μblog: engineering from the trenches

Today is the first day of the American Epilepsy Society annual meeting and I attended some lectures on sleep-related epileptic events and another one on the impact of the immune system on epileptic activity. The first part, dealing with sleep and seizures, was mostly clinical results and a strong ties with abnormal electroencephalogram (EEG) recordings during non-REM sleep and cognitive impairment in children. Several specific pathologies were discussed, the main topic being speech impediment. It is known that many children experience epilepsy-like events when they are young, however, most grow out of their seizures and go on to lead normal lives. For some, the speech impediment becomes life-long if the night-time seizures are untreated. The interesting aspect is that the children can read and write, even use sign language, but they cannot understand spoken word or speak. The consensus from the group of people at this special interest group was that it is important for children with sudden speech impediments and a family history of epilepsy to undergo night-time EEG recording to determine if their speech problems can be linked to things like autism, or if it is actually an epileptic event. If clinical epileptic activity is observed, it is suggested that the child is monitored for an extended period of time (12-18 months) to see if there is a natural improvement in their speech, and only start thinking about drug therapy if there is none. The thinking is that if the EEG signals can be brought back to normal early enough, providing they were not on course to do this themselves, the impediment would be reversed.

The second part, dealing with immune response and seizures, was somewhat out of my league. The talks were mostly covering basic science and all seemed to link increased immune response with increased occurance of seizures. Various pathways were demonstrated, mostly in animal models, from linking temperature increases to seizures all the way to demonstrating antibody binding to and blocking ion channels.

I am fairly busy at the conference so today’s IC Friday may become next weeks IC Tuesday or later. With the bad news comes some good news. The chip I have lined up is one that is very close to my heart, the Graphics Synthesizer chip from the Playstation 2. I have most of the images on my laptop already, I mainly need to stitch them, format and upload, so I apologize for the wait.

del.icio.us | digg

This very small introduction to the Central Limit Theorem is probably something worthwhile before the Shannon paper. The main point is that as we take more and more samples from a random variable, with a fixed mean and variance, the samples approach a normal (Gaussian) distribution. That is, irregardless of the distribution of the random variable, if it meets the criteria, it will behave like a normal ly distributed random variable in the limiting case. The typical application engineering application of this theorem is making the assumption that some measured quantity is normally distributed and use that assumption to define things like confidence limits and so forth. The requirements for this assumption are that the process is second-order stationary, meaning the mean and variance do not change in the window of observation, and that the number of samples is approaching infinity. The requirement for a large number of samples can sometimes be loosened since the residual differences between the sample distribution and a normal distribution can sometimes be determined. The requirement for a stationary process cannot. For example, it would be foolish to apply Gaussian statistics to a random-walk (Brownian motion).

The key message is that the normal/Gaussian assumption is typically a good one, as long as the statistical nature of the random variable under investigation is constant through the period of observation and the number of samples is large.

( sec_4_f.pdf ) ( Image is from Wikipedia )

del.icio.us | digg

I am going to the 2007 American Epilepsy Society annual meeting in Philadelphia, PA next weekend and intend to present a poster half of which will deal with multi-taper spectral analysis. I am presenting it next Monday morning, from 8:00AM to 2:00PM (poster #3.165). I could also use some suggestions for a good place to get a cheese steak.

The novelty of this presentation is that I am using orthogonal Slepian tapers to determine both the amplitude and phase of a complex transfer function, where most of the previous research deals with amplitude spectra alone. The difficulty is that the various tapers have different phase-bias characteristics, so averaging them and then computing the phase and amplitude would not work. Nevertheless, if we first compute the power in different bands, per taper, and then average, the phase dispersion will not play an important role. The second part of the presentation will deal with causing the activity of an on-going, epileptic seizure to become temporarily phase-locked to applied electrical stimulation (entrainment). The technical abstract for the poster presentation is below.

