Reflow Oven for Small Production

[flora]

When I got this I they had a larger model (twice the size) under way. So far I do not see it being available. Being larger it will probably take longer to heat.

The thing that I find very appealing about it is the simplicity. Maybe because I came from a pimped-up toaster oven. It had a PID controller, but I always needed to place the PCB exactly in the middle, attach the temp wire to a place central in the PCB and hope for the best. It mostly did the job but it cooled badly, so I always had to stay with it and open the door manually to prevent the PCB from overheating after soldering.  But still components overheated (lost their marking) while others did not solder (switches, larger fets). The white silk screen always came out slightly darker.

In comparison, with the vapor flow I place the PCB. Then turn on the fume extractor and the oven, select a profile and press start. I then go do something else, to come back 15 minutes later and find a nicely soldered PCB. Components large and small, no problem. Sometimes some solder balling, but that is easily sorted.

A while ago I blew a smart switch (TO252-5) on a proto I was torturing. This being a 4 layer PCB with a lot of thermal vias and cooling surfaces, it took a very large solder tip and quite some time to remove the component. Which made me question the feasibility of a soldering iron to get the new component on. So in the end I thought, what the h*ll, put some solder paste on the PCB, replaced the TO252-5 and put the PCB in the vapor phase oven for a second round. The results were amazing. The new component was soldered perfectly. All other SMD components were fine. PCB switches were fine. Most connectors were fine. All Molex KK connectors, the white ones, were affected. They now stand on the PCB like gravestones on a 150-year old graveyard. Still usable though. Next time I'll use the black ones. The only component that I had to replace was a thru-hole Recom DC-DC converter. Its housing was fine but it had completely melted off its legs.

For my use I like a vapor phase oven a lot better than my previous oven.

I recently purchased the same Vaporflow 275 oven to complement/potentially replace my Controleo3 and have been pretty happy with it. The solder joints are certainly much cleaner thanks to the inert soldering atmosphere.

However, it loves to tombstone my passives! Have you experienced this with yours? I've read elsewhere that similar low-end vapor phase ovens have an "anti-tombstoning" feature that turns on the cooling fan to extend the soak period, but I haven't tried making a custom profile to replicate this yet.

[woody]

I recently purchased the same Vaporflow 275 oven to complement/potentially replace my Controleo3 and have been pretty happy with it. The solder joints are certainly much cleaner thanks to the inert soldering atmosphere.

However, it loves to tombstone my passives! Have you experienced this with yours? I've read elsewhere that similar low-end vapor phase ovens have an "anti-tombstoning" feature that turns on the cooling fan to extend the soak period, but I haven't tried making a custom profile to replicate this yet.

Funny you mention it.
Using this oven I soldered a number of proto-boards (10+) with a variety of components from 0603 to 2220 without ever seeing tombstoning. Until last week. I then made another board. I usually drop the components on the solder paste and always give them a little push and shove to align them which also presses them into the paste. With this last board I did not do that for a couple of resistors (0805) that landed nicely. One of these resistors tombstoned. So I decided for myself that my old way of placing components was, although more time consuming, better.

Which makes me wonder, how do you place components?

A thing that I find more of a nuisance with this oven is solder balling. I see a lot more solder balls than when I used the home-made toaster oven. I have the feeling that this also might have to to with a too-short soak phase. My experience with the Vaporflow 275 is that it takes quite some time to reach 60 deg, after which it flies up to reflow  temp and starts cooling in under 90 seconds.

Although no showstopper solder balls are a nuisance as I have to remove them by hand. I have not tested my theory regarding longer soaking by creating a custom profile yet. First I need to make a couple of test boards for this and I am quite busy ATM. I will also take this up with Eleshop, maybe they did some testing on this.

I let you know the outcome.

[Fire Doger]

This one works perfect:
https://www.esdshop.eu/public/media/8115/hr-10-v.2.31-en.pdf

Price?

[JURP]

Approximately ~3200Eur basic version.
~3600Eur motorized door, ~4300Eur motorized door+nitrogen.

[flora]

Funny you mention it.
Using this oven I soldered a number of proto-boards (10+) with a variety of components from 0603 to 2220 without ever seeing tombstoning. Until last week. I then made another board. I usually drop the components on the solder paste and always give them a little push and shove to align them which also presses them into the paste. With this last board I did not do that for a couple of resistors (0805) that landed nicely. One of these resistors tombstoned. So I decided for myself that my old way of placing components was, although more time consuming, better.

Which makes me wonder, how do you place components?

