GB2079223A - Postal Franking Meter - Google Patents

Abstract

A postal franking meter comprises an accounting system including one, and preferably two, printed circuit boards, one 70 extending along the top of the meter and the other along one side. Print wheels are set, and, simultaneously, their values are entered by encoding switches each having stationary contacts integral with board 70 and a rotary part 46 to be set by an associated thumb wheel 1 through a rack 4 and gear 5. The accounting system includes a microcomputer and an alphanumeric display. When an item of mail is to be franked the microcomputer initiates a printing cycle, adjusts credit and tote registers and checks availability of credit and freedom from any fault condition for the next printing cycle. A door, normally sealed by the Post Office, when opened actuates a switch to allow re-funding etc. of the credit register. <IMAGE>

Description

SPECIFICATION Postal Franking Meter This invention relates to postal franking meters for franking machines.

In a franking meter a value has to be set for each item of mail fed into the machine. This value, which in the United Kingdom is usually up to 999 1/2, can be altered as required by hand. This value is automatically communicated to a mechanism that prints on items of mail an inked franking impression in accordance with the requirements laid down by International Post Offices. The value is also automatically communicated to an accounting side of the meter, including a descending register containing the sum of postage value remaining credited for the customer and an ascending register containing the accumulated sum of postage value used.

It will be appreciated that the accounting and control side of the meter can be basically an electronic system while the value setting may be electromechanical or mechanical and the printing system essentially mechanical. There have been prior proposals comprising such electronic and mechanical systems. In one of these the electronic system comprises a microcomputer and various ancillary solid state units. On the mechanical side, the values are set by press buttons and printing wheels allocated respectively to several numerical orders are set selectively by a stepping motor according to the values to be franked. The stepping motor is selectively connected to the printing wheels under the control of solenoids. This, although having benefits in some applications, involves an electromechanical system of substantial complexity and cost.

A more simple mechanical system has been proposed wherein the press buttons are replaced by thumb wheels as currently used in many mechanical franking machines each geared permanently to a corresponding one of the printing wheels by way of an individual drive bar in an arbor carrying a printing drum wherein the printing wheels are mounted. In this proposal, however, the values selected by the thumb wheels are communicated to an electronic accounting system by way of the drive bars which act on the accounting system through magnetoresistive cells only when the arbor is caused to rotate to effect a printing operation. Thus, monitoring procedures such as checks and fault diagnostics cannot be effected in relation to an input value before the actual printing of the value begins.Moreover, the input to the accounting system by the rotary printing assembly prevents the use of a satisfactory modular arrangement such that substantially all the electronic parts can be located on printed circuit boards mounted conveniently and independently of the mechanical parts and minimizing the weight and bulk of that portion of the meter.

A primary object of the invention is to provide a franking meter having an electronic accounting system and a mechanical printing system in which each value to be franked is set simultaneously in the two systems by means that are substantially more simple, economically constructed and more readily serviced than in the case of the aforesaid prior proposals and at the same time enable substantially the whole of the electronic accounting system to be mounted on printed circuit boards occupying only a comparatively small portion of the bulk of the postal franking meter.

It will be appreciated that a postal franking meter is that part of a franking machine that has to be periodically carried to a post office for replenishment of credit. The base of the machine comprises the mechanism for feeding the items of mail through the machine to receive the franked impressions. For smaller basic postal franking machines the meter and base may be a single integral unit.

According to the invention a postal franking meter for a franking machine comprises a plurality of printing members allocated respectively to the numerical orders of the maximum value to be franked by the machine, each printing member being adjustable for printing any one of a series of digits (or fractions), manually operable elements respectively mechanically connected to said printing members for setting said printing members to print required values, a printed circuit board assembly comprising solid state units, including a microcomputer for effecting substantially all accounting operations required in the machine, said operations including registering descending values remaining credited for a customer and registering ascending values of postal value used, an alphanumeric display module for displaying information and values when required, said printed circuit board assembly comprising stationary arrays of conductors serving as encoding switch contacts, said arrays being allocated respectively to said printing members for encoding numerical values to be printed thereby, rotatable encoding switch contact units mounted directly on said board assembly for traversing said stationary contacts in said arrays to select the values to be encoded, and mechanical means connecting said rotatable switch contact units to said manually operable elements whereby the operation of any said element simultaneously sets the associated printing member to a value to be printed and encodes that value.

Preferably the manually operable elements are thumb wheels and the printing elements are printing wheels, each thumb wheel being connected by a first rack and gear mechanism to the associated printing wheel and by a second rack and gear mechanism to a rotatable contact unit for encoding each value selected for printing by the associated printing wheel. The fact that the stationary switch contacts are integral with a circuit board and the rotatable contact units are mounted directly on this board means that a particularly neat arrangement can be devised, separately mounted encoding switches and corresponding connections being eliminated thereby reducing cost and improving reliability.For example, there may be a printed circuit board carrying the encoding switch contacts in addition to press button switches for controlling an alphanumeric display, this board being located along the top of the meter and connected, for example, by a flexible printed circuit or other connecting means to a second printed circuit board along one side of the meter and including the microcomputer, a memory module and associated solid state units, or the whole circuit on one pcb.

In a franking meter, it is necessary to provide means for changing over from a customer mode, that is to say a normal mode of action under control of the user of the machine, to a post office mode enabling a postal authority to adjust and reset the meter, in particular to alter the credit available to the user as registered in the memory module. For this purpose, it is advantageous to provide the meter with a casing having a door that has been sealed by a postal authority, but when opened by the authority automatically actuates a switch to change the meter from the customer mode to the post office mode and at the same time exposes a press button to be used in resetting the meter.

In order that the invention may be clearly understood and readily carried into effect a postal franking meter in accordance therewith will now be described, by way of example, with reference to the accompanying drawings, in which: Figure 1 is an exploded view showing portions of a value selector mechanism and printing mechanism; Figure 2 is a perspective view showing details of parts of the mechanism of Figure 1; Figure 3 is a vertical section through the postal franking meter; Figure 4 is an elevation of the franking meter as viewed in the direction of the arrow A in Figure 3; Figure 5 is a plan view of a stationary part of an encoding switch shown on an enlarged scale; Figure 5A is a plan view of a rotary part of the encoding switch; Figures 6 and 7 are external perspective views of the franking meter;; Figure 8 shows a portion of Figure 7 with a hinged door opened; Figure 9 is a cross-section showing some details of the franking meter; Figure 10 is a system block diagram; Figures 1 A-I 1 D are circuit diagrams; Figure 12 is a real time diagram relating to certain meter functions; Figure 13 is a block diagram showing an arrangement of software modules determining the operation and testing of the meter; Figure 14 is a timing diagrnrri relating functions of the meter; and Figure 1 5 is a chart showing a flag bank arrangement for a memory system; Figure 16 is a microcomputer data memory map.

Referring to Figures 1 to 4, the postage value to be franked is selected by manually rotating a selection of four thumb wheels 1 to the required value, the thumb wheels being allocated respectively to the four numerical orders contained in the largest value to be franked and each order comprising the digits (or fractions) that may be required to be printed for that order. Each thumb wheel 1 has an integral gear 2 which meshes with a corresponding rack 3. Rotation of the thumb wheel causes a fore and aft movement of the rack 3. Integral with each thumb wheel rack 3 is a second rack 4 which engages a gear 5 that forms part of an associated one of four encoding switches 14 described below.

