Matchbox sized 6502 / Z80 / 6809 Co Pro

[BigEd]

Indeed they are, as built. But every expansion added is another element which might push something out of spec. Even then, a lot of things just work. But in both Matchbox and Pi we have seen that we have to be careful, and that individual machines differ, even before you add the various popular expansions.

[vanekp]

is a R6502AP CPU http://www.cpu-world.com/CPUs/650x/Rock ... 502AP.html compatible
with a SY6502A http://www.cpu-world.com/CPUs/650x/Syne ... 6502A.html they are both 2 Mhz ?
They look the same to me
http://www.cpu-world.com/Compare/329/Ro ... 502A).html

[danielj]

Yup! Should be!

d.

[vanekp]

Just for interest to others that may have the same problem viewtopic.php?f=44&t=8852&start=1140#p183626 with the LX9CoPro, today I got another UM6502A processor chip put it in and no problems with the CoPro works perfect (Elite ran flawlessly) with a new chip in so I guess my old CPU is a bit dodge, as the CoPro did work with a 65C02 in in BBC but wanted the correct CPU in there as a lot of software that use no-standard op codes will not run on a 65C02.

Peter.

[fordp]

I am updating a CoPro! I am preparing one of my Copro's to pass on to another forum user so it was time to update my unit. Well I have not done that in a while. Oh boy it took a while to assemble all the physical bits but then I had to do the update.

Well it was hard to find the information so for my benefit next time I made this:
https://github.com/hoglet67/CoPro6502/w ... -the-CoPro

I hope hoglet does not mind me messing with his words of wisdom.

From this you can see I am no technical author, feel free to suggest updates or even have a go at updating the page. I have a couple of photos I will add later.

On a plus point I have a working Acorn machine turned on for the first time in ages

[hoglet]

I hope hoglet does not mind me messing with his words of wisdom.

I've no problem at all with people improving the documentation. Thanks.

Looking at github, there have been a few commits since the last release in September 2017:
https://github.com/hoglet67/CoPro6502/c ... 9...master

The first four (Aug/Sept 2018) are just me messing around testing a cycle-accurate 65C02 core. None of this is active code.

The other four (Jan/Feb 2019) fix a bug in two-byte mode. This is needed if you are using a Co Pro to run an Acorn Level 3 File Server. Most people aren't, so I wouldn't worry.

Is this board destined for Andy (anightin)? He has a Xilinx Programmer so will be able to update himself in the future.

Dave

[fordp]

Is this board destined for Andy (anightin)? He has a Xilinx Programmer so will be able to update himself in the future.

Yes it is. I can hold of sending it and update it again if he needs the very latest.

I am hoping to become a bit more active in the Acorn scene again in the next few months. I only have one machine working at the moment so need to find some time to repair all that I have.

[hoglet]

Yes it is. I can hold of sending it and update it again if he needs the very latest.

No need to do that, as I'm not planning a new release (I'm busy with other things for the next two weeks).

Dave

[jgharston]

Will there be anybody coming to ABUG South with the equipment to update the Matchbox so I can get mine updated? I think my last update was last year with the PDP11 IRQ fix.

[hoglet]

Will there be anybody coming to ABUG South with the equipment to update the Matchbox so I can get mine updated? I think my last update was last year with the PDP11 IRQ fix.

I'll have the Xilinx programmer with me.

If I remember correctly, those PDP11 changes were to PiTubeDirect, not Matchbox.

Ah yes, here they are:
- PDP-11 Co Pro - reset clears PSW
- PDP-11 Co Pro - updated client rom to 0.28
(this is why I love git...)

So we have:

PDP11:
- PiTubeDirect - v0.28 (updated 19/1/2019)
- Matchbox - v0.27a (updated 5/10/2017)

6809:
- PiTubeDirect - v1.05 (updated 5/10/2017)
- Matchbox - v1.05 (updated 2/10/2017)

But looking at the version log in the 6809 client code, both will be missing the 12-Jul-2019 update

Code: Select all

; v1.05 25-Aug-2017 JGH: OSFILE/OSGBPB/OSARGS use big-endian control blocks
;	02-Sep-2017 JGH: PRTEXT preserves X as with Z80 Client.
;	12-Jul-2019 JGH: KBDTST/KBDIN routines use pending input from OSRDCH

[anightin]

The other four (Jan/Feb 2019) fix a bug in two-byte mode. This is needed if you are using a Co Pro to run an Acorn Level 3 File Server. Most people aren't, so I wouldn't worry.

