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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


  • SMTnet
  • »
  • Electronics Forum
  • »
  • What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles?

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles?

Views: 11907

Hello, I was at a IMAPS conference last year 2016 and sev... - Mar 14, 2017 by PZappella  

#77243

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 14 March, 2017

Hello,

I was at a IMAPS conference last year 2016 and several presentations for QFN packages showed x ray images with high void levels of 30 to 50% with the best around 15% with solder paste. During the discussion it was mentioned that the auto industry is asking for < 10% voiding. Some argued that 30% was acceptable? I find even 10% voiding too high. What is really needed < 5%?

Thanks in advance for any comments

P Zappella

reply »

#77245

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 14 March, 2017

I don't have an answer but I am curious.

What data/testing shows that this voiding is detrimental? How did the auto industry arrive at < 10% ?

Pretty sure everyone has been putting down components that go into aircraft and life critical systems for many years with these voids present.

reply »

#77246

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 14 March, 2017

I've always been happy with ~25% or less combined voiding on QFN ground pads (depending on the pad dimensions).

reply »

#77250

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 15 March, 2017

@ Sr.Tech,

Any voiding means that the part of the package under the void is not forming a solid thermal interface with the PCB, therefore it's heatsinking capability is reduced.

With the drive for smaller packages with higher power handling capability (and higher operating temperatures) thermal conductivity becomes a lot more critical.

This package (10mm sq) is expected to handle up to 70 Amps, and utilises a big copper pad to dissapate the heat: http://www.nxp.com/documents/outline_drawing/SOT1210.pdf

Interestingly there is no info on any of the datasheets derating the power handling by voiding %.

reply »

#77251

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 15 March, 2017

Rob,

That's what I suspected but I would be interested to see studies that show that the voids are in fact detrimental to thermal and electrical performance. If the automotive industry is requesting <10% I would assume that someone somewhere knows something about it. There are only two things I have ever witnessed or heard about that will reduce voiding: solder paste chemistry, and vacuum reflow which I'm not sure is practical for high volume. I have heard claims that reflow profile changes will help. I have developed hundreds of reflow profiles and I have never seen a profile reduce voiding...ever. I have performed statistical studies with different profile models and void percentage data from the X-ray to back this. I have seen stencil design help, but the results are marginal at best.

reply »

#77253

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 15 March, 2017

It's well documented in higher power semiconductors where thermal interface materials are used, such as SIL pads, thermal grease etc.

There are a couple of papers on the effects of voiding on MOSFET performance, but I haven't got around to reading them yet, but you should be able to transpose the physics directly to LEDs as well.

Regarding the reduction in Thermal performance, it's real, not theoretical - the thermal conductivity of Copper is about 16,000 times better than air. However I have no idea what the thermal conductivity of a no clean flux is as that is probably going to be what is in the void.

It is starting to be taken a lot more seriously - in the last couple of weeks we have been approached with low voiding paste from our paste manufacturer, and also our Oven supplier Heller has a commercial (if not all that elegant) solution to volume vacuum soldering - http://www.hellerindustries.com/voidless.php

Suppose I'd better get reading and get the calculator out. I'll also have a word with a few apps engineers and see what they are doing on new designs regarding de-rating for potential voiding. If I get anything constructive I'll post it here.

reply »

#77254

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 15 March, 2017

From what I know, you should follow IPC standard for the class assembly that you are building. From your post somebody mentioned something somewhere - doesn't make it official information. I would not be worried about having voids, but what is their location and size. You absolutely don't want voids on your signal pads of the QFN. Big voids lead to failures. I have read several articles, claiming that small voids actually improve the strength of the solder joint. Solder paste and proper stencil design help reduce voiding.

reply »

#77256

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 15 March, 2017

Hi Etimov,

Understood that to us assembling parts that IPC is king, however it isn't neccesarily the bible for design engineers.

We've had customers demand almost zero voiding, or had to use conductive epoxy to avoid it.

For some of us solder joints are not just about strength but also used as heatsink inteface material, and the better the heat transfer the longer the life of the device.

A BGA connection can be dealing with micro amps and 0.8V, wheras a FET can be drawing 20A on 4oz copper.

reply »

#77257

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 15 March, 2017

Hello,

A speaker at the conference is where I heard that the auto industry wants < 10%. Is that real or just a would like to have, I am trying to find out. A void to heat transfer study would be a great paper.

PZ

reply »

#77258

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 15 March, 2017

Hi Phil, 25% sounds real high to me, any failures available that you can x ray and see what the void level might be? Wish I could call Ford and GM and ask what they really want for void levels.

