For the PDF version of this article, click here.
As we continue our “Then and Now” series, we once again will look at three products selected from a 1975 edition of Solid-State Power Conversion, the predecessor to Power Electronics Technology. As it turns out, there was no September or October issue that first year, so we have selected products described in the August 1975 issue. For this article, we have chosen transient suppressors, tantalum capacitors, and dc-dc converter transformers. We'll use these three components to provide a glimpse of the start-of-the-art back then, adding some perspective as to where each technology is now.
Bi-Polarity Silicon Transient Suppressor
Useful for absorbing transient energy which may otherwise cause damaging voltage spikes, nine new transient suppressors provide symmetrical protection against positive- or negative-going voltage transients. Specifications include: Breakdown Voltage (Min.) 9 to 99 V; (Max.) 11 to 121 V @IZ 375-30 mA. Peak Pulse Power (1 ms): 1500 Watts @ 25°C. BV Response Time (tBV); 0 V to BV min., less than 1 × 10-12 seconds. Dynamic Impedance: 0.7 to 35 ohms (Stock)
— Semtech Corp., Newbury Park, CA
Solid-State Power Conversion, August 1975, p. 54
“They are designed to nip everything above a certain voltage — in just nanoseconds.” That's how Ray DiBugnara, director of technology at Microsemi (Irvine, Calif.), sums up the role of the transient voltage suppressor.
In fact, suppressors are onboard virtually every jet aircraft in the air today, such as in the engine controllers. And lightning doesn't actually have to hit an aircraft, it could strike a half mile away, but if vulnerable circuits are not protected, the induced EMF could cause the engine controller to be burned out.
A typical lightning strike may last 10 µs or 20 µs, although there may be a “charge feeder” in a cloud that continues to feed more charge into the ionization path. These are the “strikers” where you see flickerings. Actually, those flickerings are still a 5-µs to 10-µs series of pulses.
The advantage of a silicon transient suppressor is its ability to respond very rapidly, in just a few nanoseconds for a bipolar device and a quarter of a nanosecond for a unipolar device. Actually, these are conventional, diffused-junction reference devices. Most supressors comprise two diodes connected back-to-back, because the line may have to be protected from both positive and negative transient excursions. They are similar to zener diodes; however, they are tested in ways that focus heavily on how well they will behave as transient supressors.
As DiBugnara sees it, over the last 30 years, requirements have changed little because so many applications for silicon suppressors were identified long ago. What's more, there are few alternatives. There are, of course, metal oxide varistors (MOVs), nonsilicon devices that are used mostly in inexpensive electronic products. The response of an MOV is quite fast, and its power density is much higher than silicon — dollar for dollar, but they do not clamp.
For example, if you want to clamp a 115-Vac line at 200 V, it is hard to get an MOV to limit at that particular voltage. So quite often, a lot of the transient signal will get past it.
“Some designers team both a silicon voltage suppressor and an MOV, using the silicon suppressor to clip the early part of the waveform, and then rely on the MOV to step in and take over, absorbing the bulk of the power during the extended part of the waveform,” says DiBugnara.
MIL-Style Solid Tantalums
The T252 Series of established reliability, extended capacitance range, solid tantalum capacitors is the MIL approved equivalent to the new MIL-C-39003 style CSR33 including low impedance life and surge current tests. Generally costing less than hermetic seal, wet tantalums, these units are claimed to excel in temperature stability, low DC leakage, shelf life and reverse voltage tolerance. All units are surge current tested for low-impedance applications, and are available in four standard military tubular case sizes with a CV range of 1.2 to 1000 µF, 6 to 50 VDC for operation from -55°C to +125°C.
— Union Carbide Corp., Greenville, SC
Solid-State Power Conversion, August 1975, p. 52
For years, product designers have been urging components manufacturers to make an effort to help the designers make power supply packages smaller and more efficient. And one of the solutions has been to raise the operating frequency of switched power supplies. Early designs were in the 10-kHz to 20-kHz region, but that soon moved up to 75 kHz, 100 kHz and then 250 kHz to 300 kHz. In fact, today you have phased switchers running above 1 MHz. But this migration to higher frequencies has meant raising the operating frequency that components encountered, and this is why the rising equivalent series resistance (ESR) of the filter capacitors became more of a problem.
John D. Prymak, applications manager at Kemet (Greenville, S.C.), recalls that his company was approached by several manufacturers of handheld products. They wanted both ac capability and battery charging so that a low ESR became a must. Could Kemet make capacitors that would be more efficient in the higher frequency range?
“In our tantalum, we took the electrolyte and replaced it with a solid material that would survive the solder reflow processing,” Prymack says.
So with these tantalums, there was no need to be concerned about temperatures or the electrolyte boiling or leaking out of the case, or causing swelling of the case over time. The solid material they chose is manganese dioxide, a material that scientists at Bell Laboratories studied when they were examining candidates for transistor technology from 1952 to 1955.
The tantalum capacitor has the highest volumetric efficiency — capacitance per unit volume — of all capacitor types. What's more, tantalum electrolytics can be supplied at high-capacitance values just below the range engineers used to specify for aluminum electrolytics, up to 1500 µF.
DC-DC Converter Transformer
A new switching transformer for use in DC-DC converter power supplies is designed to operate from a 5 VDC source and produce an output of ±15 VDC @ 175 mA. Designed for operation at a 20-kHz switching frequency, the unit is epoxy molded with plug-in printed circuit construction to meet MIL-T-27, Grade 5, Class S. This new unit augments an existing line of highly-useful converter transformers and inductors, which are shipped complete with a typical schematic and suggested component list. ($11.30/100 pcs; stock) — Microtran Co. Inc., Valley Stream, NY
Solid-State Power Conversion, August 1975, p. 53
DC-DC Converter Transformers
One might think there would be little about a dc-dc converter transformer that would change over the years. But that simply isn't so. One of the motivators driving change has been the rising crescendo of battery-operated designs and the very low system voltages that today are often the starting points, according to John DeCramer, vice president of engineering, at BH Electronics Inc. (Burnsville, Minn.).
Whereas 15 years ago a dc-dc converter was probably powered by a 12-V source, today that has dropped to 9 V, 3.3 V and even as low as 1.6 V. Minimizing losses has become much more critical. Because you are starting out at the lower end, you are driving more current and less voltage, so the resistance of the windings on the magnetics components has become very much an issue.
A few years back, 5 A to 10 A was a lot of current going through an inductor or transformer. Now it is not uncommon that someone wants 25 A to 35 A entering or leaving the magnetics. Lower voltages and higher currents force magnetics to employ larger wire gages and also square wire, which enables packing more copper into a given cross-section.
As for the core materials, they have gone through a number changes over the years as well. The ferrites and permalloys provide better performance than powdered iron, yet fall between the older powdered iron and permalloy materials with high-end, high-nickel content that exhibit the high-magnetic permeabilities. The latter has better loss characteristics than the powdered-iron cores, particularly at higher frequencies.
The permalloy, because of its nickel and metal content, is more costly but you get quite a bit more out of it than some of the powdered irons. And it is because of the lower loss that you can drive a transformer constructed with such cores much harder. The efficiency of these cores is also important. If you build a switching power supply, the losses in the transformer are going to be on the order of 10%, due to core losses, says DeCramer.