Is this board destined for Andy (anightin)? He has a Xilinx Programmer so will be able to update himself in the future.
.
.
.
But looking at the version log in the 6809 client code, both will be missing the 12-Jul-2019 update

This must be one of the best forums around! I Don't even have to write my own questions or answers for some topics

Much appreciated both

The ATOM is eagerly awaiting the Co-Pro

Update: Received today with thanks

[anightin]

Now working nicely:

[trixster]

Superb!

[KenLowe]

Nice setup! That's exactly where I want to get to. It's probably going to bankrupt me, though

Now working nicely

[anightin]

Nice setup! That's exactly where I want to get to. It's probably going to bankrupt me, though

Now working nicely

It's come a long way thanks to Hoglet, Roland and others in this great community. This photo was the Atom in June which came in a non-working state with most of it's IC's removed. This photo was after it had a bath, and after I fitted a new power connector and removed the regulators. I think the first thing I added was a AtomMMC thanks to Sir Morris then a Noisekiller from Hoglet, then a RAM/ROM board from Roland -- saved me having to source the ROMs and memory chips.

[lurkio]

Splitting this conversation off from the Ozmoo thread...

Does removing the internal Co Pro have any effect?

Will try it and report back.

Are there any other hardware addons?

Yes, my Master has a non-"Turbo" User Port SD Card reader device for use with the MAMMFS ROM; a Picton/IFEL/ctorwy multi-bank swappable battery-backed SWRAM cartridge in one cartridge slot (banks 2 & 3); and an old Peartree SWRAM cartridge in the other (0 & 1). The Kortink ReCo6502Mini copro is internal.

If you select each Co Pro in turn (using *FX 151,230,N then Ctrl-BREAK) and note which Co Pro you get, I should be able to work it out.

Here are the results of repeatedly doing *FX151,230,N with the Matchbox copro connected to my Model B, which seems to work with the copro without error:

Code: Select all

*FX151,230,N

0: Acorn TUBE 04MHz 65C102 Co-Pro
1: Acorn TUBE 08MHz 65C102 Co-Pro
2: Acorn TUBE 16MHz 65C102 Co-Pro
3: Acorn TUBE 32MHz 65C102 Co-Pro

4: Acorn TUBE Z80 64K 1.21
5: Acorn TUBE Z80 32 MHz
6: Acorn TUBE Z80 56 MHz
7: Acorn TUBE Z80 112 MHz

8: Acorn TUBE 80286 896K

9: 6809 TUBE 64K 1.00

10: Acorn 68K second processor 1024K

11: PDP11 TUBE 64K 0.50

12: ARM Second Processor 2048K [*FX 151,230,N seems to be ignored: had to switch Model B off and on.]

13: Pandora 32000 V2.00  2048 KB

14: [Standard Model B. No copro. *FX 151,230,N seems to be ignored: had to switch Model B off and on.]

15 and higher: As for 14.

Here are some photos of my copro:

Lee also has a Matchbox copro, but his looks slightly different from mine (not as many little beige caps (or whatever they are), especially near the connector):

Plus, Lee's copro seems to work with his Master without error! It flew through the Ozmoo benchmark on the 32MHz 65C102 core:

[hoglet]

I think you must have the LX9CoProCombined_20160228_1555_dmb Firmware. See:
https://github.com/hoglet67/CoPro6502/w ... h-settings

There have been a couple of releases since then - the 65C02 Core was changed (from AlanD's to Arlet's) which allowed the speed to be increased to 64MHz.

I don't think this is a firmware issue, but I'll downgrade to that version tomorrow and check.

Dave

[lurkio]

Does removing the internal Co Pro have any effect?

