Arthur Kelm knows a thing or two about power and electronics. He has built studios across the country, served as chief engineer at many others, and now consults on power and infrastructure as the owner of Ground One in Guerneville, Calif. He has a 43-year legacy that takes him back to when Aphex introduced the first 900 series modules, and he is credited with coining the term“Lunchbox” in the late 1970s when he and Niko Bolas were working at Record One.

Today, he works at Capitol Studios in Los Angeles as Director of Engineering and is overseeing the overhaul of the entire facility, and recently he joined the team that has just started the process of resuscitating the famed Chess Studios in Chicago, in its 1964 glory with a 2013 digital backbone. We asked him to comment on power and how it works in the range of 500 Series racks on the market.

Power distribution seems to be the hot topic in 500 Series today…

It’s only about in the last year or so when this talk started picking up, in casual conversations at AES or in studios. People started talking about how different racks sound different, and everybody had an opinion about which units sounded better. Me, being a power guy, I started to think, “Oh, that’s interesting.” Before, there was no comparison. People might compare an API 10-space to a Brent Averill 11-space, but no one even thought about it. Now, there are a lot more options.

So what are the primary challenges, from an electronics point of view?

Headroom. It’s all about headroom. And headroom is related to dynamic current draw. Obviously a kick drum, snare drum or any percussive instrument with really high transient requires dynamic current draw. That becomes a peak current draw on the power supply. So let’s say a module is rated at 1 amp; peak current draw with a dynamic could be as high as 4 amps for a few milliseconds. Times 4 is a good rule of thumb.

Would you recommend that’s what a buyer first look at?

I would say the first thing you want to do is add up the current draw of each of your modules and then compare that to the current rating of your power supply. That’s your first bit of homework for due diligence. Then apply the 4x calculations. This should be the current rating of your power supply. Then,if you can, find out from the rack manufacturer whether they use current limiting in their power supply. The reason for using current limiting is that in case of a short circuit the power supply shuts down. But there has to be a threshold there, right? That will be your limit for peak current draw. Most power supplies use current limiting, and most manufacturers use off-theshelf product, rather than design their own circuitry. I prefer using power supplies that do not implement current limiting. A slow-blow-fuse approach on a power supply with no current limiting is a much better approach. This will give you maximum dynamic current for high transient demands.

Are there advantages to either an internal or external power supply?

Usually internal PSUs have an EMF field issue that you have to guard against. Because it’s a transformer inside the box, depending on what module you put next to the power supply and including how well shielded the power supply is, you will have more or less hum. If the manufacturer did their due diligence, it’s not an issue. But it does raise a flag. That’s why most people went external. It takes your EMF interference issue away. But that creates a dynamic current draw problem, which is the whole basis of my conversation. Now you have a length of cable and at least two connections, if not three, between your modules and the power supply. Depending on the surface area of the connection, the length of cable and the size of wire of the cable, that has current-limiting possibilities written all over it.

What happens, then, from power supply to rack?

If we go into the minutiae, starting at the module the first thing you have in your path is your edge connector. All manufacturers use the same pin out and standard connector, but the amount of gold and the amount of tensile strength matters. So all edge connectors are not created equal. Secondly, is the connector hard-wired, or is it PC board? PC board assemblies are cheaper to manufacture, but the drawback is the size of the trace going to the power distribution. If they have really thin traces, they will limit dynamic current draw. Ideally, like in the old APIs, they were all hand-wired, bus-wired going through the connectors. When you start looking at what you’re buying, you’re looking at the number of connectors, the quality of the connectors, the surface area of the connectors…then you start looking at wiring. The gauge of wire should equal current draw. For instance, you don’t want to have 16-gauge or 18-gauge wire feeding 10 amps to your rack. It may carry the current, but there’s no dynamic current draw. You can compare it to speaker wire. The longer distance, smaller wire, less headroom, less dynamics, softer sound.

The rack itself…any difference with 2, 3, 6 or 10?

More slots in the rack would dictate the use of a heavier wire gauge or larger traces. More modules, more current draw. The whole thing is based on dynamic current draw. That’s why racks sound different. The whole goal with 500 Series power is to not restrict the module’s ability to perform the best it can. All
modules use amp blocks, whether discrete or IC. You want to supply enough dynamic current so the amp blocks can perform to the best of their ability.