In my job as an analog (system) engineer, I look at datasheets on a regular basis. There are few other ways that vendors can communicate how a component will behave under a variety of situations. The format of most datasheets for analog components are rich with lab-tested data in a variety of situations. That same multitude of data can really confuse an engineer (or others), from those just starting out to those who have been in the business for many years. Much like the “conversations” category that was introduced in the last post, I think that a regular analysis of different components from different families and vendors can really help someone figure out where to look on a datasheet for the most pertinent data and then allow that person to utilize the data however they need.

Let’s look at the data sheet for one of the most famous components of all time, the LM317, a 3-terminal adjustable voltage regulator. You can follow along from the actual file, just download it from the product page on the upper right.

As you scroll through this datasheet, you’ll see sections that are common throughout many discrete component datasheets, especially analog ones.

• Overview/General Description
  • Not a brain-buster here. This explains the basic use of the product, why you might want to use it possibly calling out the most attractive characteristic of said product.
• Features
  • This is a more specific list of the attributes the manufacturer wants to highlight. In later “Spec Sheet Dives”, we’ll look specifically at the highlighted specifications, how they match up to other manufacturers, if you should really care about them and if the manufacturer is telling the whole truth. The most interesting thing about the “features” section (in my opinion) is what they leave out, not what they put in
• Typical Applications
  • Often a vendor will showcase the most standard configuation of a part. In the case of the LM317, it’s a simple unregulated voltage in and a regulated voltage out (with some controlling discrete components). And believe me, there are a lot more complicated things one can use an LM317 for. An op amp might have show a photodiode amplifier if it operates well with high impedances and a transistor might be shown in an output stage of an audio amplifier if it was designed for high power and precision.
• Package Information
  • Parts come in a multitude of shapes and sizes, moreso now more than ever. Companies try to lock you into their part using a non-standard part shape or they might believe they have the best design for a super-small package with good heat dissipation. Whatever the case (ha!), it’s important to check to see whether or not a particular packaging will suit your needs (going to be soldering by hand? Don’t pick a BGA part).
• Ratings (temperature, current, power, etc)
  • Hopefully you are not even coming close to these extreme limits a part can be stressed to, but it does happen. If you operate outside the specifications (below), often the manufacturer will inform you that their part is specifed in a very specific way and numbers outside of that are subject to derating. If you operate outside the ratings however, good luck even talking to someone from the company. That’d be like going back to Dell with a laptop you dropped in the Artic Ocean and expecting it to work. Sorry, some things just can’t be done. So seriously, at least stay within the ratings if you hope to ever complain to a vendor about part performance.
• Electrical Characteristics (Specifications)
  • I’ll be blunt;. this section (and one other) is all that matters on a datasheet. Without the other info, it would be difficult to figure out certain things (what kind of layout you’ll need, when the part will blow up, etc), but without the electrical characteristics, you know basically nothing about how the part behaves. You would have to assume that the component is some slightly imperfect model of the ideal component, but you won’t know HOW imperfect it is (and that’s important). This section of the datasheet is almost exclusively what we’ll concentrate on in future articles, namely which specifications are the most critical to most applications.
• Characteristic Curves
  • I ignored this section when I was just starting out in electronics, only to later find how helpful the data can be. True, most of the curves utilize ideal conditions, but they can often be extrapolated to whatever condition you may be encountering. Knowing how a component will behave along a continuous axis … well, that’s what make analog components in the first place. They also point out all those weird conditions you could never really tell from the electrical characteristics by themselves. If the sheet says the voltage offset t the input of an op amp is usually 1 mV but you see the curves vary wildly over temperature, you might be reluctant to use that part in a harsh environment.
• Application Information
  • This is the section I mentioned back in the electrical characteristic section as one of the few that actually matters. In terms of usefulness to engineers (and others), the application section often designs a circuit for you to drop into your design. While it’s tempting to try and come up with a better method, what is the likelihood that you will come up with a better configuration than the people that designed the parts and talked to the most customers? The application information can also be a great source of general knowledge. In many books describing analog eletronics, the authors often quote application information as a way to learn new techniques and non-standard ways to use a component (such as using a voltage regulator in a non-standard way).

While these are not the only sections of all datasheets, these are the most common and important sections of datasheets to figure out what you’re doing with a particular component. You can always try to play around with a component until it works, but why not figure out the answer from the beginning (at least within a 95% confidence interval)? With a critical eye towards the overzealous nature of marketers and product salesmen, the reader can easily decifer necessary information for their project and how to easily compare it among multiple product offerings. If you become proficient in pulling necessary data from product spec sheets, you will have much quicker and accurate product offerings in whatever field you participate.

What about you? Have you ever noticed a section of a datasheet that you think is important that we missed here? If so, please let us know in the comments!

Sometimes the best way to learn about a subject is to be part of or listening to a conversation regarding the topic. This discussion between Chris and Mateja discusses low noise techniques for prototyping.

Mateja: Hey, did you see my tweet about breadboards?

Chris: Yeah, what happened?

Mateja: 60 Hz. It’s just everywhere!!! It won’t leave me alone!!! As soon as I take it off the breadboard though it works. I’m using an Olimex development board with the half-perfboard. So now I’ve got it on the perf with jumpers and it works well.

Chris: You try turning off the lights?

