Via-In-Paste

QFP lots of vias QFP lots of vias stencil I've heard of vias in pads, but vias in paste aren't so common. Here's a QFP footprint with a lot of thermal vias in the center pad area. I'm guessing all of those vias are necessary for heat removal, but they sure aren't friendly to solder paste as created.

Further, the solder paste stencil had good intentions too, but ended up inverted in practice. It's good to keep paste off the vias. Even inverted though, I'm not sure how the stencil would have worked. You can't have dots of stainless steel just floating out there. The vias should be capped, or better yet, filled and plated over. The proper way to create the stencil layer would be with a cross hatch as shown in the third image.Good QFN stencil b

Capping (also called tenting) vias with solder mask used to be a pretty common practice. You'd put solder mask over the via out to about 100 microns bigger than the via on the component side. Some people would put the cap on the back side, but that still leaves a lot of volume for solder to wick down into.

The capping with mask practice is kind of falling out of favor these days. For one, if the mask is too thick, it can end up lifting the part up a bit. More likely though, the caps can break, leaving an open hole for solder mask to wick down.

The optimal method is to have the vias filled and plated over at the board house. There are some thermally conductive materials that do just about as good at heat transfer as would the open via.

Duane Benson
Solder paste is people! It's people!

Modularity and Standards

Eons ago, (well, it seems like eons) when IBM designed its original PC, it took note of the success of the Apple II with it's modular expansion system - easily accessible card slots with loads of clear documentation - and added its own variety of modular expansion system. By doing so, the cost of accessories to consumers stayed low, the cost of installing or replacing said accessories stayed low and a whole new industry emerged to create compatible accessories.

Apple II I just read a Twitter Tweet ("Tweet" sounds too cutesy to me, so I'm never quite sure what to call those; maybe a "Twoot"?) from Mike Buetow that linked to an article about the latest Toyota recall. It seems that there are a couple of specific solder joints prone to cracking in the ECM (Engine Control Module) of certain models.

The last time I had any real data on the cost to replace an ECM, it was on the order of $1,500. Just scanning around the Internet, I found numbers ranging from $1,000 to $2,000. I'm guessing (I am speaking from near complete ignorance) that maybe two or three hours of that are labor at $90/ hour. That's a lot of cost in the electronics as well as labor hours that can't be used for billable hours. With so much of new cars being electronic, this issue is only going to become more extreme.

So, why can't the auto industry take a cue from the PC industry. Create a standard, easily accessible, electrical bus with standard, easy to manipulate mechanical attributes. Even if they were just standard within each manufacturer, it would still be a big improvement.

Consider this scenario: Buy a Toyota mid-size-car ECM at the local auto parts store. Take it home, plug it into a USB port on your home computer. It auto-runs a link to a specific web site. Enter your car's VIN number and the site loads firmware that matches the ECM to your car. Take the ECM outside, open your hood, flip a few latches on the water-tight electronics box, pull the old one out and plug the new one in. There you go. Done.

Instead of what is pretty much a massively expensive dealer-only operation, you have half a dozen standard bus ECMs to choose from and about 15 minutes of work that's not much more difficult than installing a new printer on your PC. And, you'd have less expensive aftermarket options as well. And, a new industry would emerge to design and build those aftermarket options.

Duane Benson
Sadly, not in my lifetime, Batman...

The Top Ten Generic Things

I'm in a bit of a ranting mood right now. That just happens sometimes. Usually it's on a specific subject, but today, I seem to have mini-rants about a whole bunch of  things. Well, maybe ten things. So here they are, ten generic things that bug me:

#4:    Not listening to customers enough. It's nice when a company has a good idea and wants to build it, but if they don't get outside of their own heads for a bit, we consumers end up with UI's that don't make any sense, features that we'll never use or products never tested under real-world conditions (see #4).

#4:    Test cycles that are too short. "Beta test the world" or "Ship it and fix it later" may get something to market sooner, but at what cost. So many companies seem to think that since "they" do that on the web, everyone should go ahead and operate that way. But what happens when the not fully tested design has a hardware problem? Where's your field upgrade then? Or what happens when the product is mission critical? Oops. Too late...

#4:    Listening too much to customers. What??? Yes. That's what I said. Most customers want way more than they need for way less than you can afford to build. You need to listen to customers a lot and very carefully, but you need to translate for them. You can't just take raw comments and try to directly put them in as product features.

#4:   "Half-gallon" containers that aren't a half gallon any more. It really annoys me to buy a Half gallon of Ice cream knowing that it's only 3/8th of a gallon.