Actuation of this encoding switch provides the appropriate value input to an electronic accounting portion of the machine. Integral also with each thumb wheel rack 3 is a value selector ring shoe 6 which shrouds and is capable of imparting axial movement to an associated one of four selector rings 7 that can slide along an arbor 8 under the control of the associated thumb wheel 1. Each selector ring 7 is fixed to a bar 9 that can slide longitudinally in a slot in the arbor 8 and carries a rack 10 in mesh with a gear 11 fixed to a printing wheel 12 in a printing drum 13. Thus rotation of a value selection thumb wheel causes a synchronised operation of the corresponding encoding switch and of the corresponding 'value printing wheel which subsequently prints the appropriate value onto the mail.

When the thumb wheels are formed with lobes as shown in Figures 1, 3 and 10 so that the spaces between the lobes on each wheel correspond to a series of digital values, rotation of any thumb wheel 1 causes a spring loaded locking bar 1 5 to detent the thumb wheel in each correct value position. As described in the specification of Patent Application 79.03987, rotation of a thum wheel 1 from one value position to the next forces the locking bar 1 5 to ride over a lobe and in so doing slides a trip lever isolator unit 1 6 behind a clutch trip lever 17, thereby inhibiting the operation of a clutch unit 1 8 until the thumb wheel has reached the next correct value position. Thus a printing cycle cannot take place if any of the value selector wheels 1 is incorrectly positioned. When a printing cycle is committed, the trip lever 1 7 is positioned so as to hold the isolator unit 1 6 against movement, thereby effectively locking the thumb wheels.

A printing cycle is initiated by a signal generated by a switch 80 (Figure 10) when an item of mail 1 9 (Figure 6) is fed into the machine comprising the meter. The signal is fed to a microcomputer which assuming the clear status described below has been verified controls the operation of a trip solenoid 20 (Figure 1) to swing the trip lever 1 7 to a position such that a clutch pawl 21 is released to the action of a spring 22 and frees a clutch release plate lobe 23 permitting a motor to drive the arbor 8 and printing drum 13. Simultaneously switches 24, 25 are allowed to open. This action only takes place after the microcomputer control has verified a clear status comprising the availability of sufficient credit and freedom from any fault condition.As the printing cycle is completed, clutch lobe 23 and a clutch lobe 26 return the clutch pawl 21 and trip lever 1 7 to their initial positions whereby the switches 24, 25 are closed and the printing drum brought to rest.

The switches 24, 25 are microswitches which, on being opened at the beginning of a printing cycle, start an accounting sequence that reads the value set in the encoding switches, adjusts the appropriate registers and reassesses the clear status applicable to the next printing cycle.

A third microswitch 27 (Figure 9) is used for selecting either a customer or a post office mode of operation of the meter. When a hinge door 28 is closed and sealed a projection 29A holds the switch 27 closed to select the customer mode. To convert to the post office mode a security seal is broken and the door 28 opened.

Four push buttons 29, 30, 31, 32 (Figure 6) are provided. When the machine is set for the customer mode, depression of the button 29 is necessary to print a high value, for example, a value greater than 99 1/2 for the U.K.; this is therefore a safety measure against printing a high value by mistake. The postal value selected appears in a display 33 when the CL button 32 is depressed in the customer mode. The amount in a credit register is displayed when C button 30 is depressed provided this is preceded by depression of the CL button 32. The amount in a tote register of the postage value used is displayed by depression of button 31 , then button 30 provided these are preceded by depression of the CL button 32.

The display is arranged to flash at 60 c/min for postal values greater than a predetermined amount and the meter will not operate unless but.ton 29 is depressed while a base trip switch 80 responsive to the insertion of a mail item into the machine is activated. The button 29 must be depressed for each value selected within the high range but for multiple high value operation the button 29 can be held depressed. When the credit available is less than a predetermined value the display shows the letters LC beside the value selected. When a still lower credit is reached the letters LC flash at 30 c/min. When credit has run out the letter C flashes at 60 c/min in each section of the display. In the case of a permanent fault the meter becomes inoperative.However, a display of the amounts in the credit and tote registers can be obtained during a safe fault condition, whilst a service test unit is required to assess information during a catastrophic fault condition.

In the post office mode depression of button 29 causes the accumulated number of items franked with a postage value greater than zero to be displayed. To increment the credit, a + button 34 (Figure 8) is depressed provided this is preceded by depressing the CL button 32, the new credit having been set by the thumb wheels 1. To decrement credit by an amount set by the thumb wheels button 31 is depressed followed by depression of a - button 34 (Figure 8), provided these are preceded by depression of the CL button 32.

The switch 14 shown in Figure 5 which is the subject of Patent Application No. 78.44793 comprises five stationary contact strips A, B, C, D, E each consisting of two arcuate portions 40, 41, joined by a radial portion 42. Each arcuate portion, as well as two common conductor rings 43, 44 are centered on a point 45 and are all on the same printed circuit board. A rotary part 46 (Figures 3 and 5A) of the switch has a pair of connected contacts 47, 48 that wipe over the ring 43 and inner arcuate portions 40 and a pair of connected contacts 49, 50 that wipe over the ring 44 and outer arcuate portions 41. As can be seen from Figure 5 the contacts 47, 48 are diametrically opposite the contacts 49, 50.

The rotary part 46 can be set in any one of ten positions indicated by O to 9 in Figure 5. In Figure 5 the contacts 49, 50 are shown bridging ring 44 and strip A41; all while contacts 47, 48 bridge ring 43 and strip C40. It will be seen, therefore, that if a circuit is completed through ring 44 and the contact strips A to E are read through one complete cycle, a 5-bit word 00001 is output. Then, if a circuit is completed through ring 43 and the contact strips A to E are read through one complete cycle, a 5-bit word 00100 is output. The words can be combined to give 00101 representing decimal 1.This system can be used to represent all decimal digits 0 to 9 as shown in the following table: Table 1 Common 44 Common 43 Combination Decimal E D C B A E D C B A E D C B A 0 0 0 0 0 1 0 0 0 1 0 0 0 0 1 0 0 0 0 1 0 0 1 0 0 0 0 1 0 1 2 0 0 0 1 0 0 0 1 0 0 0 0 1 1 0 3 0 0 0 1 0 0 1 0 0 0 0 1 0 1 0 4 0 0 1 0 0 0 1 0 0 0 0 1 1 0 0 5 0 0 1 0 0 1 0 0 0 0 1 0 1 0 0 6 0 1 0 0 0 1 0 0 0 0 1 1 0 0 0 7 0 1 0 0 0 0 0 0 0 1 0 1 0 0 1 8 1 0 0 0 0 0 0 0 0 1 1 0 0 0 9 1 0 0 0 0 0 0 0 1 0 1 0 0 1 0 Accordingly it is necessary for the outputs to pass to a device, such as a microprocessor, capable of recognising the coded information from the switch, determining for security reasons, that there is a single "1" in each word received, and then combining the pairs of words to obtain the required 2 out of 5 code which in turn is recognised as a decimal number.

