Opmerkingen/Notes PRIMAX (rebadged TECO) JEWEL 4800 scanner (SCSI)

Main Index, teco2, VM3575, ICC profile
#Scanner Info, #Testreport

Scanner Info

• Label PRIMAX Jewel 4800 Model: VM3564, internal VM356A
  
  Scanner owner PRIMAX JEWEL (rebadged Teco)
  White backing sheet
  Scan format 8.5" by 11.7" (216 by 296 mm) (A4 size)
  Interface SCSI, termination internal
  Memory buffer 8 kByte
  Optical Resolution 300 by 600 dpi (with Primax win. software 1-4800 dpi default on 300 dpi)
  
  Image Data Color: 30 bits per pixel
  Image Data grey : 10 bits per pixel, red, green or blue filter selectable
  Image Data bi-tone: 1 bit per pixel, red, green or blue filter selectable
  Voor bi-tone mode 8 dither modes selectable
  Monitor gamma value ranges from 0.70 to 3.0 (Primax win. software)
  Shudown time for automatically switch of the lamp is selectable (hh:mm:ss) (with Primax win. software and Primax SCSI card, with other SCSI card?)
  Reflective or Transparency scan is selectable
  
  Other scanners with almost the same hardware/software internal:
  RELISYS AVEC II S3 VM3564, RELISYS APOLLO Express 3 VM356A
  
  Software: SANE backend teco2.

  Return/Terug Top

Testreport

• System info: Linux, Debian (Sid), kernel 2.4.18-586tsc, processor Pentium 233 MMX
  SCSI Interface Card Adaptec AVA1515 ISA slot or Adaptec 2940U PCI slot for scanner
  cat /proc/scsi/scsi
  Attached devices:
  Host: scsi1 Channel: 00 Id: 06 Lun: 00
  Vendor: Primax Model: Jewel Rev: 1.00 Type: Scanner ANSI SCSI revision: 02
  
  
• 23-07-2002 Test performed with Sane 1.0.8, to get some information on the scanner.
  1. Test with sane-find-scanner -v -v(as root) primax-header.txt
  
  
• 24-07-2002 SCSI traces to get some information on the scanner with SCSI Tracing Tool 2.1 from Ed Hamrick http://www.hamrick.com
  1. First SCSI traces primax-scsi-traces-1.txt
  
  
• 27-07-2002 Some test done with modified teco1, teco2 and teco3 backend.
  Result
  1. Teco1 driver with patch, xscanimage will startup, sometimes a sort garbage scan is possible see also logfile
  Logfile teco1 logprimax-teco1-1.txt.gz
  
  2. Teco2 driver with patch, xscanimage will startup, scanner motor reacts wird, problem with gamma and windows value, see logfile
  Logfile teco1 logprimax-teco2-1.txt.gz
  
  3. Teco3 driver with patch, xscanimage will not startup, see also logfile for error
  Logfile teco1 logprimax-teco3-1.txt.gz
  
  4. Second SCSI traces primax-scsi-traces-2.txt
  
  
• 29-07-2002..30-09-2002 Scsi traces collected into Primax.desc file version 0.20
  also info from logfiles used and teco2.txt file.
  1. Description file primax.desc
  
• 01-09-2002 Added to file sane-find-scanner info of a Relisys Apollo Express 3 scanner,
  outside label is VM3565, inside label VM356A which is the same as the Primax Jewel 4800, outside VM3564 inside VM356A.
  
  r-apollo-header.txt
  The Relisys Apollo Express 6 (VM6565, internal VM656A) is supported by the teco2 driver.
  
  

• 09-09-2002 Some test done with modified teco2 backend (from version 05-09-2002) see diff. files. gamma routine disabled, window value 56, a sort of garbage scan is possible when teco_wait_data_read byte is changed from 80 to 19.

  Logfile teco2 logprimax-teco2-2.txt.gz
  Diff file teco2.c, teco2.h and teco2.conf teco2-primax-1.diff
  

• 29-09-2002 Debuging setup for SCSI traces: with Adaptec AVA 1515 isa card, win95 aspi2dos.sys collecting traces for the VM3575 is possible (in 16 bit mode with winaspi.dll and scorsup3.exe and in 32 bit mode with winaspi32.dll and scandy). Switching between 16 bit and 32 bit mode is possible but can generate errors. For the VM3564(A) the same configurations doesn't work yet (error TECO caused a general protection fault in module WINASPI.DLL at 0001-0000059f. Registers .... For the winaspi32.dll the error is access violation at address 00E2164B, read adress FFFFFFFF. The normal Jewel ASPI drivers is Sdr2aspi.sys a ASPI manager for MS-DOS version 1.15, makes no difference.

