How can I use my power supply’s alarm signals?

Many power supply alarm signals, such as AC Fail, DC Good, etc., utilize optocouplers or optical isolators as a means of transferring alarm signals from the power supply to the end-users equipment without direct connections. The main purpose of an optocoupler is to prevent noise, ground loops, and/or high voltages from the power supply from damaging the end-equipment to which the signals connect. Below is a typical schematic diagram of an optocoupler that consists of an LED on the input side and a phototransistor on the output side. Signals from power supply activate the LED, which in turn activates the electrically isolated phototransistor.

Typical power supply alarm signals may include AC Fail, DC Good, Over-Temp, and Inverter OK.

Fig A: These opto-isolated, open-collector alarm signals share a common ground

Fig. B: This opto-isolated alarm signal has a separate output and ground

Pull-Up Resistors - When using open-collector alarm signals an external pull-up resistor is required. This pull-up resistor needs to be selected and connected between the alarm output (collector) and an external voltage source (+VCC). The purpose of the resistor is to limit the amount of current that flows through the open-collector transistor. For example, in some applications the current should not exceed 10mA, however, always check your power supply manual to confirm the maximum allowable current and maximum +VCC voltage.

External +VCC – The +VCC voltage that is connected to the pull-up resistor(s) for the alarm signal(s) should come from an external voltage source if maximum isolation is desired. However, in some cases power supplies come with an Auxiliary DC Output, which is always present as long as the AC input voltage is present. In some applications, this Auxiliary DC Output can be used as the +VCC for the alarm signals except for “AC Fail.”

Logic Ground – If an isolated logic ground is present, it usually needs to be tied to either the (-) Vout of the power supply or to the ground of the end-equipment system.

Alarm Signal Levels - In most cases, when an alarm condition is Not present the open-collector transistor output(s) will be On (or logic Low). If an alarm condition should occur, the open-collector output(s) will turn Off (or logic High). But, different supplies can have different alarm logic levels, so you should always check your power supply’s instruction manual to determine the supply’s alarm logic levels, with and without alarm conditions.

Combining Alarm Signals – The diagram below shows an example of how “DC-OK” signals from 3 different power supplies that are mounted in a power system rack can be combined (OR’d) to form a single signal In this example, the +VCC comes from the Aux DC Output of the power system is connected to 3 separate pull-up resistors. These pull-up resistors (10K ohm) connect to an open-collector “DC-OK” output from each of the 3 power supplies. And, all open-collector transistors have a common ground connection (similar to Fig A above).

Normally, if all supplies are OK, their DC-OK signals will be in the low state (approx. +0.6Vdc or lower). Should one of the power supply’s output’s fail, its “DC-OK” open-collector transistor will turn off and that output will go high via the pull-up resistor to the +12 to +15Vdc aux supply output. This positive “high” signal will forward bias the diode and cause the combined “DC-OK” alarm output to go high (relative to the Return or Ground line), which indicates that one of the 3 supplies have failed “DC Not OK”. The indicator light on the failed supply will show which supply has failed.

Alternatively, each of the “DC-OK” signals from the individual supplies in this power system could have been connected separately to a monitoring system (without combining them). The advantage of doing this is that the specific failed supply could be identified remotely without viewing the front panel mounted indicators.

Review Hot Swap/Rack Mount Front End power supplies from TDK-Lambda

How Does Altitude Affect AC-DC Power Supplies?

Most AC-DC power supplies that meet the safety standards per UL/EN 60950-1 for ITE (Information Technology Equipment) applications are designed to operate at typical office and factory altitudes, which can vary from slightly above sea level to as high as 2,000 meters (6,562 feet). And, many power supply manufacturers provide units that are designed and rated for operation at higher altitudes, up to 3,000 meters (9,843 feet) so their supplies can be used in major cities located at higher elevations (e.g., Denver, Santa Fe, Mexico City, Bogota). Many broadcasting/communications stations/towers are located at altitudes up to 3,000 meters or higher in order to maximize their range.

