Images at a verry long focal length, where the readout speed is critical If you want to take images at a verry long focal length, where the readout speed is critical, or you need a high readout speed for other applications like observing occultations, here are some examples, of the possible readout speed, using a ST-7: ( a ST-8 has twice the number of columns, so it will take more(about *2 ) time to read out a line) It should be used SCAN, version >=1.6 and , if the computer has a Pentium processor, the option Settings (33): "clock cycle counter Pentium" should be enabled, as the clock cycle counter has a higher accuracy. At Settings(34) put in the frequency of the CPU. Use the program to test the frequency of the CPU, avalible on my home page. To test the readout speed, fill in a very long focal length (100 000 mm or so) at Settings(10), attach the camera, and then take a drift scan image(4), for example 1000 lines, A message will be shown, after an image was taken, so the readout time can be controlled. At 2 x 2 binning it will take less time to read out a line, and it has to be read out a line half as often. The readout speed is limeted a last, by the parallel port. I made some tests. I have set the BIOS to bidirectional, EPP and ECP. SCAN will work in bidirectional, EPP and ECP mode. I found no significant differenc between that modes. Testing the time to read a line, I found an other problem, if the hard disc is fragmented, the time to store a line will increase, running the program under DOS. You can use smartdrv . (Should be in the \windows\ directory) , start DOS mode and then smartdrv. ( A hard disc cash) Or defragment the hard disk. Some tests: ( the results my vary about +- 0.001 ) Pentium II, 233 MHz , 1 x 1 binning, the time to read a line in sec: EPP Windows 0.0296 Wndows, using darksubstract and flatfiledcorrection 0.032 DOS without smartdrv 0.036 ECP Windows 0.0286 DOS without smartdrv 0.0364 DOS using smartdrv 0.0278 Bidirectional Windows 0.028 DOS without smartdrv 0.036 DOS using smartdrv 0.0278 Bidirectional After defragmenting the hard disk: In DOS mode, without smartdrv 0.029 Using 320x200 screen mode instead 640x480 ( Settings(16)) 0.027 Switching off the screen (press "c" after starting an image) 0.0265 Switching off the screen and using a ram-drive 0.025 I added a "planetary mode" function to SCAN. Settings(35) Only some of the columns will be read in. The results of a test: (under windows) columns to read in: sec 1 0.0065 100 0.0095 200 0.0124 300 0.0153 400 0.0182 500 0.0211 600 0.0240 700 0.0266 Of course, the field of view will decrease. (It is required some time to dump the pixels, which are not read in, so the ST-8 needs more time for the same number of columns) Some more tests with a Pentium "Celeron", 400 MHz: ST-7: SSP mode (set in the BIOS): DOS without Smartdrv 0.03338 sec Screen off 0.03337 sec DOS using Smartdrv 0.02587 sec Screen off 0.02452 sec ECP+EPP mode (set in the BIOS): Windows: 0.024788 Screen off 0.02356 DOS: without Smartdrv 0.0334 DOS using Smartdrv 0.025875 Screen off 0.024603 2 x 2 binning 0.01594 DOS using a ramdisk 0.024708 Screen off 0.023070 To make shure, that the lines are read out at the correct time, there should added about 10%. Controll the correct read out of the lines, in time stamp file, written by the program SCAN. ( \tstamp directory, *.txt) If a line or some lines are read to late, the following lines will be read earlier. (so it should not affect the quality of the image or astrometry) Only if a lager part of the image is reat out to late, there may be an effect on astrometry.(This my happen, if you run SCAN under windows, and switch to an other program, or an other program requieres porcessor time.) More examples: Computer 486 / 33 MHz : Read out speed up to 0.045 sec per line Computer 386 : Read out speed up to 0.27 sec per line ------------------------------------------------------------------- ------------------------------------------------------------------- You can also use the tracking CCD, the readout speed of that CCD is much higher, (may be used for observing oclutations, controlling the seeing, observing variable stars (flares?) or so (with turned on telescope dirve)) here are some examples: Columns to time in sec per line: read in 190 0.0071 150 0.006 100 0.00455 50 0.0031 25 0.0022 10 0.0019 1 0.0016 ---------------------------------------------------------- ---------------------------------------------------------- There may be a corelation between the start of the columns to read in at planetary mode, and the readout speed: There are read in 50 columns. First column:(start) time in sec per line: 0 0.0031 50 0.0031 100 0.0027 140 0.0026 There are read in 25 columns. First column:(start) time in sec per line: 0 0.0023 50 0.0020 100 0.0019 140 0.0017 -------------------------------------------------------------- I have tested the same for the imaging CCD and found something different: 600 columns, first column 0 - 0.0235 sec 600 columns, first column 165 - 0.0239 sec 300 columns, first column 0 - 0.015 sec 300 columns, first column 465 - 0.016 sec 150 columns, first column 0 - 0.011 sec 150 columns, first column 610 - 0.013 sec 50 columns, first column 0 - 0.008 sec 50 columns, first column 710 - 0.010 sec 1 columns, first column 0 - 0.0064 sec 1 columns, first column 760 - 0.0089 sec