Technology

On a fairly regular basis I get asked about the equipment used for the various podcasts that Gavin and I produce (e.g., Confab and Podcasting Liberally). I, personally, have no notable background in audio engineering. When it came to shopping for our podcasting equipment, I had to piece together information from many different sources. Our goal was probably more ambitious than most, and included the requirement of supporting four or more participants from the very first recording. Luckily, only a few, minor mistakes were made on the path to achieving what we produce today.

In the most basic terms of podcasting, you need a way to record audio and convert it to a file format compatible with digital audio players. This typically means a microphone, computer (with audio recording software), and the MP3 file format. Distribution of the resulting podcast usually implies the Internet and Real Simply Syndication (RSS), but that's a whole other blog post. Let's stick with the production side for now, and start with microphones.

Following a series of recommendations and some testing, I decided on the Sennheiser HMD25-1 headphone and microphone combination. It includes a dynamic, super-cardioid microphone that, in the words of Sennheiser's marketing literature, "...has been specifically designed for use in noisy environments." They absolutely live up to the claim. Also, the benefits of a microphone and headphone combination should not be understated. For one thing, they tend to keep the microphone an equal distance from the participant at all points during the podcast recording, despite how much the participant moves their head. Although the MSRP for the Sennheiser HMD25-1 is $519.00 (USD), you can usually find them for ~$400.00 (USD). Be forewarned that the cable ends are going to be your additional responsibility - the HMD25-1s ship with bare wire ends to the headphone and microphone cables. Soldering male 1/4" jacks for the headphones, and male XLR connectors for the microphones prepared our HMD25-1s for actual use.

To improve the audio quality even further, our microphones connect to a series of tube pre-amps before feeding into the mixer. Although the mixer has built-in pre-amps for the microphone channels, their "gain" range is fairly limited. The tube pre-amps provide the necessary range, but also what audiophiles refer to as "warmth". Since each incoming microphone requires its own channel, we bought four of the 2-channel Behringer Ultragain Pro MIC2200s tube pre-amps. MSRP for the Behringer MIC2200s is $129.99 (USD), but you can readily find them for ~$99.00 (USD).

The eight, separate channels coming from the tube pre-amps feed individually into our mixer. In most traditional applications of a mixer, the input channels are all combined -- at various levels -- to produce a right and a left "main" channel (the output). Effects processors and EQ functions may also be added to modify the audio as it is being combined. The two resulting "mains" would normally be what get recorded for a podcast. In our case, we purchased a mixer that included a digital, Firewire (IEEE 1394) interface. The Firewire interface connects to our computer and delivers each of the input channels directly to our audio recording software. We not only use the mixer to combine the input channels, we also use audio software on a computer to perform the exact same task after the recording. Why? There are really two reasons. First, we need the ability to return the mixed channels to the participants' headphones in real time during the podcast recording. That's done through the mixer. Second, we want the ability to add effects and control levels on the individual channels in software, during or after the recording. The mixer and computer combination give us this remarkable flexibility.

In the very beginning, I purchased the Alesis Multimix 16 Firewire mixer. What I really wanted was the Mackie Onyx 1620 with Firewire option (~$1,300.00), but it (pictured) was more than twice as expensive as the Alesis (~$600.00). (You have to be somewhat conservative when you don't know how serious you'll be about a new hobby.) After almost eight months of regular podcasting, I felt that it was serious enough to warrant the Mackie Onyx 1620 purchase. The Onyx has several advantages over the Alesis, including visual level meters and low-cut filters for each of the eight microphone channels. It also produces better sounding audio.

For anyone considering a mixer for their podcast production, there are several Firewire and USB models from a handful of manufacturers. The number of channels delivered to your computer is the most significant difference between the Firewire and USB versions. In every model I've reviewed, the USB versions only deliver the left and right mixed channels to your computer - not each of the individual, pre-mixed channels. The lower cost of the USB versions is likely the result of this difference. Vendors need to make this difference more apparent in their marketing literature if they expect to have high rates of customer satisfaction within the podcasting community. There's absolutely nothing wrong with using a USB version, as long as you know about their limitation prior to spending the money.

The final piece of the system (excluding the computer and software) is the headphone distribution amp. The mixer has a right and left channel output, but definitely can't drive a series of headphones. To solve this challenge, I purchased the Behringer PowerPlay Pro-8 HA8000 (MSRP $179.99 / Street $140.00), which is an 8-channel headphone amp. The output from the mixer feeds into the amp, and the headphones connect to the individual distribution channels. Each channel of the distribution amp has the added benefit of it's own volume control. I eventually purchased two of the HA8000s. One is used to support the headphones of the podcast participants; one is used to support a set of headphones for listeners. This second headphone distribution amp is really important for Podcasting Liberally because it's held, weekly, at the Mountlake Alehouse. Anyone wishing to listen in on the live recording needs headphones to hear over the background noise.

