A while ago I decided to start designing my circuits using Surface Mount Devices (SMD). There are many advantages to using SMD components over through-hole components. Some of the major advantages are size and cost. The smaller size means I can make the circuit boards smaller. In addition, smaller components and smaller circuit boards means lower costs. One of the disadvantages to SMD components is they are more difficult to solder than through-hole components. The solution is to use what is called a reflow oven. The following project details how I designed and built a reflow controller based on a MAX6675 Thermocouple interface and use it to control a common toaster oven in order to reflow surface mount circuits. Check out the project here.
Hobbybotics 
Can you make the design files for the switch panel/DB9 panel/LED panel that you show in the design available, as well? Thanks
Sure thing. I’ll upload them to the webpage and Github.
Ok, check the page again. I added links to the files under the documentation section. The links will take you to a Box.net page where you can download each of the folders for the DB9 LED and Switch panels.
Thanks a lot for the quick response! Can’t wait to build it!
Sure thing. Let me know if you have any questions.
Awesome project… would like to build it. Do you have any blank PCB sets available for this project that I can buy from you? Or – are you planning to get some boards made that I can order from or along with you? Thanks.
I sure do have some blank PCBs. Let me know which boards you are looking at getting and I can send them to you.
Curtis
Oh.. that would be great! What I had in mind was to build two of them… one for me and one for a friend, so I guess I would like to get two complete sets of the boards if they’re not too much. I could email you my contact info if you would like. Thanks so much!
Shoot me your mailing address and I’ll send them to you.
Curtis
Hmmm… can’t find your email address. You have mine.. would you like to email me and I’ll reply?
Email sent to you.
Curtis
Do you have any more spare controller boards available for purchase? Or better yet… ones that are already populated? I’m sick of burning my components trying to manually control my oven!
I’m out of the current version but, I do have a fully populated previous version that is functional. It does have a repaired trace on the board that I missed in the design. Let me know if you are interested and I’ll send it to you.
Curtis
Sorry about the delay! I am interested! How can I contact you privately to discuss this?
I’ll send you an email. Just respond to that with a good shipping address. It’ll be a fully populated board minus the micro.
Curtis
Curtis… Recently built this project following your basic guidelines but with a few twists. Used the same oven too (as Macy’s was (& probably still is) running a “fire sale” on them). I built it for standalone as I only require a single ROHS profile. Another deviation is “pretty”. Yours is, mine is NOT (hand wired for one thing… and super glued the display to my PCB for another). Rather ugly in fact. but it works.
I also make the assumption that you’ve optimized the PID settings based upon this particular oven. However the defaults are giving me a lot of overshoot… 17 degrees (verified with an oven thermometer). as this project represents my initial venture into SMD I’m wondering about 17 degrees being within the profile guideline… will that kind of overshoot harm the PCB and/or associated components?
Now you’ve obviously spent a lot of time with PID algorithms. However; have you considered an alternative approach? Thermostatic control (akin say, to A/C or heat within a residence). Because PID is attempting to be proactive whereby thermostatic is of course reactive. And to become reactive some method of introducing ambient air into the mix is going to be required. BTW I’ve not taken PID out of the equation as yet (to determine if it actually is making a diff here).
So. With the assumption of too much overshoot I have a couple of choces. I can spend what is probably a significant amount of time to adjust these PID settings (via trial & error (because I really do not understand them)).. or take an alternative approach; the aforementioned thermostatic control. the problem with PID is there’s really no way to rapidly prevent overshoot from getting “nasty” due to the thermal time lag that inherantly occurs regarding a heating element. One needs to “anticipate” before-the-fact… PID needs to “learn” when to turn off the heater. Worse. the problem is exacerbated by sporting 4 heaters like this oven… more coils to store lotsa excess heat. Thus the crux of the issue is physical and unless we change the physics somehow this problem is not going to go bye-bye without a “fight”.