The four encoding switches 14 are shown in an expanded diagrammatic form in the circuit diagram of Figure 11 as eight separate 1 out of 5 encoding switches 1 4A operated in pairs. The encoding switches represented by 1 4A form part of an input module 52 with which are associated the aforementioned switches 24, 25, 27 fed through buffers 7/2, 7/3, 7/4 (e.g., SN 7407), and push button switches 29,30, 31, 32, 34, fed through buffers 6/2,6/3,6/5,7/5 and 6/6 (e.g., SN 7407). All these buffers as well as the inputs to the encoding switches are connected respectively by ten lines 53 to a binary to decimal decoder Q1 1 (e.g., SN 74145).

A microcomputer Q5 (e.g., 8049) which is provided with a ROM and into which is masked various software instructions, provides the primary control of the operation of the meter. It has a limited number of input lines P24-P27 and To that can read signals from the switches and press buttons associated with the input module 52. The total number of signal outputs from the encoding switches alone exceeds the number of inputlines. Therefore, a time division multiplexed (TDM) system is employed in which ten groups of five signals are sampled sequentially along five lines 54A-54E. The TDM scanning is provided by a binary counter Q8 (e.g., SN 7493A) and the decoder Q1 1.The counter Q8 operating under control of signals from line P22 of the microcomputer provides cycles of four inputs through a binary-coded decimal sequence of 0000 to 1001 to the decoder Q1 1 which produces a logic level "0" on each of the ten output lines 53 in turn as its four inputs cycle through the BCD sequence.

The sequence of cycles of four signals on the four outputs of OA, QB, QC, QD of counter Q8 foilows that shown in Table 2 below. It is possible for a microcomputer Q5 to be selected or for the microcomputer Q5 to be adapted for the provision of feedback means comprising lines 53A, 53B described below to enable the microcomputer Q5 to monitor the sequence from scan 2 to scan 9 inclusive.

Table2

Count CA OB OC OD 0 0 0 0 0 1 1 0 0 0 2 0 1 0 0 3 1 1 0 0 4 0 0 1 0 5 1 0 1 0 6 0 1 1 0 7 1 1 1 0 8 0 0 0 1 9 1 0 0 1 10 0 1 0 1 11 1 1 0 1 12 0 0 1 1 13 1 0 1 1 14 0 1 1 1 15 1 1 1 1 0=16 A quadruple D-type flip flop Q6 (e.g., MC 14175) is used as a four bit latch to store logic signals.

The microcomputer Q5 counts the number of signals sent through output P22 and after nine transitions initiates signals through line RD and output P02 to latch Q6 which results in logic '0' which appears on line Q4 of latch Q6 which is then transfered onto an inverting buffer 9/6. The inputs of decoder 011 change to 0000 and the microcomputer Q5 initiates another pulse through RD and logic '1' at 4D of latch 06 to alter the output of inverting buffer 9/6 to permit 08 to count again.

Several complete scans are performed by the microcomputer Q5 during which the status of all the switches in the input module 52 remain constant before the microcomputer recognises the resulting signals as valid. This guards against transient electrical interference as well as switch bounce.

The microcomputer OS interacts with the input module 52 to verify correct scanning sequences and to detect faults on lines; for example, the failure of a scan of one of the encoding switches to include two "1 s" (Table 1) or, alternatively, two "Os" according to the arrangement of the logic levels.

A buffer 7/6 shown at the lower right hand corner of Figure 11 presents the microcomputer with a "0" through line 54E at the beginning of the first scan of a sequence. Failure to do this induces an eventual shutdown of the meter.

Central to the correct functioning of the input module 52 is its interaction with the microcomputer Q5 to verify correct sequence of scanning and to differentiate between fatal and nonfatal faults. Any fault that is non-periodic and intermittent, but does not in any way influence the status of the input variables (i.e., status of the encoding switches 1 4, status switches 29, 30, 34 and command switches 24, 25, 27, 32) and does not degrade the system performance appreciably, is defined as a non-fatal fault, such as noise spike or switch bounce. A fatal fault renders the meter inoperative.

The detection of fatal faults associated with the input module 52 is achieved by taking feedback connections from outputs QB and Qc of counter 08 to respective microcomputer inputs P07 and P06.

After initi ating each scan, the microcomputer inputs these feedback states from outputs QB and QC according to Table 2. Coupled with the above fact, is the fact that the mechanical construction of the encoded switches 14 prevents the same output occurring on the same one of the lines 54 A-E in each of the consecutive scans 2-3, 4-5, 6-7 and 8-9 by way of outputs Q2 to Q9 of decoder 011, so that the microcomputer is enabled to monitor scans from 2 to 9. Aithough the above statement is true under normal operation, there is still a possibility of mis-reading the encoding switches under certain failures.For example, if the QD output of counter Q8 failed in the "0" state, the normal scan of 10 by the microcomputer OS would be allowed to energise the outputs of Q11 in the following order: 0123456701 etc., as compared to the normal order of :0123456789 etc. It can be clearly seen that this would create a wrong value for the most significant encoded switch, i.e., the encoded switch 14 (lowermost switches 1 4A in Figure 11 connected to outputs Q8, Q9) allocated to the thousands order. The status of the groups of switches 29, 30, 31, 34 and 24, 25, 27, 32 would be read as the value for the most significant encoded switch.In order to overcome this possibility of misreading, the command and status switches are arranged as shown in Figure 11.

At scan 0 through QO in decoder Q11 the status switches 29, 30, 31,34 and a deliberate open circuit at buffer 6/4, are sampled which ensures a logic "1" at line 54C. At scan 1 through Olin decoder Q11 the command switches 24, 25, 27, 32 and deliberate short circuit at buffer 7/6 are sampled to ensure a logic "0" at line 54E. (This particular short circuit is also used to initiate a trip cycle as described later). This arrangement is such that it produces a 5 bit code that is different to the 5 bit code produced by the encoding switches at all possible combinations of the command and the status switches.By creating this disparity, any misreading is averted and the microcomputer is able to monitor the scans from 0 to 9 successfully and make the meter inoperative in the event of fatal failures in the input module.

It is to be noted that the simultaneous operation of switches 24 and 25, as described above with reference to Figure 1, the inaccessability of switch 34 in the customer mode, and the fact that switch 34 can only be operated in the post office mode when the switch 27 is open (Figure 9), combine to produce a reliable input status monitoring system. If the switch 34 was operated when the switch 27 was closed the microcomputer would detect this as a fatal fault.

Description of some common faults and their detection is given below: If buffer 9/6 fails as to prevent counter 08 being reset after scan 9, no deleterious effects follow as the absence of logic "0" at line 54E at scan 1 and the reinitialisation of counter Q8 after scan 1 5 would be consequently detected and the meter made inoperative.

If the outputs of either counter 08 or decoder 011 develop a permanent fault, (either stuck at "O" or stuck at "1") the combination of invaiid feedback states from counter 08 and the appearance of incompatible 5 bit code at lines 54A-E, are interpreted by the microcomputer as fatal faults and once again the meter is made inoperative.

The afore-mentioned failure detection is also extensively used in production and in field service to pin point components which cause such failures and reduce the mean time to repair.

