T962 Reflow Oven?

[jpanhalt]

Any comments about or experience with the imported T962 reflow oven?  Is the only problem the controller?

Comments on 63/37 solder paste that is not Kester brand?  My only experience over many years is with Kester rosin core solder.

Thanks, John

[Dubbie]

I have one.
I replaced the firmware and did the other mods floating around the net.
pretty happy with it. All my boards seem to come out very well.

I initially intended to use the shipping firmware, but it really is completely unusably terrible.

[coppice]

People tend to find that the Qinsi QS-5100 is like a T962 with most of the really bad issues fixed.

[Bud]

Comments on 63/37 solder paste that is not Kester brand?  My only experience over many years is with Kester rosin core solder.

I use MG Chemical 4860P 63/37 no clean one and am happy with it.

Edit:  i prefer Kester but it is hard to find here

[jpanhalt]

Thanks for the comments.   I have been watching T962 prices for awhile.  I picked up a new one ( I hope) for less the $200 including shipping and tax and plan to implement most of the improvements to the firmware and other issues that are described on the internet.  They are summarized on GitHub.

MG solder was going to be my first option as the company has been around awhile.

[asmi]

I used it too for last 3 years, after some HW modifications and firmware change. The only problem I ever had was with connectors made by Molex, all others worked just fine. Only recently I grew out of it, so I ordered ZB2520HL oven, which I will hopefully receive in the next week or two.

[jpanhalt]

@asmi
What was the problem with Molex?  Were the footprints off, did the plastic melt, or something else.   Molex uses a variety of different plastics.  If you can recall one that failed and any other that didn't, I'd like to look into it.

 

[asmi]

@asmi
What was the problem with Molex?  Were the footprints off, did the plastic melt, or something else.   Molex uses a variety of different plastics.  If you can recall one that failed and any other that didn't, I'd like to look into it.

The footprints were/are OK. The plastic they use seems to be especially susceptible to IR reflow, so it melted

As you can see, other connectors are just fine. So now I try to avoid using Molex parts if I can, and if I must, I solder them manually after reflow. It really is puzzling when even cheap-ass Chinese parts from Aliexpress survive IR reflow with no problems, but a well-known brand ones can not If you want to know specific part numbers of these parts pls let me know - I can try to dig them out as I should still have some leftovers.
Now I prefer using Samtec connectors, their plastic is not affected by IR at all, and they have the best online services - they don't force you to fill out a form with millions of fields if you want to get some samples for fit-checks, you can order in-stock parts directly with no MOQ nonsense, they have 3D step models for most of their parts publicly accessible. But unfortunately they don't make some types of connectors I use, so I have to use others every so often.

[coppice]

So now I try to avoid using Molex parts if I can, and if I must, I solder them manually after reflow.

This makes no sense at all. They melted because of the type of plastic, not the company that moulded it. Molex use many types of plastic across their range of connectors, as does every other maker. Some plastics offer higher strength, or a greater temperature range, or greater flexibility, or very low cost, and its up to you to choose something that suits your needs.

[jpanhalt]

@asmi
I agree it is the type of plastic, which seems obvious, but a generic descriptor like "polyamide" (Molex) may not be the same a "polyamide PA6T" (JST) just as epoxy FR4 can have many different temperature ratings.   Unfortunately, my current design uses both a Molex Pico-Lock and Pico-EZmate for space reasons.  I am sure there is a JST equivalent for one but not the other.

If you have the time to find the part number for one of the Molex connectors that melted, I will dig into it.  I have a background in organic chemistry.

I might add that I modified the footprints for both by elongating the pads to allow hand soldering, but still, it would be nice to know what to watch for.

EDIT:  As suspected polyamide 6T may be quite different from a generic version:

ARLEN™ Modified Polyamide 6T
ARLEN™ is a heat resistant, modified polyamide 6T developed by Mitsui Chemicals, Inc. With a high melting point (320°C)...

