|Fringe with Greg Smith
Back Issue Article
The voices kept whispering in my ear. Greg Weaver was the first. Months before his article appeared, he handed me his hand-drawn plans for the Silver Signal Tape interconnects. After hearing the results in action, I was intrigued, but not so much as to get my soldering iron out. Next up, in March of '97, was Australian reader Shane Ceglar. He'd read my comments about the Canare interconnects I'd gotten from Markertek. He suggested I look into some of the work Jon Risch had published regarding Belden's 89259 cable. I promised Shane I'd look into that, then promptly was too overwhelmed with reviews to worry about it. He continued to politely nag me at regular intervals on the topic. The final source of DIY temptation and inspiration was a visit from the ever-popular Steve Rochlin a few months ago. He let me borrow one of his homemade digital cables. I was impressed, hooked on the concept, and Steve made it sound so easy. My new apartment finally gave me enough room to set up the old workbench again, and I've been a cable-making fool for the last couple of months.
Let's rewind that paragraph and play it back at regular speed.
Many people have been successfully constructing variants on the Silver Signal Tape design Greg Weaver has popularized. I think the best part of that project is that you can pick up every single part for the basic version in a single trip to Radio Shack. I dutifully bought some copper magnet wire (Cat. No. 278-1345A, $3.99) and four gold RCA connectors (Cat. No. 274-850, $2.59) at my local store and followed the construction plans. Trying to scrape off the insulation at the end of the red (30AWG) magnet wire with a knife seemed unnecessarily difficult. Grabbing some sandpaper and scratching the coating away is much easier, and you're far less likely to cut yourself in the process. I suspect the slightly scratched-up surface might give a better solder joint as well. Another warning I'll add is that you need to be careful not to apply too much heat to the center conductor when you're trying to slide the cable in there and get it to hold. The insulation in the jack will melt if you leave the iron there too long. Believe me, doing this to your last connector on a Sunday night 10 minutes after Radio Shack has closed is not a good way to finish up your weekend. I suggest buying an extra connector just in case.
One thing that's not really obvious from the SST article is how to deal with the tape at the ends. Running it some extra distance so that it overlaps the strain relief portion of the connector really adds considerable strength to the assembly. I ran tape all the way up to the top of the strain relief, wrapping it around as well as I could. This is kind of ugly but very functional. Total cost for this cable, $14.35 plus some packing tape and solder.
To say Belden makes a lot of wire really doesn't do justice to the huge product line they manufacture. Sorting through their catalog trying to find something that will sound good is an intimidating project. The aforementioned Jon Risch has been searching Belden's stock for quite some time, weeding out the best-sounding audio cable and posting his results to some of the Internet newsgroups. Matt Churches has organized these comments in a more permanent form as part of his listing of DIY cable resources. One of the most popular of Risch's designs uses Belden's 89259 video cable for an interconnect.
I wasn't sure at first how supportive of small-scale DIY efforts Belden was. The last time I had asked them about some cable I wanted a small quantity of, I was told that a 500' run was as short as anyone would sell me, making the whole project I'd been working on moot. Apparently the company has seriously turned around in this regard since then. Their technology development manager, Steve Lampen, wades into the audio newsgroups regularly to help out people who are trying to figure out which Belden cable is right for their application. Steve sent me some 89259 to try out. If you'd like a free sample yourself, you can call Belden customer service at 1- 800-BELDEN-1 and ask for one. They are quite glad to give out complementary 3' lengths of their cable to DIY experimenters, enough for a short pair of interconnects. For longer sections, you can get a 100' length from any Belden distributor. The typical cost is somewhere between $1 and $1.50/foot. Wandering the 'Net for a bit found Newark Electronics selling one for $141.99. A helpful fellow named Robert Hoffman recommended to me Nemel Electronics as the lowest-priced source available for 89259, at under $90 for 100'. Belden has responded to requests for purchases in smaller quantities by finding distributors who will cut their cable and sell pieces. Their "Recommendations for Home Theater and High-End Audio" page lists them. Expect to pay a bit of a premium over the regular unit cost for a smaller portion, and note you'll probably run into a minimum order cost in any case (typically $20-$50 for the kind of companies mentioned). That particular page is woefully out of date when it comes to giving website addresses for the companies mentioned, so a visit to their main Distributor Locator might be helpful as a check on the data if you have trouble.
Wait, you wanted to actually find out what was in the cable before you ordered some? OK, I can see that. The core of the 89259 consists of seven 30AWG solid copper conductors, equivalent to a 22 AWG in total. The insulator used is foamed Teflon[R], with a black Teflon jacket on the outside. Shielding is provided by a fancy copper braid. The specs are quite good, with a 78% nominal velocity of propagation, 15 ohms per thousand feet (49.2 ohms/km) resistance, and 17.3pF/ft capacitance. As this is a video cable, the nominal impedance is 75 ohms. Happy now?