The primary objective of the time divisional multiplexed scanning system is to replenish peripheral registers in the microcomputer Q5 with current status of input peripheral elements; to provide a high enough sampling rate to differentiate between transitory states of input peripheral elements (e.g., switch-bounce) and stable states; and to provide a scan rate sufficient to produce a flicker-free display in a display module 55.

it is often found in various applications of microcomputers that the input/output capability of the microcomputers is required to be increased. This increase in input/output capability is easily achieved by incorporating devices known as input/output expanders in the system.

Although the embodiment of the franking meter according to the invention shown in the drawings called for more outputs than the microcomputer can provide, this need is successfully met by incorporating the inexpensive quadruple D-type flip-flop Q6 so that the need for the inclusion of an expander is averted. The outputs of latch Q6 are manipulated by creating the desired states at the respective inputs and by the generation of a pulse at RD output of the microcomputer. Since pulses are generated at RD output of the microcomputer during memory read or port 0 (see below input operation, it is important that the microcomputer sets the inputs of latch Q6 at desired states before embarking upon such operations.This can be achieved by making lines 54 A-E in a logic "1" state if necessary and by either setting or re-setting the other inputs to latch 06.

The microcomputer OS and the latch Q6 form part of a processing module 56. The design objectives of the processing module 56 are (see particularly Figure 10):- 1. To process franking data, execute fault tolerant/safeguard and anti-noise algorithms when required and to do adequate system and self diagnosis; 2. To provide the scanning signals to, and verify the validity of the data from the input module 52; 3. To provide a trip signal via latch Q6 to the trip solenoid 20 to initiate a franking operation; 4. To provide scanning signals and illuminating signals for the display module 55; 5. To provide a signal via latch Q6 to actuate an over-voltage protection and shutdown module 66; 6.To provide signals to read from and write to or render quiescent ultra low power memories Q2 and Q4 in a memory module 57; 7. To sample inputs to verify the correct functioning of all modular parts of the machine.

The memories Q2 (e.g., 6508) and Q4 (e.y., 6508) are written sequentially and can be read either simultaneously or in sequence. The address to the memories 02, Q4 is written from Port 1 (outputs PlO-P 1 7) of the microcomputer and the data from Port O (P00--P03). This greatly reduces the dependance on critical parameters such as the address hold time and the fall time of the ALE signal.

Since several ALE negative transitions may occur before the actual read or write operation of the memory, the probability of incorrect addressing is reduced as well. The address lines are also deliberately made bi-directional as to give the microcomputer more capability of diagnosis as, for example, write then read and verify the address. The memory Q2 is written to through an inverting buffer 9/4 (e.g., LS 7405) and NAND gates 2/1,2/4 (e.g. MC 14093) and buffer 9/3. The memory Q4 is written to through an inverting buffer 9/2, NAND gates 1/1, 1/2 (e.g., MC 14093) and buffer 9/3. The memories Q2 and Q4 are read respectively through DO outputs by synchronising signals at NAND gates 2/1,2/4,1/1,1/2 and buffer 9/3.The second inputs to the NAND gates 1/1, 1/2, 2/1,214 are controlled for effecting the writing and reading operations from a microcomputer output P23 and Q2 output of latch Q6 by way of inverting buffer 9/5 and pairs of inverting gates 1/3, 1/4, and 2/2, 2/3. It will be seen, therefore that the select lines of both memories Q2 and Q4 are derived respectively from different integrated circuit packages nameiy the latch Q6 and microcomputer Q5, so as to eliminate any indefinite selection of the memories in the event of failure of either OS or Q6 and hence the probable destruction of the contents of both memories Q2 and 04.

In each memory the information is stored as an eight bit word in cells sequentially located.

Address lines AO--A8 are common to the memories, and the bits may be stored in identical address locations in each memory. The accounting data consists of only four bits. The remainder are added as part of an error detecting and restoring code. Up to two bits in error can be detected and a single error can be corrected. The specific patterns of bits based on a well known coded system are shown in Table 3 below, whereby four check bits are added to each sequence of data bits so that each such sum will remain different from all the other sums even if it contains two errors.

Table 3 Data Check 0 0 0 0 1 1 1 1 0 0 0 1 0 1 0 0 0 0 1 0 0 0 1 0 0 0 1 1 1 0 0 0 1 0 0 0 0 0 0 1 0 1 1 0 1 0 0 1 1 0 1 1 0 0 0 1 1 1 0 1 1 1 1 0 0 0 1 0 0 0 1 0 0 1 0011 Table 3 (cont.) Data Tale 3 (cont.) Check 1 0 1 0 0 1 0 1 0 1 1 1 1 1 0 1 1 0 0 0 1 1 0 1 1 0 1 1 0 1 1 1 0 1 0 1 1 1 1 1 1 0000 The microcomputer also utilises superfluous cells in both memories Q2 and Q4 as dummy locations as to perform read, write and verify operations in order to validate the integrity of both memories before embarking upon system read and write operations i.e., up-dating the memories after a trip cycle. This is also a powerful tool for diagnosing certain system failures such as a permanent fault at gates 2/1,2/4,1/1 and 1/2.

A further safeguard is provided by a comparison of the data in each memory with that in the microcomputer Q5 during use. At power up a comparison is made between the memories. At power down the microcomputer memory is discharged. However, the data in the memories is made nonvolatile in that each memory together with protection control gates has its own long life battery 58 to retain the information when mains power is removed. Decoupling capacitors C9 and C13 are provided and diodes D10 and D19 prevent reverse current flow through the batteries when the main supply is on. Resistors R19, R20, R25, R26 are pull up resistors to ensure a clearly defined logic "1" during power off states. Resistors R29, R31 are pull up resistors to provide logic compatibility between the microcomputer and memories.

The address lines AO--A8 are driven by signals from the microcomputer at its outputs P10- P 1 7. The logic condition of these lines is governed by the microcomputer programme. Data inputs Di for the two memories are taken respectively from microcomputer outputs P01 and P03. Data outputs DO are connected respectively to the microcomputer respectively at POO and P02.

Redundancy is adequately achieved by having all the interconnections from each memory independent of the rest of the system with the exception of the address lines AO-A8. The only probable failures are, therefore, safe failures. All other connections are made completely separate and the susceptible lines such as those connected to the inputs STR and WE are deliberately gated by different packages of gates as indicated above. The failure of an integrated circuit package does not lead to the desctruction of both memories. The fact that the WE inputs of the two memories originate from the different outputs WR and PROG of the microcomputer instead of the more conventional single WR output means that data corruption in both memories is prevented in the event of a WR output failure.Thus, an unusual way of accessing the memory Q2 is provided. The microcomputer writes to memory Q2 by executing the instructions that would be needed as if it was communicating with an input/output expander. The PROG signal generated by this action is gated to produce a WE signal for memory 02. This also gives the microcomputer more capability for failure analysis under various fault conditions.

