Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Uneven thermal load in PCB design.

#22614

Uneven thermal load in PCB design. | 8 December, 2002

Hi,

We have been manufacturing a PCB, however it's been quite tricky to profile, as the edges of the PCB are reflowing before the center of the PCB.

I have looked at the design, and there is a small power plane for 2.5 Volt in the center of the PCB. I assume this is taking longer to heat up compared to the edges of the PCB where this copper layer had not been extended out to.

It's now running through the process ok, however because its our own product we can get it redesigned if required. We are interested in filling that whole plane in the PCB with copper so the thermal load is more even.

It's a 6 layer board, with ground and power planes, and this small power plane is on another layer without many tracks, so it would be easy to extend the copper to fill this layer.

Is this normal for PCB's to be designed with consideration for even balance of the copper layers so thermal load is even? I must admit it's not something I had considered until we started profiling the board and noticed uneven reflow.

Regards,

Grant Petty Blackmagic Design

reply »

MA/NY DDave

#22619

Uneven thermal load in PCB design. | 9 December, 2002

Hi

Yes it is common to design for more uniform heat transfer during the soldering phase or even during the operation phase. It all depends on how bad the problem is during automated assembly or during usage.

The only difficulty is that the designers might have some circuit traces that have electrical requirements that prevent putting in extra layers.

When it gets real bad some processes call for manual assembly in the heavy copper regions.

MA/NY DDave

This message was posted via the Electronics Forum @

reply »

MA/NY DDave

#22622

Uneven thermal load in PCB design. | 9 December, 2002

Hi

Just one idea, many exist.

If your electrical designers can tolerate it, sometimes a cross hatched design is used under areas where electrical performance is important. It reduces coupling and impedance effects compared to a full layer.

Elimination of Huge Etched away areas is also good from a material(laminate) stability consideration.

The worst one I have ever seen was EONS ago in another life in an older process, where 7 oz copper was being used on the top layer. In addition to HUGE copper power regions, a dense area of fine lines and lands were used. It was a bear.

MA/NY DDave

This message was posted via the Electronics Forum @

reply »

#22623

Uneven thermal load in PCB design. | 9 December, 2002

Large copper pours create an unbalanced lay-up that is bad design practice, creating problems in fabrication and assembly. The hatching a previous poster mentioned could be an alternative.

Stepping back a bit and wondering, what's the problem? You've got a board that reflows, maybe not with the profile you'd like / extected, but reflows none the less. We're happy, happy when we can tweek a standard profile and get good flow.

reply »

MA/NY DDave

#22631

Uneven thermal load in PCB design. | 9 December, 2002

Hi

I hear you, and you are right, yet most good process engineers strive to make future life a little easier by going back to the designers and making some alterations that don't affect performance.

Now if we did a 6 sigma type analysis than this process/soldering engineer would start to increase their process Cp.

YiE, MA/NY DDave

P.S. Gee DaveF, I have to come see you some day since I go to Haverhill 3 times a week to roller skate at Skateland.

This message was posted via the Electronics Forum @

reply »

#22723

Uneven thermal load in PCB design. | 16 December, 2002

Hi,

Thanks for the info. We are reflowing the boards, but just barely. Sometimes we get the QFP in the middle of the copper area not reflowing, and it's pain. The joints are dry, and I am not sure if it's caused by lack of heat, or the flux going bad because it takes so long to get the chip up to temp.

We have been checking out a vapor reflow oven, and it's quite nice, however we are getting some tomb stoning on the 0603 parts. Nice joints though, and uses very little power. We don't cook the boards with this oven either.

Thanks for the info, and we are doing a rev on the PCB, so we will bring that layer out to the edges of the PCB so we even out the thermal load.

Regards,

Grant Petty Blackmagic Design

reply »

#22725

Uneven thermal load in PCB design. | 16 December, 2002

Grant,

On those 'dry pads', where does the solder paste end-up after soldering? On: * Pads * Component leads * Someplace else [Where?]

reply »

#22749

Uneven thermal load in PCB design. | 16 December, 2002

Hi,

It's on the pads. They look ok, but when you check the lead comes away from the solder.

Regards,

Grant Blackmagic Design

reply »

Randy Villeneuve

#22754

Uneven thermal load in PCB design. | 17 December, 2002

Grant,

The board maybe the problem but how is your oven working? Is it a convection oven and how many zones (top and bottom)? What does your profile look like? If you are unsure and if you only have one oven, send your board to Soltec or Electrovert and have them run a profile for you on one of their ovens. Just a thought.

This message was posted via the Electronics Forum @

reply »

#22767

Uneven thermal load in PCB design. | 17 December, 2002

As a starting point, consider that these component leads may not be getting hot enough. Measure the temperature profile on the leads that do not solder well to be sure: * Heating and cooling ramps are within the paste supplier's recommendations. * Component leads reach liquidous +25*C for 5 to 10 seconds.

reply »

Jade Series Selective Soldering Machines

See Your 2024 IPC Certification Training Schedule for Eptac