Yes! That did the trick! Thank you.

I removed the ReCo6502Mini internal copro from the Master, and I then reconnected the external Matchbox copro -- and suddenly I can boot the Ozmoo benchmark disc, and Ozmoo recognises and uses the Matchbox copro without error!



EDIT: So it seems that the Acorn cheese-wedge copro can coexist with the ReCo6502Mini, but the Matchbox can't..?

[lurkio]

So it seems that the Acorn cheese-wedge copro can coexist with the ReCo6502Mini, but the Matchbox can't..?

I wonder you can somehow disable the ReCo6502Mini without having to physically remove it from the Master?

The ReCo has a "supervisor" mode. Plus, there's a BASIC updater program, linked on its homepage:

http://www.zeridajh.org/hardware/reco65 ... pdates.htm

Unfortunately, I don't understand either of them!

I contacted the creator of the ReCo copro yesterday and am waiting to see if he'll reply...

[hoglet]

I wonder you can somehow disable the ReCo6502Mini without having to physically remove it from the Master?

I suspect this is an electrical/signal integrity issue, which likely wouldn't be resolved by simply disabling the Co Pro.

It's a bit difficult to diagnose this further, as the ReCo6502Mini is not open source, so I can't just look at the schematic.

Dave

[lurkio]

I suspect this is an electrical/signal integrity issue, which likely wouldn't be resolved by simply disabling the Co Pro. It's a bit difficult to diagnose this further, as the ReCo6502Mini is not open source, so I can't just look at the schematic.

Can anything useful be inferred from the fact that the operation of the external Acorn cheese-wedge copro isn't disrupted by the presence of the ReCo, whereas the operation of the Matchbox is disrupted by the presence of the ReCo?

EDIT: Or is that a silly question because the Acorn copro has its own power supply whereas the Matchbox draws power from the host machine -- and does that difference have a bearing on the problem in a way that I can't really understand because I don't know enough about the tech details as I'm basically rubbish at electronics?!

[hoglet]

Can anything useful be inferred from the fact that the operation of the external Acorn cheese-wedge copro isn't disrupted by the presence of the ReCo, whereas the operation of the Matchbox is disrupted by the presence of the ReCo?

That's kind of expected.

I hit an issue on the Electron which has similar symptoms, where the problem turned out to be noise affecting the reset signal going into the Matchbox. Because FPGAs are very fast devices, a tiny noise spike can be sufficient you trigger a reset. And if one side of the tube interface is randomly reset, everything gets out of sync, and the result is the sort of garbage you are seeing.

It's possible you have the same issue, bit that's far from certain.

An original 6502 will be much more immune to tiny spikes of noise, because the technology is much slower.

Is your Matchbox connected with a cable, and if so, how long is the cable?

EDIT: Or is that a silly question because the Acorn copro has its own power supply whereas the Matchbox draws power from the host machine -- and does that difference have a bearing on the problem in a way that I can't really understand because I don't know enough about the tech details as I'm basically rubbish at electronics?!

It's unlikely to be a power supply issue.

Dave

[fordp]

Is this board destined for Andy (anightin)? He has a Xilinx Programmer so will be able to update himself in the future.

Yes it is. I can hold off sending it and update it again if he needs the very latest.

I am hoping to become a bit more active in the Acorn scene again in the next few months. I only have one machine working at the moment so need to find some time to repair all that I have.

[lurkio]

Is your Matchbox connected with a cable, and if so, how long is the cable?

Yes, I think the cable was supplied with the copro when I originally bought it. It's about 32cm long.

[dominicbeesley]

I hit an issue on the Electron which has similar symptoms, where the problem turned out to be noise affecting the reset signal going into the Matchbox. Because FPGAs are very fast devices, a tiny noise spike can be sufficient you trigger a reset.

This is quite likely the culprit - I hit a similar issue this week getting ADFS to work with a faster 65c02/65816 (the floppy disk NMIs were a little slow/noisy causing mutliple triggers). I added conditioning of the signal in the fpga which has cured it. That won't be possible with the ReCoCo so you may need to resort to a hardware solution. If it is noise then a simple RC conditioning circuit might work but if it is due to the reset line rising too slowly then it might need some sort of schmidt trigger?