The inclusion of the aforementioned method requires another set of instructions in addition to the other set for memory 04. These two sets are mapped onto two different spaces in the ROM of the microcomputer. This gives additional safeguard against failures such as the failure of the program counter of the microcomputer. By careful choice of memory space and by adequate separation of software modules, an incorrect access to both memories due to microcomputer failure is successfully averted. The software also takes into account the failure of increment/decrement counter, registers failure, input/output ports failure and general system noise during power up and power down. The packaging is optimised to give the memories adequate protection from airborne radiation and static discharge.The packaging and power supply for the memories is described in greater detail below in relation to a power supply module 65.

The display module 55 includes an alphanumeric display consisting of typically nine blocks 59 of seven segments plus a decimal point. Each segment within a block is a variant of the traditional triode valve. A filament heated cathode provides an electron supply. The electrons are allowed to bombard an anode under control of a biassed grid. The anode comprises a shaped chemical material which fluoresces when the anode is bombarded. The nine blocks of segments are scanned sequentially at a high enough rate to prevent flicker by the action of the counter 08, the four outputs OA-OD of which are fed to a low power inverting binary to decimal converter 010 (e.g., MC 14028).The outputs of the converter 010 are fed to non-inverting buffers Q7, Q9 (e.g., each UDN-61 18). The buffer 09 enables a 24 volt input to provide sufficient current for the filaments to form a character from the seven anodes in each block 59, synchronously with scanning of the nine blocks of segments. The microcomputer OS provides signals through a group of outputs PlO-P 1 6 and the buffer Q7 that determine which grids in each block are to cause their anodes to be illuminated. The microcomputer is not capable of driving the buffer 07 directly and does so by way of buffers 5/5, 5/6 and 4/6-4/2 (e.g., MC 14050). A decimal point, when required, is signalled from output P 1 7 in the microcomputer through inverting buffer 4/1 and a high voltage (24 volt) buffer 6/1 to the blocks of segments.

A power supply module 65 is designed to provide power for the electronic components and electro-mechanical elements, to provide protection for the electronic components against mains born interference, to provide individual power supplies for the memories and to create evidence should any attempts be made to alter the status of the equipment by indirect electrical means.

In the power supply module 65, a mains transformerT1 has dual secondary windings respectively feeding bridge rectifiers B1, B2. One bridge output is filtered by a large value electrolytic capacitor Cl. This reduces the ripple content of bridge B1 output to a semi-conductor regulator Q1 which maintains a +5 volt line 81 to enable the microcomputer to complete its functions immediately following a power down condition. The regulator Q1 is a three terminal device with a heat sink, which has its reference built in. The regulator is supplied through a quick blow fuse FS 1 which has a rapid response due to a silicon controlled rectifier SCR1 in the over voltage module 66.Memory supplies VCC (M1) and VCC (M2) are derived through resistors R13, R14 and diodes D7, D9. D8 is a quick response, high energy zener diode with a very large transient power absorption capacity and is used for the protection of line 81 against breakdown of the regulator Q1 prior to the operation of the over voltage protection module 66 (which has a 200 s delay). Capacitor C7 acts as a filter as a protection element in conjunction with diode D8. The memory supplies are fed through diodes D12 and Dl 3 to prevent the batteries 58 draining when power is off. To the first order, the forward voltage drops of D12, D7 and D13, D9 mutually cancel.The regulation and ripple conditions of the memory supplies are substantially those of line 81. The simple arrangement whereby the zener diode D8 provides filtering for high frequency positive excursions on the +5V supply line and whereby the supplies for memories Q2 and Q4 are derived respectively from diodes D13 and D12, not only provides adequate redundancy of the supplies to the memories, but also prevents SCR latch up effect at memory inputs by ensuring that the input signals do not exceed the memory supplies by more than 0.3V. In the event of such a latch up as a result of either D12 or D13 failure, the memory integrated circuit can become very hot.

For this very reason, as well as to provide dissimilar environments to the memories Q2 and 04, they are packaged as far apart as the physical constraints allow. The control and data inputs and outputs of memories Q2 and Q4 are laid on a printed circuit board orthogonally to the respective address lines for the memories, and the respective supply lines are made as wide as necessary to reduce cross-talk, general system noise and to comply with proper printed circuit board layout practice.

The anode cathode potential Vf+, Vf- for the display 59 is generated by a conventional resistor zener diode combination Rl 8, Dl 4 across the regulator output. The VCC output of the line 81 supplies logic units and the microcomputer Q5.

The output from the bridge rectifier B2 provides, by way of resistor R54, zener diode D4 and capacitor C4, the 24 volt supplied for the display module 55. A diode D3 is connected from the output of bridge rectifier B 1 to the junction of zener diode D4 and resistor R54 so that an abnormal increase in mains supply voltage will cause diode D3 to be destroyed thus providing evidence of the occurrence of this abnormal condition.

The bridge rectifier B2 also supplies a trip module 82 for operating the trip solenoid 20 to initiate the printing operation as described above. A capacitor C5 is used to store the charge to operate the trip solenoid when a transistor T3 switches on. Capacitor C5 is charged via resistor R10 and a slow blow fuse FS2 which isolates the trip solenoid either under the control of the microcomputer due to defined system failures e.g., fault in input module 52, or if the transistor T3 becomes short circuited between collector and emitter.

Transistors T3 and T4 are used to switch the charge stored in capacitor C5 to operate the solenoid 20. The signal for this is delivered by the microcomputer Q5 by way of the latch Q6 when the microcomputer has noted that the input of P27 from line 54E is "0" and has monitored the condition of the two microswitches 24, 25 thereby diagnosing the correct behaviour of the printing drum. The consequent output signal at Q1 on Q6 is "1". This signal is delivered through inverting buffers 5/4, 9/1.

A resistor R24 provides current to drive transistorT4 into saturation via diode Dl 6 when inverting buffer 9/1 goes to "1" as this buffer alone does not provide sufficient drive current. When transistor T4 conducts its collector current, defined by resistor R23 and the base emitter T3, ensures that even with minimum design gain sufficient collector current flows through T3 to operate solenoid 20. The resistor R10 not only limits the charge current of capacitor C5 but also limits the current drawn from bridge B2 when transistor T3 switches on. The diode Dl 6 increases the noise immunity of transistor T4 and resistor R55 provides a discharge path to ground.

The +5 volt supply line 81 is connected by line 83 to a comparator unit Q3 (e.g., MC 3423) consisting of two comparators a constant current source and a 2.6 volt reference. The line 83 is divided by resistors R21 and R22 to provide the input for the first comparator. The input of the second comparator is connected to the current source output of the first comparator and a delay element consisting of a capacitor C12. Normally the first comparator is in the 'on' condition, clamping the constant current source to ground. Consequently the second comparator, which connects the constant current source and reference voltage, is in the 'off' condition. When the line 83 exceeds a specified voltage, the input to the first comparator exceeds 2.6 volts so that it unclamps the constant current source with the result that the capacitor C12, connected between the constant current source and ground, charges until it reaches 2.6 volts whereupon the second comparator switches SCR1 via R17.

Therefore, upon firing the silicon controlled rectifier causes the line 84 to conduct and blow the fuse FS 1. The constant current, the 2.6 volt threshold of the second comparator, and capacitor C 12 produces a delay of 200 Ms before the second comparator changes state. Therefore, the output from the rectifier B1 is short circuited and regulator Ol is cut off if the voltage of the line 81 exceeds a certain value for longer than a predetermined time.