[KE5FX]

Similar experience here.  Results from my T962 with ESTechnical mods generally range from acceptable to excellent, but the plastic used in certain parts often shows some heat damage.  JST connectors such as SM04B-SRSS-TB(LF)(SN) seem hopeless, as do Mini-Circuits parts with the CD636 plastic package such as ADP-2-10+. 

It probably has more to do with the plastic's IR absorption characteristics than anything else.  There is only so much you can expect from an oven that blasts the board with IR at extremely close range, even with added airflow.  These parts all look fine on boards that have been factory-assembled on an industrial-quality SMT line.

[asmi]

This makes no sense at all. They melted because of the type of plastic, not the company that moulded it. Molex use many types of plastic across their range of connectors, as does every other maker. Some plastics offer higher strength, or a greater temperature range, or greater flexibility, or very low cost, and its up to you to choose something that suits your needs.

I'm telling you what worked for me, and that is avoiding their parts. If you have a different experience - feel free to share with us. I don't have anything againts Molex as a company - it's just their parts tend to cause problems for me - hence my tendency to avoid them.

[asmi]

If you have the time to find the part number for one of the Molex connectors that melted, I will dig into it.  I have a background in organic chemistry.

First photo - parts 54104-4031 and 52271-0679. Parts deformed enough to become non-functional.
Second photo - 87832-1420. Part deformed significantly, but I was able to restore functionality by cutting away melted parts of shroud.

I suggest you to do a "dry run" before reflow and see if parts you plan to use will survive. Also once you receive your oven and you do all HW modifications (cold junction mod is the most important one, along with community FW), take your time to calibrate the oven. Also never place board to be reflowed straight on a drawer - it will suck a lot of heat off the board and reflow might fail because of that. I use makeshift standoffs made from M3 screws and a couple of nuts, or just a few old scrap PCBs if the board doesn't have mounting holes. This was required for most of my boards anyway due to double-sided components placement.

For solder paste, I use and highly recommend this one as it's room temperature stable, so no need for refrigeration, plus it doesn't dry out after you print it onto a board, giving you plenty of time to place all parts (this one is super important for me as often have 200+ parts per side).

[asmi]

Mini-Circuits parts with the CD636 plastic package such as ADP-2-10+.

I just recently reflowed a board with ADTT 1-6+ and it came out OK - plastic became a bit darker, but it's hardly noticeable (unless you put a clean part right next to it), and most importantly it didn't seem to affect functionality.

[Mangozac]

We too have issues with the plastic on Molex FFC connectors and certain pin headers in our modded T962. It depends on proximity to the elements so I just assumed poor heat distribution. The irony is that we don't have any issues with cheaper alternative connectors!

Overall the T962 does the job though and is good value at the price point.

[asmi]

I remember someone reported good results by placing small  foil "caps" on top of these connectors to protect them from direct IR radiation. I suspect kapton tape would also work.

[coppice]

I remember someone reported good results by placing small  foil "caps" on top of these connectors to protect them from direct IR radiation. I suspect kapton tape would also work.

A lot of connectors come with a kapton film that must be peeled off after assembly. It both avoids stuff getting on the contact pins, and protects the plastic.

[KE5FX]

Mini-Circuits parts with the CD636 plastic package such as ADP-2-10+.

I just recently reflowed a board with ADTT 1-6+ and it came out OK - plastic became a bit darker, but it's hardly noticeable (unless you put a clean part right next to it), and most importantly it didn't seem to affect functionality.

Yeah, the MCL parts get a bit scorched (which I don't like, but can usually live with) but the name-brand JST connectors melt like they're made of butter.  There's no excuse for building those out of low-temperature plastics.

[Mangozac]

A lot of connectors come with a kapton film that must be peeled off after assembly. It both avoids stuff getting on the contact pins, and protects the plastic.

That's a good idea and definitely worth trying!