All of our podcasts are recorded to disk using Apple's GarageBand software on my 17" Apple Powerbook. (I hope to be using a 17" Apple MacBook Pro in the very near future.) GarageBand's ability to perform compressor/limiter functions on a per-channel basis is the second most influential factor in making our podcasts sound good. We would have to spend another $2,000 to perform the same task in audio hardware. Gavin has been mixing and converting the shows on a Core Duo Mac Mini since the beginning of March, and I can't wait to have that sort of power in a portable form. All told, we've spent ~$6,300.00 (not including my laptop) on the podcasting gear. It has been worth every penny that we've spent to hang out with good friends for smart conversation on a weekly basis.

Go forth and podcast!

The new high-density servers in our data center have posed quite a serious cooling and power consumption anxiety. Within the span of two years, our typical 1U server purchases have more than tripled in BTU output (from 221 to 780)¹ and more than quadrupled in amperage draw (from .54 amps to 1.91 amps)¹. Despite how well the data center might have been designed, these demands are technically beyond its inherent capacity. Special steps must be taken for those cabinets where we have a high concentration of the newer servers.

One recent step was to optimize the cool airflow within the cabinet enclosure. After some trial and error with several different options, we settled on the Delphi Enclosure Blower -- a 2U, 250 CFM device uniquely designed for this particular application. The cooled air in our data center is supplied at the base of the cabinets through a raised floor, and exhausted into the room through the top of each cabinet. The Delphi Enclosure Blower directs a portion of the cooled air up the inside front to ensure that adequate cooling reaches the intake for all the equipment in the cabinet. Delphi's research claims a 15^o F improvement for equipment at the top of a cabinet (where traditional hotspots occur). Although our temperature monitoring equipment is not configured to measure the front of our cabinets, I do not doubt the claim's validity. My anecdotal experience leads me to believe that Delphi's research is pragmatic.

Improved exhaust flow remains a challenge for the cooling project. The 500 CFM fan at the top of the cabinets aren't powerful enough to expel the heated air so as to keep the temperature differential between the base and the top under 20^o F. Using the basic rule-of-thumb that the cabinet exhaust fan should have a rating roughly equivalent to the sum of all the housed equipments' exhaust fans, we're about 400 CFMs shy of what we really need.

¹ The noted BTU and amperage values are nominal, and can represent as little as 50% of the peak BTU or amperage ratings of the equipment. In practical terms, the nominal values are accurate for regular operation.

For a couple years now I've been seriously considering building a portable solar power system for my laptop (a 17" Apple Powerbook G4, 1.67GHz). The two motivations behind such a project are to achieve a real understanding of photovoltaic technology and the power consumed by my computer. I use the qualifier "considering" because I've yet to find a source for all the necessary components. Let me explain.

Most consumer solar power systems consist of one or more solar panels, a charge controller, and one or more batteries. The solar panel(s) supply power to the charge controller, which regulates the storage of that energy in the form of a battery charge. The battery energy can then be used when direct sunlight isn't available to the solar panel(s). Connecticut Solar sells just such a package (60 watt, 16 amp) for $795, and it's specifically marketed for laptops and consumer electronics applications. The "rub" is that you're unnecessarily complicating the issue with such packages. You shouldn't need a charge controller and battery when your laptop already has those components. Another downside is that these systems only supply enough power to run the laptop - not charge its internal battery.

What I would like to build is a portable solar power system for my laptop that eliminates all of the duplicate components. This should require only one or more solar panels and a DC to DC converter. The challenge has been to source a DC to DC converter that can take power from a set of solar panels and produce ~2.65 amps at 24 volts DC (the exact output rating of my AC-DC laptop power supply). Any ideas?

You might suffer from an addiction to latte art or roller coasters, but I have an addiction to a certain category of gadgets. My particular "fix" comes in the form of gadgets that perform an automated data acquisition task. (Geek Alert!) An example is the real time temperature monitor I recently installed at our data center. The latest "kick", however, is Garmin's new Forerunner 305. Having previously owned two of Timex's Ironman speed + distance systems (one with, and one without heart rate monitoring) and several other handheld GPS units, I can authoritatively state that Garmin has finally succeeded in an odyssey strewn with antecedent carcasses.

"Such strong words," you say? I'll let you in on a couple dirty little secrets about most portable GPS units. After spending between $300 and $500 on each, you'll learn to your chagrin that they only work in the middle of a desert on a clear day for the ten minutes that their double "A" batteries hold out. That's right. They're so pathetic that trees and clouds will reduce them to nothing but a battery consumption apparatus (which, ironically, they are exceedingly good at). Did I mention high-tension power lines? How about outside tall city buildings?

Don't get me wrong. GPS has held enormous potential. It's just that those of you, like me, who have purchased model after model need to start being honest about their shortcomings. Or, like me, you need to toss the lot of them and buy the Garmin Forerunner 305. (No, I'm not being paid my Garmin.) It's compact, easy to use, and reliable. I'm searching for something I don't like about it.