#4:    Not considering the whole story. This is where the law of unintended consequences comes in. Okay, we want to reduce our consumption of fossil fuel so we subsidize corn ethanol. Fine except by doing so, we tie a major food staple in developing nations to the volatile price of filling giant SUVs. People go hungry because of it.

#4:    Rushed design cycles. Yes, we, ourselves, contribute to this by reducing the turn-times for electronics assembly, but I'm not really talking about the assembly phase. More about the design, layout and kitting. (and test - see #4) We all need to chill a little and take some extra time to run a few more tests, double check the component footprints and make sure we've done a thorough job of it.

#4:    More science and less hype. No one can really tell if global warming is man-caused or not. I'm sure the real data is floating around somewhere, but everyone talking about it has a personal agenda. There's so much pseudo-science and political ranting thrown about that anything that an interested citizen might use to come to an informed conclusion is obscured by all of the exaggerated and faked material.

#4:   How about some electronics-targeted legislation that actually makes sense from a technical and social perspective. As with things like global warming in #4, there's too much hype, too much cash-based lobbying and not enough actual understanding going into some of these laws that affect all of us in the electronics industry.

#4:    Allocation. It really annoys me. Related into this is the proliferation of specialized chips. There are a seriously larger number of varieties of every form of chip you might imagine. That's great for design, You can pick the microcontroller that pretty meets your exact specifications, or just the right buck/boost controller. That's cool, but I think it also makes forecasting and the allocation of foundry time simply crazy. That can only exacerbate the supply issues that cause parts to go into allocation mode.

#4:   Missed opportunities due to personal-agenda based hype. So many people want to replace fossil fuel so they bend reality and call the electric car the green replacement to gas cars. Then everyone is disappointed that they can't drive 600 miles with just one or two five-minute fill-up stops. They focus on far too far into the future and make everyone dismiss as hype what is otherwise a perfectly viable technology. Market electrics as a second car. It's not the main car for trips and the ultimate in convenience. It's the run to get a gallon of milk car, the back and forth to Jr.College car, the "I'm going to a friend's house" car. Market electric cars like that and they are 100% viable right now.

I'm not sure which of these things bug me more or less than any other, so they all tie at Number 4.

Duane Benson
Have a nice day

What, what, what?

Unknown cathodeQuick. Pretend this is the footprint for a diode. Which way does the cathode face?

Okay, if you can't figure out which side gets the cathode, how about the anode. Which side does it go on?

Hah! It's a trick question. You can't tell based on the limited amount of information available on this PCB footprint mock-up. Sadly, this image is all too frequently real. In theory, with a properly constructed centroid file (See yesterday's post), we really shouldn't need any marking on a PCB at all. And, if everything is properly done, and if no gremlins are swimming through the air, we won't need any marking.

However, when you get the thing back in your shop and hand it off to a technician to measure current and signal quality and slew rate and all of that junk, or if something does go wrong and it needs some troubleshooting, those missing or ambiguous markings can make a big difference.

I'd go ahead and put a little diode symbol there or maybe an "A" for anode. You can use a "C" for cathode too but just make sure it's clearly identified as a diode so it won't be mistaken for a rogue capacitor.

Duane Benson
No. It's not a contest. Just a question.

Updated Centroid Documentation

Passives orientation r2 A little housecleaning is usually a good thing. Here at Screaming Circuits, we try to be as flexible as possible and we'll do a lot of different things - standard and non-standard. But we really should, when passing on documentation, give out the standard form of data. And that's what housecleaning has done for us today.

I got a comment on an old blog post calling out an error relating to our centroid (AKA XY rotation / pick & place file), so I went back and cleaned up the blog post and linked to a PDF we have describing our centroid file requirements. It matches IPC-7351A now. And that kind of match is a good thing.

Duane Benson
Matchmaker, Matchmaker,
Plan me no plans
I'm in no rush
- actually, we're always in a rush.

Mask Control

I wonder. Is solder mask difficult to control in most CAD packages? Or do we just not need to control it very often so we forget? Take this little footprint here. Oval pad It looks like someone just used a little flood fill to create the thermal pad rather than creating a new custom footprint. That would have been fine except that the flood fill area has solder mask on it.


In Eagle, if you want to keep the mask off of an area that would other wise have mask, you draw a polygon in layer 29, tStop (or 30 bStop for the bottom) over the are you want to keep mask off of. Not difficult, but not necessarily obvious either. With Sunstone PCB123, you pretty much do the same thing with the SS Top or SS Bottom layers. I don't know about any other packages, but I would guess it would be a similar approach.