To make an interconnect out of this, you'll first need to strip the cable into sections. While a real man would do this using only a rusty pocket knife, I wussed out and bought a tool for the purpose. Radio Shack's "Economy" Coaxial stripper (Cat. No. 278- 247, $6.99) makes short work of this job (they also sell a more expensive variant). It will probably take you a few tries to figure out the weird twisting pattern you need to use with the tool, because if you don't follow the directions just right you end up cutting rather than stripping. Since you'll need both left and right cables, I suggest alternating between the two when you're practicing and tossing away screwed up sections, else one might end up shorter than the other.
The Radio Shack gold connectors turn out to be perfect for this application. You want to take off about 1/2" for the center conductor, which happens with part B of the Economy stripper. You can safely take off a bit extra because it's easy to cut down to size if it's too long after you try placing the seven-conductor wire in the center hole of the connector. A regular wire stripper or some other cutter is needed to pull 1/4" of the white insulation off to expose the copper center for soldering. After you get this part right, part A of the stripper will pull off the outside insulation. Do this about 5/8" from the end. Make sure you don't cut through the braid at this point. If you do, you'll have to cut off the exposed portion of the cable and start over.
If you're attaching the second connector to any cable, always make sure you've pushed the strain relief and screw-on portion on before soldering. You'll also note the gold connectors come with a little piece of black rubber tubing designed to prevent loose pieces of the ground-shield wire from touching the center conductor. While these worked very well at insulating the SST interconnects, they don't do you much good with the 89259, as they don't really fit around the center properly. Solder the center copper core to the center pin on the connector. Then, pull the braid down around the ground, cutting away any loose strands, then solder around the outside. This is critical to get right because the braid is quite glad to get loose and short out the center if you let it. Screw the connector back together, and you're all done. What I really like about the resulting cable is that it's very durable. I could easily make a stack of these and give them to people, expecting them to survive. I can't really say the same thing about the SST interconnects, which are a bit too fragile for a ham-fisted guy like me, even when I'm trying to be careful.
I love the smell of solder in the morning
The third part of my DIY inspiration was from Steve Rochlin's Enjoy the Music site. I borrowed a sample of the RCA-to-RCA Max Rochlin Memorial cable (MR1) from Steve and refused to return it to him. I highly recommend reading his assembly instructions to get a feel for a more perfectionist approach to building DIY cables than I've outlined here, complete with cleaning solutions and a very professional look. As for what I think of the sonic virtues of the MR1, note the conspicuous lack of any other digital interconnect in the associated equipment part of this column. 'Nuff said.
After weeks of having the Radio Shack SST and Belden 89259 cables sitting around in various systems, I finally found some time to organize a shootout for these audio interconnects. A popular benchmark for the DIY market seems to be the products from JPS Labs. I think at least half the cable construction testimonials seem to mention how their design "whups JPS's ass." Since the $295 JPS Superconductor interconnect has been firmly seated as the reference cable in my system for quite some time, that made it the obvious target for my own comparison.
I started out with the Belden 89259, which I had zero preconceived expectations about. Listening to "The Ubiquitous Mr. Lovegrove" from Dead Can Dance: Into the Labyrinth [Warner Bros 945384-2] was quite enjoyable, and I was rather impressed with how the modest cable performed. Switching to the JPS Superconductor restored a touch of authority in the deep bass, but the big difference was at higher frequencies. Throughout the track is a constant brushed-cymbal part which snapped into clear focus with the Superconductor, while it had been merely a background sound with the 89259. The overall presentation was more dynamic and open. I had listened to two minutes of the track with the 89259 before turning the system off to swap cables. With the JPS interconnect, I blinked and five minutes had gone by. The song was over, and I'd been too lost in the music to remember I was supposed to switch cables.
Grabbing the copper magnet wire version of the signal tape I'd made, I threaded it through the mess of wires in the rear of my system to try next. Now, I'll admit I was walking into this one with a bit of a prejudice. You look at these tiny wires and you can't help but think there'll be no bass. I realize I've been slowly brainwashed into thinking you need big fat wires for bass. Amazingly, the low frequencies were actually better than the 89259. The overall level of bass balance to the rest of the music was similar to that of the Superconductor, but the DIY cable wasn't quite as tight on the bottom end. My second surprise: I'd expected a really great treble. Wrong again. The critical cymbal part I'd picked out on the last cable swap sounded even worse with the signal tape than with the 89259. The Belden cable doesn't quite match the crystalline clarity of the JPS cable, but it was clearly better than the signal tape.
Mixed results all around. The Superconductor was the clear winner in every area, with the DIY designs having their own characteristics. While neither home-brew is going to unseat the $295 JPS cable in my system, their performance is phenomenal considering the construction cost and simplicity. It would be quite embarrassing to pit either against all but the best cables in the sub-$100 range.
There you go. Two designs, under $20 each, with easy construction and minimum acquisition hassle. And once you get comfortable with making your own cables, there's a well-traveled path toward doing better. While Greg Weaver suggests the Radio Shack gold jacks are just about optimal for the SST design (he's tried other connectors up to $80/pair without improvement), I still haven't tried making the silver version. I'd expect that would greatly improve the high-frequency weakness I felt was present in the copper version. On the other hand, I'm told the Belden 89259 greatly benefits from better connectors. Also, some feel that removing the 89259's copper core and twisting a pair together yields an even superior cable. I'm skeptical about winding my own cable, but I'll try it anyway. I've been getting surprised a lot lately, so why not? Meanwhile, I haven't even started to walk down the road where I use better solder, develop some sort of good crimping technique, or chemically treat the connectors for better sound and less corrosion. I may beat the JPS cables myself before it's all done. But is that really probable? Talking with cable manufacturers gives an interesting perspective on the limits to what you can achieve in a DIY setting.
The Cable business
One of the things that's always limited my exposure to the DIY audio community in the past is the vehemence many members of that clique have toward cable manufacturers. While I'll freely admit that there's a substantial amount of "wire fraud" going on in that part of the industry, there are quite a few companies that easily justify their high price tags based on the work that goes into their products. It's just not always obvious what you're paying for. Some companies actually break their costs out and let you order their parts. For example DH Labs sells their excellent RCA connectors, spades, and wire so that hobbyists can build their own cables, typically at around a 25% discount below the retail for the finished product. It's very clear in a case like that what your dollar is going toward.
The big secret of the cable industry is that the materials involved often cost very little relative to the retail price of the final product. Frequently the expensive part of the process is the termination, the actually melding of the individual parts into a finished whole. This can require expensive machinery and highly trained individuals. For example, a technician named Mark assembles the speaker cables at JPS. He'd been working in the cabling field in other industries for about 12 years before moving on to audio. Even given that background, it still took six months of practice before Mark was confidently terminating cables using all of the techniques JPS had developed in the years before he worked there. I don't know about you, but I'm lucky if I can set aside six hours to practice my cable-assembly technique.
Another interesting tale comes from Nirvana Audio. Nirvana starts by buying raw copper, insulating/jacketing urethanes, and Teflon. A number of factories that specialize in one particular process (i.e. drawing copper or assembling wire strands) receive the evolving materials in succession, as Creamer has yet to find any single source with all the required tools, machinery, and skills in-house. After the wire itself is built, three Long Island factory teams handle the complicated set of termination procedures Nirvana calls for. This requires another vast array of equipment and mil-spec-trained technicians (all currently with at least 17 years of experience each in making exotic cables for aerospace, medicine, and other such industries). Because the parts that make up each
Nirvana speaker cable are shipped among so many places during their construction, a non-trivial portion of the $695 to $1095 you might pay for one of their cables goes toward postage. Stephen tells me it would be a major cost savings if he could get all the stages of his assembly pipeline done near where the company is based, but at the moment the closest facilities that meet his standards are spread all along the East Coast.
A skeptic would suggest these guys were playing psychological games with their customers to fool them into thinking they were getting a better cable. It's impossible to talk about $1000 cables in most crowds without somebody yelling "snake oil." I happen to believe that there's a place for cost-no-object craftsmanship in every field, as long as the prices are justified by the work and equipment involved. After looking into cable construction as a business, I can understand why some designs are priced so high. I find it hard to believe talented designers like Joe Skubinski and Stephen Creamer are chasing after phantom improvements when they devote their lives to improving the quality of the cables they manufacture, working on every tiny detail in an attempt to optimize the cable as a system. As to whether these efforts justify their cost, that's a subjective call every listener has to make. I'm happy to use $20 interconnects I made myself when that's appropriate, and equally happy to utilize $1095 speaker cables if that's what I like for a system.
Radio Shack speaker update
Just as last month's Fringe was being assembled, I learned that the Optimus PRO-X5 speakers I was recommending were being discontinued. This was good in one regard because it meant a number of people got a very nice little speaker for a very low price as part of the close-out (typically under $45/pair including sales tax). As I promised, I'm still planning on publishing some upgrades for the speakers, which will probably show up in January. I'm still sorting through the new product line, which hasn't finished appearing at my local Shack yet. There doesn't appear to be a direct replacement, but one of the other models may end up being a good value. I'll keep you posted.
Copyright © 1998 SoundStage!
All Rights Reserved