The first comparator in the unit Q3 is arranged to drain current through diode D17 and resistors R27 and R28 so that normally the voltage across a diode D1 and the base/emitter junction of a transistor T5 is not sufficient to turn T5 on. However, when the capacitor C12 charges the potential across D18 and T5 rises until it is sufficient to turn transistor T5 on. Thereby line 85 is rendered conductive and the NAND gates 2/1 and 1/1 ensuring the master control of the memories are shut down. Therefore, the memory data is protected before the power supply is interrupted by blowing fuse FSl.

The power supply to the regulator Ol can also be cut off by the microcomputer Q5 acting through the latch Q6 and a connection 86 between Q3 output of Q6 and the device 03.

A power fail/reset module 87 is provided to ensure that the microcomputer and memories operate safely during the appearance and disappearance of power and also to provide an advanced signal to the processing module 56 ensuring the correct sequence of events if there is a mains power failure.

Redundancy of the circuit elements is provided to ensure that a single component failure within the module 87 does not prevent correct failure signal generation. Thus the line 84 from the power supply module 65 is connected to zener diodes D1, D2 in parallel to provide current to bias respective transistors T1, T2 into saturation. The bias current of D1 (D2) creates a voltage across resistor R1 (R2) to provide a base current for T1 (T2) via resistor R3 (R4) which protects the transistors from excessive drive. These diodes and resistors are chosen to ensure saturation of T1 (T2) under all normal power up conditions. A "O" at the collector of T1 (T2) is converted to a "1" in a low power inverting buffer 10/1 (5/1).The outputs of these two buffers are combined in a NAND gate 11/1 to interrupt the microcomputer action at port INT by way of an inverting buffer 10/2.

The output of 10/1 (5/1) is connected via resistors R7, R11 (R8, R12), diode D5 (D6) and capacitor C2 (C3) to one input of NAND gate 11/2 (11/3). This is a Schmitt trigger NAND gate that ensures that slow changes on its inputs produce fast clear transitions on its output. When the output of buffer 10/1 (5/1) changes from "0" to "1" at power up, C2 (C3) changes via R7, Rl 1, (R8, R12) until the threshold of gate 11/2 (11/3). The output of 11/2 then changes from "1" to "0" because its other input is held by resistor R15 to the +5 line voltage at "1".The transition of 11/2 from "1" to "0" allows capacitor C2 to discharge via resistor R7 and diode D5 until the low threshold is reached and the output of gate 11/2 alters from "0" to "1". The discharge time of C2 is thus quicker than its charge time. The output of gate 11/2 is inverted in an inverting buffer 5/2 to produce the "1" level for X microseconds after power up and the "0" level Y microseconds after power down. The output of the inverting buffer 5/2 provides the other input ,to the gate 11/3 which does not change state until both inputs are at "1" so that it is the slower of the two timing circuits Rl 1, D5, C2 and R12, D6 C3 that is dominant at power up.This differs from the power down condition when it is the timing circuit that discharges faster than controls gate 11/3. The output of gate 11/3 is inverted by an inverting buffer 10/3 which feeds, on the one hand, a low pass filter and on the other hand, another gate combination 11/4, 5/3. The signal from gate 5/3 goes to an input CLR to clear the latch Q6.

The low pass filter comprises capacitors C6, C10, Cm 1, inductors L1, L2, resistor R16 and diode Dl 1 and its output is connected to the reset input RESET of the microcomputer. Clearly there is a predetermined delay between the interrupt signal through buffer 10/2 and the clear and reset signals through buffer 10/3. The primary purpose of the filter is to remove any spurious transitions from gate 11/3 or 10/3 that may occur as the +5 volt supply to the system rises to its normal operating level. The danger of a false reset signal arising prior to an interrupt signal is avoided. The filter also provides good filtering of general system noise, "cross-talk" and radiated bursts of noise at higher frequencies.

The input of gate 11/2 connected to R15 also provides a convenient point to produce a reset signal during fault finding or servicing without having to switch off the meter.

The power fail/reset module also provides adequate component failure detection, thus ensuring a complete shutdown of the meter to protect the memories. This is achieved by a combination of circuit design, with the utilisation of few locations in the memories to store component dependent timing flags, and by the execution of a fault tolerant/safeguard algorithm by the microcomputer.

In general, the redundancy of the circuit elements is provided to ensure that a single component failure either does not prevent the emergence of the correct failure signal INT at switch off, or does not prevent the generation of the RESET signal during switch on and switch off. There are some failures which may disable the microcomputer and the memories indefinitely but since these failures are not in anyway deleterious to the contents of the memories, they are considered as safe failures.

As in the memory modules, the gates of different packages are arranged in such a way that a failure of a single gate or failure of a complete package does not produce deleterious effects. The timing flags assigned to combinations of components such as R7, Rl 1, D5 and C2, along with other variant flags, are stored in the non-volatile memories Q2, Q4 before shipment of the meter. These flags are received by the microcomputer at switch on and compared with the flags stored in the ROM and any mis-match is interpreted as a failure. If there are no failures, the flags are negated and written back to memories.At switch off, the time between the emergence of a power failure signal at the INT input of the microcomputer and a RESET signal at the RESET input of the microcomputer is divided into segments assigned to the aforementioned flags respectively. At switch off, these time segments are simulated inside the microcomputer by the timer/counter in the microcomputer and generate an alarm inside the microcomputer after the elapse of each time segment.

When there are not faults and when power is switched off, fhe microcomputer, after elapse of each time segment, re-negates the assigned flags in the memories to their true states. It is, therefore, clear that the premature emergence of the RESET signal would have prevented the microcomputer from re-negating one or more of the aforementioned flags. Thus the un-modified flags would be detected at switch on and interpreted as failures and necessary action taken by the microcomputer.

The absence of the RESET signal after a predetermined time is also interpreted as a failure.

The component values of R7, R1 1, C2, R8, R12 and C3 are selected such that they provide a temporary cushioning effect under some failure conditions, i.e., the microcomputer and the memories are initiaiised safely even if there are some multiple failures such as D5 and D6 becoming short circuited.

To enhance the failure detection, an additional output feeds into the microcomputer at P05. This line has its' origins at gate 11/1 and enables the microcomputer to do adequate diagnosis. The diagnosis relies on the fact that failures of any component such as Tl , T2, D1, D2, 10/1 5/1, 11/1 and, to some extent the failures of R1, R3, R5, R2, R6 would manifest as a fault at the output of 11/1. The output state of gate 11/1 has a different significance at various time frames during system operation i.e., power-up, power-on and power-down. Any anomaly is again interpreted as failure of the module, for example if the microcomputer continuously sampled a logic "1" at P05, but both INT and RESET were absent for a predetermined time.

It will be seen that Figure 11, in addition to circles shown connected to various lines in the circuit which indicate memory and service interface test points, there are rectangles representing electronic weigh scale interface connections, and a number of broken circies which represent switches, for circuit isolation during fault diagnosis. Each memory Q2 or Q4 is provided with an interface connector to enable diagnostic equipment to be connected for direct interrogation during a fault state or for diagnostic purposes.

