Improving a T-962A Reflow Oven Part5

[@safjkjlkasnhjlhafsnj]

Thanks for your excellent contribution to upgrade the T962 owen. Your article is the first which put attention at the physical design fail in the construction, fail no fancy temperature controller will deal with. (Of course a cold junction temperature controller will deliver better accuracy). I also noticed that the onscreen watch is not exactly. 1:00 minute for the oven, is 1:35 min on my stop watch. This, combined with the quite long factory solder profile, can boile away the flush. Thanks again Jerry 😊

[@gregoryfink7368]

My interest in your modifications is not waning. I am very interested. I think you relocated one of the two thermocouples. what is the location of this relocated thermocouple? What is the RPM of the recirculating fan? Do you have a mechanical profile of the recirculating fan blade? Do you have a mechanical layout of the IR heater baffles? You referred to these baffles as being wider in the mid point. Thank you for your diligence/methodology seen in your videos.

[@johnleclair663]

Hello Jerry. I just absolutely loved this series - especially running experiments, gathering data, analyzing data, repeat. Outstanding analysis. Thanks. I have modded my T-962A with cold-junction and one of the open source firmware images (Unified Engineering I believe. I changed the firmware mainly for a better user interface. Within the past month, I re-fllowed 3 boards with 484-balls FPGAs. Of the three 1 functioned correctly, 2 did not. Actually, one of the failed boards worked after reflowing it with my Quick hot air rework station - probably out of shear luck. Fortunately, I bought "disposable" FPGAs for 1/3 the price off a Chinese ebayer knowing the re-flow would probably be a disaster. I bought a pair of cheap temperature meters with thermocouples - the two could not agree on the temp on any surface - or themselves to try and see what was going on. Then I discovered your series. However, duplicating any of your work with Amazon temperature meters/thermocouples would be impossible. So I would love to follow a few of your improvements. However, I have a few questions that might help others too. I do not recall (my apologies if you did mention) seeing any approximate measurements for your heat baffles. I understand yours were wider in the middle. However, I would be happy if you could share an approximate width value. In addition, any hints on the motor you used and get an idea of RPM ? Thanks for your time.

[@JerryWalker001]

Cheap thermocouple meters are fairly useless as they are inaccurate and inconsistent. My setup was fairly complex to get good results and was calibrated against a TTi7 reference meter (accurate to 0.001C). The modifications I ended up retaining (although I tried many others) were, 1) Baffles 2) Fan 3) relocate sensor 4) Door seal. I cannot give enough detail in a comment so I will post a video shortly explaining the modifications. It may be possible to further improve the oven but as these modifications give me accurate and consistent results which track the set profile then I see no real value in doing this considering that the oven will still have limitations.

[@jameswolf297]

@@JerryWalker001 As I have been considering the purchase of a T-962A in the very near future, I am well aware of the problems users of this machine have been dealing with, and all of the solutions being offered by others. It is amazing how much equipment is still coming out of China with the same problems present, as when the product was first introduced. I don't mind taking on the risk of buying a product with shortfalls, if I can see a path forward. None of these modifications are difficult nor expensive to do. I may still apply the open source firmware fix, but first I want to see how well the unit operates with the heating compartment issues having been addressed. Fix one issue at a time, then re-evaluate. Thank you for taking the time to apply your knowledge and resources to this problem. You are on the correct path. I look forward to your concluding video(s).

[@MrOrangeman18]

would love to see an update on this, is there one coming anytime soon? steve

[@TomislavDarlic]

Thanks for this excellent series Jerry. I've been using this oven for years with updated firmware and junction compensation. Pretty much gave up on soldering lead free regular solder with it as I needed to raise the temperature so much in order to reliable reflow all board that the parts of the boards were burned. I now exclusively use bismuth solder paste with this machine. It's lead free and with much lower reflow temperature. Once see your upgrades complete I'll probably do something similar myself.