Now to use thermostatic control one needs to physically modify the oven (nothing new there). I’m not using it for any other purpose. A bit of short (aluminum can) duct work attached to a small VAC fan (mine is 75CFM), an additional small SSR, a coupla “flappers” (to ascertain that air goes & stays where its supposed to) and I’m in business. Lets use an example:
Say we have a setpoint of 180 degrees… and we’re ramping up but we don’t want significant overshoot or undershoot. what to do? turn on the fan at say, >= 184 degrees to immediately mix ambient with heated (residual heat within the PCB itself will prevent radical changes there). when the temp drops below that turn the fan off. turn it back on again if it >= 184, off at 183. that way if it continues to fall below 180 the heater comes on… but NOT at the same time as the fan, see? By experimentation we can control the rate of flow of ambient air being introduced by adjusting the flappers. The idea here is to keep turning the thermostatic fan on & off in cycle-like fashion until the coils have cooled sufficently to eventually leave the fan off…. and to prevent severe undershoot. Thus the thermocouple output is directly controlling air temp by keeping it hovering about the setpoint (by cycling an instrument with a much faster reaction period than heater bleed-off).
This also will help at cooldown… no reason to open the front door. because with the thermostatic & convection fans on and the heater off it won’t take long to cool down (as the heated air will be forced out the exterior flapper via air pressure)… (and in fact I’d probably pulse this area so as to not allow cool down more rapidly than the typical ROHS profile indicates). The end result is a lot of heated air “pulsing” due to the relatively rapid on-off sequencing of the thermostatic supported fan. however that pulsing should be relatively close to the setpoint (rather than 10% or so away).
yeah lotta work on the oven here. however anyone that is tackling a project such as making their own probably has inherited enough skills to also make it happen. Finally I’ll post back here when all is complete as the idea may interest others that have otherwise been staved off by the presence of PID. if you want pics of all this send me a place (I don’t do social media). probably be a month or more though. busy like all of us.
First, thank you for tackling this project based on this design. Second, the PID routine is optimized for my oven and even though you have the same type, your settings will more than likely differ.
The method that your are proposing will work but, as you say, will require additional modifications to the oven. Now, I believe they call the method you describe as “Bang-Bang” whereas there is an upper and lower set point. Some versions introduce hysteresis to slow down the amount of on..off cycles. The inherent flaw in bang-bang method for this application lies in how slow the heaters react. This is called thermal run-off and is affected by such factors as ambient temperature. You talk about this in your explanation. Thermostatic systems usually introduce a time and temperature buffer to allow for some variations. This keeps the system from bouncing around the set point.
Some overshoot is tolerable but, that’s usually dictated by the requirements of the components on your board and the reflow profile for the solder paste you are using. I’d say an overshoot of 17 degrees is really high.
Adding an external fan/ducts can help pull air out of the heat chamber as you explain but, may require as much tuning as the PID routine to dial in. Also, you may run into issues of getting consistent results with each run.
I say go for it and please, post your results. I’m very interested in how this turns out. Let me know if you need any help with your project.
Curtis
Curtis… you’re certainly correct regarding a very busy little fan there (its ball bearing though and designed for rapid cycling). another small downside is if one wants to use the oven for its intended design that would be a no-go (I’ve even taken the liberty to cannibalize the original controls). the only additional setpoint would be a 5 degree (not set in concrete) overshoot level (would rather it be a percentage but alas, no real numbers allowed here). I’ve kept your original code intact as a copy but now my friendly KISS ax comes out! btw I commend your presentation procedures as a great stepping stone. and very intuitive indeed. the code is also commented just enough to tell everybody what’s going on but nearly enough to teach BASIC! anyhow It will probably be a coupla months of tinkering as the physical changes are largely trial & error as I’m a long way from becoming a scientist in this regard! anyhow will post the results. and didn’t look as though 17 degrees of overshoot was gonna be a good idea… just too much profile intolerance.
Thanks for the constructive comments. I’ll definitely keep an eye out for your project.
Curtis… moved along quicker than anticipated! but butchered the back of the oven (the area that has the “bump” in it). removed that portion & installed a flat metal shield over same. I then placed a 3/4″ piece of plywood and mounted a 75cfm fan to it (mainly to separate it from the big heat). Deleted PID & substituted thermostatic control. its not quite as tight as your graphs but stays close enough to the setpoint (about 4 degrees either way) so its not an issue. also have the advantage of completely unattended usage once the go button is pressed. what happens here is that the fan blows the front door open just enough to vent hot air. but I tilt the unit 2 degrees downward to facilitate that (its either that or start messing with the door return spring). a 100 CFM fan probably would work a bit better however there’d be a tad more undershoot (but less overshoot thus I could adjust the off/on point a coupla degrees as now the heater is off when temp > setpoint, off when even, on when <. the thermostatic (exhaust fan is of course, opposite).
