Sensor on board vs air in reflow oven...

[alank2]

Hi Everyone,

I was searching around to look at others reflow oven projects I found the controleo2 on eBay and started taking a look at their oven tips.  One of the things they recommend is not having an on board mounted temp sensor but they show it in air.  I've reflowed a few boards with my oven/pid software and it _is_ a pain to get the sensor to sit on top of the pcb and not move and make sure it doesn't touch anything grounded, etc., so I decided to do a test run.  I had a sensor in air and this was the one I fed my pid loop a and then I also monitored the pcb with a sensor pressed against the pcb surface using a small spring.

Here is a graph I did:



You can see the blue line (air) following closer to the setpoint than the pcb sensor which is a bit more smoothed out.  There were times at the top end that the pcb got hotter than the air temp and this is my concern, I don't want to cook anything.

What are your thoughts?  Do you mount the sensor on a pcb or do open air?  It is a pain to deal with, could I just mount a sensor to a dummy pcb that I realize may not always be the same size as the one I'm reflowing?

ALSO, 2nd question, they lined the inside of their oven with this stuff:

http://www.amazon.com/010396-Reflect---GOLD-Tape-Roll/dp/B0039Z5TYU/ref=sr_1_1?ie=UTF8&qid=1430534568&sr=8-1&keywords=reflect+a+gold

Is it any good?  Would this make the heat losses less and give more even heating?

Thanks,

Alan

[mjkuwp]

I don't think it is great to leave the sensor in the air.  I have also been searching for the best answer to this problem.

What I do.

 - I only do SMD boards, no thru-hole.
 - I place the boards on a copper sheet that I tarnished so that it is mostly black in color
 - I stuck the thermocouple onto the copper sheet and cover it with aluminum tape.  I've also tried soldering the thermocouple to a hole in the copper with extra-high temp solder but couldn't get this to work well.
 
- I place my boards close to the thermocouple and observe the feedback temperature during the cycle at the same time as watching the pcb to catch the temp at which the solder flows.

What I have noticed is that on really small boards the solder melts almost perfectly at the same temp as solder melt temp.  On larger boards I have seen as much as 15C different but typically it is 5C or 10C.  In other words, the feedback will show a copper sheet temperature of 195C when the board flows.  whatever the feedback temperature is, I make a quick adjustment in my head as to when I will open the door.

My solder joints always look really good and i have not burned any components.

If you want to get really obsessive you can solder a thermocouple directly to the PCB with silver solder.  I've done this and of course you get nearly perfect temperature feedback.  However, this is really not a practical approach to making boards and I don't think it is necessary.

It can also work pretty well to connect a thermocouple to a spare blank PCB and place this next to the PCB you are building.  I did this method for a while but eventually decided I didn't like the look of that one PCB getting cooked over and over again.



here is my write-up on how I made the copper black.

http://themzlab.tumblr.com/tarnishing

[alank2]

What are your thoughts about just using a pcb with some higher temp soldered.  I've got plenty of them laying around.  Why didn't you want to cook it again and again?

Any thoughts on the gold insulation tape, (a) would it keep more heat in and allow for faster heating?  (b) would it make the heat more even?  I've got a convection fan type oven and I run the fan...

[tggzzz]

Any thoughts on the gold insulation tape, (a) would it keep more heat in and allow for faster heating?  (b) would it make the heat more even?  I've got a convection fan type oven and I run the fan...

What evidence do you have that you need to heat the oven more quickly?

You can get "heat shadows" whether you use radiant or convective heat. Component choice, board design, board orientation, and experimentation and experience will affect the severity of that effect. It would be unwise and presumptious of me to remotely diagnose your setup.

[alank2]

You ask some good questions tggzzz.  My oven is 120V and it can't get up to heat as fast as the profile asks, but I'm not sure that is a big deal.  So I don't know that I have any evidence...  What I have does work, but if it can be improved easily and be more even that would be a good thing.  I've baked about 30 boards so far and a bit of smoke comes out right at the end, but nothing is burned.  Maybe it is the flux burning off.

My biggest question is, could i get a pcb that is in between the size of the two I usually will be doing, 5x5 cm and 10x8 cm and solder the temp sensor with higher temperature solder.  Would that likely be workable?

[alank2]

I keep thinking about the best way you can attach a TC to a pcb, but because I have a sheet of aluminum underneath it, I can't put the TC through a hole or it will touch the metal and grounding causes it to fail.

What about putting a small hole in the pcb just big enough to be detent.  Then perhaps an inch away from the TC I put a small weight to give it some downward push.

My biggest problem is that TC moving from where I put it when I close the oven door on the wire, etc...

[Pjotr]

What about smd PT100/PT1000 resistors? Those are not that expensive. You can design in some on the board for a test IMHO the only valid way to measure board temperature where it counts most.

[IanJ]

My findings after building my own reflow oven and playing with all sorts of ideas and methods:-

- Mount the thermocouple in air, about 15mm above the centre of the board.
- Sit the board on a thin perforated sheet (I used stainless steel). A sheet peppered with 8mm holes will do.

Why?
If you sit the board on a dense material like a thick slab of aluminium then your oven will be working hard to heat up that slab, not your board.
The perforated metal sheet allows better air (heat) flow to both sides of the board. It also helps to stop the metal sheet sucking heat away from the board.
Mounted 15mm above the board the thermocouple will pick up a reasonable average of the heat of the board below it, rather than a spot.

Also, I insulated the outside surfaces of my oven (the internal oven chamber, not the outside cover) with some oven insulation. It helps to speed up the initial part of the profile which was slow on mine.

Ian.