An important feature of the sOlid state circuitry shown in Figure 11 is that it can all be incorporated on one printed circuit board but a preferred solution is to use two printed circuit boards 70, 71 as shown in Figures 3 and 4 located on top and on one side of the meter where they are readily accessible and occupy very little space. The two printed circuit boards are joined by conductor 72 that may be provided by a flexible printed circuit board or some other connecting means. This particularly neat arrangement is primarily due to the fact that the contact strips of the encoding switches are integral with the printed circuit board 70 and the rotary parts 46 of these switches are rotatably mounted directly on the board 70.This printed circuit board also carries the push buttons 29 to 32, the alphanumeric display and the associated solid state units while the board 71 carries the microcomputer, memories and associated solid state units.

A socket 73 is provided for external access to the electronic system and a socket 74 is provided for mounting the franking meter on the base of the machine which carries the means for transporting and controlling the items of mail to be franked.

Turning now to what may be described as the software side of the system, so that the actual organisation of the system can be appreciated, the microcomputer Q5 has a ROM that incorporates meter software modules that maintain the various working registers which are mainly relevant to the timing aspects of the information gathering and output. The remainder of the ROM maintains .information which collectively describes the total system status. Within the system status a distinction can be made between the metering status and peripheral status.

The metering status is maintained by registering information contained in a RAM of the microcomputer Q5 pertaining to the monetary standing of the customer, e.g., credit, postal value used, number of mail items franked, as well as the meter's integrity, e.g., fault conditions. This information is copied into the RAMs Q2 and 04. The peripheral status is concerned with the communications link between operator and machine which depends on the thumb wheels, push buttons and the display.

The contents of the peripheral status registers in the microcomputer provide an interpretation of these peripherals at any instant. Since change of status of these peripherals is unpredicatable from the point of view of the microcomputer, this interpretation is the result of continuous scanning.

Decisions pertinent to the meter functioning are based primarily on the information in the peripheral status registers and also on the status of flags in a flag bank shown in tubular form in Figure 1 5. The software reacts instantaneously to changes in peripheral status, whereas a flag is considered only when a go/no go situation arises. A map of the flags and other areas of the microcomputer appears in Figure 1 6.

The software hierarchy is shown in Figure 1 3 which shows there are five main modules. The real time relationship between these is shown diagrammatically in Figure 1 2. Here the arrows indicate the direction of progression rather than time flow. Progression around arrowed lines is considered to occur in zero time.

Referring to Figure 12, the "power-up module" restores the metering status in the microcomputer to that which existed just prior to the last power down, i.e., the input from the battery maintained memories.

The "scanner module" gathers information from the peripheral systems and outputs status information to the display. This module primarily maintains the peripheral status registers and performs diagnostic operations. Timing aspects of the scanner module are indicated in Figure 14.

The "account module" controls the input to and output from the metering status whilst also maintaining records pertinent to the customer's account history.

The "memory input/output module" controls transfer of data between the microcomputer data memory and the external support memories Q2 and 04.

The "power down module" ensures that the meter status, as it was just before a power failure being anticipated, is properly stored in the support memories Q2 and Q4.

Diagnostic routines operate continuously during power on within the scanner module to protect the metering status against pollution and accounting error. For the support memory, the diagnostic is effected by data coding techniques involving the addition of checking bits and also upon comparison operations with the microcomputer's data memory. The check bits are removed upon memory read and are such that at least 3-bit errors are required within a word to have a chance of degenerating into another valid word. In the comparison check each support memory is independently comparable with the microcomputer data memory.

The thumb wheel diagnostic relies for its integrity upon the encoding switches described above.

The software distinguishes between invalid codes which are non-fatal (e.g. intermediate wheel positions) and those which are fatal (e.g., short or open circuits). Non-fatal codes inhibit the franking cycle without causing the meter to enter a "fault" mode as would a fatal code.

The diagnostic performs a checksum calculation on the microcomputer's programme memory and in addition does test arithmetic calculations that emulate those performed during a franking cycle account.

The power fail circuit diagnostic makes use of the power status inputs to confirm, or otherwise, the power on or power fail condition.

Claims (33)

Claims

1. A postal franking meter for a franking machine, the meter comprising a plurality of printing members allocated respectively to the numerical orders of the maximum value to be franked by the machine, each printing member being adjustable for printing any one of a series of digits (or fractions), manually operable elements respectively mechanically connected to said printing members for setting said printing members to print required values, a printed circuit board assembly comprising solid state units, including a microcomputer for effecting substantially all accounting operations required in the machine, said operations including registering descending values remaining credited for a customer and registering ascending values of postal value used, an alphanumeric display module for displaying information and values when required, said printed circuit board assembly comprising stationary arrays of conductors serving as encoding switch contacts, said arrays being allocated respectively to said printing members for encoding numerical values to be printed thereby, rotatable encoding switch units being mounted directly on said board assembly for traversing said stationary contacts in said arrays to select the values to be encoded, and mechanical means connecting said rotatable switch contact units to said manually operable elements whereby the operation of any said element simultaneously sets the associated printing member to a value to be printed and encodes that value.

2. A franking meter according to Claim 1, in which said manually operable elements are thumb wheels and said printing members are printing wheels, each thumb wheel being connected by a first rack and gear mechanism to the associated printing wheel and by a second rack and gear mechanism to a rotatable contact unit for encoding each value selected for printing by the associated printing wheel.

3. A franking meter according to Claim 1 or Claim 2, in which said arrays of stationary switch contacts and their associated rotary contacts are so allocated to said manually operable elements that for each such element two electrically connected portions are provided, these portions being respectively allocated to two 1 out of 5 codes for each decimal digit, whereby a rotatable contact unit can be set to select two 5-bit words respectively in each portion, each word containing one and only one "1" (or "O" for complementary coded switches), the stationary contacts being arranged to be electronically scanned under control of the microcomputer to derive an encoded decimal digit by combining said two 5-bit words.

4. A franking meter according to Claim 3, in which for each printing element an array of stationary encoding contacts is provided and in which for each array, the stationary contacts comprise on a flat surface five contact elements each consisting of two connected arcuate strips centered on and at different distances from the centre of rotation of said rotatable contact unit, with each outer arcuate strip overlapping the inner arcuate strip of the adjacent contact eiement, and two common rings also centred on said centre of rotation, said rotatable contact unit being arranged, for selection of each decimal digit, to bridge a corresponding one of said outer arcuate strips and one common ring, and also to bridge an inner arcuate strip of another contact element and the other contact ring.

5. A franking meter according to any one of the preceding claims, in which the microcomputer is arranged to verify a clear status comprising availability of sufficient credit and freedom from any fault condition prior to operation of said printing members to print a selected value.

6. A franking meter according to Claim 5, comprising means for signalling to said microcomputer the introduction of an item of mail into a machine carrying the meter, said microcomputer being arranged thereupon to initiate a printing cycle and start an accounting sequence comprising reading the value set in the encoding switches, adjusting registers of credit remaining and postage value used and reassessing the clear status applicable to the next printing cycle.

