I've recently purchased a popular T962A SMD reflow oven. The reviews are split on this oven - some find it to be quite acceptable, while others find it completely useless. After much researching, I decided to give it a try.

(Feel free to skip this part if you already have the oven.)

T962A is an inexpensive IR (infrared) SMD reflow oven. It has a microcontroller to control temperature and timing (profile), so you can basically put populated PCBs in, push the button and simply wait for a few minutes. Then the PCBs are done.

There are some issues with this oven, but my experience is rather positive overall. For my purpose this oven works well.

However the problems are:

1. The oven produces horrendous smell.
I'm not taking about the smell of solder paste melting. The oven itself produces the smell even without anything in it.

2. The heat distribution is not very even.
This is not an issue with leaded solder, but becomes a major problem with lead-free solder. As lead-free solder requires much higher temperature to reflow, the extra temperature needed to ensure the solder reflow on all parts of PCBs can burn the parts in the hot spots. Even though I'm lucky enough to not fry any PCBs with this oven, some parts did come out looking burned.

I want to share what I did to improve on those issues.

Step 1: Why So Smelly?

Like any oven, I expected some odor/fume to come out on the first few runs. But this oven is no ordinary oven in that respect.
Upon its first run, it produced something that I have never smelled before. It was so toxic, that it gave me a bad headache. And the smell did not go away even after a few runs.
So I decided to open it up and see what's causing the odor.

Opening the hood was quite simple, just pull the drawer out, and remove a few screws. Then the top is open.
Then there was the source of the problem - masking tape! Masking tape is used to prevent electrical shorts between the aluminum foil used to cover the insulation material and the controller PCBs. Even though there's heat insulation material beneath it, the tape gets hot enough to produce that nasty odor. So I removed all of the masking tape, and replaced it with Kapton tape, which is heat resistant.
(Sorry I forgot to take the picture of the masking tape before removal. You can see some pics here though.)

This mod took care of the odor completely.

Step 2: Improve the Heat Distribution

Looking at the inside of the oven, got me thinking - is there any way to improve the uneven heat distribution?

Infrared reflow oven has an inherent weakness - because the infrared heat is absorbed at different rates depending on the surface color, black ICs get heated up quickly, yet metal parts heat up slowly. Temperature profiles are used to give some "soak" time to let the temperature even out between different parts of the PCB. However that can only help so much.

Modern reflow ovens mostly use convection technology. Convection ovens use moving air to deliver the heat. This method distributes heat much more evenly then IR method.

T962A has a cooling fan in the back. This fan sucks air into the oven to cool down the inside. The heated air comes out of the bottom of the oven, which has many slots.

So I thought, why not use this fan to give some air movements during the heat up, to help distribute the heat?

Looking at the controller closely, it is revealed that:
  1. Heating elements are controlled by a SSR (Solid State Relay) via 5 V logic voltage from the controller.
  2. Fan is control in the similar mannor - logic signal controlling a triac via an opt-coupler.
Which means that I should be able to sense whether the heater is on or off, and turn on the fan as I wish.

I connected four wires to get the necessary I/O out of the controller board. and routed the wires out. Connecting the switch to the fan controller line proved that it works.
(If you are doing this please see comments on the photos - there are some importent information in there.)

Arduino to the Rescue

Now I got the signal that tells me whether the heater is on, and a line that lets me turn on the fan when I want. I believe by blowing just the right amount of air when the heater is on would improve the heat distribution. So I took out my Arduino, and wrote up a simple code. The code basically senses if the heater is on, and if it's on runs the fan at a low speed. To run the fan at a low speed, I chose to do super low frequency PWM, which ended up being turning on the fan for 0.1 to 0.2 seconds, and turn off for 0.8 - 0.9 seconds. I did it this way instead of using Arduino's PWM output, because the fan controller may not respond well to the signal going on/off hundreds of times a second. I'm using 15% duty cycle to run the fan.

By running the fan slowly enough, the hot air inside is moved around to help distribute the heat. The heater is strong enough to still heat up quickly.