Of course, just making the footprint with the library package editor would take care of it too, but sometimes it's just more expedient to take a footprint that's close and mod it up with a polygon or something similar.

Duane Benson
Is Oregon like a Polygon?
No, because it hasn't "gone" anywhere

All Leadless

It wasn't too terribly long ago that just about any design could still be built all thru-hole. Okay, maybe it was a little longer ago than that. Once the big CPU chips stopped showing up in PGA (pin grid array), thru-hole PC motherboard possibilities went out. Then when blue-tooth and Zig-bee came around, most if not all of those chips came out in BGA, LGA or QFN forms - no thru-hole. Anyway, it's not too difficult to do pretty much any design in all SMT now, but what about all leadless?

DFN-8 Okay, we don't really consider passives to be leadless, but they kind of are. So, we have all of our passives in a leadless like form. Now all we have to worry about are the chips.

I'll start with a Microchip PIC18F4550 in a QFN44 package. It's got built-in USB, so I don't have to worry about a separate USB chip. I'll load up a bootloader and it will all be happy. Wireless will have to wait for version 2.0. This is going to control a two side-by-side wheel platform scooter type thing, so I'll need a gyro and accelerometer. Digi-Key just sent out their "techzone" mini-catalog/magazine featuring just some of these type parts. I'll take the Analog Devices ADXL345 three-axis accelerometer in LGA form-factor.

I only need to worry about pitch and yaw, so a dual axis gyro should be fine. I'll try out the ST Micro LPY550AL in a 5x5mm QFN package. For voltage regulation in the prototype, I'll use a Linear Technology LTC3642 in a 3x3mm DFN package. It has a 3.3 volt output and can accept 5 to 45 volts in. That gives me the flexibility of powering off of a dedicated battery pack or off the scooter main battery.

All LGA or QFN/DFN. The only problem is soldering up the prototypes and next half-dozen or so units, for all of my friends, after that. I'm not going to stick those things in a toaster, and I certainly can't hand solder them like I could with the old thru-hole or TSSOP and SOIC chips. Oh. Wait. I work for a company that does that.

Duane Benson
Fight Uni!

Global Shifting

Friday morning, I walked my way to the office in a slight drizzle. It was overcast and cool in the morning, eventually warming to near August temperatures later in the day. We had a long, mild winter, a long, cool spring and it looks like we're having a very short, not terribly hot, summer. Yet, the statistics (probably) don't lie. Global warming seems to be very real everywhere but here. The peat bogs around Moscow are burning, giant ice islands are breaking off of Greenland, and it's like 900 degrees over on Venus. All that, yet still nice and mild here in Oregon.

I think the same kind of things may be happening in manufacturing. By some accounts, it's all doom and gloom for North American manufacturing. We seem to be losing all of our manufacturing to other continents. But, maybe we aren't losing all of our manufacturing. Maybe it's just a shift. The high volume, low-value add stuff is likely gone, but to me, it appears that there's a thriving manufacturing industry for low-volume high-tech items. And not just in the contract manufacturing area.

There seems to be a re-emergence of small companies that are performing their own assembly in house too. They outsource what makes sense to outsource and in-source what doesn't. Adafruit and Sparkfun are two examples. The Arduino folks (in Italy) are building locally-to-them too. The Beagleboard from big-player Texas Instruments, is built in North America. Screaming Circuits and other small-volume companies like us seem to be doing quite well these days. Is it possible that the North American manufacturing industry isn't dieing, but is just changing? That's what I think.

Duane Benson
"Comet due to explode the earth at 9:42 this evening. Details at 11:00"

Does Angle Matter

It's standard practice to avoid joining small PCB traces at 90 degrees, but instead to join them at an angle. But, does it matter for thick traces?

Right angle traces Here's some 20 or 24 mil traces. Is it really going to matter, with a pad-size trace, what angle the joining trace hits the other one?

Convention would have you do something like the alternative layout on the right. Either like "A" or like "B". But, is it really necessary and worth the extra timeNot right angle traces required to do that?

Part of me really wonders and another part of me says, it shows attention to detail and implies that the entire design was produced with the same care. It's that elegance in design thing.

The other question I have relates to "B", in the image on the right. Does it matter which direction the 45 degree trace intersects? Does it matter based on the direction of current flow or does it matter at all?

Which of the three illustrated techniques do you prefer and use?

Duane Benson
Winslow Leach says hi.

« July 2010 | Main | September 2010 »