[@JerryWalker001]

You may find this video of interest as it shows results using lead free paste. kzbin.info/www/bejne/r6PFdoJuoJd8jdk

[@joyztik]

Hello Jerry. Awesome work you have done here. Im sure most, if not all T-962A owners have thought at least once about modding the chamber to get better performance out of it. I am really looking forward to an update explaining all mods, especially the lamp baffles. Those, I think are key to removing the hotspots which, in my case, prevents me from using the oven for larger boards. I really appreciate your invaluable efforts. Keep it up, and you just got a subscribe.

[@joyztik]

Hello Jerry. I know you have been busy. Just wanted to know if you had plans to update us on the final results and list of mods you came up with. Id love to give them a try.

[@symeonkrizias2534]

Hello Jerry, sorry for insisting on your description regarding the re position of the 2nd sensor.In your detailed reply to Loic G , you say we have to use a tube fastened on the side wall,this means that there must be a hole made suitable to accept the tube and in order to keep it fastened ,must ''solder'' it using argon or something similar?and then ,close to other end and a distance of 50mm from the sidewall so the sensor will protrude from the other end...this means that the internal tube diameter must be chosen to accept it tightly......otherwise how it can be fastened? is there any particular reason not fastening it in its new position as it was originaly fastened?

[@MrTon4eff]

The easiest way to simulate a conveyor for this oven I can think of is to rotate the pcb, like in the microwave oven... :) I'm using a modified toaster oven for my pcb's and even it's smaller, I have the same problems with the hot spots... The biggest problem with direct IR heating is dependency from the color of the boards and components - darker heats faster. I'm using halogen lamps as a heat elements, maybe changing them with quartz heaters will improve the situation - longer IR wave should heat deeper in the materials. Experiments continue... :D Thank you for the ideas...

[@JerryWalker001]

I did try rotating the pcb but it does not work very well as the centre does not change position relative to the heaters and also the max board size is very small (as the diagonal needs to fit in the drawer. The best solution I found was to fit a motor beside the rear fan and use a crank and link (spring loaded plunger to allow the drawer to open) to push a pcb platform back and forward. The crank should be just under 1/4 the length of the spacing between the heaters and this gives an average which is close to constant as each part of the pcb spends the same amount of time (on average) at each distance from the heaters which simulates a continous conveyor. A cycle speed of around 12 per minute works well. I tried this and it does improve things but as I was getting good results with just a fan and difuser I am now using just that and it works very well.

[@stephenprovost7030]

Jerry: outstanding series of creating a more stable oven for reflow soldering. Have you created a document on the "simple" changes everyone can follow. Also do you have a follow up video since nov. 2019? thanks

[@JerryWalker001]

Thanks. No documents, only videos. I did not make any further videos as interest seemed to be dropping but I do mention the final upgrades I used in this video kzbin.info/www/bejne/pJfCdnSkgtJ4bbM

[@rajeshbalihar8691]

Sir , I have a reflow oven but sometime it temperature get high and sometimes it temperature get low. Sir please you can explain me what is the reason to be get high and low temperature.

[@gregoryfink7368]

When you installed the recirculating fan, you removed the existing sensor. Where did you reinstall it?

[@vitsedlacek7387]

hi Jerry, many thanks for this series and sharing your knowledge! I wonder about the ideal cooling rate - based on all the reading I did, I've figured that 2.4-4degreees celsius might be the sweetspot for good balance between strenght and longetivity and I've got a little frustrate that I am barely achieving 1 degree rate, ... it is reassuring it doesn't seem to concern you and maybe it's not such a big of problem as I thought! :) Would you care to elaborate, please?

[@Pops180]

How effective would the added fan be on a smaller PCB? Doesn't the large PCB (with thermocouples) effect the airflow? Just curious if your fixes would inherently be more effective with a very large board vs a very small one. Also, what size baffle did you settle on under the heating elements?

