Thursday, September 1, 2016

Comparing Power Supply MTBF Numbers


One subject that confuses specifiers of power supplies is comparing MTBF numbers from different manufacturers, who often use different standards for calculating the number of hours.  There are many well written articles going in to great detail available on the internet on the calculation of MTBF, but this blog article will attempt to simplify things.

MTBF (Mean Time Between Failures) is the mean time between successive failures, and only really applies to a part that will be repaired and returned to service.  So if the up-time of the power supply was a year in each case below, then the MTBF would be ½ x (1 year + 1 year).

A low cost power supply will probably not be repaired and if it is under warranty, it will normally be replaced.  In this case, the numbers to look for would be MTTF (Mean Time To Failure), but that figure is not widely stated.  Usually life testing of a large number (to cut the test time down) of power supplies is used to calculate that. 

The MTBF number is often thought to be the minimum (guaranteed) time before a failure; that is certainly not the case!  Reliability “R” is based on the probability that a piece of equipment, in our case a power supply, will operate satisfactorily for a given time period “t” (based on specified conditions – for example ambient temperature and output load).   

For random failures, the probability that a power supply will survive to its calculated MTBF is just 37%, no matter what the MTBF number is:

R(t) = e –t /  MTBF = e-1  = 0.368 (when t = the MTBF number)

To complicate things further, a variety of methods are used to calculate MTBF.

Prediction Standard
Applications
Disadvantages
MIL-HDBK-217F
Provides failure rate data and stress models for parts count and parts stress predictions. It also provides models for many environments ranging from ground benign to projectile launch
Hasn’t been updated since 1995, gives higher failure rates of commercial parts than is seen in actual product life
Telcordia SR332
Gives three prediction methods based on parts count, lab testing and field life
Narrow ambient temperature range
RCR-9102
Produced by JEITA - Japan Electronics and Information Technology Industries Association.
In each update component failure rates (FIT) have been changed, particularly fans
Issued in 1994, based on MIL-HDBK-217F
RCR-9102A
Issued in 2000, based on MIL-HDBK-217F (Notice 2)
Includes SMT parts & pcbs
RCR-9102B
Issued in 2006

Usually for commercial power supplies, the figures are calculated at 25oC, ground benign or fixed

Taking TDK-Lambda’s RWS150-B series as an example, the calculated numbers are as follows:

RCR-9102             444,089 hours

RCR-9102B          218,172 hours

Telcordia              2,235,743 hours Ta=25

Telcordia              1,063,230 hours Ta=40

It can be seen from the above numbers, that there is a 10-fold difference between RCR-9102B and Telcordia, and more than a 2 fold difference between RCR-9102 and RCR-9102B.  Several customers have asked why our newer products calculated using the JEITA method appeared to be less reliable than older products, but did not know the significant impact of the updated, harsher standard.
Engineers should be more concerned about electrolytic capacitor and fan life (if used) as these are the typical failure modes.  Many manufacturers are showing expected capacitor lifetimes in their reliability reports.  Below are the plots for the RWS150B, which was designed for long capacitor life.  As a note, some manufacturers show similar plots, but state in small print that the convection cooled power supplies had external forced air cooling applied.
 
 

 
Power Guy

Thursday, May 26, 2016

Constant Voltage, Constant Current Battery Charging


There are three common methods of charging a battery; constant voltage, constant current and a combination of constant voltage/constant current with or without a smart charging circuit.

Constant voltage allows the full current of the charger to flow into the battery until the power supply reaches its pre-set voltage.  The current will then taper down to a minimum value once that voltage level is reached.  The battery can be left connected to the charger until ready for use and will remain at that “float voltage”, trickle charging to compensate for normal battery self-discharge.  A typical example would a low cost auto battery charger for home use or basic back up power systems.  This method enables fast charging rates and is suitable for lead acid types, but not for Nickel Metal Hydride (Ni-MH) or Lithium-Ion (Li-ion) types.

Constant current is a simple form of charging batteries, with the current level set at approximately 10% of the maximum battery rating.  Charge times are relatively long with the disadvantage that the battery may overheat if it is over-charged, leading to premature battery replacement.  This method is suitable for Ni-MH type of batteries.  The battery must be disconnected or a timer function used once charged.

Constant voltage / constant current (CVCC) is a combination of the above two methods.  The charger limits the amount of current to a pre-set level until the battery reaches a pre-set voltage level.  The current then reduces as the battery becomes fully charged.  This system allows fast charging without the risk of over-charging and is suitable for Li-ion and other battery types.

Smart charging involves the use of a micro-controller to compensate for temperature rise and adjust the charge current and charge time accordingly to the battery specifications.  This extends battery life and is used with Li-ion battery types.  This battery management circuit or unit can be fitted externally to the charger.  A number of the power semiconductor manufacturers offer control circuits to perform this function.

