Historial issues:
02/02
I've been playing with the ethernet chip and usb for the past two days
and I'm having trouble. Here's what I am seeing:
- the usb supply voltage to the connector needs to be +5; it's currently
switched from +3.3 and this is not working. I replaced the switched +3.3
with +5 and things get better; I believe the assabet uses switched +5
- the D+ and D- lines are switch; on the schematics USB_MINUS connects
to D+ and USB_PLUS connected to D-;
- I'm fairly convinced one of the signal lines, D+ or D- does not
connect to the SA-1111. This may be due to a bad connector. Since I
dont' have a flex connector to use between the cpu board and the basic
board, I can't diagnose this problem; If you get me a flex connector
I can check it out further. Or, if you can check that would be great.
- The interrupts from the UCB1200 and CS8900 are not working
correctly. It's not a problem with the SA-1111/SA-1110 interrupt line
as USB interrupts seem to be fine. The problem seems to be that the
CS8900 interrupt line is not connected to GPIO_C(2).
If I make this pin an output I do not see any interrupts. If I make
it an input, I see constant interrupts, but if I ask the SA-1110 for
the status of the line it says it's 0. I have it programmed as a
rising edge interrupt. Interestingly, if I turn on the interrupt in
software but don't enable the CS8900, I see sporatic interrupts on
this line, somewhat associated with keyboard input. I can't explain
that. I know that GPIO_C(1) is connected because I can turn the LED
on and off. Also, while I am getting the constant interrupts I look
at the CS8900 INTRQ0 line and it is zero with no activity. This make
me wonder if the GPIO_C(0) is floating - again, it may be a connector
problem.
I'm seeing similar things with the UCB1200 interrupt.
02/18/02
My main problem right now is that my system is somewhat unreliable.
The CF and PCMCIA connectors seem to be very touchy. If they are not
in exactly the right position the system won't boot. This makes
debugging really hard. I was hoping to get a PCMCIA flex connector and
maybe a new basic board. I will try and clean up the connector solder
pads but I'm not sure this will help...
11/27/01
Ethernet board working, with address decode and buffer hacks
10/22/01
Current status
CPU BOARD:
memory, cpu, serial port 3 working
flash supported via jtag program and linux kernel (mtd)
USB function not tested
BASIC BOARD:
VGA output working
CF and PCMCIA working
+12 not functional on PCMCIA
USB host does not work; signal levels seem wrong, usb pwr not high enough
serial port 2 untested
i2c untested
PERIPHERAL BOARD:
Ethernet chip does not work; can't seem to select registers
PS/2 mouse & keyboard work (use very short cable)
audio out works; some issue with back-to-back audio
audio in untested
touch screen untested
IR untested
10/01
wavelan on pcmcia works; have a stable file system and ram disk
got ps/2 mouse & kb to work; some keyboards would not work; required
very short cables from basic board to socket.
can't get usb or ethernet to work;
09/18
Pinout of the connector on the Adapter board
__
|
GPIO_18 GND
GPIO_17 LCD(0)
GND .
GPIO_B(1) .
GPIO_B(0) .
L_BIAS .
L_PCLK .
L_FCLK
L_LCLK LCD(7)
GPIO(9) GPIO(2)
. .
. .
GPIO(6) GPIO(5)
|
-
9/01
got pcmcia working with new pcmcia board and by removing cable from
cpu to basic board. can now boot from cf or pcmcia.
7/8/01
fixed gpio problem; cf now sees microdrive. ide-cs.c has an incorrect
device id string, however; /etc/pcmcia/*.conf looks for "ide_cs" and the
driver specifies "ide-cs"...
wired up vga cable; can't get vga output to sync w/monitor. signals looks
reasonable on scope (didn't verify timing, however).
7/7/01
New cpu card fixed SKIP problem;
PCMCIA & CF, voltages are correct but status reg is zero; something
is resetting sa-1111 via gpio...
7/5/01
SA-1111 SKID register is not reading back correctly; suspect h/w
7/4/01
kernel boots; ramdisk works; sorting many unix driver issues out.
7/2/01
kernel boots
6/12/01
- check bcs_n(5) on ethernet card; is this signal connected to a buffer?
5/29/01
- Several of the schematics do not have pin numbers on them and this
makes the task of mapping connectors difficult and error prone.
- There is no documentation on the use of the small buffer board. I
assume I need to use the small buffer board to connect between the
JTAG on the CPU board and a PC parallel port.
[yes - the small buffers board is needed to access the JTAG ports]
- The small buffer PC board is not "plated through" (i.e. the vias
have no metal in them). Connectors need to be soldered on both sides
of the board and vias need to have wires solder through them for this
board to work.
[true]
- The order of pins on the J1 connector on the small buffer board does
not allow a simple ribbon cable to be used with headers on each end,
since the signals do not line up. If the signals on J1 where placed
in a slightly different order a simple header-to-header ribbon cable
could be used to connect the CPU card to the small buffer board.
- The USB function connection (SA1110) should have a pullup to +3.3 on
D+ which is switched via software. This is very important for proper
functioning of the USB function/client. Basically, the host will try
to enumerate the device before the SA1110 has booted unless the pullup
is disabled until the software is ready.
[aside: is there currently any pullup on D+?]
- Both the peripheral and basic boards have power connectors which
have +3.3, GND, +5, GND connection but it's not clear which end if
which or the exact order of the connections.
the order seems to be +3.3, GND, +5, GND, from left to right looking
down; one of the holes in the connector has no notch in the bottom
right-hand side, so viewing the connector "head on" you see:
+-------+-------+-------+-------+
| | no | | |
| notch | notch | notch | notch |
| /| | /| /|
+-------+-------+-------+-------+
+3.3 gnd +5 gnd
- The grounds on the basic board do not appear to be connected.
- There is a question about the power "monitor pads" on the basic board
and the peripheral board. Do they need to be shunted?
[yes, they do]
- wrong buffer chip supplied with "small buffer board" kit
[true]
POWER ON
- I can't seem to get proper voltages on the cpu card. I don't think
the power and ground planes are right on the "basic" board. Certainly
there does not seem to be a ground connection between the power
connector on the "basic" board and the cpu board when they are
connected via the "pcmcia bus" connector.
- The Vddi power regulator on the cpu card does not seem to start all
the time. I am not sure that the "REG_EN" and "PWR_EN" signals are
driven anywhere. Sometimes I do see 1.75v out of the regulator - but
it seems to go away after a while. This is odd. Most times I see
no voltage at Vddi after powering up the board set.
- If I connect all 3 board up my power supply "crowbars" at over 0.5Amps.
I assume these boards don't need more thant .5 amps... Something appears
to be unhappy.
I tried powering the cpu board from the peripheral card only but this does
not seem to supply Vddx and Vcc_m.