7527 Code History
7527 MICROCODE PROBLEMS/FIXES
This file describes the bug fixes incorporated into each
successive release of the 7527 Data Collection Terminal's
ROM microcode.
------------------------------------------------------------------
A1) MAIN MICROCODE
NOTE: See also "A2) CO-PROCESSOR CODE" below.
Version V01.22 / 1.22M 01 Nov 96
**** IMPORTANT NOTICE ****
SEE WARNING UNDER FIX #2 IN
THE 01.21 VERSION...
1) If the display contrast was adjusted via the keypad during
normal use of the terminal, then upon the next power-on
restart the 7527 would clear all memory and require re-setup
and download. This was caused by missing update of the set-up
checksum when the contrast was adjusted in this manner.
Workaround: Go into the setup mode after adjusting the
contrast and make some change to any setting (say, beeper volume),
set it back the way it was, then exit set-up. This causes
set-up to re-calculate its checksum (taking into account the
new contrast set point).
Version V01.21 / 1.21M 19 Jan 94
**** IMPORTANT NOTICE ****
SEE WARNING UNDER FIX #2 IN
THIS VERSION!!!!
1) Fixed 2 bugs which gave about a 5-10% chance of loosing
your download of Extended Terminal Services and all your
data files each time the power to the 7527 was cycled.
This level is recommended upgrade for installs where there
are frequent power failures or the user frequently turns off/on
the machine.
2) Fixed bug in application (e.g. ETS) code space cyclical
redundancy checking algorithm which was preventing the CRC
from finding corrupted downloads or otherwise corrupted
applications.
*** WARNING *** Due to the CRC checking in the microcode,
this level is incompatible with older versions of DCC/2,
Extended Terminal Services, Data Collector, and AIX Data
Collector/6000. You must update your system software:
** All of the below need Extended Terminal Services 1.01i
or later.
** DCC/2: Release 1.02 and/or service level UN50609
AND install the PTF for DCC/2 which includes DCCLOAD.EXE
dated 9/15/93 (1.01i comes with the DCC/2 PTF diskette).
** Data Collector: 1.12 plus a PTF from the support center
** AIX Data Collector/6000: 1.00 plus a PTF from the
support center.
** Other, non-IBM applications which download ETS and
have followed properly the instructions in the ETS
manual for use of the .MAP file for calculation of
values in the downloaded program ACS will be OK.
Improper values in the ACS (code start segment and code
size) will result in rejection of the application file
and a cool-reset of the 7527.
See 1.14M description regarding content of 1.21M.
Version X01.20d
1) Fixed bug #1 as shown in V01.21, but did not include
CRC checking fix (#2) from that level, so you do not
need any special ETS, DCC/2, Data Collector, etc. level
to use this level microcode.
Version V01.20 / 1.20M
1) No substantive changes over 1.16 unless you are using
PAR Microsystem's 'CIMPrint' product.
See 1.14M description regarding content of 1.20M.
Version V01.16 / 1.16M
1) The 7527 will occasionally lose all files and ETS during/after a power
loss or failure. Several additional areas of the NMI code and re-start code
were enhanced to avoid this loss of files.
See 1.14M description regarding content of 1.16M.
Version V01.15 / 1.15M (BETA)
1) The 7527 will occasionally lose all files and ETS during/after a power
loss or failure. Several areas of the NMI code and re-start code were
enhanced to avoid this loss of files.
See 1.14M description regarding content of 1.15M.
Version V01.14 / 1.14M (BETA)
1) Support previously available only in 1.11a for RTS/CTS pacing of modems
attached for host connection to the RS-232 port has been imbedded with
all the fixes prior to this version and after 1.11a. Version 1.14
has regular use of the RS-232 port, while 1.14M forces the RS-232 port
to use the modem pacing (even if the RS-422 port is defined for host use).
Thus, downstream (non-host) devices attached to the RS-232 port when
using 1.14M will need to support the modem line pacing also.
The 'M' version imposes modem control line requirements as
follows:
RTS is activated when the 7527 wants to send characters
CTS must go active in response before the 7527 will transmit
DTR is activated upon startup and remains active
CD must be active in response to DTR.
You may wrap RTS back to CTS and/or DTR back to CD on the 7527
if CTS and/or CD is not provided by the attached equipment
Version V01.12k (BETA)
Test of imbedded RTS/CTS support for RS-232 Host Port connnection.
See the 'M' suffix versions described above.
Version V01.12j (BETA)
1) Occasionally (1-5% of the time), input from a laser scanner would not result
in any observable behavior from the terminal; i.e. no indication of a good
read or a bad read. This was traced to failed communication between the main
microprocessor and the co-processor within the 7527. This failure has been
remedied.
Version V01.12i (BETA)
1) Sometimes, a CMD 0, which is supposed to cause a warm reset, would appear
to fail to reclaim all of memory. A cold reset was required to correct
the condition. This would result in a failed re-download of ETS.
This has been fixed by performing a destructive memory test over all
memory regions except that area which contains the configuration
parameters. That area is tested in a non-destructive fashion. Also, we
reinitialize all BIOS variables from the ROM data area, in the same
manner as during a cold reset.
Version V01.12h (BETA)
1) After a variety of circumstances, the communications subsystem of the 7527
would fail to respond to polls, until it was cold reset. This can result
in getting DCC0642 errors at the PC running DCC/2.
