9 Feb
2011
9 Feb
'11
8:23 p.m.
Hi guys,
I've just finished adding a rudimentary keyboard handler to my 3B1
emulator (http://www.philpem.me.uk/code/3b1emu). Problem is... it
doesn't quite work.
I can get into the test software, and if I hit a key it'll drop into the
main menu. Problem is, when I get to the main menu, it doesn't seem to
recognise the commands I'm entering. That is, I can enter a menu item
number (e.g. 6 <CR> for Goto Subsystem Menu), it echoes to screen
correctly, but then the screen clears and I'm thrown unceremoniously
back to the main menu, instead of getting passed along to the subsystem
menu... This happens with all the possible menu selections, and the
"expert mode" command ("s4test").
Is anyone aware of any quirks with the test software which might cause
this type of behaviour?
Does anyone happen to have a 3B1 and an RS232 protocol analyser (or a
logic analyser with RS232 decoding)? It'd be really handy to have some
dumps of the data sent and received by the 6850 ACIA (the keyboard comms
controller chip) when the keyboard is initialised, and when various keys
are pressed...
(If I had a 3B1 of my own, it'd be on the table downstairs with a
keyboard sniffer cable and my HP logic analyser mainframe plugged in...)
Thanks,
--
Phil.
classiccmp at philpem.me.uk
http://www.philpem.me.uk/
10 Feb
10 Feb
2:18 a.m.
In fact I got one in home, Philip, together to my workdesk. It works, but
I'm not sure how could I help you. Let me know, please.
Greetings
Sergio
2011/2/10 Philip Pemberton <classiccmp at philpem.me.uk>
Hi guys,
I've just finished adding a rudimentary keyboard handler to my 3B1 emulator
(http://www.philpem.me.uk/code/3b1emu). Problem is... it doesn't quite
work.
I can get into the test software, and if I hit a key it'll drop into the
main menu. Problem is, when I get to the main menu, it doesn't seem to
recognise the commands I'm entering. That is, I can enter a menu item number
(e.g. 6 <CR> for Goto Subsystem Menu), it echoes to screen correctly, but
then the screen clears and I'm thrown unceremoniously back to the main menu,
instead of getting passed along to the subsystem menu... This happens with
all the possible menu selections, and the "expert mode" command
("s4test").
Is anyone aware of any quirks with the test software which might cause this
type of behaviour?
Does anyone happen to have a 3B1 and an RS232 protocol analyser (or a logic
analyser with RS232 decoding)? It'd be really handy to have some dumps of
the data sent and received by the 6850 ACIA (the keyboard comms controller
chip) when the keyboard is initialised, and when various keys are pressed...
(If I had a 3B1 of my own, it'd be on the table downstairs with a keyboard
sniffer cable and my HP logic analyser mainframe plugged in...)
Thanks,
--
Phil.
classiccmp at philpem.me.uk
http://www.philpem.me.uk/
7:49 a.m.
On 10/02/11 07:18, SPC wrote:
In fact I got one in home, Philip, together to my
workdesk. It works, but
I'm not sure how could I help you. Let me know, please.
The easiest way I can think of to get a keyboard-data dump would be
something along the lines of:
- Attach an RS232 protocol analyser (or a logic analyser with RS232
decoding, e.g. the Salae Logic) to the RX and TX pins on the 6850 ACIA chip.
- Plug the keyboard in, insert the Diagnostics disc (Foundation set disc
1), and boot the 3B1.
- Wait for the memory check screen, count to five, then press SPACE.
- Pick one of the sub-menus (e.g. the Subsystem menu) -- type in its
number, followed by RETURN.
- Press 0 <RETURN> to get back to the main menu
- Type p4test <RETURN> to get to the Expert screen
- Exit the Expert screen and reboot (7 <RETURN>).
- Stop the logic analyser and remove the Diags floppy.
This should give me more than enough information to fix the bugs in the
keyboard driver...
The reason I suggested the Saleae Logic was because it stores transition
data to the host PC, and has onboard RS232 decoding (not to mention the
fact that the Logic software will run without the hardware and still
allow you to use the protocol decoders...)
A HP analyser should serve the same purpose -- set it to Transitional
Timing mode (so it doesn't waste all the ACQ memory on blank space),
trigger on falling edge, pre-trigger buffer to 1% (which if memory
serves is the minimum on the 16500B), and save the output as a CSV. I've
got an app (somewhere!) which can convert a HP 16500B CSV to a VCD file
which I can then load into GtkWave to analyse.
Thanks,
--
Phil.
classiccmp at philpem.me.uk
http://www.philpem.me.uk/
4945
days inactive
4945
days old
2 comments
2 participants
participants (2)
-
classiccmp＠philpem.me.uk -
spedraja＠ono.com