Here the description and solution of some problems that can (but
not must) occur with FX.

Because this is a lot of text, it could be wise to print it out.

- General instability also at 16MHz
- General instability at 32/36/40MHz CPU clock.
- FPU problems
- IDE problems
- RAM/SIMM problems
- SCSI/DMA problems
- Heat problems and resulting instability.

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
- General instability also at 16MHz

-- Maybe the DSP doesn't like the 50MHz clock.

   Solution: The clock wire that leads to the DSP has to
   be connected at the right (hanging in the air) side of L102.
   Then the DSP gets the system clock (32/36/40MHz).

-- Maybe the main clock of 32MHz (taken out of the falcon) is a bit
   weak, so that it is disturbed by the FX.

   Solution: Dissolder the red wire at the right pin of L102 and connect
   it (with a longer wire) with the right lower connection of the 32MHz
   oscillator near the DSP. This is the source of the connection which has
   to be separated for the DSP accelerator.

-- Shorten all wires as far as possible.


++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
- General instability at 32/36/40 MHz CPU clock

-- Different setting of the 123I-buttons in the CPX.
   In order to work properly at doubled CPU clock (32/36/40MHz) the
   Falcon needs to have a certain phase relation between system clock
   and CPU clock. The functioning gap is so narrow that a difference of
   8 nano seconds can lead to immediate crash.
   This correct phase relation is not the same in each Falcon.
   To make acceleration possible, there are different delays can be activated
   in FX to shift CPU clock from system clock.
   Button 1 delays approx. 9ns
   Button 2 delays approx. 4ns
   Button 3 delays approx. 2ns
   Button I inverts the clock (i.e. 12.5ns at 40MHz)
   With that the phase relation can be adjusted very fine. If the machine
   for example crashes with the default '123I' adjustment, you may play
   around with that setting. '' (no button activated) may be worth a try.

-- Since V1.1: Setting DM2 in the CPX.
   Through tolerances it is possible that switching the clock does not work
   cleanly and clock jitter can happen. This can lead to instabilities
   or to 'flashes' on the screen.
   Since V1.1 it is possible to activate a different mode of clock switching
   with the DM2 button. This leads to a performance loss of approx. 1-2%.

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

- FPU-problems:
  Should the FPU work not at all or unreliably (check this with
  appropriate software) there are 2 possibilities:

-- FPU fails already at 32/36 MHz:
   16MHz-FPUs should not be clocked higher than 25MHz.
   If you use such a FPU you should get the clock not from L102
   but from the upper connection of R217. Then the FPU always gets
   half the system clock (i.e. 16/18/20MHz). Alternatively the wire
   can also be soldered to the free upper right connection of the
   25MHz oscillator (U41) unter the IDE hard disk.

-- FPU fails at 40MHz system clock:
   Probably the clock amplitude is to small.
   Solution:
   - Shorter wire from 68-Ohm resistor to R221
   - Use a smaller resistor to the FPU instead of 68 Ohm. (57, 47 or
     39 Ohm). For unknown reasons (ATARI...) it can then be that the
     screen may be disturbed by some flashes, when pixelclock=system clock.

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

- IDE-Problems:
  For the IDE-drive in the Falcon being controlled by a quite strange
  circuit it is possible that, under certain circumstances (mostly
  older Seagate drives with 40-65MB), these drive will not deliver
  proper data.
  This may be tested by pressing the Escape key on desktop while
  a window from the questioned drive os opened. Watch the total number
  of bytes displayed in the windows info-line. When this number changes
  sometimes, don't trust the drive and don't start programs from it.
  A better test is the program MEMTEST.TTP enclosed with Kobold, the
  fast file copier.

  Solution (in V1.01):
  By activating DM1 in the CPX a clock reducement for IDE accesses
  can be activated. This is a rather brutal method which could also
  lead to crashes...
  In case that should work, test your drive carefully again.

  Other solution:
  Install the miniprogram IDEAL.PRG in the AUTO-folder behind
  FXDRIVE.PRG.

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

- RAM/SIMM problems:

-- If FXDRIVE announces no or too few memory at boot time, check the
   configuration in FXCONF. Remember to SWITCH OFF the Falcon after
   changing the configuration!!

-- If FXDRIVE announces defective SIMMs, the error message gives a hint
   on the defective memory bank and the bit error mask.
   If only one or two bits are set in the mask it is probably a contact
   problem of the SIMM sockets. Gentle(!) bending down and cleaning the#
   contacts should help.
   Are even 8bits at once '1', the SIMM itself is not working properly.

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

- SCSI/DMA/Sound problems:
  Occur mostly at 40MHz system clock and can be checked via MEMTEST and
  appropriate SCSI drive.
  If the clockpatch using the 74F04 is not installed this should be done
  first!

  Solution:
  As described in DMA_PROB.GB

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

- Heat problems and resulting instability

  By a too strong rise of CPU temperature the critical phase relation
  is changed, which can lead to increasing instability. This has got
  nothing to to with 'overheating' the CPU. In a properly designed
  68030 system that would do nothing...

  Possible solution:
  1) remove and let away the upper shielding iron.
  2) remove the Falcons fan from the original position (which is quite a
     useless one) and move it between ROM-port and STE-joystickports, so that
     it can blow below the FX thereby optimally cooling the CPU. The necessary
     +12V can be found at the right connection of C110 above the right foot
     of the power supply. There you can solder the red wire of the fan.
     The black wire can be soldered to the border of the mainboard.


 Georg Acher / BlowUP  11.11.95