A thing that I find more of a nuisance with this oven is solder balling. I see a lot more solder balls than when I used the home-made toaster oven. I have the feeling that this also might have to to with a too-short soak phase. My experience with the Vaporflow 275 is that it takes quite some time to reach 60 deg, after which it flies up to reflow  temp and starts cooling in under 90 seconds.

Although no showstopper solder balls are a nuisance as I have to remove them by hand. I have not tested my theory regarding longer soaking by creating a custom profile yet. First I need to make a couple of test boards for this and I am quite busy ATM. I will also take this up with Eleshop, maybe they did some testing on this.

I let you know the outcome.

Good to know you've seen it too. As I understand the rapid ramp of vapor phase is at least partly to blame for the tombstoning, and may have been one of the reasons the technology was phased out in favor of convection as SMT became more mainstream. I'm sure some profile tuning can mitigate it...

My boards have parts ranging from 0402 - 0805 (imperial) and the 0402s are especially prone to tombstoning (at least half were affected.) I hand-place all the parts and press them down into the paste pretty firmly. Though with larger batches the paste loses a bit of tackiness by the time they are reflowed, since stuffing takes so long!

I thought maybe my board layout was exacerbating the tombstoning, since I don't have thermal reliefs on this particular design; but after inspecting some boards it seems there's no rhyme or reason to which side of the passives reflowed first.

I have noticed a few solder balls, but nothing outrageous. Pretty comparable to the Controleo3. I should mention that I'm using TS391AX 63/37 paste, so the 230C peak temp is overkill. But, Eleshop assured me other customers use leaded with Galden LS230 without issue. I'm just happy that there's no graping or cold joints with this process-- my Controleo3 had issues with that on some boards.

[dkonigs]

I have noticed a few solder balls, but nothing outrageous. Pretty comparable to the Controleo3. I should mention that I'm using TS391AX 63/37 paste, so the 230C peak temp is overkill. But, Eleshop assured me other customers use leaded with Galden LS230 without issue. I'm just happy that there's no graping or cold joints with this process-- my Controleo3 had issues with that on some boards.

When I first got my Controleo3 oven, I used TS391AX and SMD291AX paste (and some similar formulations), and I constantly had solder balls on all of my resistors and capacitors.  I also frequently had bridges to clean up after reflow.

But since I switched to using GC10 (and probably a slightly tweaked profile too), the problem has almost completely gone away.

SAC305 (GC10, et al) solder doesn't look quite as pretty as the leaded stuff when it cools, but it seems to actually work quite well.

[danielm]

I just bought iTECH RF-A250 for 350EUR including shipping to Slovakia. Will update on how it works. It should be similar to RF-A350 (both IR  + hot air), just a little bit smaller.

[MR]

Hi,

do not use any oven with infrared heating, it needs to have hot-air(convection) heating. IR will have different effect on dark and bright components, so you will never get reliable results and profiling will always be a problem. Power should be as high as possible, because you will never get too much heating on single phase, be sure that heating ramp with PCB inside is 1°C/s or more. At least 2,5-3kW is needed, so it needs to be for 208VAC.
I know many Chinese manufacturers, but the only oven in this price range that is hot air and powerful enough is T-937M from various manufacturers, Puhui i.e.
Buy a used LPKF ProtoFlow S (not S4!), this will be your best choice. It's reliable and with very good soldering results.

 

I totally agree with that.
Either you know a really good oven by yourself or don't spare the money it will save you from a lot pain seriously.

A very sweeping statement.. We ran an IR, conveyor based oven for years - one of these https://www.mekko.co.uk/ovens-for-reflow/ . Certainly it has its drawbacks but component colour was not really an issue with FIR elements, its biggest enemy was mass. Chonky inductors. SM transformers, 240pin FPGAs, already sub-optimal unbalanced layouts, these things were either basically impossible to reflow or came out with tombstoning. Your day to day boring circuits however, were fine and you would find many of these issues with an underpowered "convection" oven too with similar workarounds - cranking the temperature and longer process windows.

I have a 4K EUR oven and a 20$ pizza oven... the advantage of both of them is that they have a window to look inside. The 4K oven strictly follows the temperature curve as it has enough power as well - it's just hit and go absolutely reproducible.
The pizza oven - well it depends on the human controller.

The chinese 300$ IR oven which I have bought and modified in the past did not work out well for me, it didn't melt the lead free solder below some components it was not reproducible.

So for small volume / prototypes there are quite a few options.

