A quick jump to a power conditioning device may not solve the problem and actually may create more of a problem.
Let's begin the examination of power conditioning by considering voltage regulating equipment and equipment called "line conditioners." This latter term is generally accepted to mean the combination of a voltage regulator with a noise rejection system, such as a shielded isolation transformer, for the protection against common mode noise and the enhancement of voltage stabilization.
The products in this category (voltage regulators and line conditioners) were among the first recommended by manufacturers of sensitive equipment to assist in protecting data processing systems from outside power influences. In the early days of the applications, little consideration was given to load/source interactions. In some cases, the resulting "conditioning" was worse on the output than the power source waveshape at the conditioner's input. In time, it became apparent to manufacturers and end users alike that just another "box" in the circuit might not be necessary, even though it appeared to give additional protection.
One example of this over zealous approach to voltage regulation came by way of a new regulator said to have a range of correction of + 10% to -40%. When we asked the salesman if that lower limit would extend the low-voltage limit of the power supply (designed already for -12%), he said it would make the low end -52%! We tested the theory only to find that the protective device for the circuit would trip each time we lowered the input voltage to the product more than 25%. This is what we would expect: On a constant load, the current will rise in proportion to the drop in voltage. At the point of 125%, the proper protective device will trip.
Consider that almost all of the sensitive power supplies used in digital electronic systems are designed with the finest voltage regulator available, one having a range of operation of almost 20% from the top of the voltage limit to the bottom. Also, it responds extremely fast. The CBEMA (Computer Business Equipment Manufacturers Association) curve indicates a + 6% to - 13% range for many devices. If these units have such a fine regulator, one that can handle any variation within the listed range, then why do we need to promote an additional voltage regulation product in line with this unit?
This question seems to beg an answer that we may have "oversold" the voltage protection concept without realizing what we have done. Granted, these voltage protection products are the least expensive way of providing some form of "buffering" ahead of the digital system; this initially was very important to manufacturers wishing to avoid problems with the "outside power influences." As the power quality industry becomes more alert to the specific needs in each of the categories, we may be able to make better application recommendations.
Our first, and oldest, form of regulation is a style of transformer designed to supply a stable output voltage through all current loading within the rating of the product. This is a ferroresonant technology known as a Constant Voltage Transformer (CVT) or a Saturable Magnetic Device. The general outline of this concept is shown in Fig. 1. This transformer product is designed with a core that is intended to saturate; in other words, it's designed to actually accept energy in the core steel, flooding the iron core in such a way that the result is a flat voltage - stable across the range of load currents. Within the ratings of the device, the voltage will remain level from light load to heavy load. This is a simple and very good older technology that's still in use today. And it's very economical.
A note of caution here: The design of this product is based on sine wave load characteristics dating to a time when harmonic interaction was not yet upon us. As such, you should not try to apply this technique into the face of a high harmonic load spectrum without considerable oversizing or design modifications by the manufacturer to assure harmonic content compensation.
We also need to remember that there is an upper limit to the load range of this device; if we exceed that limit, this product may actually reduce its output slowly to zero in order to self-protect the circuit. This performance trait, known as current limiting effect, can be avoided by sizing your maximum load approximately 20% below the full load rating of your regulator. This type of oversizing also will help in the handling of harmonics, if they're present and haven't been compensated for by the manufacturer.
Finally, remember that because a CVT uses energy to "saturate" its core, its efficiency will be very low if your loading is very light. Normal efficiencies for fully loaded units will be in the high 80% area, and up to 92%, while efficiency under light loads might drop to 45% to 55%. By sizing your load somewhere in the middle, you'll be able to operate under most conditions, giving up some efficiency points to an average of 75% to 80%.
Tap switching regulators
A more modern design of a transformer product for voltage stabilizing is a tap switching regulator, as shown in Fig. 2. This device uses the same concept, but in a fully adjustable configuration with diodes and transistors, to perform automatic "stepping" between the taps by sensing the output of the transformer.
In the first production, tap switching was done on the secondary; however, this method produced electrical noise. As a result, the design was changed so that the switching is done on the primary; this design remains today and uses the nominal six tap settings of the unit.
The design is based on a regular magnetic core rather than a saturable one, and thus has all of the properties of a good isolation transformer. It has excellent overload capability, like a regular dry-type transformer; it has the same low internal impedance, which is good for interfacing between a high impedance source and a high impedance load; and it does not introduce harmonic distortion to the system.
A tap switching regulator uses sensing of the output voltage to signal the primary tap change to adjust for the variances in input supply. It can make those adjustments at the current zero crossing in order to avoid electrical noise, and it has a response time of three to five cycles. While this speed of response seems slow by comparison with the speeds of computers, it's just this difference in speed that commends this technology. If this product is used in a circuit where there is a "fast" regulator inside the CPU product, then the slower regulator will not interact with the higher speed product; the two will "handshake" well together.
Efficiencies are normally higher and remain high from zero to full load, as one would expect of a standard dry-type transformer. The efficiency of the saturable product, as per our previous discussion, falls off rapidly at partial load ratings and requires an assessment of loading for proper sizing.
In both cases outlined above, we can add an electrostatic shield to make the regulator into a combination product. This is one of the best application moves that can be made when a voltage regulator is required in the circuit. With the shield, as seen in Fig. 2, the product now becomes a common mode noise rejecter and a voltage stabilizer.
A form of voltage regulator, combining stabilization with the supply of harmonic current, is known as an active power line conditioner. This technology, which was developed through the Electric Power Research Institute (EPRI), has the promise of a further "combination" solution for the harmonic-requiring load.
Watch for the August issue installment, where we move onto the next level of power conditioning: The resurgence of motor-alternators to provide not only voltage stabilizing, but also full mechanical isolation from input to output.
RELATED ARTICLE: WHAT HAVE WE LEARNED SO FAR?
For a quick review, remember that we break down power quality problems into the following four general categories.
* Ground loops, which are associated with wiring and grounding methods. These unintentional loops lead to the circulation of common mode noise, which disrupts or damages a signal system. A sign of this trouble would be when our power quality problem has no correlation with any power system events.
* Transient phenomena, which are associated with lightning strikes, power switching, coupled bursts of radio frequency (RF) noise, etc., causing high-speed, high-frequency disturbances.
* Harmonic interactions, which are associated with the flow of low order power frequencies, the odd multiples of 60 Hz, and normally "requested" by the load device as a spectrum of currents to satisfy the operating characteristic of the load power supply.
* Power conditioning problems, which are associated with the energy supply. These problems are characterized by voltage instability, limited capacity, poor frequency, source disruption, poor power factor, etc.
Notice that power conditioning is mentioned last. While our goal in this series of articles is to expand upon the various "conditioning" tools and techniques, we must remember that the choice of "conditioning" must be able to pass the test of the other three categories listed above. For instance, we normally won't take care of a ground loop problem with a power conditioner, since the tool in question does not address wiring and grounding concerns. Likewise, we may be able to ask the chosen power product to protect against transients, but we normally won't have a product that will accomplish this. In fact, quite the opposite may be true; the power technology itself may need transient protection! In a similar manner, we must remember that our product choice in the conditioning area must be able to handle any harmonic currents in the system and must not be a contributor of more of those currents to disturb the power source further.