[@JerryWalker001]

The board size has no bearing on the radiant temperature profile. That is the point of adding a fan. The only thing that should be affecting board temperature is radiant heat which should be evenly distributed and conduction or convection should be minimised. This is the function of the fan and baffles. The goal is to create even radiant heat over the area of the cabinet and not rely on the board to conduct the heat. I run boards from 15x28mm to the boards I show in the video and they all now reflow very well.

[@Pops180]

@@JerryWalker001 i see, for a second i thought the added fan was underneath the tray. Do you recall the baffle size for the IR heaters?

[@Pops180]

@@JerryWalker001 also, did you ever try to upload new firmware and the cold junction fix? I was able to get very consistent results.

[@chrismr3972]

One thing I thought, how much of the improvement is due to the board acting as a duct for the air flow from the fan? For example, if your instrumentation board had some large holes in it that would simulate more circumstances - like having lots of separate boards, or just one smaller one in the middle.

[@JerryWalker001]

Good question. This is why the fan 'pulls' the air up rather than blowing it down. The air then spreads out and flows down the sides of the chamber and reenters underneath the tray. This in theory should eliminate the major board effect. I have already tried it with smaller boards and it works better as the air can flow more freely but I wanted to show worst case in the videos. I am actually carrying out far more testing than I am showing but will go over anything that ends up being relevant to the final design once I get there. As this is an IR system and not really an 'air' heater then the air circulation is to prevent hot spots rather than be an active part of the thermal cycle. The general idea is to draw air 'across' the board rather and down onto it so this will avoid big differences between small boards and large boards although there will still be better mixing with small boards but I figure if it is better than I am demonstrating then that will not be a bad thing.

[@Conservator.]

Hi Jerry, Have you already planned an update video for theT962A? I don’t mean to sound impatient, I’m just curious since you’ve already shown such excellent results. Yesterday I finally received the power supply for my RD6006 and installed it so guess what video’s I’m going to watch next. 😉 I’m happy to be your 609th subscriber ;).

[@msp2402]

Hi Jerry, I watched all the parts about the T962A from you now and I see you achieved some major improvements with your modifications. You are also explaining everything that you did with a lot of background but, please correct me if I'm wrong, you didn't share more than the temperature results. For example, you experimented with the size of the baffles but didn't share the actual sizes that worked. I am not a native English speaker so maybe I just missed these pieces of information, but it would be very interesting for me. Or do you have a webpage where I can find them?

[@paulgallagher2937]

I've got 2 of these. The first one was just IR heating. The consistency was terrible. I took a chance and bought the model with the forced air circulation. It was actually worse. I've managed to improve it somewhat by modifying the shape of the blades and moving the temp sensor away from the edge of the furnace. Its really a shame they put all this time and effort into a fairly nice product, but then don't take the extra time to tune it to make it perfect. I'm probably going to give up on high temp lead-free solder with these machines.

[@JerryWalker001]

I think they miss understand the problem. The modifications I made are not really a fan as it is not 'blowing' air around. It is a 'mixer' to create turbulence. In a later video I show the profile of this. It also rotates fairly slowly. I now get excellent results while still using the original controller and firmware. There are other changes you need to make at the same time which I also mention in the videos. For example do not put the boards on the metal base but raise it a few mm to prevent heat transfer.

[@paulgallagher2937]

@@JerryWalker001 You also need enough air flow to heat up larger components like USB plugs, etc. Its not just a matter of having equal air temperature, you need to transfer the heat into the PCB and the components. SO yes, you can mix the air to get same temperature, but that will not always give you good reflow results. The big pro machines have rather strong blowers. Not so strong as to move the chips, but enough to transfer heat.

[@JerryWalker001]

@@paulgallagher2937 It does not use air to heat the parts, or at least is should not. That would just cause uneven heating, It is an IR heater and not an oven so any air movement is only to prevent heating by air. This is how it achieves fast heat and cool times. It should not be trying to heat up the inside of the chamber as this will cause all sorts of problems. It is best to have zero air flow during heating or the profile will be all over the place and will never follow the programmed profile.