• 06-10-2002 Scsi traces collected and placed into Primax.desc file version 0.35
  

  Stil problems with collecting scsi traces, now also apollo, avec windrivers installed. When apollo driver is started scanner is recognized as a Apollo Express 3 but fails when the selftest is started. Because of installed avec and apollo drivers, scandy will startup in 32bit mode but no complete scan is possible

  1. Scsi traces scsi-traces
  2. Description file primax.desc
  
  
• 18-10-2002 Primax Jewel header text version 1.011.01 primax-header-1.txt
  

• 19-10-2002 First b/w scan with Primax Jewel 4800, 150dpi, color reference chart [5-8], sane 1.0.9pre1, see logfile and scan , problems some black lines in scan, not a complete A4 sheet is scanned, blocks of data is shifted. In teco2.c file gamma status = 0 is added, teco_wait_for_data return value changed from == 0x80 to > 0x17.
  1. Logfile logprimax-1.txt logfile from 150dpi scan
  2. Scanfile scan-primax-1.png scan 150dpi
  

• 20-10-2002 2nd b/w scan with Primax Jewel 4800, 300dpi, color reference chart [5-8], sane 1.0.9pre1, see logfile and scan , problems some black lines in scan (calibration 10200, 2,), In teco2.c file gamma status = 0 is added, teco_wait_for_data return value changed from == 0x80 to > 0x17. with GIMP checked with 50% OK, with 20% etc. NOK. It seems that 0x18 lines are available but with 150 dpi 0x33 lines are read and placed and subtracted from the bytes to read or something like that. With 300 dpi 0x18 lines available but 0x19 read. For 300 dpi i have now almost a complete A4 scan (150 dpi only a part is scanned)
  1. Logfile logprimax-2.txt.gz logfile from 300 dpi scan
  2. Scanfile scan-primax-2.png b/w scan 300 dpi

  Values voor calibration (2550, 12), NOK yet
  Changed value 0x17 to 0x02 for scanning grey and color mode.
  Grey scan in 300 dpi is possible, output is visable but NOK(gamma, calibration, position of data), for 150 dpi output is garbage.
  Color scan in 300 dpi is possible, output is visable but NOK(gamma, calibration, position of data, color problems), for 150 dpi output is garbage.
  

• 26-10-2002 Diff file teco2.c, teco2.h and teco2.conf teco2-primax-3.diff
  update description file primax.desc
  
  

• 27-10-2002 MKSCSI_GET_DATA_BUFFER_STATUS added to teco_fill_image, update diff file teco2.c, calibration subroutine disabled, for color scans, color sequence has to be changed and color skewing has to be checked and changed, 150 dpi scans are now possible

  teco2-primax-3.diff
  

• 28-10-2002 Done some color dpi value tests, 150, 225 and 300 OK, see teco2.c file teco2-primax-1.tar.gz

  
  
• 30-10-2002-31-12-2002 Update description file primax.desc
  

• 03-11-2002 Update teco2-primax.c file, calibration routine NOK yet, calibration value 6666 added to remove vertical lines, gamma routine disabled otherwise program will hold (waiting on gamma status?), teco2-primax-2.c.gz
  

• 08-11-2002 Update teco2-primax.c file, calibration routine with 8 bit mode patched yet, gamma routine disabled otherwise program will hold (waiting on gamma status?), teco2-primax-2.c.gz
  bw, gray and color scans are possible, color scans only in 150, 225 and 300 dpi.

• 24-01-2003 Update for 150, 225, 300 dpi value changed from (1, 0, 2, 1, 1) to (1, 0, 2, 1, 2), patch added for calibration routine with value 16432, see file, tested with sane 1.0.10-pre2, gcc 3.2
  Change and check patch from Alex Wulms:
  994c994
  < tmp_buf[j] = ( 4206639 * dev->def->cal_lines ) / tmp_buf[j];
  > tmp_buf[j] = 0x1000 - (tmp_buf[j] / dev->def->cal_lines);
  For the 356A instead of 4206639 value between 14432 and 12432.
  teco2-primax-3.c.gz

  Further test with value 14432 and 12432 gave as result:14432 background is white, 12432 background begins to change into light pink, further test with 14000 gave good result.

• 30-01-2003 For tmp_buf[j] = colsub - (tmp_buf[j] / dev->def->cal_lines); when colsub =0x110 also the result of the scan is improved.
  