[48X24X48X]

I have the larger IR desktop from Puhui. It does work for lead free but within a smaller area than the entire tray size but it suck big time with plastic connector. My hacked toaster oven works perfectly with lead free but has size limitation (any bigger sized toaster would be too big volume for the heat to sustain). There's 1 unit from HWGC that runs on convection but much pricier than Puhui ovens, would want to try that but it's not so desktop size and it runs close to the 13A limit.

[IanJ]

I am using GC10 (there are two ball sizes available) with a T-937 for production of my products. The 937 is streets ahead of the 962 as it has an active exhaust.
The 937 is still a cheap Chinese oven so learn its bed profile and from there great consistent results.

Ian.

[Pinkus]

I am using GC10 (there are two ball sizes available) with a T-937 for production of my products. The 937 is streets ahead of the 962 as it has an active exhaust.
The 937 is still a cheap Chinese oven so learn its bed profile and from there great consistent results.

Ian.

I also have a T937 and I can only confirm this. If you have problems, you should run the oven at a lower temperature with some seconds added to the profile.
With very high plastic connectors (e.g. RJ45) I sometimes place a piece of aluminium foil over them to protect the top from melting. I never had problems with usual 5-8mm high connectors like JST PH etc.

[n4te]

I have the Whizoo ready-to-run. It works well and I think hard for other options to compete given the price and size.

I am unable to use a single profile as fire and forget, as different sized boards/components change the cook time. I created a profile for TS391AX that is the longest I ever want to cook, for my largest boards or when cooking multiple, then I have to watch near the end of the cook and stop it manually.

Given that, I think the OSS software UI could be improved a little. When stopping manually, it could open the door. It could have a button to extend the cook, since if it hits the end but the solder isn't melted then your board is ruined. Not having such a button means I have to use a long running profile and if I am not paying attention and let it cook to the end, it would overcook a small board.

New TS391AX works great but I had problems with older paste. Even paste that is not yet expired but is near the bottom of the jar doesn't work as well as from a new jar. With NC191SNL no lead paste it was even worse, unless the paste is quite new some board are ruined. That makes me scared to try other no lead pastes, even though I hate lead paste. I'm very careful but it's super easy to accidentally get it everywhere.

[trampas]

I have tried IR drawer ovens and they are crap, even with all the mods they are crap.  I ended up sending my T962 to recycle bin.  Even with IR shields I would get hot spots on PCB, burnt areas and then unsoldered areas.  Was  the T962 was total junk and waste of money.

I modified a 1/4 size oven with insulation, convection fan and heaters using the Controleo3 controller.  It works well, the Controleo3 PID is a joke, but I was able to get it roughly follow profile with some hacking.  I found the oven works well for leaded soldering but lead free I still have issues.

For lead free I found that the thermal mass of the PCB makes a big difference, specifically the Controleo3 only has one thermocouple.  As such you have to manually configure the overshoot and soak times to compensate for the thermal mass of PCB.  I wish there was a second thermocouple where you could attach thermocouple to unpopulated PCB and leave in oven such that you have one thermocouple for PCB and one for air temperature.  Then the controller could dynamically change soak times, while not exceeding air temperature.   Note I am considering doing another PCB design of the Controleo3 to add this or maybe even moving to Raspberry PI.  I would really like to have a web interface, and hence a Raspberry Pi will most likely win.  Anyone interested on helping with this?

Note something that is very important is to get an oven with exhaust.  Either exhaust it outside, or run it through a carbon filter.  The fumes from the reflow process, especially when you over bake a PCB are bad. 

Also note that I double sided PCBs are a problem.  I have to setup custom jigs to hold the PCBs to reflow second side.  I am not sure how the conveyor units do this, I assume someone must sell a holder or something?

On the double sided lead free, I am finding that I can get first side to reflow great, as well as the second side. However when I reflow the second side, the first side gets burnt coloring like it is over heated.  I have not root caused this yet, but suspect it is because of bad thermal control and air getting too hot.

If you want a good reflow oven, especially off 110V here are few things I learned:
1. For 110V operation with 15A circuit you really need to have low thermal mass.  This means no metal inside heat chamber and good insulation.  1800W is barely enough to get 1 degree C per second ramp rate unless you have low internal thermal mass and good insulation.  Better to just go with 220V and be done...   
2. You want convection heating,  you want all the air circulating to have same air temperature all over PCB.
3. IR is evil, it might work on conveyor systems (not tested).  However in a stationary oven the IR does not work as it can cause uneven heating.
4. You need more than one thermocouple so you can measure air temperature and PCB temperature.
5. Do not buy the cheap gold insulation as mentioned on the Controleo3 website, that cheap stuff stinks even after years of use. I regret this purchase and really can not fix on my current oven.
6. You want external vented, again I can not stress this enough!  Do not consider any oven that does not have means to exhaust fumes outside.