Let me explain.

The Forerunner 305 is an oversized wristwatch combined with a traditional heart rate monitor chest strap and, optional, cadence sensor (for cycling). I've been using it for running, cycling and hiking. In those activities, I most often have the unit's display set to indicate event duration, distance, pace or speed, and heart rate. Unlike pedometers or bike computers, GPS provides truly accurate speed and distance measurements. (Elevation calculation is a bonus that pedometers don't even pretend to deliver.) The advantages, however, aren't limited to the course or trail. A provided computer interface enables transfer of the acquired data to a web-based training log and analysis service (subscription required: www.motionbased.com). That's where the trifecta of GPS, data acquisition and Web Services reach a mesmerizing crescendo.

Take, for example, the following log of the past three running events at Greenlake. They were all fairly slow and short runs, but you can begin to image how such a log would aid in training over the season.

Motion Based logs all of the data acquired by the Forerunner, and combines it with weather data, a graphing system, and a handful of mapping systems. For instance, I can see that this morning's Greenlake run weather was noted as "scattered clouds at 1000 feet" (from the Boeing Field weather station). The average, low and high temperatures, relative humidity and wind speed were also recorded. Sitting like a maraschino cherry on the top of all this delicious data acquisition dessert is the Google Earth feature. (Your jaw will drop when you first see a Google Earth tour of an event recorded with the Forerunner.)

The image to your left is just a still photo of the Google Earth tour. If you want to experience the real thing, here are a couple Google Earth files from my training log: Greenlake (running), Lake Washington (cycling), and Rattlesnake Ledge (hiking). [You'll have to right-click on the links and save the files to your computer. You will also need to download the Google Earth software from http://earth.google.com.]

I'm not even scratching the surfice of Motion Based's data analysis features with this blog post. You can combine the data in several different ways to produce line, distribution and pie charts. Here's a simple one showing heart rate and elevation of this afternoon's hike to Rattlesnake Ledge.

The Garmin Forerunner 305 receives my strongest endorsement.

When planning on being abroad for more than a week, and traveling with electronics that require direct power or battery recharge, you'll have to address the issue of power adapters. Since most consumer electronics have auto-switching power supplies (i.e., support for 100-240 volts and 50-60 hertz), the biggest concern is the type of plug required for power outlets in the country or countries you will be visiting. I'm trying to eliminate any unnecessary equipment, so things like the iPod power adapter stays at home -- it can charge while connected to my Powerbook. As for the Powerbook's power supply, Apple sells a World Travel Adapter Kit with all of the plugs you might need. I expect to only need one type of adapter, which will be the only one I take on the trip.

You should always make sure and check whether the power supply supports auto-switching. Plugging a 110 volt power supply into a 220 volt source will destroy the power supply. Trust me on this. I once blew up a $2,000 laser printer by plugging it into a 220 volt outlet on a Norwegen-built ship that I worked on.

It's a good rule of thumb to assume that any non-portable consumer electronics lack auto-switching power supplies. In such cases, you'll need a step-down transformer, rated at the approprate wattage.

The nifty temperature sensor (mentioned in a previous post) showed up on Thursday, and was installed at our data center today. This is when the geek in me really shows. Here's a device that combines an embedded server (probably Linux-based, although I haven't checked for certain) and a data acquisition component (in this case, one or more temperature sensors). The result is a compact, rack mountable device that will display real time temperatures for multiple locations. It will also log the temperatures at a variable time interval, which defaults to every 5 minutes.

I've rack mounted the device in the top of the cabinet, and run the second sensor to the bottom (a distance of approximately 72 inches). The important thing to know is that the cooled air flows through the raised floor; therefore, the differential between the two temperature sensors will represent how much heat is added by the equipment in the cabinet (i.e., servers, disk arrays, switches and firewalls). So what are my expecations?

First, the cool air supplied through the raised floor should not fluctuate much. Fluctuation in the temperature of the supplied cool air would indicate a failed cooling unit or thermostat.

Second, there's a possibility that a measurable differenital in temperture could coincide with the daily network traffic patterns. Now that would be interesting. Also, there could be a spike in temperature during the scheduled backup procedures when the multi-terrabyte nearline array is running at full clip.

The small sample taken following installation today indicates an average differential of 22^o F. That puts the temperature in the top of the cabinet at 82^o F. (Does anyone other than me think 82^o F is a bit too high?) I realize that the upper tolerance for the equipment is 104^o F, and that I'm measuring the exhaust temperature (not the temperature of the air flowing through the equipment), but additional equipment planned for the cabinet will push the temperature even higher.

As with most new information, there are new questions to consider. Additional temperature sensors might need to placed in the cabinet to test for air circulation patterns. The cool air should be flowing from the front to the back, in addition to flowing from the bottom to the top. If this isn't happening, the equipment could be starved for cooled air.

I'll post graphs once I have a large enough sample size to work with.