• 04-02-2003 Update description file primax.desc
  

• 14-02-2003 Update patchfile teco2.c teco2.h and teco2.conf for VM356A (Primax Jewel 4800, RELISYS APOLLO Express 3) and VM3564 (RELISYS AVEC II S3)
  Only Primax Jewel 4800 is tested (only colormode 150, 225, 300 dpi is OK)
  Code for APOLLO Express 3 and AVEC II S3 is patched in files but status untested
  teco2.conf with patch for VM3564/VM356A
  teco2.h with patch for VM3564/VM346A
  teco2.c with patch for VM3564/VM356A

• 19-03-2003 Update patchfile teco2.c for VM356A (Primax Jewel 4800, RELISYS APOLLO Express 3) and VM3564 (RELISYS AVEC II S3)
  Only Primax Jewel 4800 is tested (color table colormode 50 dpi is NOK, and colormodes above 300 are stretched), also VM3575 is tested
  Code for APOLLO Express 3 and AVEC II S3 is patched in files but status untested, patch for VM6575 colormode is not (yet)included
  teco2.c with patch for VM3564/VM356A

• 27-04-2003 CVS version 26042003 tested, the patches for the VM3564/VM356A are included in the teco2 backend and will be released with SANE version 1.0.12.
  

• 25-05-2003 Some checks done on 300 and 600 dpi value, for 300 dpi change values 1 0 2 1 2 to 1 0 2 1 4, for 600 dpi change values from 1 0 2 1 2 to 1 0 2 1 8 (line 200 and 211 teco2.c sane 1.0.12).
  

• 30-05-2003 Scan made with Sane 1.0.12, teco2, 600 dpi, 1 0 2 1 8, test horizontal lines, still some color skewing on color 600 dpi mode
  test-hor-8-short.pnm file (1.2Mbyte)

• 05-06-2003 Scan made with Sane 1.0.12, teco2, 600 dpi, 1 0 2 0 0, test horizontal lines, color skewing NOK, with 1 0 2 1 8 best value
  co600-raw0-10200p.pnm file (2.4Mbyte) co600-raw1-10200p.pnm file (2.4Mbyte)

• 09-06-2003 Update teco2.c Reason: bug around line 2740
  dev->def->color_adjust[i].resolution != dev->x_resolution;
  should be
  dev->def->color_adjust[i].resolution != dev->y_resolution;
  
  For the VM356A to get a better scan result use values from below:
  {25, 1, 0, 2, 0, 0}, {50, 1, 2, 0, 0, 1}, {75, 1, 0, 2, 1, 1}, {150, 1, 0, 2, 1, 2},
  {160, 1, 0, 2, 1, 2}, {225, 1, 0, 2, 1, 3}, {300, 1, 0, 2, 1, 4}, {305, 1, 0, 2, 1, 4},
  {310, 1, 0, 2, 1, 4}, {315, 1, 0, 2, 1, 4}, {375, 1, 0, 2, 1, 5}, {380, 1, 0, 2, 1, 5},
  {385, 1, 0, 2, 1, 5}, {450, 1, 0, 2, 1, 6}, {455, 1, 0, 2, 1, 6}, {525, 1, 0, 2, 1, 7},
  {600, 1, 0, 2, 1, 8},
  range from 305 to 600 ok, but wrong y proportion
  

• 20-06-2003 updates files (not yet in cvs) teco2.h -white_level added
  updates file teco2.c -custom gamma disabled for VM3564 and VM356A, no gamma function for these scanners
  -SANE_TECO2_CAL_ALG0 values added for 0 and 1 mode for VM3564 and VM356A, default is 1
  -white_level added for VM3564, VM356A and VM3575 (SANE_TECO2_CAL_ALGO mode 0 and 1)
  -color dpi tables update, correct 600dpi color scan for the VM356A and VM3575 possible (only scaling problem left)
  -values for the VM3564 not tested
  teco2-20062003.h
  teco2-20062003.c
  teco2-20062003.conf

• 21-06-2003 updates teco2.c file (update some text color table, files not yet in cvs)
  teco2-21062003.c
  

• 25-06-2003 updates teco2.c and teco2.conf file (update some text table and name scanner selection added used in frontend for example xsane and quiteinsane, files not yet in cvs)
  teco2-25062003.c
  teco2-25062003.conf
  

• 20-07-2003 updates teco2.c, teco2.h and teco2.conf file are placed in CVS.
  

• 15-01-2004 For info updates etc. see teco2 page
  Return/Terug Top

Copyright © 2002 - 2006. Alle rechten voorbehouden, Revisie: 13 November 2006.