The result is an improved heat distribution, and seemingly more accurate temperature reading. I think the air movement helps the thermocouple (temperature sensor) inside the oven to sense the air temperature.

More testing and tweaking might be needed to get the most out of this modified oven. But so far I'm very happy with this.

<p>Thanks for this. I saw it pre-purchase, and it made me comfortable enough to buy. I'm surprised that in 2 years they have not addressed the masking tape issue. Must be tight margins. </p><p>Opening, removal of masking tape and replacement with caption tape, and reassembly worked like a charm. Front screws were a bit fussy... they are stripped-ish. Reassembled well. First fire-up - still pretty smelly.</p><p>I'll do another run and see how it goes..</p>
<p>Kapton tape worked like a charm. I didn't smell anything nasty when I fired up the oven for the first time. Maybe I'll try the fan controller next -- but you know what would be kinda cool? A light.</p>
<p>I did your improvements on the oven with the exception that I cut the connection from the fan - line, put the one side to an other input of the arduino and an other output from the arduino to drive the fan, as you did. In the software I did also the check, that if the heat is off, and the fan is on, I switch on the fan constantly</p><p>Now I did two measurements, first I put a PCB in with a underlay of about 6mm. On this pcb I put a temperature probe. During the process I also took time and compared with the time of T962. I realized two points:</p><p>1st: the temperature on the display is about 20 - 30 Kelvin below the measured temperature on the board. That means, at a display temp of 160&deg;C I had about 180 - 190&deg;C on the board. The higher the temp, the smaller the diffrence. At about 250&deg;C, the board temp was about 265&deg;C.</p><p>2nd: the time for the process took on the display 8mins. I took a time of 12minutes which is 1.5 times more!!</p><p>At the company where I work, we have a Protoflow E reflow oven from LPKF. We have really good results with lead free soldering. I put in that soldering curve, which gave us really good results....<br>First heating up to 160&deg;, holding the temp, than put in the PCB and as the temperature reaches again 160&deg;, it pre-heats the pcb of at this temp for about 160seconds. Than it goes up to 260&deg; during 100seconds, that means a ramp of 6K per 10 seconds. At 250&deg;C we need to open the door at the LPKF, but the origin curve goes than down with a ramp of 10K per 10 seconds.</p><p>About the much higher temperature: Has anybody made the same experiences that this temperature is this higher than the displayed? If this is really ture, I need to take down the whole diagram of about 20K. And about the time, I need to shorten the whole time range at factor 0.66.</p><p>This to much of temperature might be THE reason why so many people get burnt components during lead free soldering with curve 3. I have tried this curve, had also burnt components and bad solderings.....</p>
<p>I just bought one of these ovens. I'm sure glad that that you wrote this up</p>
<p>I would like to know why if you have firmware for the unit you also need an arduno to control the fan. seems to me if you have working source code you can just pwm the fan as you see fit.</p><p>I just bought an oven and got it today and am thinking about doing a full schematic on the controller board.</p>
The controller project is hosted here: https://github.com/UnifiedEngineering/T-962-improvements
<p><a href="https://github.com/UnifiedEngineering/T-962-improvements" rel="nofollow">https://github.com/UnifiedEngineering/T-962-improv...</a><br>This should be a link now. It wasn't entirely obvious to me that it wouldn't automatically turn the URL into a link when posting from my mobile device earlier.</p>
My alternate firmware project that is.
<p>That looks very impressive. I will check into it. Thank you very much for sharing!</p>
I got inspired by your Kapton tape mod and got carried away a bit so I wanted to share my findings here.. :) At least it will show people how the hardware is wired to the controller MCU. Everything but the buzzer is operational (and the buzzer is a simple known gpio pin)