PZ

reply »

#77259

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 15 March, 2017

Hello Rob,

I totally agree with you. Is this a real problem that is being shoved under the rug or just whatever voiding we get from a belt furnace is good enough. I think vacuum reflow can do much better but is a batch approach and not continuous like a belt furnace. So does it boil down to quality or cost/time of assembly. A paper relating voiding to thermal heat transfer would be a great study.

PZ

reply »

#77261

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 15 March, 2017

Hi PZ,

What, sort of like this one?

http://creativeelectron.com/papers/EconomicsofLEDVoiding.pdf

It even has the handy calculation for the reduction in thermal performance with voiding.

reply »

#77262

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 15 March, 2017

Rob, I believe that is for die attach voiding not solder interface.

reply »

#77263

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 15 March, 2017

Rob, Absolutely agree. Customer precedes standard. As you mentioned design engineers, I think that PCB design has the biggest influence on the voiding. Everything else we do is trying to fix bad PCB design.

reply »

#77265

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 15 March, 2017

Customer precedes standard, but customer must be educated and informed on what is achievable with their design and what is accepted as industry standard practice.

My guess is that the negative effect of those voids is negligible. {Voids at the solder interface on the thermal pads}

Here is a fact that I do know. People get nervous when they see voids in an x-ray image.

reply »

#77269

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 16 March, 2017

It's the same thermal issue whether it's a die or package. It's different if it is just a signal trace, but if it's a heatsink pad (LED, FET, Motor driver, PSU IC etc).

I think the voiding issue boils down to what the part is actually doing. The original comment was regarding automotive, and the key issues in automotive are reliability and wide temperature range at a low cost. The cooler the part runs the longer it will last, however not designing with IPC acceptable voiding in mind means that the fall out will be higher so the cost will be higher.

reply »

#77276

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 16 March, 2017

Hi Rob,

Thanks for the LED paper. I like the void area equations but I am puzzled, the paper did not say if the solder was a solder paste that would generate a lot of voids, or a fluxless preform or sinter material. The measured voiding levels are awful. $9.7% down to 35% back in 2014. I can usually get below 5% with flux free preforms and with certain pastes < 10% void levels. Here I am using a vacuum pressure oven and not a belt furnace. So I assume these LEDs were soldered on a belt furnace. And this was acceptable? back then? Unbelieveable.

PZ

reply »

#77279

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 17 March, 2017

I think the main driver is MTBF data driven by the Car OEMS, mainly german makers pushing this onto suppliers in supply chain. Vacuum soldering and void reduction is a hot topic for the leading electronics suppliers in Germany at the moment. Is not related to and IPC type standards, it is more driven from internal IP of car OEMs. Yes power electronics and led would be main applications I would suspect. Rehm Thermal Systems is leader in inline reflow ovens offering proven solutions for several years due to their high exposure to the Auto Industry demands and this is their strongest , main market share. Vacuum Soldering can also be done in Vapor phase, again this was championed by Rehm before most other vendors, due again being close to where lower and lower void quality requirement demands are coming from...

Paste flux does contribute so much to reducing voids and more work in this area could get there, but vacuum soldering combined with a low voiding flux is as of today the only way yo meet some of these tighter QA needs.

Changing profiles agree has minimal effect if none and some paste deposition optimisation can help more in larger paste deposits, not so much on small parts. Parts them selves contribute a % of chance voids also etc.

Vacuum soldering is not new in packaging industry been around long time, just now it may get push down the production stage line so to speak.

reply »

#77653

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 31 May, 2017

In IPC it is said that the void % should be established between you and the customer. There is no number. There is only requirement for balls on BGA. I have a customer that agreed to 40% on D-packs which is okay for both us and them.

reply »

#80330

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 28 April, 2018

Hi, please go through IPC 7093, its only for design and assembly process implementation for bottom termination components. its says up to 50% voids will not reduce ant electrical and thermal performance.

reply »

#80337

What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles? | 1 May, 2018

In the past I have worked on 600 Amp FET Blocks. The rule was no voids. The circuit board would be heavy copper. 4 times normal plating time. Not particularly high resolution on the board say 20 mils. To help with the heavy plating. The boards were then screwed on with heat sinking compound to 10 x 40mm Aluminium Busbars. Which were in turn screwed and heatsinked to the steel panels in the metal enclosure. Each box would be set up for 3000 Amps. This requirement was a product spec, and I am sure we could have switched much higher currents. These products were tested for temperature rise on the busbar. So any dodgy soldering would be picked up!

Voiding ruins the whole equation, as heat can't get away from the component anywhere near as effectively.

Some of these products are running in the Pilbara (up to 60C Ambient) for nigh on thirty years!

sarason

reply »

  • SMTnet
  • »
  • Electronics Forum
  • »
  • What is the QFN or Device Voiding Levels Needed for Automotive Industry for Power Devices for Electric Vehicles?
SMT fluid dispensing

Reflow Oven