ran into another issue though. this oven is on a par with your graph included within the project. but it takes too long to come up for a ROHS profile. looking at insulation here. what's to prevent one from stuffing the sides of this thing with fiberglass insulation? then using a wrap around the rear, top, bottom? that would help to radiate the heat that's being lost through the outer metal envelope back to the interior. should speed up the ramp. 1300 F degrees is the melt point of fiberglass insulation so that's not an issue. ROHS wants the unit below 217 degrees within 330 secs from start and as you notice in your graph that's not gonna happen with a 1500 watt oven without some extra "uumph" from somewhere. installing nichrome wire is one option but now we're looking at a 220VAC oven (the other phase would drive the nichrome). I'll go with the insulation idea first.
In deleting the PID I didn't require the setup any longer for it thus installed a fan "test" menu choice in its place so one wouldn't have to run to setpoint only to discover that something may be amiss. will post back here regarding the insulation concept.
Insulation should work really well to keep the heat in. Should help ramp the temp curve up faster. I initially looked into adding insulation to my oven but didn’t want to tear it apart to fully add it.
Curtis
Curtis… right. that’s going to be a pain no doubt about it. but like you mentioned it should work quite well. hoping for something around 30% improvement as right now an awful lot of thermal is simply passing through all this metal surrounding the envelope. yeah the big concern is the ramp up as cooling off is not an issue as one can simply open the door if it remains too slow.
Curtis… carrying on here regarding insulation. You mentioned the difficulty of stuffing insulation inside the box and I’d have to agree. However I’m not using it for anything else and my idea of pretty is “does it work”? Now the box is still “pretty”. but it doesn’t work (at least not for ROHS). So. why not remove the outer envelope & simply discard it? Then build a plywood box with the remaining oven raised 3.5″ to accommodate batt insulation being installed on the underside. Set the oven within the box & stuff IT all around with 3.5″ batt insulation. The only part of the oven with no insulation would of course be the front door. Anyhow that’s what I’m going to do. This is being kept in the garage anyway along with my other ugly contraptions that nobody ever looks at!
Anyhow thermally it has no place to go except to stay within containment until time to release it. Thus the ramp would be slow at first but then would increase rather exponentially (they don’t want more than 2.5 degrees per second though at any point (but if that happens I’ll pop the thermostatic fan to contain the ramp)). yeah totally trial & error here as thermal engineering I don’t even know how to spell it. Finally though I’m going to have to install a larger fan as venting is not quite quick enough (without opening the front door). Going with a 6″ inline all metal 240 CFM. About a third of that will be “cut-off” because I don’t want venting above the grate where the PCB sits. Even so we’d be moving at least twice as much air as we do now so I’d expect tighter thermal control along with a much faster ramp up due to all that insulation. will let you know how all this works out… probably be a few weeks though.
Have you considered putting the opening for the fan on the top of the enclosure? Heat likes to rise and venting out the top might help a bit. I also agree with gutting the outside enclosure and putting it in a wooden box stuffed with insulation around the sides. I hadn’t considered much on lead-free profiles in my design. Part of my reasoning was the max output of the oven was right at the heat boundary necessary for many lead-free profiles. My plans was to design an oven from scratch using independent heating elements. That’s part of the reason for the redesigned controller I’ve been working on. I intend to build it out of aluminum with two chambers separated by insulation. That’s why I am curious as to how your project turns out. Do post some pics of your progress.
Curtis
Have you considered putting the opening for the fan on the top of the enclosure? Heat likes to rise and venting out the top might help a bit. I also agree with gutting the outside enclosure and putting it in a wooden box stuffed with insulation around the sides. I hadn’t considered much on lead-free profiles in my design. Part of my reasoning was the max output of the oven was right at the heat boundary necessary for many lead-free profiles. My plans was to design an oven from scratch using independent heating elements. That’s part of the reason for the redesigned controller I’ve been working on. I intend to build it out of aluminum with two chambers separated by insulation. That’s why I am curious as to how your project turns out. Do post some pics of your progress.
Curtis
Hello Curtis!
I am very interested in your project. There are a couple of questions.
As I understand the compiled firmware for PIC16F877A, the article in the photo it was he. Also in the source code in include. Then the article typo? “I chose the 16F887A for this project as I have plenty of them available.”
