HP 16702A mouse

[dkasza]

Hello everyone!

I recently picked up an HP 16702A Logic Analyzer (option 003, 16717A modules, really nice) and it appears to work fine using a remote session (VNC), but I cannot get things to work on the built-in display. When it boots up, it claims that there is no mouse connected even though I have a PS/2 mouse (USB mouse that came with a PS/2 adapter) connected to the back. I checked it on another computer, and the mouse appears to work fine with its PS/2 adapter. Unfortunately, I don't have a "real" PS/2 mouse on this continent.

Has anyone else experienced the same issue? Are these units picky when it comes to external mice and keyboard? Is there any workaround?

Thanks,
Dan

[krivx]

I think most of those USB/PS/2 mouses will "listen" for handshaking and select a protocol from there, the adapters are just passive wiring. It's no too surprising that you need a "real" PS/2 mouse. They seem to cost about 2-3$ on eBay or Amazon though.

[dkasza]

I think most of those USB/PS/2 mouses will "listen" for handshaking and select a protocol from there, the adapters are just passive wiring. It's no too surprising that you need a "real" PS/2 mouse. They seem to cost about 2-3$ on eBay or Amazon though.

Yeah, that's what I figured.
Still, I'd like to make sure that these things work with most PS/2 mice. I don't want to order a mouse from ebay just to find out that this device does not like that one either.

[free_electron]

These machines cannot handle 3 button or scroll wheel mice.
You need an old scool 2 button ps2 mouse with a mechanical ball.
The same for the keyboard. The only keyboards that work well are the original hewlett packard curved keyboards.

[dkasza]

These machines cannot handle 3 button or scroll wheel mice.
You need an old scool 2 button ps2 mouse with a mechanical ball.
The same for the keyboard. The only keyboards that work well are the original hewlett packard curved keyboards.

Thanks! This is exactly the information I was looking for. I found a mouse on ebay already. I'm less concerned about the keyboard because this device has a QWERTY keyboard built-in, and I can use a remote session if I need to type something long.

[Sigmoid]

This is extremely interesting. By everything I know about the PS/2 mouse protocol, there should be no compatibility issues whatsoever. My first thought was that the new optical mice might draw more current than the HP can supply, but even that is standardized...
Info on the protocol: http://www.computer-engineering.org/ps2mouse/

As for dual protocol mice, even normal PS/2 ones need to be initialized by the host... So either there is some interesting PS/2 history hidden here, or HP simply went with a slightly proprietary, cut-corners approach in implementing PS/2.

[mwilson]

These machines cannot handle 3 button or scroll wheel mice.
You need an old scool 2 button ps2 mouse with a mechanical ball.

That's not strictly true, it's just likely a side-effect of other compatibility problems.

My 16702A works great with a Microsoft IntelliMouse Explorer + USB-to-PS/2 adapter that I got many years ago. It's a scroll-wheel optical mouse (with the scroll wheel click acting as the third button). I also recently tried it with the last (recently discontinued) version of the MS IntelliMouse using an adapter and it worked as well. The new "Microsoft Basic USB/PS2" mouse, however, does not work. And I even have an old "real" PS/2 mouse from Dell which the 16702A doesn't recognize.

So it seems to be pretty random in terms of which mice it supports and which it doesn't. As others have said, it is also picky about keyboards.

[free_electron]

the machine is  -very- picky. it has something to do with the timing.
the hp branded mice and keyboards form that era work fine. other ... not so.
i had a 16702a and it took me at least 10 keyboard tries before i found one that worked. i later traded for a 16702b and the problem still exists there.
These machines are NOT pc based but use a PA-RISC processor running HP-UX with a X-client.

Anyway , mine is ging out the door soon. i got a 16900 now.

[Martin Hodge]

=== Reopening an old topic to add additional info for future searches ===

It appears that all that is wrong with these units is that they are missing pull-up resistors on the PS/2 ports. There are even unpopulated pads for them directly behind the connectors. (see image) All that I had to do was add 1k resistors in the four locations and the unit recognized an old el-cheapo PS/2 mouse and keyboard. (As expected, a USB mouse or KB thru a PS/2 adapter did not work.)

[The Dilettante]

=== Reopening an old topic to add additional info for future searches ===

It appears that all that is wrong with these units is that they are missing pull-up resistors on the PS/2 ports. There are even unpopulated pads for them directly behind the connectors. (see image) All that I had to do was add 1k resistors in the four locations and the unit recognized an old el-cheapo PS/2 mouse and keyboard. (As expected, a USB mouse or KB thru a PS/2 adapter did not work.)

Thanks - that was quite helpful and came just in time for my newly arrived 16702B! 
I took a couple of 4k7's and can now use a Compaq keyboard (model KB-9860) and Microsoft IntelliMouse v1.3a.