We know of the following potential reasons:
1. A watchdog reset while a network packet is being processed
2. A failed download and subsequent internal warm reset while a
network packet is being processed
3. Power fail while a network packet is being processed.
2) With certain magnetic stripes, the record was incorrectly decoded.
We determined that the algorithm was not enforcing the leading quiet zone
in front of the start sentinel character. This allowed the start sentinel
to be detected in the middle of the bit stream. It was just chance that
the next several 5-bit bytes formed a valid record. After all, the 5-bit
LRC has a 1 in 32 chance of being right.
The fix consisted of ensuring that a quiet zone immediately preceeds the
start sentinal. We chose a minimum of 8 zero bits and will verify up to
15 zero bits of quiet zone. The standard calls for 20 bits at the leading
and trailing edge, and 15 bits between redundant records.
3) While sending data INTO the 7527 RS-232 port from another device,
characters would occasionally be lost.
A situation was discovered in the interrupt handler in which a write to an
index register (which selects the next register to address) could be
followed be a reentrant interrupt from the other channel thereby losing
the value of the index register. The sequence was made a critical section
by disabling interrupts during the port register access.
Version V01.12g (BETA)
1 Modem control register status' incorrectly reported with
ComGetMdmCtl.
Workaround: Contact hot line for sample CFR code.
2 If a terminal misses the NAK to a transaction that it was
meant to receive (i.e. due to comm line noise, etc.), it
may interpret the next ACK/NAK it sees on the line (even if
meant for another terminal) as being its own. This could
result in a lost transaction if a ACK was used instead of the
missed NAK.
NOTE: Extended Terminal Services "Transaction Release Mode"
(using command K) will prevent any transaction from being lost
entirely -- it will only be delayed by a few seconds.
Workaround: Use "transaction release mode" whereever possible.
3 The J9 (configuration query) command returns incorrect data for
the model 002.
The second parameter of the J9 command, 'display type' as documented
should return the following:
00 = 2 line display
01 = graphics, normal video
02 = graphics, reverse video
However, a Mod 2 terminal returns 01 every time regardless of the setting
of the display mode of the terminal.
Workaround: None.
Version V01.12f (BETA)
1 Fixed printer support, which was broken in 1.12c, d, and e.
Workaround: Don't use version 1.12c, d, or e.
Version V01.12e (BETA)
1 Magnetic stripe decoding improved read rate (especially when
using magnetic wand, where reading may not always start at
begining of the stripe).
Workaround: Be sure to scan complete magnetic stripes, starting
in the synchronization character area.
2 Fixed window where power shutdown may leave timer interrupts
to microcode routines disabled (routines which use TmrSetInt,
which includes ETS's DI-initiated transaction monitoring).
Workaround: None. Machine must be cold-started.
Version V01.12d (BETA)
1 Code 128 decoding when Shifts between character sets were in the
symbol.
Workaround: None.
Version V01.12c (BETA)
1 Eliminates operation of DO points upon power-up and Shift-Enter-Clear
key sequence (Note: Printer support was broken in this beta version,
and was fixed in 1.12f).
Workaround: None, besides don't tell operators about un-documented
Shift-Enter-Clear keypad reset function.
Version V01.12a (BETA)
1 Improved bar code reading when colored overlays on the bar codes
makes the first bar appear to be very large.
Workaround: Add 'set-up' bars to bar code to trigger the proper
'white area' response from the bar code scanner, or make badges
with the security overlay extending all the way to the edge of the
badge.
Version V01.12 (BETA)
1 Power interruption may cause loss of network communications.
A small window of mis-opportunity existed that could leave
communications disabled is power failed at just the wrong time.
Workaround: None. Machine must be cold-started.
Version V01.11a
Enhancements/new features:
--------------------------
A Uses RTS/CTS pacing when the RS-232 port is used as the host port.
This is specifically done for broad-band modem support.
See 1.16M for latest version.
Available by request only.
Version V01.11
1 If a terminal application program is loaded, and if
the user trys to enter KEYLOCK mode during the initial
power-on screens (Copyright, Version number, etc.)
the 7527 will crash, losing the application program
and data, and possibly requiring removal of the battery
to recover.
2 USD-2 bar code decoding does not work with any character
above "J" in the bar code. Workaround is to use the nearly-
identical 3 of 9 bar code symbology
3 Long Code 128 bar codes will not decode. If the symbol length is more
than approximately 32 characters, the checksumming does not function
properly. Shorter Code 128 codes will work properly.
4 MSI bar codes with Modulo 11/10 checksums will not decode.
5 Customer Diagnostics for DI/DO has been "enhanced" so that it sets all
points to input and constantly updates the "1"/"0" status, so you
can flip your input points and see the transitions.
Version V01.10
This was the original shipment version for the 7527
A2) CO-PROCESSOR CODE
Version 20.00
1 Intermittently, a 7527 will stop receiving data through a magnetic hand
scanner. The hand scanner lights will remain on, but no data is
transmitted when a stripe is scanned. Powering the device off and back
on corrects the situation.
Workaround: The cause appears to be radio frequency emmission from
nearby radios, which causes an overflow of interrrupt to the coprocessor.
Radios should be removed from the area.
====================================================================================