D

[lurkio]

I heard back from the creator of the ReCo6502Mini internal copro board. He pointed out that, as his Guide indicates, the on-board Tube ULA is always active -- unless you short the two small holes in the board, whereupon the copro will effectively be disabled. However, you're apparently only supposed to do that if you're preparing for a "software upgrade" -- but I still thought it was worth a go.

So I connected the two holes in the board together, using a short piece of copper wire, and then I reinstalled the ReCo in the Master. I then connected the Matchbox to the Master's external Tube socket and found that the Ozmoo benchmark now runs on the Matchbox without error (because the internal copro seems to have been "physically deactivated")!

I asked the creator of the ReCo if it would do any harm if I somehow connected two wires to the two "software upgrade" holes in the board and attached a switch (to the other ends of the wires) which I would toggle to enable and disable the ReCo manually.

He replied, "It's sort of OK, yes. Not what it's meant to be used for though." He added, "Don't fit long wires. This may also be the issues that the Matchbox thing is causing. Just put in a 2-pin header, and use a jumper to 'enable' and 'disable'."

[hoglet]

Hmmm, this still doesn't make a whole lot of sense to me.

When the external tube is being used, the internal tube's nTUBE signal is held in the inactive state, and as a result theReCo6502Mini should never drive the data bus. I don't really understand what more this jumper is doing.

I take it you have double-checked that removing the jumper from the ReCo6502Mini immediately causes the issue to recurr?

Dave

[lurkio]

I take it you have double-checked that removing the jumper from the ReCo6502Mini immediately causes the issue to recurr?

No, I hadn't! I just tried it now: I removed the short piece of wire that had been connecting the two holes in the internal ReCo copro board. The Matchbox was still connected to the external Tube socket. I then switched on the Master and booted the Ozmoo benchmark prog. It still ran without error on the Matchbox -- even though the ReCo was back to its original "unshorted" state, which had been causing so many problems before!

Weirder still, I then did a *CONF.INTUBE to try to switch over to the internal copro, which is a config that had worked before. But now, suddenly, the command prompt never appeared after a Ctrl-Break! The machine just hung on a flashing cursor after printing the ReCo boot message, followed by "BASIC" on the next line. Then, sometimes, duplicates of the ReCo boot message would start appearing seemingly at random down the screen..!

No matter what I did, I couldn't get the BASIC prompt to appear. I tried switching off, disconnecting the Matchbox and its cable, and switching on again. No joy. So I just switched the Master off and removed the internal ReCo altogether.

Am v confused!



EDIT: Is it possible that by shorting the upgrade holes without actually performing an upgrade I put the ReCo into an undefined state or something..?! (Oops?!)

[hoglet]

EDIT: Is it possible that by shorting the upgrade holes without actually performing an upgrade I put the ReCo into an undefined state or something..?! (Oops?!)

Sorry, I have no idea.

One of the reasons I avoid closed-source hardware is that when something goes wrong, you are beholden to the original developer to help sort it out. I'm not comfortable being this reliant on others....

Dave

[lurkio]

Hang on. My mistake! Oops. I'd forgotten that the hang-without-ever-reaching-the-BASIC-prompt is a sporadic(!) symptom of having the DataCentre connected without powering it on at the mains. (What is it with me and mains power?!)

Now that I've switched the DataCentre power on, the BASIC prompt does appear. (Mind you I only remembered to do this after spending a frantic few minutes trying to reflash the ReCo ROM -- i.e. doing the software upgrade for which those little pinholes were intended! -- only to see multiple attempts fail with ROM verification errors! Argh! It worked eventually though. Why? Search me!)

I've now got the internal ReCo6502Mini installed in the Master and the Matchbox copro connected to the Master's external Tube socket, and I can seemingly switch between them at will and run the Ozmoo benchmark on either of them, without error! Which was the goal I was originally trying to achieve, and which I have now achieved, without quite understanding how I've achieved it!