Altitude affects the design of power supplies since ‘air’ is used as an electric insulating medium (aka, dielectric) in the construction of power supplies, as well as most electronic devices. The density and dielectric strength (insulating property) of air is very good at sea level, but at higher altitudes, the thinner air loses some of its dielectric strength, which needs to be compensated for. Switchmode power supplies operate off of high voltages (inputs of 90 to 265Vac) and internally generate even higher voltages (400Vdc or more), which need to be insulated and contained to prevent high voltage arcing or breakdown within the supply, and to protect the end-equipment and operating personnel.

The drawing below shows a cross section of a typical printed circuit board (PCB), which is comprised of copper electric conduction paths that our chemically etched on an insulated (dielectric) fiber board material (e.g., FR4, woven fiberglass cloth with epoxy resin), plus electronic components that are not shown in this drawing. As can be seen, the fiber board and air, combined with the distances between the etched conductive traces are the primary insulation mediums for the circuit board.

Drawing Credits: Mammano B, ‘Safety Considerations in Power Supply Design, Underwriters Laboratory / TI

• The term ‘Clearance’ refers to shortage path between the two conductive parts (circuit traces, components, etc.), measured through air.
• The term ‘Creepage’ refers to the shortage path between two conductive parts measured along the surface of the insulation (PCB, insulating materials/barriers, etc.).

What does this have to do with altitude? Since ‘air’ gets thinner (reduced barometric pressure) at higher altitudes and becomes less of an insulator, the PCB and component layouts have to be designed with sufficient safety spacing distances to prevent high voltage arcs or breakdowns between conductors and/or electronic components.

For example, typical power supply design practice may allow 8 mm spacing distance between primary and secondary circuits and 4 mm spacing distance between primary and ground. These spacing distances will vary depending upon the voltage levels between conductors and components and the expected humidity, temperatures, pollution levels, and attitudes.

For those power products that must be approved per the Chinese CCC organization (required to export supplies into China), the new Chinese Safety Standard GB 4943.1-2011, which is similar to UL/EN 60950, requires strict specs for creepage and clearance distances. As of December 1, 2012, the primary-to-secondary clearances must increase by a factor of 1.48 to qualify the supply for operation up to 5,000 meters, since many regions in China are located at high altitudes. The alternative for CCC certified power supplies is that they must clearly marked with a warning label that states that the power supply must be used below 2,000 meters (see table below).

The base design altitude for ITE power supplies is 2,000 meters. However, as mentioned before, as the altitude increases, the air becomes a poorer insulator and the spacing distances have to be increased per the following table (assuming an 8 mm clearance at 2000m).

Altitude (meters)	Barometric Pressure (kPa)	Multiplication Factor for Clearance	Resulting Clearance (mm)
2000	80.0	1.00	8.00
3000	70.0	1,14	9.12
4000	62.0	1.29	10.32
5000	54.0	1.48	11.84

As can be seen from this table, if a power supply is to be operated at 5,000 meters, its conductor/components clearances must be increased by 48% compared to a supply designed for 2,000 meters.

The other major effect of high altitudes on power supplies is that the less dense air does not conduct heat as well. To compensate for higher altitudes, power supplies need to be derated, or employ larger heat sinks, or have increased forced air flow, or a combination of these to insure proper cooling. In addition, the power supply must be designed with the proper conductor and component clearances as discussed above.

In summary, whenever an application requires that a power supply must operate at altitudes above 2,000 meters (6,562 feet), always check with the manufacturer to determine if this is acceptable, or if an alternate model that is designed for higher altitudes is required.

Efficiency Calculations for Power Converters

A power converter’s efficiency (AC-DC or DC-DC) is determined by comparing its input power to its output power. More precisely, the efficiency of the converter is calculated by dividing the output power (Pout) by its input power (Pin). The Greek symbol Eta “η” is usually used to represent “Efficiency.” Here is the formula for determining a power converter’s Efficiency (η).