[@paulgallagher2937]

@@JerryWalker001 The problem is, IR heating for reflow is really, really bad. It cannot heat all components evenly. And, it will have a scattered pattern of hot/cold spots. You can get reflow on one area of a PCB and 25 mm away, no reflow at all. And shiny components such as large connectors, won't reflow at all with IR. You would need to cook the rest of the PCB to get them hot enough. These cheap ovens use IR because its cheap, not because its correct. The convection fan is an add on to try and improve a poorly designed product. Its a compromise. I got mine working, but barely. Once you start relying on air to heat the components, then you must have significant air flow. I mean, you can put a can of soda in your fridge, and it doesn't instantly get cold. Large amount of forced air will greatly reduce the time to get it cold though. So its basically a hybrid. IR is heating the dark components, and forced air is helping to heat the components with low emissivity. When I get rich and famous I'll buy a proper flow through forced air machine. With IR heating, you can't even trust your temperature sensors. That is the temperature that the sensor is, but a component with different emissivity could be very different. In the end, you just have to experiment with the profiles to get the desired results, and not worry too much about the profile.

[@JerryWalker001]

@@paulgallagher2937 I disagree and have proven this with my own oven which as I say now works perfectly and gives very good results. Even our professional ovens struggle with some components using lead free paste. For example we have to add some parts manually, such as SM transformers. That is not unusual and fully expected. However simply blowing air around an IR chamber will not help as the air is always going to be cooler than the heated parts. Professional hot air ovens have air heaters. Yes the IR pattern is important which is specifically why I added the baffles which I explain in some detail in the videos. I also show a few runs where the oven now exactly follows the profile across the entire area and that is all the oven is expected to do. It is then a matter of creating a profile which is suitable for the board you are trying to solder but this only works if the oven reliably follows the profile. Even then you often have to compromise but the smaller parts are generally closer to the board and so subject to more board cooling which tends to even things out. Either way my original point is that simply fitting a 'fan' will not help and is likely to make things worse as the air is not heated. My solution is Not a fan but a mixer and baffles.

[@glencook9379]

Hi Jerry, really interesting series, pity it ended so abruptly, although I appreciate it was a lot of work. Can I ask a couple of questions that may help other people watching the series. 1) I take it you ended up with 2 baffles in the centre of the machine after all your testing, what was the final width of the baffles. 2) The additional fan added above the heating elements, what length fan blade did you find optimal and what rotation speed? Cheers!

[@glencook9379]

Whoops, sorry, I've just read your reply to Loïc G. Thanks for the informative series :)

[@Difool80]

Hello Jerry, First of all, thanks for this serie of videos. I loved the experimentation and data analysis you did. As well as the very simple modifications you did to get such impressive improvements ! I see in another video the conveyor idea was finally too complicate for not so big improvement and the only useful modifications are : 1. Baffles 2. Air recirculating fan 3. Sensor relocation 4. Door seal 5. Cooling fan baffle However, is it possible to get some more details about it ? - What is the baffle width you ended with ? - How did you fixed it ? (The distance from the IR heaters should probably be as much important as the baffle width) - What kind of fan blade shape did you end with ? (Maybe you could post a picture ?) - What kind of fan did you used ? (RPM could be enough but if you have a reference ...) Did you also get rid of thermocouple cold-junction compensation ? I don't remember hearing about this in your videos. I should receive a T-962A+ soon and I plan to do the same experiments you did using a Seneca Z-8TC thermocouples input module in order to improve it too. Thank you very much !