<p>Thanks very much for this tutorial. Was good tech review before I bought one, just saved me to make the entire house smell bad for that useless masking tape. Anyway I made your mod with the Arduino, works fine definitely the temp in the oven is more even across the surface. However my oven still burns PCB's!. The issue is more likely to be caused by timing, as varying the reflow graph I get different outcomes ( sometimes burned, sometimes solder did't melt). I first tried to change the crystal as hoping that will fix the issue, (originally has 11mhz mounted on) 12mhz improves but not enough, 14mhz the display will not display characters properly. At the end I gave up with the idea of fixing the timing, and went back trying creating a new reflow graph including the time offset. So far no luck in creating an ideal graph for Leaded 63Sn/37Pb paste, with this oven. There is any chance that you can share with us your adapted reflow graph, showing time Vs Temp, that will be really helpful. </p><p>The T962-A is a such good oven, is a shame that they made a mess with the software. I noticed that the board as all the capabilities to drive that oven properly, and luckily there is an ISP connector for programming. I hope that maybe one day someone will create an updated firmware for it... </p>
I've put some effort into an alternative firmware for the LPC MCU present in the original controller. I have also enabled the fan to run during the reflow cycle, and it definitely helps as long as it doesn't run too fast. Reaching peak lead-free reflow temperature is challenging when fan runs faster than minimal speed.
Great someone finally put the effort in doing that. Did you solve the timing issue? I think that the best way to solve the problem of no airflow, is to add an internal fan (as conventional cooking ovens). Any chance you can share the firmware with us?
Firmware linked in the comment above. Timing is proper, but the UI needs work, that's why I'm not posting the hex-file for direct programming yet as it's not a direct replacement for the original firmware (for instance I have to recompile in order to switch profiles). Feel free to have a look in the GitHub repository!
<p>Here's the pic.</p>
I'm glad that it works for you. I did have to do a lot of tweaking of the profile to get it just right. Photo of my profile is attached.
<p>Hi,</p><p>maybe it is worth to think about an new Controller instead of a hack. New Controller, USB, so no nasty Buttons, cause the KBD is ... and have easier way to select/create profiles and control the fan/heater ...</p><p>With best regards</p><p>Gerhard</p>
<p>You made my day, I was ready to give up, it didn't work with lead-free paste.<br>I have the little one (T-962), and there is no SSR in this model, but the card seems to be the same. I control now the fan as you do, by putting 0V (through a 1k resistor) to the pin #2 of the optocoupler and I read the state of the heater at the same place as you do.</p><p><br>Tests still in progress but turn the fan on at slow speed during the entire process seems to be better than turn it on only when the heater is on (in my case anyway).</p>
<p>So I made this mod to my T-962 oven and it all works as expected. The problem I'm having is that it's affecting the profile way too much. After all it's sucking in cold air. I've slowed down the fan as much as I could but it's still too much. When the fan is running on PWM, the oven can't reach the required temperature and when it does get within 10 degrees, which automatically shuts off the heating elements and therefore also the fan, the temperature will always overshoot the target. I did a quick plot of a profile on an empty run and you can see it here: <a href="http://imgur.com/9SMPlkw" rel="nofollow"> http://imgur.com/9SMPlkw</a> (the Heat, Fan, and PWM Fan plots are simply on/off indicators, not like the temperature ones.) You can see, during ramp up, how every time the heating elements shut off (and therefore the fan as well), the temperature will raise and overshoot the target temperature.</p><p>So my next step now is trying to figure out a way to not suck in cold air in, but somehow recirculate the hot air that's already in the oven. I'm not sure how well it would work if I close the intake vent during the ramp up, and open it up again during the cool down phase. Maybe I need to find a third fan to install somewhere inside the oven so it only recirculates the hot air that's already inside, as opposed to using the cooling fan to recirculate the air.</p>
<p>I did the modification nearly a year ago, but lost the images somewhere. Anyways it works good. I used a Freeduino for FAN control, but Now I am thinking about replacing it with 555 timer circuit. :)</p>
<p>Thanks for the info about the masking tape. Man, did that stink! </p>
<p>Worked Perfectly with the Digispark USB Dev. Also had to change LED = 1 for rev 4 board. The pullup resistor is a must of course, but now I need to do proper calculations for the timer been off. On a PCB In the center of the lamps (T962) im getting a max temp of 293 which is seriously bad for component life. Thanks LEDartist </p>
