temperature begins to fall (I think by one degree).  I buzz the buzzer and open the door.  I could probably trigger this when the temperature hits the target *or* starts to fall without quite making it to the target -- that would speed up the cooldown by a few seconds.

So, I basically press "Start" and wait for the buzzer.

My post with the explanation and some plots:
https://www.eevblog.com/forum/projects/finally-starting-on-this-convection-oven-gt-reflow-oven-conversion-project/msg3573697/#msg3573697

[mindcrime]

Yeah, I mentioned this a couple of times, but probably not clearly.  I have three  temperature trigger points:

I don't think there was any problem with you not being clear enough... it's more that there has been a lot of information shared in this thread over the last couple of weeks, and I don't have all the details of everything that was said stored perfectly in memory. 

My post with the explanation and some plots:
https://www.eevblog.com/forum/projects/finally-starting-on-this-convection-oven-gt-reflow-oven-conversion-project/msg3573697/#msg3573697

Yeah, I remember that now. I'm definitely planning to consult that as a reference as I start getting serious about implementing the profile logic.

[mindcrime]

Project Update 06-04-2021

Earlier today I made my first test run with the software attempting to do something that at least roughly approximates the "Kester Standard Profile" as seen here. And while the results are far from perfect, you can start to see the rough outlines of how this can be molded to match their curve.

In the end, I decided not to bother with hand-coding the timings, and do something closer to what @fourfathom was talking about. I may not have done it exactly the same way, but I think the general principle was about the same. Basically I took the Kester profile, worked out the max time and max temp for a given phase, and have the code treat each phase as ended when either the time or the temperature exceeds the boundaries of that phase. For the SOAK phase, since the elements have been turned off at the end of PREHEAT, I'm relying on thermal inertia to continue to carry us up in temperature. That works, but I think I need to stop PREHEAT a few degrees earlier than I do. As it stands, my SOAK phase winds up being really short.

Also, ignore the cool-down phase of the graph, basically. I accidentally moved the test board with the thermocouple a lot when I opened the door and the whole thing fell out of the oven, and I hand to fumble around a few seconds to find something to use to pick it up and put it back.

Anyway, here's a chart, and the raw data if anyone is interested in that sort of thing.

[ Specified attachment is not available ]
 kester_trial2.png

[mindcrime]

Here's another run of that same profile code, but without the fumble on the cooldown side.

 temp_data_oven_06_04_2021_kester_trial_3_with_comments.txt (2.42 kB - downloaded 42 times.)

And a run that tweaks the break between PREHEAT and SOAK down a few degrees. Now the SOAK phase is a little longer and this looks pretty close to what I need, except thermal inertia is now not enough to keep the temp up throughout the SOAK phase. But I could modulate the heating elements a little bit in this phase and I think I could get this very close to "just right".

More experimenting to do...


 temp_data_oven_06_04_2021_kester_trial_4_with_comments.txt (2.43 kB - downloaded 32 times.)

[fourfathom]

For the soak phase at what temperature did you shut off the heater?  How long after that did the chamber temperature start to fall?  As you zero in on this, remember that the temperature won't start climbing again the instant you turn the heater back on -- there is a delay as the heating elements warm up.

It does look like a bit of duty-cycle control on the heater would help you during the soak phase.  I have a duty-cycle option in my oven controller, although I don't use it for reflow.  My duty-cycle technique uses a one-second interval, with the heat ON or OFF one second at a time.  I have a "Bresenham" algorithm (quite simple, really) that keeps the on/off cycle spacing as fast as practical.  For example, a 30% duty-cycle results in a repeating on/off (1/0) sequence of "0010010001", rather than "0000000111". The thermal inertia of the system is enough that this probably makes no practical difference, but I have a thing for Bresenham, having used it a lot in frequency synthesis and other problems.  I'm using the 1-second rate since I don't want to stress my SSR more than necessary.