η = Pout / Pin

For example, the efficiency of a converter that provides 500W of output power (Pout) and requires 625W for the input power (Pin), would be 80% (500W/625W=0.80). In this case, the input power exceeds the output power by 125W or 20%, which is lost/wasted power. Therefore, 20% of the input power is converted to heat energy that must be removed from the converter by some means of cooling (conduction, convection, and/or radiation).

Since all power converters have inherent conversion losses, the output power is always less than the input power. Most often, the manufacturer of the power converter specifies its efficiency and maximum output power on the product’s datasheet. When the efficiency (η) and output power (Pout) is known, the end-user can determine how much input power (Pin) will be required and how much power will be wasted (Pwaste) and converted to heat energy under full load conditions.



Here are the formulas to determine Pwaste and Pin with sample calculations using the examples listed above.

Pwaste = (Pout/η) – Pout
Pwaste = (500W/0.80) – 500W = 625W - 500W = 125W

Pin = Pout + Pwaste
Pin = 500W + 125W = 625W

Obviously, with a higher efficiency converter, Pwaste is reduced. Using the example above, but with an improved efficiency of 90% (instead of 80%), here are the revised calculations:

Pwaste = (Pout/η) – Pout
Pwaste = (500W/0.90) – 500W = 555.5W - 500W = 55.5W

Per the examples above, by employing a more efficient power converter it reduces Pwaste from 125W to 55.5W, which provides a substantial savings to the user in both electric energy and cooling costs.

Here are alternate formulas for calculating the factors associated with power converter efficiencies:

Pin = Pout/η
Pwaste = Pin – Pout
Pwaste = Pout (1/η - 1)

In some formulas, Pwaste is referred to as Pd, where “Pd” means the power dissipated (in the form of heat) within the power converter. Pwaste = Pd.

When dealing with AC-DC power supplies, not only is “Efficiency” important, but so is the power supply’s “Power Factor.” Information about the effect and importance of the power factor in power supplies is covered in the following article.

Power Factor Correction

http://power-topics.blogspot.com/search/label/Power%20Factor%20Correction

More information about power converter efficiencies and cooling methods/techniques can be found at these web links:

Power Converter Efficiencies

http://power-topics.blogspot.com/2011/06/power-supply-losses-and-impact-of.html
http://us.tdk-lambda.com/lp/ftp/other/cost-savings-high-efficiency.pdf
http://us.tdk-lambda.com/lp/news/2012_release05.htm

Cooling Methods

http://power-topics.blogspot.com/2009/01/what-size-fan-do-i-need.html
http://us.tdk-lambda.com/lp/ftp/Other/cooling_bricks_ecn.pdf

What is a Limited Power Source?

There are many technical details regarding Limited Power Sources (often referred to as an LPS) covered in the IEC60950-1 safety standard, involving a variety of applications, but I will cover just the basic aspects regarding AC-DC power supplies in this blog article.


Several safety standards refer to IEC60950-1 for a wide range of applications, and one of those referrals involves the use of Limited Power Sources.

What is, and why are Limited Power Sources important? Simply put, if a piece of electrical or electronic equipment supplying DC power to external devices is to be installed by a third party, such as an electrician, the risk of wiring fires & electrical shock needs to be minimized. That electrician will not be expected to know all the potential fault scenarios and use the appropriate cable thicknesses and insulation to cover those hazards. By using a Limited Power Source, the system wiring can also be reduced, saving cost.

If a Limited Power Source is used, then the electrician’s job is simplified, even if there is a (single) fault inside of the power supply.

