WO1994022084A1

WO1994022084A1 - Mains failure back-up

Info

Publication number: WO1994022084A1
Application number: PCT/DE1994/000114
Authority: WO
Inventors: Wolfgang KÖCHEL; Shmul Rubens; Wolfgang KÜNKLER
Original assignee: Siemens Nixdorf Informationssysteme Ag
Priority date: 1993-03-23
Filing date: 1994-02-04
Publication date: 1994-09-29
Also published as: DE4309369A1

Abstract

Back-up of mains power failures for any length of time in personal computers by auxiliary batteries. Use of power management and securing the machine status on a fixed disc.

Description

Power failure bridging

Technical field

The method relates to bridging network failures in desktop computers, in particular personal computers.

State of the art

For their operation, computer systems require a continuous supply of electrical energy in order to maintain the content of volatile memories, in particular the main memory and the condition of the central unit. If the energy is provided from the network, the contents of the main memory and the state of the central unit are lost in the event of a network failure. This can include the input work of several hours that has to be repeated. As long as the inputs are transferring e.g. If information is available on paper, it is simply repeatable. In particular, desktop computers, including personal computers, are used to an increasing extent for direct input, data being generated by the input activity of the operator and not previously existing on another medium. Data can therefore be lost due to a network failure, the data of which can be lost Reconstruction is complex.

It is therefore known to provide emergency power supplies for personal computers as well. Accumulators are used that are kept continuously charged and take over in the event of a power failure. The maximum bridging time is determined by their capacity. For long bridging times in the hour range, expensive, heavy and voluminous accumulators are required, which cannot then be accommodated in the housing of the desktop computer. It is an object of the invention to provide a power failure bridging that can be integrated into the device for any length of time with little effort.

Presentation of the invention

The power-saving technology developed for network-independent notebook computers is used to solve the task. In particular, the memory contents and the state of the central unit are saved on a hard disk in phases of inactivity. The invention consists in adopting the power-saving devices of a notebook computer in the desktop computer, buffering the voltage supply and using the power-saving devices for bridging power failures. Thanks to the energy-saving technology, small accumulators can be used for long buffer times, and a backup on the hard disk serves for buffering for any length of downtime. Instead of accumulators, batteries are also possible, which have the advantage of a longer service life. By using a system manager interrupt, the function of bridging the power failure is independent of the active operating system. Further details and refinements of the invention are given in the exemplary embodiment.

Brief description of the drawings

It shows

Fig. 1 is a block diagram for a desktop computer with supply buffer

Detailed description of the invention

Fig. 1 shows a block diagram of a table computer with power supply NT, supply buffer VP, central unit ZE, cursed Main memory RAM, basic program BIOS in the non-volatile memory ROM, and a non-volatile secondary memory 19, in particular a hard disk, for example with an integrated controller. Not shown are the screen and keyboard, which are not buffered by the supply buffer, because it is not necessary to continue operation during a power failure.

When a power failure begins, a signal 20 is generated by the power supply unit NT, which on the one hand activates the supply buffer VP and on the other hand activates the basic program BIOS at a control input 21. For the sake of clarity, this control input is drawn on the non-volatile memory ROM, which contains the basic program BIOS; the execution by an interrupt control, which then causes the central processing unit to activate an interrupt processing routine in the basic program BIOS, is generally known.

As a result of the power failure, a program part stored in the basic BIOS program is activated, which has access to a parameter memory 18 and, according to the parameters available there and which can be set by the user of the desktop computer, carries out various versions of the power failure bridging.

The central units used have a system manager interrupt, which is managed independently of the interrupt table managed by the operating system. This enables the basic program BIOS to carry out the processes described below independently of the special operating system. In particular, immediately after the power failure, the central unit ZE and main memory RAM can be brought into a power-saving mode, which allows bridging short power failures by means of an accumulator or a battery that is small compared to known solutions. A first embodiment consists in that the basic program begins immediately on the hard disk 19 with the saving of the machine state, that is to say in particular the content of the main memory RAM and the state of the central unit ZE. An area is reserved for this on the hard disk. Further memory areas, not shown, for example the content of the video memory, belong to the machine status and are also saved. The successful backup is noted in the parameter memory 18. Following the fuse, the supply buffer is switched off via the output 24 and the connection 25. This mode is advantageous if accumulators are used and if power failures frequently follow one another so that sufficient capacity is retained for the subsequent backups.

After the mains voltage is restored, it is determined during initialization, for example by the marker in the parameter memory 18, that a valid backup is present on the hard disk 19, from which the state of the machine at the time of the mains failure is restored and the interrupted program is continued .

As a variant of this embodiment, the basic program BIOS activates a power-saving mode in the central unit ZE and the main memory RAM after the fuse initiated immediately after a power failure via the output 26 and the connection 27; and turns off the hard drive 19. This mode and its implementation, in which, for example, the clock frequency of the central unit ZE is reduced and the main memory RAM is switched to pure refresh operation, are known from portable computers which are independent of the network. The supply buffer VP is designed such that sufficient energy is available on the hard disk 19 at least for the described backup. The energy that is not used is used to buffer the state of the central unit and main memory in power-saving mode. If the mains supply is reached before the end of the buffer time returns, the backup on the hard disk is not used; rather, the energy-saving mode is deactivated and normal processing is continued. This variant is advantageous if network failures occur rarely and usually only individually.