[mindcrime]

One more try. This graph doesn't necessarily look real pretty, and visually it's different from the Kester profile, although part of that is a scale mismatch. But as best as I can tell, strictly speaking, this would be compliant with their constraints (or very close) even if it doesn't track the ideal graph perfectly. And of course there's plenty of room to continue tweaking and improving the algorithm. I think though, if I stuck a board in the oven and ran this, I'd probably get a usable board out of it.

I would test that theory, but the only pre-made boards I have handy are some SOIC-16 to DIP adapter boards, and I don't have any SOIC-16 parts here, that I can find. For the most part, the only SMD parts I have around are ones I ordered by mistake when I wasn't paying attention and clicked the wrong part number or whatever. I do have some TSSOP-16 shift registers here, but that doesn't help.

Oh well... I have purposely ordered SOIC-14 boards and chips just to use for testing this oven, and they will probably be here Monday. So I can keep tweaking this software over the weekend, and maybe test some boards early next week.

 temp_data_oven_06_04_2021_kester_trial_4.txt (2.25 kB - downloaded 35 times.)

[mindcrime]

For the soak phase at what temperature did you shut off the heater?  How long after that did the chamber temperature start to fall?  As you zero in on this, remember that the temperature won't start climbing again the instant you turn the heater back on -- there is a delay as the heating elements warm up.

It does look like a bit of duty-cycle control on the heater would help you during the soak phase.  I have a duty-cycle option in my oven controller, although I don't use it for reflow.  My duty-cycle technique uses a one-second interval, with the heat ON or OFF one second at a time. 

This most recent run used 150 as the trigger point to run the heater off at the end of PREHEAT. The one before that was 147, IIRC. The big change this last time was exactly to add some duty-cycle control. It's a dumb algorithm, but basically I capture the temp each cycle and store it in a "lastTemp" variable. During the SOAK phase, if the temperature ever drops compared to lastTemp, I give it a 5 second nudge of heat.
That has gotten me a little closer to where I want it, but still plenty of room to tune and experiment.

Code: [Select]

116,145.83
118,149.03
120,151.84
SOAK: Turning oven OFF
122,154.62
124,157.16
126,159.62
128,161.96
130,163.93
132,165.64
134,167.30
136,168.76
138,170.00
140,171.10
142,171.94
144,172.79
146,173.45
148,174.10
150,174.65
152,175.12
154,175.45
156,175.87
158,176.11
160,176.43
162,176.63
164,176.70
166,176.74
168,176.73
SOAK: Turning oven ON
SOAK: Turning oven OFF
175,175.05
SOAK: Turning oven ON
SOAK: Turning oven OFF
182,174.84
SOAK: Turning oven ON
SOAK: Turning oven OFF
189,175.41
191,175.83
193,176.37
195,177.06
197,177.64
199,178.23
REFLOW: Turning oven ON
201,178.57
203,178.77


I have a "Bresenham" algorithm (quite simple, really) that keeps the on/off cycle spacing as fast as practical.  For example, a 30% duty-cycle results in a repeating on/off (1/0) sequence of "0010010001", rather than "0000000111". The thermal inertia of the system is enough that this probably makes no practical difference, but I have a thing for Bresenham, having used it a lot in frequency synthesis and other problems.  I'm using the 1-second rate since I don't want to stress my SSR more than necessary.

Huh. Interesting. I'm not familiar with Bresenham algorithm... I'll have to go read up on this.

[coppercone2]

i use a fuzzy logic 16 zone controller, I did not want to deal with all that programming and possibly EMI and stuff inside of a toaster oven

[I_Saldana]

You lot are all doing it wrong !!!

Turning reflow ovens into toaster ovens should be the essence of this recurring theme.



tsss. *smh*

[coppercone2]

I almost posted this yesterday, but the problem is I only have 1 ssr on my oven. I could require it for 2ssr to make that. Possibly worth doing because the SSR is $50 and its 2 hours work instead of 400$ for a toaster

it would need a bit of brazed brass or stainless tubing leading into the oven, but I don't know how much I want to mix bread an PCB residues. probobly a bad idea

[mindcrime]

OK, time for some real testing! Time to bake boards!!

My first effort...