An example of a CVCC charger is the TDK-Lambda EVS series.  The output voltage and the charge current can be set by two potentiometers and the output characteristics are shown below.  The transition between constant voltage and constant current is automatic.

As an example, consider a 24V battery system (with a maximum float voltage of 28V) and discharged down to 15V.



When the discharged battery (at 15V) is connected to the power supply, the battery will start to charge at the pre-set constant current level.  The current will remain constant until the voltage rises to 28V.  At this point the power supply will transition to constant voltage mode and the current will decay to zero when the battery is fully charged.

The charge current is controlled to avoid overheating and the float voltage limited to avoid over-charging.

A typical application for the EVS being used with a battery management unit is shown below.

 
Under normal conditions, when AC is present, the electronic switch would be closed and AC would be connected directly to the end equipment.  The EVS power supply will charge the battery via the battery management unit and transition to constant voltage mode when complete.  In the event of an AC power interruption, the switch would connect the battery and DC/AC inverter to the end equipment.  If the power interruption was extensive and the battery was to approach a fully discharged condition, the switch would isolate the battery to avoid a damaging deep discharge.

The EVS power supply can be used with the EVS-RP module to avoid the battery discharging into the power supply when the AC supply is not present, or under a fault condition.


EVS300, EVS600 and EVS-RP Module

Monday, April 18, 2016

What is a Power Supply Standby Voltage?


The standby voltage is generated by a power supply circuit within the main power converter.  This became widely used in 1995 when the ATX specification was published to allow a desktop computer to be put into a sleep-mode to save energy.  The standby voltage supplies a small amount of power to the motherboard enabling the computer to quickly restart, rather than performing a full, lengthy, boot cycle.  The term “standby” is often confused with an auxiliary output, which has a different function.

A standby voltage is generated by a separate switching circuit and is not affected by the use of the remote on/off signal or even an overload condition on the main output of the power supply.  A typical block diagram is shown below.



Figure 1 Block diagram of a typical power supply with a standby output

The main and the standby switching converters share the high voltage output voltage (typically around 380Vdc) from the rectifier & PFC circuit.  This saves cost by not duplicating the rectification and filtering components.  It can be seen that they are independent of each other and the remote on/off control is only applied to the main converter.

The auxiliary output is supplied from an additional winding on the main converter transformer.  If the main output is turned off by the remote on/off, the auxiliary output will also turn off.  An auxiliary output is often used to power an external cooling fan if the power supply has a forced air cooling rating.  In this case if the auxiliary output is not present when the power supply output is inhibited, it does not matter as the main converter will not be providing any load and will not require additional cooling.

Figure 2 demonstrates how the various outputs and function interact with each other.  If AC power is removed for any significant length of time (10-50ms), then of course all the outputs on the power supply stop functioning.

 



Figure 2 – Timing diagram

Many power supply designers also use the standby converter to power any “housekeeping” circuitry on the output of the main converter.  This allows an “enable” type remote on/off to be offered, where the signal is pulled low to activate the main converter.  Without a standby circuit, an external voltage has to be applied to the remote on/off to inhibit the power supply.

Manufacturers of mid to high power converters with a standby voltage will often state the off-load power draw, or off-load power consumption, with the remote on/off activated from the standby voltage.

Power Guy

Monday, February 29, 2016

An alternative to isolated DC-DC converters

Traditionally when several voltages (5V to 24V) are required in a system, either a multiple output power supply is used or a single output “bulk” supply with isolated DC-DC converters.  For voltages lower than 5V (0.6 to 3.3V) the electronics industry has migrated to using multiple non-isolated DC-DC converters, often referred to Point of Load or POLs to drive FPGAs powered from a bus voltage between 5V to 12V.
With low power (typically less than 300W) dual, triple or quad requirements in the standard voltages of 5V, 12V, 15V and 24V, a single AC-DC power supply is used.  These are cost effective and readily available.
For medium power requirements (350W to 1500W), often the choice is to use a modular power supply like TDK-Lambda’s NV, Vega or Alpha series.  As the term “modular” implies, they are put together using pre-assembled modules and are available with short lead-times.  All the outputs are conveniently put into one package.


TDK-Lambda’s Vega series

Another choice is to use a single output AC-DC power supply with board mount isolated DC-DC converters to produce additional outputs.  These readily available converters range from around 10W to 60W, can accept input voltages of 12V, 24V or 48V and supply single, dual or triple outputs, with output voltages of 3.3V, 5V, 12V and 15V.