[n4te]

@trampas When your double sided boards burn on the bottom, are they touching the oven tray? I am using a Controleo3 prebuilt. I prop my boards up with old blank PCBs so they don't touch the tray. Same when doing double sided. When double sided with heavy components (eg SMT ethernet jacks), I stack old PCBs so the heavy components are supported.

I don't see how an oven with exhaust would work, unless you meant to exhaust the air around the oven. Exhausting the air in the oven would kill the temps. I setup a step ladder outside, put a long piece of tile across the rungs, and place the oven on that. It's eye level and the fumes waft away outside. I would never use the oven indoors.

[trampas]

I use some aluminum scrap or PCBs to prop up the boards for double sided.

The exhaust is to remove the smell.  Currently what I have is an exhaust fan outside the oven, like a range hood that is exhausting while oven runs.  Then when oven is done and door is opened it exhausts that as well. 

I would like to keep the oven sealed and have exhaust port that opens inside the oven and then blow fresh air, maybe by air compressor, to cool off boards after reflow.  The idea is to keep all the fumes in box, or exhausted outside.

What I found is for 110/120V operation with 20A breakers you need really good insulation.  So I basically have no metal inside the oven other than wire tray to minimize thermal mass.  Everything in the oven is ceramic fiber insulation coated with the gold reflective tape.   Then I use a  convection oven heater element and fan, I added secondary heater elements to oven to get higher ramp rates.   This works really well.

My next oven I would build similar but with an exhaust port and use the better gold foil.

[pastaclub]

I am considering doing another PCB design of the Controleo3 to add this or maybe even moving to Raspberry PI.  I would really like to have a web interface, and hence a Raspberry Pi will most likely win. Anyone interested on helping with this?

I am co-author of the uReflowOven project: https://github.com/dukeduck1984/uReflowOven-Esp32-Micropython. Similar to Controleo I guess (although I've never used that product).

uReflowOven is an open-source reflow controller with PID, touch screen and Wifi. It's based on an ESP32 and written in MicroPython. Several people have built one and they use it with toaster ovens and hot plates. If you want to make your own controller, you might be able to save a lot of time by starting with this and then modifying it to your needs. It has some quirks, and the PID could be better, but again, it's open source.

[n4te]

uReflowOven looks interesting! I like that the screen doesn't stick out, as the Controleo does. The Controleo build shows more insulation. Controleo measures the air temp while it looks like you are attaching to the PCB? That could make a huge difference.

Controleo is also OSS. The build guide and software is provided. They sell kits to build your own or you can buy a readymade oven. Have you considered something similar for uReflowOven?

I understand "NIH", but it sucks to duplicate work. Merging the efforts of both projects could benefit them both.

My readymade Controleo3 has had some issues: a wire was disconnected on arrival. An easy fix, but confusing. Then recently an SSR died. It was a bummer because many cooks got wrecked before I figured it out. The fix was more extensive, as the oven is riveted together. It required some doing, but I replaced all 3 relays with brand name SSRs (Crydom). Whizoo was responsive about the issues and made everything right in the end.

[pastaclub]

I like that the screen doesn't stick out, as the Controleo does.

Enclosure-wise you can do whatever you like since it's more or less just a firmware project with some information how to wire it, but we don't even have an approved PCB. The circuit is rather simple and both Kaiyuan (the other author) and I use perfboards with soldered wires and since those work, there was no need to replace them Someone made a PCB in KiCAD at some point and I added that to my next JLCPCB order, but I found that the touch screen traces were missing, the gerbers were different from the schematics and the person who made it didn't answer, so we didn't make this the "official" board.

Controleo measures the air temp while it looks like you are attaching to the PCB?

You could put the thermocouple wherever you like. I think Kaiyuan puts it in a through-hole of the board in his oven. My build is a hotplate where the thermocouple sits within the plate, so it measures the temperature of the plate, which of course will differ from the temperature of the board which is exposed to room air. But it's not an exact science in any case. The plate itself doesn't even heat evenly, so it takes some tweaking until you get good results. It's certainly not plug and play, and I wouldn't do 400-pin BGAs with it, but for many applications it's good enough.

They sell kits to build your own or you can buy a readymade oven. Have you considered something similar for uReflowOven?

No, I think we both have no ambitions to make it a product. There are many out there already. I think the appeal here is that you can easily change the firmware if you don't like something. I hope more people will improve it.