Second, I can not compile the firmware using MicroCodeStudio, Pic Basic Pro 2.6 and MPLAB does not compile, got error 118 asm line 423: Owerwriting previous address contents (2007)
Regards
Alex
Alex, I sent you an email a while ago about this error. Did you get it working? If not, let me know.
Curtis
I no receive e-mails.
I began to understand myself and realized that the reason the configuration file, removed it and put compile and configuration manually.
Excellent construction, currently engaged in processing roster.
Curtis… still working on it. Anyhow used your noted oven for this project but ramp up as I eluded to earlier was way too slow to support an ROHS profile. I decided to use the oven itself because basically I needed three things 1) convection fan assembly 2) tempered glass door 3) some kinda basic enclosure. I also scrapped PID but after doing that it turned out to probably be a bad idea overall. The big advantage though is cool down (> 3C/Sec) as I’m using a fairly large thermostatic fan. the downsides though is this: mechanical complexity. if you’re not into mechanical stuff fabricating the “flapper” assembly (to protect the under mounted fan from excess heat) is a real pain. what happens here is airflow opens the flapper, no air flow and gravity closes it. anyhow I fabricated the (2) flappers out of a cheap aluminium throwaway turkey basting pan (which covers approximately 4″ X 4″ air flow area).
The convection assembly has been moved to the rear of a new metal box installed within the main cabinet. the original lateral coils have of course been removed as well. one of the problems with the originally sized oven is its just too big of an area for the default coils that came with the unit to handle. Thus I fabricated a box from sheet metal (Lowes, HD) that is 9″ deep, 6″ wide, 6 1/4″ high and installed this centered latterly within the cabinet. To heat this much smaller area I tried a ceramic heating element of 1500 watts. well that did NOT work at all (I think it was defective to be honest about it because I dropped the thing a coupla times but did NOT test it after these mishaps.. duh!).
What I found that works great though is a simple dual heat gun. In fact this thing works so well that I’m using the 750 watt winding! The 1500 watt winding was ramping up way too fast… over 5C/per sec! but with selecting the smaller coil it takes about 120 seconds to reach 200C… well within the ROHS ramp profile.
As mentioned in previous posts I’m using thermostatic control to limit over/under shoot and there’s another advantage here. These are high powered windings but also lose heat very quickly. Thus overshoot is only a coupla degrees. I lose about 5 degrees on undershoot but this is well within the profile. yeah this is obviously the result of losing heat more rapidly than building it back up after turn on. Cool down is also very rapid (but not TOO rapid) meaning basically you can walk away (NOT recommended) once you hit the “go” button. btw the outer cabinet never gets hot enough to be uncomfortable to touch. no insulation either!
Anyhow I STRONGLY suggest one uses these very economical heat guns ($15 at Harbour Freight) and cannibalize same (to extract the coils). I would also suggest though to stay with PID… and go with a MCU with more flash as there is almost no wiggle room with the one you used.
-Lee
Fantastic. I hadn’t considered heat coils from heat guns. I’ll have to experiment with them. Thanks Lee and I would like to see your setup.
Curtis
Curtis… yeah these coils are QUICK (unlike ceramics there isn’t much mass)! with PID I doubt you’d get drift much more than a coupla degrees C or so. the gun I cannibalized was from harbour freight. most of these have 2 settings, 750, 1500 watts. they’re designed of course with a switch that places one or the other (but not both) across the line. however guess what? I can turn them BOTH on at times. those times are 1) ramp up from 200C to 230C under reflow 2) when temp drops below setpoint. this is not an issue on a 20 amp breaker as both coils are only on for relatively brief stints.
btw once cannibalized you need to be careful handling the coil set. cut the attached wires as far away from the coils as you can manage & then solder & heat shrink any extensions to your control electronics (copper stranded only) that you may require. hopefully the original wires are long enough to extend outside your box & if so its not likely that the insulation will melt from that point on. only a very small amount of the insulation melted off the original wires… and that was right up near the coils.
and be careful with coil support as its relatively easy to break the coil supports that the coils are wound around. I use thin nichrome wire for this (26 gauge) after carefully drilling some small holes for same.
Note that this setup is NOT “pretty”! there probably isn’t a square corner ANYWHERE (other than the original oven). and the controller looks absolutely abysmal if I do say so myself. however this IS a one-of-kind deal here and I’m not putting the thing up on a pedestal just for display. it just needs to WORK good enough to bake PCB’s without the necessity of spending thousands. I suppose the best way to see it is via utube but I’ve not gotten around to that yet as I’m still working out some software issues. not easy with an unfamiliar freebee IDE!
