Broken New AP5 Fixed & Electron Mouse testing

[1024MAK]

I have a couple of working, unmodified Electrons, a couple of Acorn Plus 1's, somewhere I have a mouse for a Beeb user port and also a track ball. I have a new AP5 (but I have not extensively tested it yet). I also have bought various I2C modules to play with (but have not yet found the time, and have stored some and then forgot which box I have stored them in ).

Duncan, if you want me anything to Cambridge, let me know.

With regards to the mice, keep in mind there are two slightly different pin-out / connection schemes for user port mice / track balls. Some software supports both, some, only one. All mice that connect to the user port are "bus" mice. So with a suitable passive adaptor lead, Atari ST, Amiga, Archie or the very rare IBM compatible PC bus mice can all be connected and used. Or in theory, any of the modern PS/2 or USB active adaptors (although some modern USB mice may have too good a resolution, so may not work very well in practice).

Mark

[Elminster]

With regards to the mice, keep in mind there are two slightly different pin-out / connection schemes for user port mice / track balls. Some software supports both, some, only one. All mice that connect to the user port are "bus" mice. So with a suitable passive adaptor lead, Atari ST, Amiga, Archie or the very rare IBM compatible PC bus mice can all be connected and used. Or in theory, any of the modern PS/2 or USB active adaptors (although some modern USB mice may have too good a resolution, so may not work very well in practice).

I have a pic controlled Amiga to usb board (from 2009), that is then re-cabelled for userport, works fine now I switched y pins.

It is smalkymouse2 that has issue. Which is the one that should work without issue.

Mouse used on both is one of the first usb optical mice from Logitech. I keep it because it pretty much works on anything.

[Elminster]

Duncan, if you want me anything to Cambridge, let me know

I suspect all we would need is. Another smallymouse2, a real amx mouse, an extra AP5, and electrons with and without MRBs. I should think that would narrow it down. But you never know the two Dave’s might have sussed it by then anyway.

[daveejhitchins]

Just posting this from a PM i've just sent:

"It may not be needed (the MRB, for the mouse to work), however, I'm seeing a Power-On anomaly when I'm using BeebSCSI 'sometimes' with a MRB fitted. Maybe worth a try (removing the MRB)?"

Dave H

[Elminster]

Today’s task is free off 6 key, test i2c mod and maybe remove MRB. Need to go shopping and do chores first.

As an aside can you disable tube on Electron. Plug and removing pitube is a pain. If not will have to stick a switch across power supply to level shifter or pi.

[simoni]

SmallyMouse2 uses the AMX pin-out (as this works for AMX utilities as well as Domesday). The board also accounts for the extra sensitivity in USB mice and intelligently changes this to the appropriate quadrature rates.

On the 'user-port' side of things, the output is simply quadrature - just like a real mouse. The only possibility I can think of that would cause it to act differently to a PS/2 converter is power. USB is more sensitive to the power levels (generally requiring 4.8V or more to work) - if the user-port it's connected to doesn't supply that (or ramps up slowly on power up) you could potentially see an issue.

Testing the theory is easy though, just power the SmallyMouse2 board from something else.

[Elminster]

Should be easy enough to test.

An aside. I can confirm, as I just tested, that the i2c 4 pin header, that shares userport 1, is working fine with the clock module (same one Martin used in ic2 documentation).

[MartinB]

Yay! I win.....





.

[1024MAK]

I've got a smallymouse2 (from Bas) a few years ago. If / when I find it again, I'll put it aside to take to Cambridge.

Can't say I remember testing mine on an Electron though.

Mark

[1024MAK]

In another thread, Bas talked about the combined current for the SmallyMouse2 and the USB mouse causing problems when used on an Atari ST. I was supprised by this, as the +5V supply available at the mouse/joystick ports is fed from the main motherboard +5V supply (the PSU being more than capable of supplying extra current). Of course, there is cable and PCB track resistance plus a bit of resistance from the optional filter that is fitted on some keyboard PCBs to the +5V line.