Mateja: I’ve tried that before, but not with this project … what do you think is worse, fluorescent or incandescent lights for noise?

Chris: Fluorescent hands down,those things are nasty. But they throw off a lot more than 60 Hz. What are you working on?

Mateja: I’m writing a driver for an RTC module for our sensor platform, but I’m using a dev board because we have some hardware revision issues that we have to work out with the sensor board. So I’m using an Olimex dev board and a handmade breakout board for the RTC. I couldn’t get the clock output of the rtc to stabilize on the breadboard because it was picking up the 60hz noise and trying to trigger off that at the crystal input. The crystal would never stabilize.

Chris: Oh yeah, that’s a stinker of a problem. Putting a crystal on a breadboard in general probably isn’t advisable.

Mateja: Really, breadboards are useless for anything above the audible spectrum. I guess low duty cycle digital works too but only for blinkenlights, not communication stuff.

Chris: Have you ever heard of the dead bug method?

Mateja: What’s that?

Chris: Not sure who came up with it, but Bob Pease wrote about it a bit. It’s how he creates circuits to reduce noise and preserve signals in high frequency systems. He flips chips over and then puts the plastic part down on a grounded piece of copper; then the leads are sticking up in the air like a dead bug.

Mateja: How does he connect to the leads?

Chris: He skywires them.

Mateja: I’m not familiar with the term skywire.

Chris: Oh, just running a wire from one component to another; so they’re “wires in the sky”.

Mateja: Like wire wrapping or with female to female jumpers?

Chris: Solder.

Mateja: That’s probably the safest but it’s not that flexible.

Chris: Flexible, like “easily changeable”?

Mateja: Right, it’s not easy to modify.

Chris: True, that’s one of the drawbacks,but when you’re a solder cowboy like myself, it’s not too big a problem. Plus you can’t beat the connection between two soldered components. If you do stuff on a bread board, pressing down on a component can literally change the circuit because of how it’s contacting the ground plane.

Mateja: You’ve probably shaved a couple years off your life with all the flux fumes.

Chris: Haha, don’t I know it. Again, worth it.

Mateja: OK, fine. You convinced me I’ll solder it.

Chris: Yay! You’ll be better off, especially for a crystal.

Mateja: Oh the crystal is soldered down, I have it on a tiny little single sided PCB that has pins like a DIP package.

Chris: Oh that’s good. Before I forget, here’s a good explanation of the dead bug technique:http://www.lpkfusa.com/articles/prototyping/edn_2_96.pdf. It starts on page 2.

Mateja: Yea I see it, I’m going to have to look around the shop for a copper clad board. How do you adhere the IC to the copper?

Chris: Basically every IC has a ground pin, so you bend those down and solder them directly to the copper plane.

Mateja: Oh dear.

Chris: Oh yeah man, welcome to my world!  But soldering them should hold them on there and provide a really short path to ground too.

Mateja: You only get to bend those things once…

Chris: Yessir. Twice if you’re really careful. Expensive components you’re using?

Mateja: Shoestring budget more like.

Chris: I mean, it was easier back in the days of DIP packages, easier to bend those leads down. But you don’t really need to bend them. What kinds of packages are you using?

Mateja: This one is TSSOP.

Chris: Well, you could always just solder in a jumper wire or a 0 ohm resistor, that’ll do in a pinch and no bending involved.

Mateja: OK, thanks for the advice, I’ll break out the hot iron.

Chris: Sounds good!

A Sept 3, 2009 article from Printed Circuit Design and Fab: Designing for Power Integrity

Article Summary:
Sometimes problems that appear to be related to signal integrity are actually power integrity problems.  This article talks about how you need to be sure to consider the frequencies at which decoupling caps work well and to consider the amount of copper actually available to carry current after holes’ antipad and any other voids in the plane.  Noise at a digital IC’s power pin can be as problematic as noise on the digital lines connected to it.

Electricio.us is just in its infancy, but is capable of becoming a great resource to EEs (and non-EEs) the world over. We are looking for bright engineers with a knack for writing to contribute technical, useful articles for publication. Our hope is to become a central location for electrical engineering knowledge, in whatever form it might take. In the future we could see features on this site such as:

• A message board (short term)
• Networking capabilities (short term)
• Link sharing (medium term)
• Archiving of important articles based upon a ranking system (medium term)
• Resume pool (medium term)
• Collaborative projects (long term)
• Consulting services (very long term)

If you are interested in joining the writing team for this site, please send an email to chris@electricio.us. Also please feel free to tell others about the opportunity presented here. While writing was never associated with the great EEs of the past, we live in changing times. Engineers need to be able to get their thoughts to others quickly and efficiently. Those that are best able to communicate will find that they are awarded with the most work and the best jobs in the coming years. This site will be a resource for many and a showcase for many other talented writers and engineers. If you would like to be a part of it, please send an email to the address listed above. If you have any other ideas you would like to see implemented on this site, please leave your thoughts in the comments.

PS. For any StumbleUpon members who happen to hit this site, if you are interested in contributing, please send an email to the above address with “StumbleUpon” in the title. Thanks!

Electricio.us is meant to be a platform for multiple industry professionals write about their experiences in Electrical Engineering. The knowledge contained within this site will be a reference point for many new and established engineers for both professional and hobby work. Please come back to see what articles have been added and to read the profiles of some of the top minds in Electrical Engineering today.