Screaming Circuits: April 2009

Short-Run Parts

IC strip sm I'm so confused... My head is spinning...

Don't panic. Locate your towel and I'll do my best to explain some of it. I'll start with taped parts.

So what do we really need when we ask for parts from you? Our Short-run service asks for parts on reels, partial reels or continuous tape. The good news is that we still just need the number of parts you need for your job plus a few percent. If you need 100, just order 105. You don't need to order 5,000. We just need it in one continuous strip that's at least 12" long for short-run. (Prototype orders can still use tiny little cut strips though)

When we buy your parts for a turn-key order, we'll just order a continuous strip or parts. And we don't need leader either. It doesn't have to be on a reel. It can be though. Just ask for a reel with only the number of parts you need on it (plus 5% for spares). (0201's need 50% spares. Sorry)

DigiKey is especially good at this. You can order a "DigiReel" with a custom number of parts on it, or you can ask for a custom length cut strip. That's what we do when we buy turn-key parts for you. For our Short-Run, we do need a strip that is at least 12 inches long though. That's the only real "got'cha". (not even really much of a "got'cha" if you ask me)

Digikey strips 500

So, don't panic. Order just the parts you need (+5%) and save money with our Short-Run production.

Duane Benson
But, can we assemble from a mobius strip of parts?
Call Matt and ask him

0.4mm pitch BGA pads

At 0.5mm pitch and larger, we generally recommend non-solder mask defined (NSMD) pads for BGAs. The NSMD pad will allow for better adhesion with the solder being able to grip the sides of the copper in addition to the top surface. Even down to 0.5mm pitch BGAs, most part manufacturers still say to go with NSMD pads. Granted, you need to make sure your fab house can do a good job of mask registration. It's no good to have pads that are half SMD and half NSMD. We've seen that.

Once you go down to 0.4mm pitch BGAs, though, things get different. We're starting to see more and more of these. Some CSP (chip scale package) or WSP (wafer scale package) BGA and LGA parts are starting to show up in 0.4mm pitch, as well as some bigger parts like the Ti OMAP processor.

At that size, you start to see more risk of solder bridging with NSMD pads, as in the left side of the illustration below.
0.4mm pitch SMD BGA pads 
This is a somewhat exaggerated view, but it shows the potential problem here. With such small geometries, you probably aren't going to be able to route escape traces on the top layer either. You'll have to put blind vias in the pads and escape through the inner layers. By the way - NO OPEN VIAS IN PADS. The pad needs to be solid metal with such a part. No exceptions. Don't try using solder mask to cap the vias at this pitch either. Plate over it. It has to be all metal.

You can finds some more detail here in a design guide written by Ti for their OMAP processor. Pages 8 through 12 talk about pads, mask and vias. Obviously, you should consult similar materials written specifically for the part you are using, but Ti did a great job of covering all the issues here so I use it as a good general piece.

Duane Benson
Are you there Moles? It's Edgar.

Beavers and Ducks at ESC last week

Up in Oregon, we have our rivalry between the OSU Beavers and the U of O Ducks. Everybody's got some kind of rivalry and most of the times you exhibit at a trade show, you'll end up with some competition around. This one takes it to the extreme though. Microchip and Atmel are of course as rival as rivals get. In the microcontroller community, the M vs A discussion is full of passion and absolutes. It's like AMD vs. Intel or Mac vs. PC.

At this year's ESC show last week, the two companies are right across the aisle from eachESC 0409 MicroAt 002 (Medium) other and ESC 0409 MicroAt 001 (Medium) seemed to be having some "fun" with it. Atmel has a poster quoting the Microchip CEO complimenting Atmel. Microchip has a big poster quoting a survey the shows declining favor for Atmel. Goofy.

Personally, I suspect that they both make great products, but it would help if there was actually some way to tell which is best for what without getting into a pseudo religious argument. I've asked that questions before of folks that use one or the other, but I haven't ever run across anyone that uses both and can give me a fair and un-biased overview of where the strengths and weaknesses are of either company's parts. Anyone that uses both want to chime in with some plain and dry toast, I mean plain and dry thoughts?

ESC 0409 MicroAt 003 (Medium)

Duane Benson
Two chips, no salsa

ESC, Day the Last

I've run across a few interesting things here this year. The best demo has to go to Freescale with their robotic air hockey player. Robots are cool in general, but when you can throw together enough processing power in a small, easy to use package like they did, such that a robot can compete with a human, that's a good piece of silicon.

The other really interesting thing I've run across a couple of times is the idea of open source hardware. The open source movement is long well established in software, but I don't think it's really found its niche in the hardware arena. Maybe it has now.

Folks from beagleboard, Project Sun Spot and a few others stopped by our booth to chat. The basic idea is that these groups have put together viable and high quality eval boards and have made the schematics and layout open source. You can download CAD files or just the Gerbers and the BOM and build your own just as they designed it. You can take the schematic and layout files and modify them to suit your application. You can manufacture and sell it. Or, you can buy a completed board off the shelf some place.

Eval boards used to be just an example to get a new chip up and running. Then to really make it work well, you'd have to roll your own. Not so with these things. And if you aren't an expert in high-speed design and layout, you don't have to buy an off the shelf board that may almost but not quite fit your application. You can just use the difficult parts that someone else already created and add your customization on to the edge.

Keep a watch on this movement. There might be something big here and it might just shave a month off your next embedded design project.

Duane Benson
Direct from show central...
I'm tired.

ESC, Day Two on April One

In past years, ZigBee, BlueTooth and WiFi have been hot wireless topics here at ESC, and those radio standards are still around. But the big news this year in wireless seems to be the re-emergence of TRF (Tuned Radio Frequency). Both Atwater Kent and Philco are demonstrating TRF systems in their booths. Atwater Kent is just down the aisle from us (or up the aisle, depending on which way you are facing) in booth 262 and Philco is doing invitation only demonstrations in meeting room C8.

I haven't had time to stop at the Microchip (booth 416) or Atmel (booth 316) booths to see if they are developing software stacks for the AK or Philco TRF systems. If anybody has done so, feel free to pass me a note here.

Bluetooth is of course widely used in consumer devices and ZigBee is mostly used in custom embedded applications. TRF looks to be targeted more along the lines of applications that would have been ideal for WiFi or WiMax, such as broad-based information distribution. The TRF systems typically will not fit a star or point-to-point model, but more of a one-to-many or server to many client connectivity model.

One of the primary advantages of the TRF is in spectrum utilization. While most other standards are set to a very specific frequency and have bandwidth limitations to go with that specific target, the TRF will vary the bandwidth with the frequency. This allows for a greater number of receivers to be set in place and allows for multiple transmission stations to coexists with simple adjustment to the individual tuned frequency. Further, with a set of relatively easy to manage adjustments, receiving units can be converted to receive analog data from any of available transmitting stations within a reasonable range.

AtwaterKent booth Unlike WiFi which has usable range of maybe a couple of hundred feet, in a field, going down hill with the wind at your back, or WiMax which can achieve city scale range with a large set of transmitters, TRF requires only a single transmitting station to reach receiving units for miles and miles. It is also a scalable system. Increasingthe power of the transmitting unit will increase the specific range. It is subject to the inverse square law like any other propagating signal, but through appropriate power devices, ranges in the dozens and even in some cases, the hundreds of miles can reliably be achieved.

If you have a chance, stop by the AK and Philco booths. Tell them Duane over a Screaming Circuits referred you. They're giving me a kick-back for everyone I send over.

Duane Benson
Danger! 50,000 Ohms

« March 2009 | Main | May 2009 »