[jpanhalt]

@asmi
I checked into the three Molex part numbers and found nothing striking.

#54104 4031 is polyamide (PA46) or Nylon 4/6 which is one of the higher temperature Nylons/polyamides (Nylon is a polyamide)(melting point 295°C).  The actuator is polyphenylene sulfide (PPS)

#52271 0679 is also polyamide (PA46) and PPS for the actuator.

#87832 1420 is described simply as polyamide.

The first and last connectors are black; the 4031 is tan.  Color may have something to do with sensitivity in IR heated ovens versus hot air.  Also, construction details such as PCB attachment legs size and relationship to the bodies may be a factor.  Another common "plastic" used is described as "liquid crystal polymer."  They are also said to be more heat resistant than earlier/cheaper plastics (https://www.steinwall.com/wp-content/uploads/2016/05/Liquid-Crystal-Polymers-LCP.pdf ).

As suggested by others, a reflective shield  (e.g., Kapton or aluminum foil) over sensitive components may help.  I use little pieces of aluminum foil to protect parts when using hot air.  It doesn't stick like tape, but is so easily formed that it stays put.  I will definitely try that.

My T-962 should be here in a day or two.  I appreciate the input from all of you and have a little more confidence in using the modified T-962 instead of a toaster oven.

Regards, John

[asmi]

@asmi
I checked into the three Molex part numbers and found nothing striking.

Yea, that's why I prefer just trying it out and seeing what happens. It's always better to sacrifice one part to find out that it can't withstand IR reflow, then to find this out during actual reflow of assembled board and then trying to remove it from completed board. Some of my boards have quite a lot of parts on them, so removing (usually) large parts can be quite a headache.

My T-962 should be here in a day or two.  I appreciate the input from all of you and have a little more confidence in using the modified T-962 instead of a toaster oven.

Take your time to calibrate the thermocouples once you install community FW (it allows setting gain/offset for each probe independently). Also like I said above, consider using room temperature stable solder paste - it's a bit more expensive, but to me it's totally worth it because it doesn't need to be stored in the fridge, and it doesn't dry out once printed. Also use first few boards to tune reflow profile such that you will have boards fully reflow, but don't burn anything up. The oven tends to lag a bit behind the program, so include a soak phase if you reflow multilayer boards with power/ground planes - these can take in quite a bit of heat.
Oh and remember to have some sort of arrangement to remove smoke/fumes that come out from the back of the oven during reflow (typically they come off once the reflow goes from peak temperature into chill-down and fan on the back turns on).

[jpanhalt]

Thanks for all the good advice.  I can see it's going to be a journey.  Tried to download "Flashmagic" to load the hex file, and Google blocked it as "dangerous," malware, or pup.

Need to look into that more as my only experience has been with PIC's, and for a one-time effort, it is hard to justify buying anything.

[KE5FX]

Thanks for all the good advice.  I can see it's going to be a journey.  Tried to download "Flashmagic" to load the hex file, and Google blocked it as "dangerous," malware, or pup.

Need to look into that more as my only experience has been with PIC's, and for a one-time effort, it is hard to justify buying anything.

That's interesting.  Flash Magic is a pretty common utility application.  What site did you try to download it from?

[jpanhalt]

That's interesting.  Flash Magic is a pretty common utility application.  What site did you try to download it from?

First downloaded from FlashMagicTool.com (http://flashmagictool.com) in early afternoon.   Just tried again (11:25 PM EDT).  File is FlashMagic.exe, 37.9MB .  Downloads, but if one tries to open it, there is a delay of a few minutes, then message,

"FlashMagic.exe man be dangerous so Chrome has blocked it." appears. If one clicks on more info you just get a general message as stated below.

Edit: The link to Flashmagictool.com was on electrosome.com  (https://electrosome.com/flashing-lpc2148-serial-isp-bootloader/ )

[arvidj]

What appears to be the best firmware to flash?