Curtis… this morning I fine-tuned the testing procedure watching carefully for anomalies while recording temps vs time within the preheat and reflow zones. Ramp to 150C from a cold start took 80 seconds while total time to 200C took 140 seconds (minimum is 120 seconds). Thus the total pre-heat time was also 140 seconds (this must be somewhere within (120-240 seconds).
Reflow… ramp from 200C to > 217C took 25 seconds. The recommended time on this ramp is 2.5C/sec +/- 0.5 sec. Near as I could discern via eyeball (don’t have a graphical output here as this is a standalone design). is a 17 degree rise in 25 seconds (which equates to approx 1.47C degrees per sec). hmmm a bit slow here as the profile indicates a minimum of 2C per second. Now here I could switch to the 1500 watt coil but tell you what the ramp then would be way too quick!. The other thing I could do is to is alternate coils every few seconds to achieve the desired ramp… or I could get real elaborate here and go with using the bigger coil via PWM.
Well for now I’m gonna do nothing at all regarding this reflow ramp. The reason is that some ROHS profiles don’t even mention this ramp in their specs… and I’m really not that far off. Once 217C is reached the setpoint is 230C which then holds between 225-232C for 60 sec. Once that time is reached the cool down phase is very rapid due to the size of this thermostatic fan. as within 10 seconds from ending reflow we’re back under 200C.
The specs on cooldown states we cannot decrease temp at > 6.5C/sec rate or we could crack the PCB! My PCB’s are relatively small so that is not likely and not only that but the rate I have here is more like 4C/sec. This is wiithout opening the oven door & with no insulation. Cool down in fact would be way too quick if I DID open the door (at the ending beep) thus I leave it closed until the temp reaches 60C. no doubt about it though there is a significant advantage to using some type of fan here at cool down.
Continue to understand, everything works except -BAKE- after starting this function does not work, the temperature does not change. The program could not figure out what could be the problem?
That’s odd, the bake function is designed to ramp up to the set point temperature and stay at that temperature for the specified time frame which is set to 2 hours for the default setting. Bake is used to dry out any moisture in components before placing them on the PCB to be reflowed. What issues are you experiencing?
Curtis
I understood the purpose of this function. But when start the heating temperature on the display does not rise – stay, the relay is turned on and the heater continues to heat up. Tried to launch the model in Proteus, the same thing. Reflow, test sensor works fine.
Let me take a look at the code for this function and get back to you. Does the reflow function work like it should?
Curtis
Alex, are you getting any warnings when you compile the code or are you using the already compiled hex file I provide with the project?
Thanks,
Curtis
As Curtis indicated that’s exactly what that function accomplishes. I didn’t need it though so instead of that I instilled a “Test” that verifies that all the parts within the oven are in fact, working. I run the test for 15 seconds on each component (2 hot air gun business ends (each test is separate), convection fan, thermostatic fan). however if you’re building this thing stay with PID (which doesn’t use a thermostatic fan). The advantage to going with a thermo fan though is cooldown (as it gets the PCB out of reflow in a hurry (faster than simply opening the oven door). however the hardware complexity is a real pain so simply open the door at beep time (and adjust the actual reflow time to compensate for the slower cooldown).
Yep, (what Lee said) could not have said it better. Now, some ovens have a built in fan and the current cool down routine keeps the fan spinning until the temperature within the oven is around 40 degrees Celsius. This cool down temp is adjustable. You still need to open the door at the beep.
Curtis
Curtis
Reflow works fine.
No warnings, i try you hex to.
Well Alex, you are correct in that the Bake function is not working. I have confirmed it and I’m fixing the code. I’ll upload a revision with a fix shortly.
Thanks for pointing this out,
Curtis
Alex, a new version of the firmware is up on my Github site. It’s labeled reflowPID5H.
Let me know how this works out for you.
Curtis
Huge thanks! All works fine now.
Whith best wishes
Alex.
Curtis… my unit has a pretty strong thermostatic fan that I run on cooldown. and we have a max spec there (ROHS) of 6.5 degrees/sec and I’m right on the edge of that if I open the door at beep. so I simply keep it closed. my box is only 6 X 6.25 X 9 or about 335 cubic inces so it doesn’t take long to replace hot air with cool. anyhow waiting on a better convection fan thus still haven’t tested these ramps. tracking says it oughta be here Monday so we’ll see.