As a further development of this variant, the buffering after the backup is limited by a timer in the central unit ZE, which also functions in the power-saving mode, by switching off the supply buffering VP via the connection 25 after the set time. A complete discharge of the battery or rechargeable battery can thus be avoided, so that a second power failure following the first power failure can also be buffered. Instead of a timer in the central unit ZE, a monitoring circuit for the remaining capacity of the battery or the accumulator can also be used, which then sends a signal to the basic program via a connection 22 and the input 23.

The variant with timer allows a particularly simple supply buffering for devices with only one operating voltage of e.g. 5V through a nickel-cadmium battery with 4 cells, which is connected directly to the supply voltage of 5V via a parallel connection of a Schottky diode and a charging current limiting resistor, this connection being broken up, for example, by the main switch. The power failure signal is supplied by the power supply; the supply buffering is activated automatically by the Schottky diode.

In a second embodiment, in the event of a power failure via connection 20, the supply buffering VP and the power failure handling are activated in the basic program BIOS. Initially, the power-saving mode is now activated via the connection 27, but there is still no backup to the hard disk 19. By means of a timer, which is also effective in the power-saving mode, or a monitoring circuit for the remaining capacity in the supply buffer VP, which activates the input 23 via the connection 22, the power-saving mode is deactivated again in the event of a prolonged power failure, the hard disk 19 is started and the contents of the main memory and the state of the central unit are saved in the manner described . The supply buffer VP is then preferably switched off via the connection 25.

In order to achieve reliable bridging of the power failure even under unfavorable boundary conditions, in a further variant, when using batteries after the mains voltage has returned, use is initially blocked by the basic program BIOS for a certain time until the batteries have reached a charge which Allow a new backup to the hard disk and only then will the restored state of the hard disk be restored. During this time, the energy-saving mode is activated in order to have the entire power of the power supply available for recharging the battery.

In a third embodiment, the power failure signal is not used to store the machine status, but rather the interrupted program is initially continued. The machine state is not saved until signal 22 from the supply buffering arrives. This operating mode is useful if the power outages usually last only briefly and you want to continue computing.

Due to the low energy requirement of the described power failure bridging, batteries can also be used instead of accumulators. On the one hand, batteries, such as lithium batteries, are subject to very little wear when not in use and have a lifespan of up to 10 years, while accumulators can fail after one year, particularly if used rarely and in warm devices. There is also a finding for batteries the remaining capacity is reliably possible via a voltage measurement.

Since the selection of the different variants depends on the area of use, all variants are preferably offered by the basic program BIOS and selected via an entry in the parameter memory 18 which can be changed by the user.

Claims

1.Bypass for network downtimes in desktop computers, especially personal computers, with power supply unit (NT), central processing unit (ZE), volatile main memory (RAM), non-volatile secondary memory (19), a non-operationally changeable basic program (BIOS), a supply buffer (VP) and a power failure signal (20) preferably generated by the power supply unit (NT), the power failure signal (20) being connected to an interruption control, which is not influenced by the operating or application program located in the main memory, which is used by the basic program (BIOS) Backing up the machine status, in particular the main memory (RAM) and the status of the central processing unit (ZE), on the secondary memory (19) in a form by means of which the basic program (BIOS), after the end of the power failure, the main memory (RAM) and can restore the state of the central unit (19) from the backup on the secondary memory (19).

2. Bridging according to claim 1, wherein after the backup of the machine state, the central unit ZE or the main memory RAM are put into a power-saving state and, if the mains voltage before the expiry of a timer or the arrival of an early exhaustion of the supply buffering (VP) indicating signal (22) returns, the power saving state is ended and the fuse on the hard disk (19) is discarded.

3. bridging according to claim 1 or 2, wherein the central processing unit (ZE) or also the main memory (RAM) are put into a power-saving state and the backup of the machine state after the expiry of a timer or the arrival of an early one before the backup of the machine state Exhaustion of the supply buffering (VP) signal (22) is carried out.

4. Bridging according to one of claims 1 to 3, wherein the supply buffer (VP) has a switch-off input (25) which is activated by the basic program (BIOS).

5. Bridging for network downtimes in table computers, especially personal computers, with power supply (NT), central processing unit (ZE), volatile main memory (RAM), a supply buffer (VP) and a power failure signal (20) preferably generated by the power supply (NT), the Power failure signal (20) is connected to an interruption control which, preferably via a basic program (BIOS) which cannot be changed operationally, puts the central unit and the main memory into a power-saving mode, and continues processing after the power failure signal has ended.

6. Bridging according to one of the preceding claims, wherein the central unit (ZE) has an SMI input for a system manager interrupt and the system manager interrupt is independent of and without interference of an operating program using the central unit and the interrupt means used SMI interrupt routine in the basic program that secures the machine status.

7. Bridging according to one of the preceding claims, wherein the basic program (BIOS) is assigned a non-volatile parameter memory (18) in which at least two of the following operating modes can be set in the event of a power failure:

- First mode, in which the operation of the central unit (ZE) and main memory (RAM) continues unchanged; - Second mode, in which the central unit (ZE) or also the main memory (RAM) assume a state of reduced power consumption;

- Third mode, in which a backup of the machine status on the secondary storage (19) takes place.

- Fourth mode, in which the machine state is saved on the secondary memory (19) and then the central unit (ZE) or main memory (RAM) assume a state of reduced power consumption.

- Fifth mode, in which the central unit (ZE) or main memory (RAM) assume a state of reduced current consumption and, if the power failure continues, the machine state and the main memory content are backed up to the secondary memory.

- Sixth mode, in which the signal (22) for early exhaustion to secure the machine status on the magnetic disk (19) leads.