Cheers,

Henry.

[Gerhard_dk4xp]

My 16702B came with an Intellimouse 1.2 / "PS2 compatible" and a Labtec keyboard and they work both.
(green/purple plug)

regards, Gerhard

[Hamster]

That's great! I shall add the resisters to fix it

However, i have the Branded HP Keyboard and mouses for my 16700A,16700B,16702A, 16702B units .

[SeanB]

I still have 2 elderly well used Intellimice PS2 ( still have the coated balls) lying around. White and need cleaning, and probably have gunk on the insides still.

[Martin Hodge]

I did not have reliable results with standard 4k7 ohms. I had to go all the way down to 1k before the waveforms were remotely square on the scope.

[ogoun]

Hi, time to revisit this topic..

I have a 16702b, pimped out (replaced all fans with low noise noctuas, option 003 memory (main and video), scsi to CF card and high speed CF card instead of rotating hdd, and all works well, except the annoying fussiness with mice and keyboards..

So time to investigate further.

Definitely add the 1k pullups, as already discussed.

Here is what I found helped:

Change the 4 data line filter caps near the 6 pin ps/2 mouse/keyboard connectors from 10nF (way too large!) to 1nF.

This will make the ps/2 comms edges and levels good for just about all keyboards and mice.

HOWEVER..

I still had a keyboard (IBM SK-8840, with integrated trackpad and trackpoint) that would not work. Signal levels were good, and other keyboards worked fine.

So, at first I thought it might be the supposed requirement of HP-UX to use a keyboard that supported code set number 3, and I decided to capture the keyboard boot up sequence on my mso-x 4054a

Results were surprising.

hst = 16702B logic analyser
kbd = attached keyboard

First, a working keyboard..

(power up)

0xAA    (kbd to hst)   Keyboard passed BAT (power up) self tests.

0xF5    (hst to kbd)   Disable scanning
0xFA    (kbd to hst)    ACK

0xF2   (hst to kbd)    Get Keyboard ID
0xFA    (kbd to hst)    ACK
0xAB    (kbd to hst)    Keyboard ID = 0xAB83
0x83    (kbd to hst)

0xFF    (hst to kbd)    Reset keyboard/perform BAT tests
0xFA    (kbd to hst)    ACK


0xAA    (kbd to hst)    Keyboard passed BAT (power up) self tests.

0xF0    (hst to kbd)    Set scan code to "scan code set 2" Looks like the scan code 3 stuff was a furphy
0xFA    (kbd to hst)    ACK
0x02    (hst to kbd)    0x02
0xFA     (kbd to hst)    ACK

0xF0    (hst to kbd)    Get current scan code (should equal what we just asked for)
0xFA    (kbd to hst)    ACK
0x00    (hst to kbd)    0x00 (request scan code)
0xFA     (kbd to hst)    ACK
0x02    (kbd to hst)    0x02 got what we expected

0xED    (hst to kbd)    Set LEDs
0xFA     (kbd to hst)    ACK
0x02    (hst to kbd)    Set NUM lock LED on
0xFA    (kbd to hst)    ACK

0xF4    (hst to kbd)    Enable scanning
0xFA    (kbd to hst)    ACK                                All done now, working keyboard..


For the non working keyboard, this is what I found..

(power up)

0xAA    (kbd to hst)   Keyboard passed BAT (power up) self tests.

0xF5    (hst to kbd)   Disable scanning
0xFA    (kbd to hst)    ACK

0xF2   (hst to kbd)    Get Keyboard ID
0xFA    (kbd to hst)    ACK
*** 0xAB    (kbd to hst)    Keyboard ID = 0xAB84
*** 0x84    (kbd to hst)

AAAnd.. That's all she wrote! No more keyboard comms, no keyboard attached.

So, unless the HP-UX ps/2 keyboard driver sees a keyboard with ID = 0xAB83, it will ignore the attached keyboard.

This should be a fairly easy fix.. Just patch the binary of the ps/2 driver to bypass this ID test.. I will leave this as an exercise for those more skilled than I in such matters

Cheers,

Pete

[gslick]

=== Reopening an old topic to add additional info for future searches ===

It appears that all that is wrong with these units is that they are missing pull-up resistors on the PS/2 ports. There are even unpopulated pads for them directly behind the connectors. (see image) All that I had to do was add 1k resistors in the four locations and the unit recognized an old el-cheapo PS/2 mouse and keyboard. (As expected, a USB mouse or KB thru a PS/2 adapter did not work.)

Just for future reference the next time someone else comes along looking at this (someone was recently asking about this elsewhere), here are images of the 16700-66503 I/O Board as manufactured by HP / Agilent without the four PS/2 pull-up resistors populated in the pads provided on the PCB.