Mark

[1024MAK]

USB is more sensitive to the power levels (generally requiring 4.8V or more to work)

Okay, who's gonna tell those Chinese producers who pump out cheap thin USB cables, and various other stuff that results in USB voltages that sometimes fall to around 4.25V!

Mark

[Elminster]

Yay! I win.....

Yep. I hate to admit it but I now have upurs, i2c and eeproms working on master and electron. You will have to invent something else for me to break now.

[MartinB]

I believe ’Elminster-Proof’ is to become a new engineering standard post-Brexit...

Incidentally, did you test I2C with a legacy 6522 or a WDC device? I had a note to self to try this because I’m not sure if the data and control line input circuits have internal passive pull-up resistors like the Rockwell devices - I assume the presence of these resistors when specifying that no external pull-ups are to be used....

( A fundamental of the I2C bus is a Vcc pull-up on the data and clock lines - usually external on one of the bus slave devices but not [normally] needed on Beebs and Elks because of the 6522 electrical characteristics.)



EDIT : Loads of edits coz that how my mind works

[Elminster]

I believe ’Elminster-Proof’ is to become a new engineering standard post-Brexit...

Incidentally, did you test I2C with a legacy 6522 or a WDC device? I had a note to self to try this because I’m not sure if the data and control line input circuits have internal passive pull-up resistors like the Rockwell devices - I assume the presence of these resistors when specifying that no external pull-ups are to be used....

The clock module is running on the new WDC 6522. I hope to do some more testing with other modules if/when I get some.

[MartinB]

[Elminster]

Just posting this from a PM i've just sent:

"It may not be needed (the MRB, for the mouse to work), however, I'm seeing a Power-On anomaly when I'm using BeebSCSI 'sometimes' with a MRB fitted. Maybe worth a try (removing the MRB)?"

Dave H

Not the MRB, pulled it this evening, was a pain as don’t know where I put the MOS, stole it from another Electron, and as they are both issue 4’s I took its ULA as well. (For my overheating issue, not put that in yet in case removing MRB fixes that).

With MRB gone that leave just issue 4 Electron, AP5, wdc 6522 and smallymouse2.

Will try Simon’s different power supply next. I could put in myelin diag cart, or Hoglet’s ice-t but I suppose. But that sounds like hard work reviewing the results

[hoglet]

Will try Simon’s different power supply next. I could put in myelin diag cart, or Hoglet’s ice-t but I suppose. But that sounds like hard work reviewing the results

The best way I've found to debug a hang issue like this is with the £12.50 logic analyzer / 6502 Bus Decoder.

You just capture a few seconds of bus activity that includes the hang, feed it through the decoder, then all is revealed. Once you've done it a couple of times, it's actually dead easy. As long as you have a decent text editor that can deal with 700MB text files (e.g. emacs on Linux).

The big advantage over ICE-T65 is it completely non-intrusive.

Do you have a Linux machine to hand?

Dave

[Elminster]

I have 3 logic analysers (1 installed in the myelin cart is your one), 1 scope, 3 multimeters, ice-t, logic probes.

Can’t say I am that great at using them.

I have a Mac and a pi-Top. So macOS and Raspbian native. But windows 7, 10 and 3 or 4 flavours of Linux as VMs.

Edit: And one of Simon’s Cartridge breakout boards.

[daveejhitchins]

Not the MRB, pulled it this evening, was a pain as don’t know where I put the MOS, stole it from another Electron, and as they are both issue 4’s I took its ULA as well. (For my overheating issue, not put that in yet in case removing MRB fixes that).

That's good, at least we're down to a basic machine now . . .

Dave H

[-B-]

Hi All,
Another one here having problems.

I downloaded Stuart's .dsd image, extracted the ROM file and made an EPROM.
This appears to work as I can turn the AMX mouse on and off.
I have installed it in ROM slot 3 (AP5), in case that makes a difference.