[@JerryWalker001]

The baffles were just 0.8mm stainless steel wire. I mounted them by drilling small holes 10mm below the bottom of each heater and then fitting a short spring on one end (This stops them sagging as they get hot). They are thin and are only intended to create a 'shadow' directly under each heater. The front sensor is left in its original location but lowered to just below the heater centre line. The door seal is bolted to the top of the door and fits snugly into the chamber opening when the door is closed to totally seal the door. The cooling fan baffle is a piece of 0.5mm stainless steel sheet and it is 150mm wide and fitted inside the chamber in front of the fan inlet grille. It is only attached at the bottom and the top ends 5mm below the inside 'roof' of the chamber. This is to cause the incomming air to blow around the chamber rather than directly onto the pcb. The second sensor is moved from its original location in the centre of the chamber and fitted in through the side of the chamber so that its tip is 10mm lower than the heater centre line and 50mm from the chamber wall (A piece of stainless steel thin walled tube 35mm long is used to support it (similar to the original but from the side and not the top). The most important part is the 'fan'. This is Not intended to work like a normal fan but is intended to generate turbulent air motion within the chamber. The motor is as shown in the video but must be mounted on 25mm spacers to keep it well clear of the chamber. The fan blade is fitted to an extension piece I made which fits onto the motor shaft and positions the fan blade mid way between the chamber 'roof' and the heaters. It was made from 5mm wide stainless steel and the length is as long as possible without hitting either of the sensors. It is twisted as close to the centre mounting point as possible to opposing angles of approx 85 degrees (ie almost an impeller). The motor is driven via a 3V regulator I wired to the internal supply on the control pcb and this gives a motor speed of approx 500 rpm, It should not be too fast. I did Not modify the firmware at all. In some of my later videos you will see that the results are profiles that are not only very close (+/- 2C) of the programmed profile but also very even across the chamber. I hope this helps. I did not take any photos at the time as there did not seem to be much interest. This oven has now been run for a total of approx 120 hours since the modifications and it has worked perfectly.

[@Difool80]

@@JerryWalker001 Thanks for this prompt and detailled answer !

[@JoeBlow-zr2ru]

@@JerryWalker001 Such a great video series ... thank you. It helped me make the decision to take a chance on one of these ovens. One thing I have become confused about ... the original long narrow aluminium strips that you hung under the elements ... these were replaced with a single 0.8mm dia wire under each element? I was trying to find the dimensions, but now I'm concerned that I've misunderstood, because that seems like it would create an almost negligible shadow effect?

[@soldermaster4737]

First of all I enjoy your videos, so keep up the great work. It appears to me that you are concentrating on the enclosure and hardware setup of the T962A and NOT the micro-controller. Many other KZbin Videos looked at the micro-controller and left the enclosure and setup alone. The micro-controller on the T962 and T962A has major problems, one of which is that the thermocouples have no cold junction compensation. Another problem is the firmware is very poor and doesn't work very well, if at all. Many KZbin people have flashed the micro-controller with new open source firmware and added a cold junction fix to the circuit board. The best fix that I have seen is a project on Tindie: www.tindie.com/products/pnoxi/t-962t-962a-reflow-oven-controller-upgrade/ This circuit board attaches, via magnets, to the top of the case and it has either two or four thermocouple inputs. It only requires minimum changes to the micro-controller circuit PCB (Move the two thermocouple leads to the new add on PCB and solder one wire to the original PCB, cut a USB hole in the side of the top case). It also flashes the micro-controller with new firmware. Check it out.

[@JerryWalker001]

Yes I have seen the various upgrades and I mentioned them in my first video. From the testing I have done I do not think that any amount of firmware upgrades can properly correct the problems I have seen so far. There is a fundamental issue with the temperature distribution. I will be looking at the controller and firmware in later videos but you may be surprised how well the original controller works after correcting the issues with the chamber. The controller can now follow the programmed profile very closely. That is the main reason I am focussing on the chamber to begin with as it will have by far the greatest effect on improvements. Simply messing with the firmware is akin to polising a Turd. The main problem I have with the controller is the poor key response and menu system.

[@MrOrangeman18]

@@JerryWalker001 i agree, you have quite clearly demonstrated in your videos that a well thought out design/modifications to the main carcus of the chamber is really all that it needs, your graphs are looking very smooth now with your modifications and you havent even touched the firmware yet. nice job jerry