Hi. Really enjoy your work and would like to build one. Have your same oven. Do you have any of the boards left you could sell me? Appreciate the consideration, Steve
Steve… please note that this oven will NOT work regarding a ROHS profile! for one its way too big and secondly the heating coils are spread over all those cubic inches. I had to completely redesign this thing to make it work for ROHS due to these limitations. but it is NOT pretty as only the outer shell remains. Hopefully Curtis’s newbie design will run ROHS on a subsequently scratch built oven. another possible issue is uneven heat distribution regarding this wussy convection fan (which he correctly leaves on all the time via his firmware).
do this: take a coupla pieces of WHITE bread and cut the crust off of it. Place it in there and toast same on a setting that turns on both the upper and lower coils (forgot which settings those would be as my controls have long sense been removed). watch the toasting process carefully and when the top is a medium brown shut off the oven wait to cool and have a look at the bread. is it perfectly even on both sides in all areas? take a bit or two. it should be crunchy all the way through… not just on the outside (I never actually tried it on the unmodified oven). however I suspect they’ll be issues here. I did this on my redesign (which took MONTH’s of experientation, btw… see my posts above) & there’s absolutely NO difference anywhere on the bread.
the problem with ROHS is the degrees per second that are required. this stock oven just does not have nearly the uumph required. Also note Curtis’s profile pics. they are absolutely dead-on what I got on my original oven. OK for leaded solder but way, WAY too slow regarding ROHS. here’s what happens on the ramp ups: at first everything goes hunky-dory, ok? but as the oven heats up it gradually loses its ability to “keep-up” (by maintaining the profile). so what happens when you need it the most (increasing heat-per-sec) the profile maintenance goes south. now if you’re building stuff only for yourself or locally then it doesn’t really matter. but if you’re planning on selling your stuff globally ROHS is an absolute must have.
Finally. Is/was Curtis actually SELLING this? seems to me its a DIY project. another clue that he was NOT selling was the amount of detail that he posted here regarding this project (right down to providing software source). that fact alone saved me MONTHS of work! Thus like a lot of really smart people in DIY land he believes in sharing his work so others like us may benefit.
One final note. if you’re planning on building this the MAX6675 has been depreciated after reaching EOL. there’s a drop in replacement for it though; the MAX31855
-Lee
Lee is pretty much spot on with his comments. I would correct one part though…the MAX31855 is not exactly a drop in replacement for the MAX6675. Take a look at both data sheets and you’ll notice that you don’t tie T- of the thermocouple to ground for the MAX31855. I know I’ve been moving a little slower than expected on the upgraded board but, I want to make sure it can be produced as cost effective as possible and work has been taxing me lately.
More to come.
Curtis
On Sat, Oct 12, 2013 at 11:47 AM, Hobbybotics
Steve, Curtis… Ah! missed that. yep there’s a note on pin 2 about NOT tying T- to ground (while the MM6675) says to please do so). so like Curtis mentions it is NOT a direct drop in. also wondering about Vcc regarding this part. yeah that’s an issue as well. 3.3VDC nominal on the MAX31855 and 5VDC nominal on the MAX6675. So even though actual operation and pinout is the same we got 2 pins here that differ a bit electrically. Yep gotta be careful before you go changing things! I still design with some 5VDC TTL but its only because I happen to have hundreds of those parts basically left over from oldie days.
Hello Curtis,
I have read your outstanding article and want to built the Reflow Controller myself. Do you sell the PCB’s, or should I have them made from the projectfiles?
Best Regards,
Robert
PA1RBH
Robert… I built this controller basically following Curtis’s lead. I didn’t use PID but I strongly suggest that you do. The problem I had though was with the oven retrofit itself… No way that oven is gonna follow an ROHS profile as it has way too many cubic inches to heat even with the 1500 watts of coils. Another issue is with the wussy convection fan. it is way too wimpy to circulate the heated air properly.
What I did was to build an oven inside of THAT oven which cut the cubic inches down to a 6 X 6 X 8″ area… Then I mounted a really strong convection fan (purchased separately). finally I dispensed with the oven coils completely and went with 2, 1500 watt heat gun front ends. it is truly ugly but… it follows an ROHS profile with no more than about a 5C over/undershoot. now that is a lot of power so what I did was power each one off of a different phase (you’ll blow the breaker otherwise). Fortunately I have 2 different circuits in our bedroom area that also happen to be on 2 different phases so it was a simple thing to run a 20 amp extension to one of the phases.