TDK-Lambda’s CCG series of 25mm x 25mm 30W isolated DC-DC converters

When the requirement is for a higher power (100W or greater) second, third or fourth output, the DC-DC converter choice becomes more limited and because of the power involved, heat dissipation is harder to manage.  Cost can also become an issue.  Utilizing technology developed from the low voltage output Point of Load non-isolated converters, higher output voltage non-isolated converters are now being considered.
Without the constraints of input to output isolation, high performance “buck” (step-down) converters with very high efficiencies can be achieved.  With less waste heat, package sizes can be minimized and costs reduced.
TDK-Lambda’s i6A24014A033V, for example has the following specifications:


Input voltage:    +9 to 40Vdc
Output range:    +3.3 to 24Vdc
Output power:   Up to 250W
Output current:  Up to 14A
Efficiency:         Up to 98%
Package size:     33mm x 23mm



As a note, these types of (step-down) buck converters cannot supply a voltage higher than the input.


Although these types of converters have no input to output isolation, the AC-DC power supply will have, in accordance with the safety standards IEC 60950 / 60601.
Below is a typical application using the i6A:



Power Guy

Thursday, January 21, 2016

Department of Energy Level VI energy efficiency standards for external power supplies

I recently received a question from one of our sales people about to what extent the new Department of Energy’s Level VI will affect our customers, and asked me to comment on it.  As usual with my blogs, let us look at the background.

The power supply industry, in particular those who manufacture external or adapter power supplies, has been aware of the US Department of Energy’s legislation on the efficiency standards for External Power Supplies (EPS).   This legislation was made final on April 11, 2014 and comes into effect February 10, 2016.  The intent is to reduce waste energy both from off-load operation and normal operation.

Details of this lengthy, but detailed, final ruling can be found on this link: http://www.regulations.gov/#!documentDetail;D=EERE-2008-BT-STD-0005-0219

I remember many years back when energy efficiency standards for power supplies were first discussed.  Initially the reaction was “it is only a few Watts, why bother”, but with the staggering number of external power supplies now being used (and it is expected to grow in future years) those few Watts soon adds up to billions of dollars in electricity and the associated environmental pollution.

Most people leave their laptop/tablet/phone chargers plugged in 24 hours a day, and that applies to numerous gaming consoles and other electronic equipment.  Power supplies continue to draw power when not supplying load and legislation has been introduced to set (decreasing) limits year on year by multiple bodies.  In addition that legislation has gradually increased the minimum operating efficiency – this is measured at four loading levels; 25, 50, 75, and 100 percent of maximum rated output current.

There has been though, some debate and confusion about what types of EPSs are actually covered by the legislation.  This is significant as the DoE ruling forbids the imports of these types of power supplies after the February deadline if they do not meet the new efficiency standards.  It is made clear that EPSs for some medical applications (those requiring FDA approval and listing) are exempt.  Spares are also excluded from the import ban.

Looking at the final ruling web-link provided above, it states in section III General Discussion, B. Product Classes and Scope of Coverage, 1. General:

An “external power supply” is an external power supply circuit that is used to convert household electric current into DC current or lower-voltage AC current to operate a consumer product.

1. Is designed to convert line voltage AC input into lower voltage AC or DC output;
2. is able to convert to only one AC or DC output voltage at a time;
3. is sold with, or intended to be used with, a separate end-use product that constitutes the primary load;
4. is contained in a separate physical enclosure from the end-use product;
5. is connected to the end-use product via a removable or hard-wired male/female electrical connection, cable, cord, or other wiring; and
6. has nameplate output power that is less than or equal to 250 watts.

Section 2: Definition of Consumer Product” is where the DoE noted that some companies have made comments questioning the vagueness of the term.  Schneider Electric commented that the definition of consumer product is “virtually unbounded” and “provides no definitive methods to distinguish commercial or industrial products from consumer products.”

The DoE ruling refers to an EPCA (Energy Policy and Conservation) document that defines a consumer product as:

 “any article of a type that consumes or is designed to consume energy and which, to any significant extent, is distributed in commerce for personal use or consumption by individuals.”  For clarification, manufacturers are advised to consult this document: 

https://www1.eere.energy.gov/buildings/appliance_standards/pdfs/cce_faq.pdf

To answer our salesperson’s question - one thing is for sure, embedded (installed internally to the end equipment) and DIN rail power supplies are not affected by this legislation.  It only applies to EPSs that are contained in a separate physical enclosure from the end-use product.

Does this affect an EPS designed for and sold for use with commercial or industrial products?  I think there will still be some debate on that, but there is strong evidence in the final ruling that they are not covered and hence exempt.  The document refers to “household electric current”, “personal use” and “consumption by individuals”.  It is very clear that if an EPS manufacturer is producing a product that could likely end up in your home, it has to abide with the legislation.

As a note, TDK-Lambda has launched a number of external power supplies that comply with Level VI efficiency standards.  TDK-Lambda’s new industrial products also have low off-load power draws and efficiencies in excess of 90%.

Power Guy



There was an error in this gadget

Popular Posts

TDK Corporation - Americas | Press Release