( nchernyy-abstract.pdf )

del.icio.us | digg

It is already Thanksgiving and I am still posting chips from the iPhone. This is a mystery chip from the iPhone with limited information on the package and “PMA19″ on the die. After taking a close look at the surface at 400x total magnification, I still couldn’t see any hint of the origin of the chip. The only guess that I could make is that it it might be an RF modulator/demodulator with a built in DSP. If anyone has some insight, feel free to post a comment.

pma19-4x-stitch.jpg

Update: Here are two pictures of the passive component as per request.

del.icio.us | digg

I found this surprisingly well-written manual for I2C serial communication protocol today. In short, this is a fairly popular message-based protocol that can be found in many embedded systems in consumer electronics, test and automation and automotive fields. There are low-speed alternatives and the structure of the protocol is fairly user-friendly making it a good option for hobbyists. There also schematics available on-line for rs232 and USB to I2C adapters available on-line like this open-source platform.

( an10216_1.pdf )

del.icio.us | digg

This weeks selection is the PCF50605 from Philips (now NXP).  This is another power management IC aimed at portable electronics. This is considered to be a fairly flexible, single IC solution providing fine power control. Stitches are included of the analog areas for convenience.

pcf50605-4x-stitched.jpg

pcf50605-10x-stitched.jpg

pcf50605-40x-stitch.jpg

del.icio.us | digg

Looking through webserver logs, I noticed that some people were using a search engine to find this blog. They remembered part of the url, but not the whole thing. This is understandable since microblog is pretty generic and it is not even the domain name. To make things easier, I am forwarding all of the traffic from ENGR.BIZ over here, so until I come up with a good idea for that domain name, it can serve as an easier-to-remember alias for microblog.routed.net.

del.icio.us | digg

The inaugural paper for the Journal Club is titled “Power-constrained high-frequency circuits for the IBM POWER6 microprocessor” by Brian Curran et al. and is published in the November 2007 issue of the IBM Journal of Research and Development. I have much respect for the whole POWER micro-architecture, consequently, I am interested in learning a little bit about their design methodology which lead to a near-5GHz core logic clock rate. The IBM design team responsible for the POWER6 applied a three-direction strategy to achieving this performance goal: cutting edge technology, manual circuit optimization and thorough testing.

The processor was designed at a 65um manufacturing node so various technologies needed to be employed to keep leakage current to a minimum and thereby maintain an acceptable power usage. The first method involved using silicon-on-insulator (SOI) which reduced back-gate current due to parasitic capacitances and can CMOS latch-up. The processing steps to implement SOI are well understood, however, extra care must be given to design layout as it is no longer possible to drive the back-gate by connecting the whole substrate to a fixed potential. Another technological advance employed was the use of dielectrics with low relative permittivity between traces to further reduce transmission line effects and the associated propagation delay of interconnects. Since less energy is stored in the dielectric material between interconnects, this also reduces power consumption.

From a design stand point, the goal of the team was to distribute the clock properly and to maintain the latency of the core logic circuits below “13FO-4”. Propagation delays, loading and transmission line effects play a very important role in the 5GHz regime. It was very interesting to see how multiple layers of buffers and clock delays were included to guarantee that clock pulses would be synchronized around various cells while maintaining an adequate slew rate. The 13FO-4 latency means that each processing cycle had to be accomplished in the time it would take for a signal to propagate along a chain of thirteen inverters that were loaded with four devices each. This is the criteria which allowed for a 5GHz core logic clock rate. It was mentioned that threshold voltages were tuned, probably through ion implantation, to minimize leakage while maximizing speed.

Simulations, being the last major piece of the paper, were less interesting as they relied mostly on proprietary tools. The piece that may have been important for readers was the iterative cycle of debugging and performance tuning. Going from schematic overview to transmission line calculations to back-annotation, to placing and routing made some sense.