I then copied the files from the .dsd image to BeebSCSI but the software won't start, either with or without the SmallyMouse2 connected.

I have an Issue 4 Elk, Prime's MRB, Plus 1, AP3/4, AP6, AP5 and BeebSCSI mini.

Looking forward to a solution.

Neil.

[Elminster]

What error do you get when you try to start it?

My config was the same except I ran it from gosdc and a data centre (and no ap3/4). Beebscsi not built yet, but think others used them okay.

It doesn’t work with pitubedirect but you don’t mention having one of them connected.

[-B-]

What error do you get when you try to start it?

My config was the same except I ran it from gosdc and a data centre (and no ap3/4). Beebscsi not built yet, but think others used them okay.

It doesn’t work with pitubedirect but you don’t mention having one of them connected.

No errors, it just hangs the whole thing.
I don’t have a Pitubedirect (Yet).

[Elminster]

I usually switch MRB to normal mode, srload the Rom, Ctrl break then shift break (then plug mouse interface back in).

It gives error if pitube installed, hangs if smallymouse plugged in, does odd random things if gosdc plugged in.

Can you remove ap3/4 just to confirm that makes no difference to starting it.

[-B-]

I’ve removed AP3/4 and put MRB into normal mode with no difference.
I am trying to load from BeebSCSI, has anyone else tried it this way?

Unplug SmallyMouse2, switch on Elk.

*MOUSE ON
CTRL-BREAK
*DIR UTILITIES
CH.”RUN-ART” or *EXEC !BOOT

Hangs

[Elminster]

I don’t suppose you have any other electrons or plus 1’s to try? Or anything else like a gotek that could plug into the ap3/4 etc to run software from?

[MartinB]

*MOUSE ON
CTRL-BREAK
*DIR UTILITIES
CH.”RUN-ART” or *EXEC !BOOT

When you run RUN-ART, this is what's being executed...

Code: Select all

   10 REM * ART LAUNCHER *
   20:
   30 *MOUSE ON
   40 *POINTER ON
   50 *E-ART

...so in order to try and narrow things down, after entering your *DIR UTILITIES, try typing each of the above lines as single consecutive manual commands and see at what point the hang occurs.

Type these in turn...

*MOUSE ON <Return>
*POINTER ON <Return>
*E-ART <Return>

[-B-]

I've now also got a vanilla Elk, Plus 3 and Plus 1 with the AP5, BeebSCSI and SmallyMouse2 installed.
The PC I use to make disks from images doesn't want to boot today - typical, as I've just paid for a ADFS explorer key for it.

*MOUSE ON <Return> ok
*POINTER ON <Return> ok
*E.ART <Return>

I briefly see a line of gobbledygook then the whole screen moves down a few lines and the cursor goes to the top right.

Same thing happens to the other setup (without AP3/4 installed and MRB in normal mode).

The only difference I see in the two systems is that the one fitted with MRB and AP6 shows the mouse status on startup and after a <BREAK>.

Neil

[Elminster]

When you get gobbled gook it is often a corrupt image. I had that several times. Did you use the dsd or ssd version of the package? Not sure on beebscsi but on the gosdc you have to manual tell it the dsd file you are putting on is interleaved or you get garbage. Doing a directory seems to work but it is corrupted.

At a guess that is what happened here.

[-B-]

When you get gobbled gook it is often a corrupt image. I had that several times. Did you use the dsd or ssd version of the package? Not sure on beebscsi but on the gosdc you have to manual tell it the dsd file you are putting on is interleaved or you get garbage. Doing a directory seems to work but it is corrupted.

At a guess that is what happened here.

That is entirely possible. I used BeebEm to access the files on the .dsd, used the ROM file to make an EEPROM and copied the other files to the BeebSCSI SD card using ADFS Explorer.
So something could have gone wrong along the way.

[Elminster]

Somewhere in that thread it points to an ssd version of elkamx, you could give that a go.

Or try to work out issue with dsd version