Some examples of testing that the power supply manufacturer will do, to determine the maximum output current and power meet the “Limited Power Source” requirements are:

1. Check the power supply at maximum rated load
2. Increase the load of the power supply until it is on the verge of overload
3. Simulate an internal fault by shorting the current-limiting resistor
4. Short out the opto-coupler that provides feedback to the control loop

The conditions for a “Limited Power Source” AC-DC power supply are:

1. For a power supply rated at 30V or less, the following must be met even with a single fault condition:
    a. The output current must not exceed 8A
    b. The output power must not exceed 100W
2. For a power supply rated above 30V, but not exceeding 60V:
    a. The output current must not exceed 150 ÷ Vout
    b. The output power must not exceed 100W

This graph shows the limits for Limited Power Sources:

Taking some examples from TDK-Lambda’s DSP series of low profile DIN rail power supplies:

 

The DSP60-5 is rated at 5V 7A – That would meet (1b), but when put into overload it would certainly not meet the (1a) limit of 8A. This model is not listed as being a Limited Power Source.

The DSP60-12 is rated at 12V 4.5A – Overload current would be around 7A which meets (1a) and the maximum power is ~ 84W (12V x 7A) which also meets (1b). This model is listed as being a Limited Power Source.

The DSP100-24 is rated at 24V 4.2A – Right out of gate this unit exceeds 100W, and so it is not listed as a Limited Power Source.

The DSP100-24/C2 is rated at 24V 3.8A – This model actually has a special current limit & over power circuit which strictly limits the output current and power under a fault condition, so it meets the <8A requirement of (b) and the limit of <100W. Interestingly though, a single fault on the control circuit made the output rise to 30.8V and so it now falls under the 30-60V limit of 150 ÷ Vout = 4.8A maximum current (2b), which it passed. Therefore, this model is listed as a Limited Power Source.

 

Test results performed by the safety bodies, such as UL, CSA or TUV, are normally found in the CB report for the power supply.

 

Minimum loads and cross-regulation on multiple output power supplies

One subject that our Technical Support team frequently gets asked about is minimum loads on multiple output power supplies, so I thought this would be a good subject to write about.


On a low cost, low power, multiple output power supply, the datasheet will often state that to maintain regulation, a minimum load has to be applied to one or more of the outputs.

To explain why, here is a block diagram of one such simple triple output power supply.

On the middle right hand side of the diagram are the three output windings from the transformer.

On Channel 1 (+5V), the output from the transformer is rectified and filtered to provide a smooth, DC output. If that output voltage is not at the set voltage, say due to a load change, the power supply will automatically correct itself. It does this by sensing the output voltage, comparing it to an internal reference, and feeding back a signal to the control circuit via the opto-coupler. The control circuit will then adjust the pulse width of the converter accordingly. The regulation on this output is typically 1 to 2%.

On Channels 2 & 3 (+V and –V) though, it can be seen that there is no feed back to the control circuit. These outputs are referred to as “semi-regulated”. If Channel 2’s load were to increase for example, the output would drop slightly, but there would not be any automatic correction. That voltage drop is specified by the load regulation specification, typically 3 to 5%.

With respect to minimum loading, if there is little or no load on Channel 1, the output will still be at the set voltage, but the switching converter pulse width will be very narrow. The output voltage on Channels 2 & 3 drops fairly dramatically at those narrow pulse widths, particularly if the outputs are supplying their full rated load. An output voltage of 12V may drop to 8V.

Conversely, if the full rated load is applied to Channel 1, but Channels 2 & 3 are not loaded, the voltages on 2 & 3 will rise, and a 12V output could deliver over 14V.

The effect that varying loads on Channel 1 has on the “semi-regulated,” Channels 2 & 3, is many times referred to as the “cross regulation” specification.

Manufacturers of power supplies specify a minimum load requirement on Channel 1, usually 10%, to warn the user. Minimum loads may also be specified on Channels 2 & 3 to promote a better regulation specification.

Operating without a minimum load will not normally cause a power supply to fail, but can stress the user’s equipment.

Some products like TDK-Lambda’s MTW series employ two converters to improve power supply regulation, one to supply Channel 1 & one to supply Channels 2 & 3. Note that both V2 and V3 are sensed by the control circuit.