... was pretty abyssmal. 

That's OK though. The oven did it's part just fine, and it looks like I just put WAY too much solder paste on the pads. Let's try again:





Still a couple of bridged pins there, but I did a better job of controlling how much paste I laid down on the pad. Sadly I'm doing this by hand with just a syringe, as I don't have any stencils for these little breakout boards. I could probably do a better job as well if I pulled out my actual microscope instead of just trying to do it with a simple magnifying glass.

Still, I think once I get my technique down a little better and get a better feel for the correct amount of paste, I'll have no problem producing usable boards using these SOIC components. Going to anything smaller than that would probably be nigh on impossible without using an actual stencil, but that's OK. If I ever send out to have boards fabbed, I'll get a stencil if I'm using any very small form-factor parts. And even if I mill a board at home on the CNC mill, I could probably mill out an associated stencil as well. Maybe not out of stainless steel, mind you,  but at least something to get the paste on the board.

[mindcrime]

This one came out looking OK. The initial reflow job left one solder bridge, but I was able to go in and clean that up by hand. I think the overall process netted what would be a usable board, if I actually needed this for something. Now I just  need to keep practicing and really nail getting the exact required amount of solder on the pad before placing the part.

[fourfathom]

For my complicated boards I use a stencil, but for one-off simple stuff I use a toothpick or small wire to transfer solder paste to the pads.  I can't use the syringe that came with my solder paste, the needle is too large to dispense small-enough drops of paste.  Perhaps a smaller-diameter needle would work, but the one I have already takes too much finger pressure to operate easily.  When hand-pasting the 0.5mm QFN packages, I essentially drag a thin line of paste across all the pads on a side, and the melting solder wicks to the pads and leaves a clean gap in-between.  It's extremely seldom that I have a short or open.

Bur a stencil is really the way to go for any significant board.

[coppercone2]

I just cured another varnished buzzy HP transformer in mine

[mindcrime]

For my complicated boards I use a stencil, but for one-off simple stuff I use a toothpick or small wire to transfer solder paste to the pads.  I can't use the syringe that came with my solder paste, the needle is too large to dispense small-enough drops of paste.  Perhaps a smaller-diameter needle would work, but the one I have already takes too much finger pressure to operate easily. 

Yeah, I know what you mean. It's hard to be precise with the syringe, because you have to apply so much pressure just to get the solder out. 

When hand-pasting the 0.5mm QFN packages, I essentially drag a thin line of paste across all the pads on a side, and the melting solder wicks to the pads and leaves a clean gap in-between.  It's extremely seldom that I have a short or open.

I had thought about trying the "lay a line across the row of pads" approach. When you do that, do you just put the line on and then place the component? Or is it necessary to take a razor blade or something thin and divide the line between the pads?

I may give that a try tonight after work. That or try the approach of just globbing some solder out on a scrap piece of board and then transfer it in small dabs with a piece of wire, or something along those lines, as you mentioned above.

Bur a stencil is really the way to go for any significant board.

No doubt! 

[fourfathom]

I had thought about trying the "lay a line across the row of pads" approach. When you do that, do you just put the line on and then place the component? Or is it necessary to take a razor blade or something thin and divide the line between the pads?

I may give that a try tonight after work. That or try the approach of just globbing some solder out on a scrap piece of board and then transfer it in small dabs with a piece of wire, or something along those lines, as you mentioned above.

I use the wire or toothpick to lay/drag the line of paste across the pads.  The syringe I have puts out too much for that approach (perhaps for larger geometry pads it would be OK?).  No need to clear the pad gaps, the solder wicks into the pin/pad junction and automatically clears from the inter-pad gap -- as long as you don't use too much paste.  Sorry, I don't have any descriptions better that "enough" and "not too much", but I've just been eyeballing it and it doesn't take much solder to make a good junction when there is close pad/pin alignment.  Getting the paste to stick to the wire / toothpick / pad seems to be a function of paste temperature -- too cold and it just doesn't stick, too warm and it is too runny.  "Comfortably cool room temperature" seems about right.