<p>I have the T962 and like dzak, the arduino worked for the fan. However, the fan doesn't go on during cool down mode anymore. It works with the arduino controlling it but no longer with the stock controller. Any hints? Worst case I can always add a switch to the circuit to manually turn it on.</p>
<p>Do you have a 1k ohm resistor in series in the fan control line? You need one.</p><p>(see 5th photo under step 2)</p>
<p>Very nice! I purchased a T-962 (the smaller sibling of the T-962A) and the Arduino hacks works there too. I ripped out the masking tape and replaced it with kapton tape. I used a tiny, cheap Digispark (which uses an ATTINY85) instead of an Arduino Uno for the fan control. A little shrink wrap to insulate the USB connector and a little more hot glue to hold it in place (who is going to notice a little more in there, eh?) and it works like a charm. I had to change the LED pin to &quot;led = 1&quot; for the Digispark. Everything else just worked.</p>
<p>I was wondering Colorado_rob, How is that ATTINY85 working out for you in the system regarding the Main intake fan at the back regarding the even temperatures and temp profiles. I have the exact the same version of reflow oven and I find the temps seriously crazy. On Profile 1 with fluke DMM temp sensor I reach up to 277 degress (C) (LCD screen thinks it's doing a great temp job with false readings)</p><p>which is crazy, although it doesn't burn the boards, I am worried about the lifetime of the components attached to the PCB over the heat madness. There is another problem about hot, and cold spots, will this mod actually help it ?</p><p>Thanks for your reply Rob..</p>
Thanks a lot for your instructions - helped me a lot!<br><br>Recently I also finished a review about the T-962a - maybe you are interested in it: http://blog.petrockblock.com/2014/01/10/reflow-oven-t-962a-a-review/<br>
thanks for the info; i have ordered a T962 ; which in fact is only 800W ; this model is NOT using the solid state relay to power the lamps, but is in fact using a BT139 triac to power on/off the heaters. the board is the same as your just that the empty spot on your board is filled with the BT139 and some resistors <br>
Good to know. Thanks!
I can confirm that the hack works fine also with the T962 version; there is just a little space problem to put the ardunio board ; i glued it to the top and protected top &amp; back with some kaptop tape. <br> <br>Mine also had masktape underneath the isolation instead of just on top; stripping that solved the horrible smell aswell! <br> <br>thanks for the nice mod, and your code! it's simpel but effective!
Nice article. Just received mine and sure enough, masking tape inside. Have Kapton tape already on order for another project, not arrived yet though. Anyway, need to add one very important thing: Safety. I checked mine and suspect all will be the same. The panels are coated with paint which acts as a good insulator. An earthing check showed the earth tag to not be connected to the back panel chassis. Having fixed that, there was also no continuity from the back panel to the rest of the case. This was fixed by grinding away the paint under a couple of screws each which go attach the back to the bottom and top chassis halves respectively.
I was wondering what width and approximately how much Kapton tape you needed for this modification?
I used wide tape (about 6 inch wide), but narrow tape should work just fine. Not too much tape is needed. I'm guessing about a yard.<br>I think you'll have to buy a roll anyway right?
Yeah, I was just trying to figure out how much of a roll, based on the width and length. <br> <br>You posted in May, is it still working well for you? How often do you use it? What's the largest thing you've baked? <br> <br>Thanks. <br>Don
It's working pretty well. Once the temp profile is tuned in, this oven produces pretty reliable results.<br>I only use it once or twice a week on average. The largest PCB I've done is about 5 x 5 inches, two of them together.<br>
i need some information for t962a <br> <br>We are a computer and mobile phone repair shop and we repair laptop and mobile phone motherboards, for example GPU on the laptop main-board we fix by applying certain heat by hot air gun and applying flux on it reflowing process, and liquid damaged mobile phone main-board we do ultrasonic cleaning and then heat the main-board to reflow. <br> <br>My understanding of this machine is that while doing the reflow manually we can just put the laptop mother board in this oven set the right time and heat setting and it will reflow the soldering on the ic and chips? am i right? if this is correct i will buy the machine and is it compatible with voltage in new zealand? <br> <br>hoping to hear from you soon? <br>