Fig. 2: Two converters (click to enlarge image)

Here is a link for more details about the MTW series: http://www.us.tdk-lambda.com/lp/products/mtw-series.htm

Other products like the NV175 series employ post regulators on each output, completely eliminating the minimum load requirement. Although this does add cost to the power supply, it removes any concern for the user and helps with system flexibility. Here is a link to the datasheet for NV175 series: http://www.us.tdk-lambda.com/lp/ftp/Specs/nv175.pdf

Wide range adjustable power supplies

SWS1000L Series

We had a customer that wanted a 1000-watt power supply to drive a medical centrifuge.  The only catch was that the speed of the DC motor that drives the centrifuge had to be varied and they were going to do this by using a step down dc-dc converter to change the voltage applied to the motor from 12V to 60V.

We suggested TDK-Lambda’s SWS1000L-60 power supply, which has, as standard, a Programming Voltage (PV) Input.  By applying from 1 to 5.5 volts to this input, the output of the power supply can be varied from 12 to 66V.

The connections are very simple, using the PV input and common terminals on the front panel of the power supply (see diagram).

Alternatively, the Programming Voltage could have been derived from the SWS1000L’s 12V auxiliary output, which in this case could be connected to an external potentiometer to provide the variable dc input.

This simple solution saved the customer both design time and money.  More information about the SWS1000L power supplies can be found at this web link http://www.us.tdk-lambda.com/lp/ftp/specs/sws600_1000l.pdf

The PV function is also available as on option on TDK-Lambda’s HWS series 300-600W models and is standard on the 1000-1500W models. More info can be found here http://www.us.tdk-lambda.com/lp/ftp/specs/hws1500.pdf

Advantages of Conduction-Cooled Power Supplies

Most mid- to high-power supplies use fans to help dissipate the internal heat that is generated as a result of imperfect AC to DC conversion efficiencies.  Since fans are electromechanical devices, they reduce the system’s MTBF and add to the required maintenance expenses.

Attached is a photo of a power supply that operated for many years at a postal depot where mail is handled and sorted automatically.  As can be seen (after the fan was removed) paper fragments and airborne dust contaminants were pulled into the supply by the fan and eventually caused a blown fuse. 

As might be expected, the proper maintenance program for any fan-cooled power supply calls for the periodic inspections of the supply, with the fan removed, and the replacement of the fan with a new one.

A new breed of conduction-cooled power supplies has been developed that do not depend on fans for cooling.  Instead, the required cooling is accomplished by conducting the internal heat loads to an external metal structure or enclosure, which act as a large heat sink surface.


The second attached photo shows TDK-Lambda’s new CPFE1000F series, which are conduction-cooled, 1,000 watt AC-DC power supplies.  (A 500 watt version is also available.)  All heat is conducted to the supply’s aluminum plate, which is designed to easily mount to a metal enclosure or cold plate for cooling.  More details and specifications for these power supplies are at this web link: http://www.us.tdk-lambda.com/lp/products/cpfe-series.htm

In some applications, these conduction-cooled devices are mounted to liquid cooled cold plates that are made of metal with internal serpentine channels through which a liquid circulates while removing the unwanted heat.  The net result is that the system operates with a substantial reduction in audible noise, reduced maintenance costs (no dust build-up and fan wear-out), and an enhanced MTBF.

Recently, I visited a Television Broadcasting Station that consumes about 100 kilowatts of power.  At this location, in separate areas, was a traditional fan-cooled system as well as the latest generation system, which uses conduction-cooled power supplies and RF amplifiers that are cooled via liquid flow cold plates. During the operation of the traditional system with fan cooling, the audible noise was so loud that personnel within 100 feet of the system had to wear hearing protection devices.  By comparison, in the other area where the new system with liquid cooling was operating, the noise level was so low (similar to an office environment) that no hearing protection was required.