Please feel free to contribute your thoughts on this paper, my interpretation or another paper that would be an interesting read in the comments section. Lets look at Claude Shannon’s paper titled ‘A Mathematical Theory of Communications’ as suggested by Adam. As the full paper is quite long, we may want to look at only the first thirty pages in detail. Those that want to brush up on their mathematics before attempting the paper should start on page thirty-two.

2007curran-power-constrained-high-frequency-circuits-for-the-ibm-power6-microprocessor.pdf

del.icio.us | digg

I came across this interesting post at the Mouse Print blog and thought, that I have found a compliment to feature creep. The keen observation was that the Dial soap company introduced a grip feature to the soap and added “all day odor protection” and reduced the size of the bars from 4.5 to 4.0 oz. From reading the comments, the Mouse Print readers were unhappy and one recalled the days when soap was bars were 5.0 oz. Those readers, and I am sure some other consumers, seemed to feel that the soap was less effective because of reduced size and thereby less of a value for them.

This line of reasoning to me was odd because soap, and detergent for that matter, has very little of the main active ingredient, called surfactant, in relation to the overall weight or volume. This is the chemical that has both hydrophobic and hydrophilic end groups and is can be used to allow water to act as a solvent for fats. This chemical also facilitates soap bubbles by reducing the surface tension between water and air and forming a air-surfactant-water-surfactant-air interface.  Anywhere from 1% to 5% is probably sufficient to get a nice, creamy lather and a clean wash. Most of the remaining mass of the soap is can be made up of something like a paraffin wax and lotions with trace amounts of anti-bacterial chemicals and deodorants. The point that I am trying to make is that we can have a much smaller bar of soap do the same job as a larger bar of soap by keeping the amount of surfactant in the two bars about the same and by altering the waxy substrate for increased longevity. The addition of odor-fighting chemicals on the package seems to indicate that at least some of the soap chemistry was changed, so it may be possible that other chemistry was also altered and the slightly smaller bar of soap will still last for just as many showers as the larger one. A single milliliter of Tween-80, a popular commercial surfactant, is probably enough to clean a load of laundry, however, people still have the mentality that adding more detergent will give a better wash.

A topic closer to the thoughts of many engineers may be the Everex TC2505 being sold at Walmart. What is interesting is that the mid-sized tower contains a mini-ITX mainboard with a Via C7-D processor, something that can fit into the same volume as four or five DVD cases. It is suggested that Walmart performed some market research and concluded that people still see size as an indication of performance for a desktop computer and would therefore think that a smaller system would be inferior.

My conclusion is that changing an existing design characteristics, like size, is very difficult and doomed to rejection by the consumer. The best hope is to try to launch a new product and hope to put positive spin on the improvements. In 1985, the Coca Cola drink recipe was altered resulting in a backlash and a subsequent re-introduction of Coke Classic. Diet Coke, however, has received a warm reception from many. As far as the soap goes, a whole new application will have to be designed. Although body wash is fairly new, we are already ingrained with the idea that we need about a tea-spoon of it to wash up. I guess the future could be some soft device with built in electronics and microfluidicts. It would be the size of a good, old fashion bar of soap and will feature a button and a LED. When we press the button, the LED will blink, the microfluidics will dispense micro-liters of surfactant and deodorant to the device’s surface and the device will wiggle or click a little letting us know that something useful has happened. We will lather and wash with it just like normal soap today and think back to the days where respectable soap bars weighed 5.0 oz.

del.icio.us | digg

This is another USB power management/battery charger IC from Linear and a follow-up to the LTC4066 displayed earlier. This seems like an earlier version of the same design, which makes sense, since the 4066 was from the iPhone while the 4055 is from a fourth generation iPod. The designs look similar, but we won’t fully know what changed until the datasheet for the LTC4066 is publicly available.  The 10x stitch is of the whole chip while the 20x stitch is a closeup of the control electronics on the bottom of the chip.

ltc4055-10x-stitch.jpg

ltc4055-20x-stitch.jpg

del.icio.us | digg