There is a lot of information about this oven - just google T962. My experience is that if you limit the use to lead-tin solder and avoid large components like USB jacks, this oven works well.<br> <br> I'd look for a better oven for what you are trying to do personally...<br> <br>
Thanks for this teardown of the T962A! I am looking at getting one, this is better than any review I have found!
Hey I just post my reply to this post, but It just got vanished somewhere, so I am reposting my comments. <br>I got you. But I can say the IR heating part is doing there work quite well, because I use an additional set up for Temperature Measurement using an Industrial Thermocouple+ Arduin+LCD display. What I found the heating element is actually working w.r.t Temperature increment resolution (Which is 1-3 degree C rise in temp per Sec) which is required for reflow process. All issue comes with Uneven heat distribution. Yah you are right , I will try out your Fan trick, n find if it solves the heating issue. Thanks for your support. <br> <br>Oh yes I will ask her to leave comment or query on your blog , in fact your blog is the best resource to know about your work.. :)
Is it.. Did you set up your own temp profile? Have you tested it? I hope Customized profile does not show up That time difference. I tried to bake two WiFi board using this oven, but due to this time difference the whole process ran over time , as result WiFi modules pop out n module got dead :( . Then I learn for this reflow process timing is very imp factor, even if you bake little more just for 2-3 secs components get damage seriously. So please share your results on the customized temp profile. :) And yes I am A Fan of your Led Art work. I showed your work to one of my Friend , who is an Interior Designer, and she totally blown away. She asked me if you would like to share your art work process from your blog.
I did set up custom profile, but the preset one worked pretty well for lead-tin solder. I just wanted to see if I can improve it.<br> <br> The built in timer is simply slow. It's the same when you use custom profile as well, so you can simply treat each 10 sec step as 15 sec step in creating your profile.<br> <br> I also did not find the&nbsp;timing&nbsp;to be too critical. With lead-free solder, I tweaked the profile to keep the peak&nbsp;temperature&nbsp;time as short as possible, but&nbsp;temperature&nbsp;accuracy (which I don't have the tool to&nbsp;verify) and heat distribution are more important.<br> <br> I'd attribute your problem of burning parts to uneven heat distribution (hot spots), and&nbsp;temperature&nbsp;error. This oven can be running much higher temp than what it's telling you.<br> Also taller parts get too much heat because they are closer to the heater. My fan trick might help a bit.<br> <br> Thanks for showing my work to your friend! Please have her email me or leave comments on my website, and I will answer her questions.<br> <br> <br> <br>
Hello Akimitsu Sadoi. good to know you have found a solution of that nasty smell. I bought the same model oven few months back. But during running this oven I found a serious defect in heating process. As per the profile configured for re-flow soldering the complete process should be completed in 7-8mins, but when I ran the process It took nearly 12 mins in actual .And a very strange thing is you can actually monitor that time delay error on its GLCD screen. I did some test running to monitor the re-flow graph , and I found this delay was happening through out the process.I hv more detail information regarding this issue n wud like to share . But you have also found way to improve heat distribution. I will try it on my system, hence will update you if dat solve that delay problem. <br> and All the best for your Maker Faire trip. :)
Thanks for your comment.<br> The timer is indeed very off - takes about 1.5 times what the panel says. So I took that into account in creating my temp profile.<br> <br> I think it's an usable enough machine to be worth the price, but could be a lot better.<br> <br> Please let me know more about your experience.<br> <br> Aki<br> <br>
muy didactico muy buen trabajo te felicito!!!!!

About This Instructable




Bio: I am an electronic artist living in Brooklyn, NY. I work with LEDs and microcontrollers to create beautiful objects.
More by ledartist:T962A SMD Reflow Oven Fix/Hack Color Organ Triple Deluxe II Aurora 48 - 48 RGB LED Sequencer 
Add instructable to: