A couple months back, I was reading one of the alumni bulletins from Northfield Mount Hermon School, which I attended 35 years ago. I was poring through a whole bunch of not-too-interesting social stories about people whom I did not know when suddenly I ran across a technical story written by Mike Hannon, another NMHS alumnus who lives in Lompoc, Calif. He was explaining how some new incandescent lamps manufactured by GE, Philips, Sylvania, and others are now able to provide high efficiency as well as improved life. And how do they achieve these advantages?
Well, first of all, they are halogen lamps. Instead of the filament boiling off, it boils off and then back on, so the filament has a good chance of lasting several thousand hours. This is good technology, but nothing new. The real kicker is that the filament is enclosed in a new dichroic filter, which is selective to energy at long (infrared) vs. short (visible) wavelengths. Every lamp's filament generates some energy in the visible wavelengths, and a lot more heat (radiation in the infrared region). If all this energy is allowed to radiate, the bulb puts out a lot more heat than light. You have to dissipate 100 W, for example, to keep your filament hot while you get a mere 6 or 8 W of visible light. However, if we had a specially engineered coating to let the visible light out and reflect the heat radiation back in, we might be able to keep that filament hot with only 70 or 60 or 50 W, yet still keep the same light output. That's not necessarily more efficient than a fluorescent lamp, but if you want a bright, intense lamp rather than a wan, weak light source, then there may be places where these new halogen-plus-IR lamps have real advantages. GE deserves a lot of credit for engineering these improved "dichroic" lamps and bringing them to the market-place.
AND, Mike Hannon deserves a lot of credit for trying to educate people on a real opportunity to save money and energy. I mean, I've been trying to keep an eye out for these bulbs ever since I first heard of the concept about 6 years ago, but I had not found them until Mike pointed them out to us.
Mike didn't just recommend buying these lamps; he did a lot of evaluation on the best ways to apply them. He studied the tradeoffs of lamp life vs. brightness. Of course, most people are aware that if you want the highest efficiency, you run the lamp at a fairly high voltage. But this degrades the life of the bulb according to a 12th-power law. If you take a 115-V bulb and run it at 125 V, your brightness is way up, and the efficiency (lumens per watt) rises significantly, too, but the bulb life is way down. A couple of times, I have lived in a house where the line voltage was up near 124.5 V. It was not uncommon to lose two light bulbs in one day. I kept a calibrated ac voltmeter and I would check the line voltage quite often. The line got up ever so close to 125 V-but it never exceeded it. If the voltage got as high as 125.1 V, I would have complained to the Light & Power people and forced them to come out and adjust the taps on their step-down transformer.
Conversely, if you want to get really long life from incandescent bulbs, you run them at relatively low voltage. The easiest way to do that is not by fooling around with Triacs or transformers, but rather to put a single silicon rectifier in series with the bulb. This cuts the power dissipation by about 65% and the light output by about 82% (19 dB), and bulb life is extended enormously. This is what you may want to do if lights are located in really rotten, inaccessible locations.
The disadvantage of poor efficiency will be well offset by the savings from the reduced frequency of bulb replacement. Some of the new halogen bulbs are also available with 130-V ratings, which is another good way of matching your requirements to the available power.
Mike did many experiments in this area. He found that some of the new bulbs had problems with singing, vibrating filaments, an often annoying problem that can cause early bulb failure. He found that putting a full-wave bridge in the lamp wiring can feed dc with a lot less ac component to the filament, banishing the vibrating filament problem.
However, with one particular lamp, the Sylvania CAPSYLITE, the lamp brightness seemed to go way up with the bridge. He took measurements and confirmed that the power drain was greatly increased. What's going on here? He figured out that in this lamp, a single rectifier was already built-in. In normal operation, the lamp ran on half-wave rectified power. But when you inserted the full-wave bridge, the lamp ran nearly twice as much average current-extremely bright and extremely hot.
Mike also observed that some of the cheap "miracle button" rectifiers aren't very reliable, and can fail in a mode with bad local overheating. So if you buy some of those inexpensive rectifier disks that get stuffed down in the socket, keep an eye on them for the first few hours....
Mike also did lots of research on the startup characteristics of these lamps. Because the filaments run SO hot, the dc resistance when cold is even lower than on ordinary incandescent bulbs-as low as 1/12 of the hot value. So the in-rush current can be quite large, say, 6 A on a 50-W bulb. This can slightly shorten the life of the bulb.
Again, the solution is fairly simple. You can buy In-rush Current Limiters (ICLs) from Keystone Carbon Corp., if you want to buy OEM quantities of them. Orif just a dozen are needed, you can order them from Digikey (call 1-800-344-4539) for about $1.39 each (part number KCL013L-ND). These thermistors come with ordinary leads, rather like a fat ceramic disk capacitor, about 0.42 in. in diameter. To install them, you have to solder them carefully into your lamp's wiring. This particular model, which I find suitable for 50-to 100-W bulbs, provides an in-rush current of barely 2 A, so the lamp takes many milliseconds to warm up gradually. I am buying a dozen of these, because they will improve the life of every 100-W bulb in my house.
When we add this ICL to the new halogen lamps, that should improve our lamp life out from perhaps 2000 hours toward 3000 or 4000 hours, in normal service. That's a rather good thing because right now these lamps cost somewhere from $7 to $11 at retail. If we can make them last 6 times as long as an ordinary 100-W bulb, the cost of the halogen bulb (per hour) is still relatively high. Then why bother? Answer: for the improved energy efficiency. If I can run a 50-W halogen bulb at 42 W and get slightly more light than a 100-W bulb, then over a 3000-hour lifetime it can save 3000 hours X 58W, which costs (at 8 cents per killowatt hour) about $13.
So there's where you will break even, and even start to save money. Eventually, when manufacturing costs fall, these halogen/dichroic lamps will be even more cost-effective.
Of course, nothing competes in a vacuum (pardon the expression). The modern fluorescents have already come down in price. The recently announced Philips magnetic-induction bulbs will be coming along the learning curve, too. Five years from now, will there be a clear-cut winner for every application? Maybe, maybe not. But it will be fun to see how the competition works out!
So, you should run right out and buy a dozen of these GE Halogen-IR 60-W 60PAR38/FL/HIR bulbs (GE's part number; product order code is 18626), or the slightly more compact 50-W 50PAR30/FL/HIR, or a Sylvania CAPSYLITE, at your corner hardware store, right? Heck, no. First of all, if there are two in stock, you'll be lucky. If there are none, you may be able to wave this column at the manager and convince him to special order some for you, and to stock them. But, will you need a whole large number of them?
If you leave a light on many hours per day, that becomes a good candidate for installing a (relatively) expensive high-efficiency lamp. If you leave it on all the time, your break-even time on the investment can be as short as six months. But if there's a lamp you use just a few hours a week, you will not get your investment paid back for several years-no point in fooling around with that. So, until the price drops, you might want to buy just one or two to check out the situation. If you find the light from the fluorescents adequate, these lamps are already well developed along the learning curve.
On the other hand, do you dislike the idea of electronic ballasts humming in your ear as you sit reading? Then the newer halogens may find favor. I have a couple of each. I'm trying to evaluate where my preferences lie.
I must point out that right now the best halogen/dichroics are in the form of a floodlamp, with a 38-degree beam. I'm still trying to get some lenses to spread out the beam to be more like an ordinary lamp's spread. The floodlight is not, by itself, exactly the light pattern I want.
If, as you shop, you see some that are semi-spot or spotlights (Model /SP/), don't buy one unless you're prepared to use an 18-degree cone of light. One engineer says you can spread out the light to a wider pattern by sandpapering the face of the bulb. This might give a nice diffuse source, but I would not sandpaper any glass that sees that much stress and heating. However, placing a separate piece of frosted or sandpapered glass in front of the bulb should do the trick. I'll try that soon.
An additional caution: The bulb's shipping box has a rather weak warning, which states that if the outer glass bulb breaks or cracks at all, you should throw the bulb away, even if it appears to work okay. Why? Two reasons: the inner halogen capsule emits lots of ultraviolet rays, and you could get a bad sunburn, or go blind if the light shines right on you. Also, the failure mode of the halogen capsule can sometimes be a blow-out, and you need the outer glass for protection.
Additional comments-Mike Hannon points out that using a dimmer is one of the nicest things you can do with one of these lamps. As you turn it on, the lamp intensity comes up gradually, which is ideal for its life. You won't need any ICL. These halogens do like to run on a dimmer, which fluorescents refuse to do. However, if you turn down the dimmer too far, and the capsule gets too cool, lower than 250ºC, the filament material will refuse to reevaporate back onto the filament. A dark film will then appear on the inside of the capsule, and you'll have to turn the voltage back up to get the bulb back to full efficiency again (that's a reversible mode of degradation of efficiency). This might even happen if you use a half-wave rectifier-the capsule might not run hot enough for proper halogen action. Further, running a filament on dc can help accelerate a failure mode called "notching." This causes the boiling off of the filament to be localized rather than uniform, so the actual increase of filament life at low line voltages isn't as great as predicted theoretically. So when the magic rectifier peddlers claim, "extends bulb life 100 times," they're just blowing smoke, as you suspected. They're extrapolating into regions where they have never taken any data....
On the other hand, if you're doing industrial lighting, you may be lighting stairwells or hallways, 24 hours per day, and if the lights are virtually never turned off, an ICL would have no advantage (however, the extra couple volts dropped in the ICL will cause the halogen bulb to run a little cooler, for a little longer life).
I went researching at a Radio Shack. They don't sell thermistor-type ICLs, and they don't sell little rectifiers assembled in the magic-button format. What they DO sell is a little Triac assembly, (Part #61-2726, two for $4.25). It claims to save 10% of power for an "undetectable" decrease of brightness. But it does start up the SCR with a small (55%) duty cycle, so it effectively acts as a current limiter. After it warms up, the Triac fires at a high duty cycle, so the lamp runs at substantially full voltage. So, there's one more game you can play to improve longevity.
I guess it would be fair for me to say that at this state of application, these new high-efficiency halogen lamps are, like 'most every new product, still in need of some planning, engineering, and experimentation. But in the long run, there will surely be other formats, other shapes, other products, that will be very convenient and appealing and efficient. Savings electricity by buying more efficient, lamps is one of the easiest and most effective ways we can save energy and money, and the opportunities to do this are expanding rapidly. I'm getting my feet wet now (of course, when changing bulbs, you do not literally want to have your feet wet...), starting to find out which places in my house can benefit from more light with less electricity. I'll be darned if I buy two 22-W fluorescents just to throw a decent amount of light on my workbench. And people who already use a lot of flood lamps for institutional lighting can, right now, start taking full advantage of these products to save on energy and on the cost of lamp replacement. If these are used in an air-conditioned location, every watt of heat saved means your air conditioner will have an easier job, too! And, unlike the magic-button rectifiers that improve the bulb life while ruining the efficiency, these new halogen lamps provide long life plus dazzling improvements in efficiency.
All for now./Comments invited! RAP/Robert A. Pease/Engineer
Address: Mail Stop C2500A, National Semiconductor, P.O. Box 58090 Santa Clara, CA 95052-8090