7. A franking meter according to any one of the preceding claims, in which said printed circuit board assembly and mechanical means connecting said manually operable members to said rotatable switch units and to said printing members are contained in a casing having a door arranged to be sealed by a post office authority, the casing when sealed serving to prevent unauthorised tampering with the accounting or printing action of the machine and a switch being located in the casing adjacent said door with means automatically operable on opening the door to actuate this switch for changing the mode of action of the meter to a post office mode enabling the authority to amend the credit registered under the control of at least one press button exposed by opening said door and of said manually operable members acting through said microcomputer.

8. A franking meter according to any one of the preceding claims, comprising press buttons manually operable to control the display of information in said alphanumeric display module, including information stored in a memory module, under control of said microcomputer.

9. A franking meter according to any one of the preceding claims, in which said printed circuit board assembly comprises two printed circuit boards joined by flexible conductors on other suitable connecting means, one said board extending over the top of the meter within a casing for the meter and the other said board extending over one side of the meter within said casing.

10. A franking meter according to Claims 8 and 9, in which said top printed circuit board comprises said stationary contact arrays and said rotatable switch units and switches operable by said push buttons and said alphanumeric display module, while said side printed circuit board includes circuits comprising a memory module and said microcomputer.

11. A franking meter according to any one of Claims 1 to 8, in which said printed circuit board assembly comprises a single printed circuit board.

12. A franking meter according to any one of the preceding claims, in which said stationary arrays of coding switch contacts are incorporated in a single printed circuit board.

13. A franking meter according to any one of the preceding claims, provided with a scanner driven under control of the microcomputer to enter values set by said encoding contacts in duplicate random access memories, said microcomputer being arranged during each franking cycle to up-date the memories and compare the memories as a check on their status.

14. A franking machine according to Claim 1 3, in which each memory is provided with a battery to support the memory in the event of power failure and is also provided with an interface connector to enable diagnostic equipment to be connected for direct interrogation during a fault state or for diagnostic purposes.

15. A franking meter according to Claim 3 or Claim 4 or to any one of Claims 5 to 14 when limited to Claim 3 or Claim 4, in which each array of encoding switch contacts comprises five contacts and in which five input lines by which signals are transmitted from the encoding switches to the microcomputer are respectively connected to the said five contacts in each array, the rotatable dual ,contact unit in each encoding switch being connected by a respective pair of scanning lines to a binary to decimal decoder having four inputs to which a binary-coded decimal sequence is fed from a binary counter under control of the microcomputer for the purpose of delivering two successive signals to each rotary unit, such that each rotary unit is scanned in sequence.

1 6. A franking meter according to Claim 1 5, in which at least one group of individually operable switches comprises switches respectively connected to at least some of said five input lines, said group being connected by an individual scanning line to said decoder.

1 7. A franking meter according to Claim 1 6, wherein in each cycle of the decoder, the total number of signal outputs from the encoding switches and the group or groups of the individually operable switches exceeds the number of input lines, the binary counter being arranged to provide time division mutliplexed scanning by way of the decoder.

1 8. A franking meter according to Claim 1 6 or Claim 1 7, in which feedback connections are taken from selected outputs of the counter to the microcomputer whereby the microcomputer is enabled to monitor the sequence of the scanning of the switches by way of the decoder.

1 9. A franking meter according to anyone of Claims 1 6 to 18, comprising at least one group of four switches connected respectively to four of said input lines and arranged to be scanned by a single scanning line from the decoder, the fifth input line being connected to provide a short circuit to said scanning line, or arranged in open circuit in relation to that scanning iine, and the arrangement being such that the encoded output from the four switches together with said fifth line is different from any of the encoded outputs from the encoding switches.

20. A franking meter according to any one of the preceding claims, provided with two independently packaged random access memories each independently connected to batteries under power fail condition to provide non-volatile operation independently of one another and provided with common address lines driven by the microcomputer and the same lines are connected via buffers for operating the alphanumeric display, each memory being provided with independent data input and data output connections and control connections enabling data to be written into and read from the memory by way of the microcomputer.

21. A franking meter according to Claim 20, in which the address lines are bi-directional.

22. A franking meter according to Claim 20 or Claim 21, in which the access of one memory is initiated under control of a gate operated from the micrcomputer and access to the other memory is under control of another gate operated from a latch controlled by the microcomputer.

23. A franking meter according to any one of Claims 20 to 22, in which each memory is provided with additional cells arranged to be used as dummy locations for performing test operations for validating the integrity of both memories before up-dating the memories and also for general system diagnosis.

24. A franking meter according to any one of Claims 20 to 23, in which the write control connections respectively of the two memories are taken from different outputs of the microcomputer, thereby preventing data corruption in both memories, the instructions for the generation of control signals for the two outputs being mapped onto two different spaces in a read-only memory in the microcomputer.

25. A franking meter according to any one of Claims 20 to 24, in which the data input and output lines are laid on a printed circuit board orthogonally to the address lines and the memories are each packaged and located apart.

26. A franking meter according to any one of the preceding claims, comprising a power fail/reset module having redundant circuit elements arranged so that failure of one of said elements does not prevent appropriate generation of interrupt and reset signals at the microcomputer in associated with the interruption and iniation of the supply of power to the meter, timing flags, associated with combinations of the said circuit elements, being stored in non-volatile memories, being at initiation of supply of power compared with flags in a read-only memory in the microcomputer, such that mismatch is interpreted as a failure, the flags being negated at various time frames after interruption of supply of power.

27. A franking meter according to Claim 26, in which the power faiVreset module has an additional output derived from signals dependent upon the state of components in pairs of redundant circuit elements by a gate which is connected to the microcomputer and which has an output state that differs at differing time frames during the operation of the meter, any anomaly being interpreted in the microcomputer as a particular failure of the module.

28. A franking meter according to any one of the preceding claims, in which the input/output capability of the microcomputer is increased by a D-type flip-flop serving as a latch manipulated by creating desired states at its respective inputs and by the generation of a pulse at an output of the microcomputer.

29. A franking meter according to any one of the preceding claims, in which independent memory supplies, with single fault isolation are provided respectively for two independent memories, the supplies being derived from one voltage regulator driven through circuits arranged to be connected to a mains voltage supply, the said regulator also being connected to supply other logic elements arranged to avoid the creation of any SCR latch-up effect in the input circuits of the memories.

30. A franking meter according to any one of the preceding claims, in which a diode is connected between a point where the voltage varies in proportion to mains voltage and a second point where the voltage is stabilised so that an abnormal increase in the mains voltage will cause the said diode to be over-loaded and destroyed thus providing evidence of this abnormal condition.

31. A franking meter according to any one of the preceding claims, provided with means whereby at switch on, a reset signal is given to the microcomputer only when the mains voltage exceeds a certain threshold level, and whereby, if after exceeding the said level and achieving a normal operating condition the mains supply falls below the said level, a power fail condition exists leading to the franking meter becoming inoperable.

32. A franking meter according to Claim 13 or any one of Claims 14 to 31 when limited to Claim 13, in which each memory is provided with a battery to support the memory and in which the memories and batteries are protected from over voltage by a fuse associated with means whereby the master control of the memories is automatically disabled prior to interruption of power supply by the fuse thereby protecting the memories.

33. A postal franking machine substantially as hereinbefore described with reference to the accompanying drawings.