I actually designed a really nice reflow oven but never got around to building it as I became involved with the design of a small, fast PnP machine. These go hand & hand with the oven as its a very, VERY tedious process to place these tiny components by “hand” while working under some type of magnifying glass. My reflow unit is standalone as I don’t use native windows but rather defer to linux. Microcode studio does not run under wine but it does run under Oracle’s VM. You need the windows CD though…XP or win7 (to install to a VM). yeah I know a bit off topic here just trying to render a bit of assistance.
-Lee
Hello Curtis,
I’ve read your excellent article about the Reflow Controller. Great Stuff!!
I would like to built one myself and have a question:
Do you sell PCB boars or ashould I use the files to order a PCB somewhere?
Best Regards,
Robert
Robert, Thanks for the interest. The quickest way to build the current project is to download the files and have some boards made. I have a new version in the works that I have been working on for some time now that has more features and better functionality. I don’t have a final eta yet as I have a number of tweaks still left to finalize. I will be selling the new version and releasing the design files also.
Thanks,
Curtis
Hey Curtis.
I would just like to ask if you used a Bootloader on the PIC16, and if, which one.
Thx for your great work.
I sure did and it is the one that is available through Microcode Studio that comes with PicBasic Pro 2.6.
Curtis
Hello from Bulgaria , great project. I want to do this project, my question is, with what software are designed scheme and board. I opened these files in Altium Designer 14 but not working. My idea is to recreate it in Proteus and create a new one sided board.
Yep, the design of this board was when I first started using schematic/PCB tools when good free options was not available. I designed this project with Express PCB software that is free to download but, the board files are locked to their applications. You do have the option to purchase the gerber files after you have ordered at least one set of boards from them. I will be transferring this design over to eagle cad.
I’ve found eagle to be needlessly complex and also NOT very intuitive. diptrace is a much better deal. however their autorouter is awful if one gets into a high pin count. however you have the capability of using many alternative routers. one should never design using a proprietary fab as that locks you in. also Express PCB is very expensive (try sitopway.com)… 15 days from sending your files until the fab is in your hands though. plan ahead and save BIG bucks for an excellent fab. btw… Kevin is the “man” no doubt about it! this outfit also seems to always send a few “extra” PCB’s of your design. they also sport absolutely the cheapest S/S stencils I’ve ever seen! AND they work! Hello, Curtis. been awhile.
Yep, you’re right. I do not prefer proprietary software. At the time, ExpressPCB seemed to be the best option for me since the tools was free and easy to use. Eagle CAD was a beast to learn and I lost interest really quick. I have since learned Eagle and developed a love-hate relationship with it. I love it for the extensive libraries but, I do not prefer the UI workflow (even though I now understand it). I’ve been using DesignSpark also since it is easy and free. I’ll probably stick with Eagle in the end as it seems to be used by a broader audience.
I’m back in the mix now and got some updates for some old projects and lingering projects. Documentation is slow going though.
Curtis
Curtis… no getting around it those huge eagle libraries can save one a buncha time. the workaround in diptrace are their awesome tools for rolling your own. however in saying that once you master the learning curve AND pay for it eagle certainly works. also very widely used as you mentioned. yeah same here regarding projects. I’m working on a precision (25-200 micron) DLP printer and the docs I already know are gonna be a nightmare. And same thing past projects never seem to catch up in this area. Patent issues are gonna be a hassle so I’m gonna kit the thing as a workaround… even more detailed docs.
I’ve also designed/built a 50 micron (plenty good enough for 0603 & TFQP-64 sizes parts) PnP machine that I’ve never gotten around to document or kit (use it in house here but programming a 500 pin design via gcode is a real pain… takes days!). another reason for not kitting it is its parport controlled (rather legacy) and only runs under a specially compiled kernel via linux (called “linuxCNC”). another disadvantage its pricey due to low volume sourcing on LOTSA parts. all this means its impractical to sell. yeah my projects are just too damn BIG to make any serious $$$ with but they keep me very busy anyway. and finally got to thinking DLP so I can print most of these parts rather than having to painfully fabricate or purchase.