Building Rolands new Atom design...

discuss both original and modern hardware for the bbc micro/electron
Post Reply
User avatar
Multiwizard
Posts: 1882
Joined: Wed Jan 11, 2012 9:03 pm
Contact:

Re: Building Rolands new Atom design...

Post by Multiwizard » Thu Feb 12, 2015 8:03 pm

I was curious so I had to try this, all the wires with screen... :-k

For me this is the best option for the best picture Quality, perhaps a bit less beautiful to look at, but so be it. 8)

(I had bend the 8 pins on the GODIL 90 degrees, so if you don't, the cable can even be a little shorter. :D )


Greetings, Wim... :-)
Attachments
P1090248.JPG
P1090258.JPG
P1090265.JPG

User avatar
hoglet
Posts: 9237
Joined: Sat Oct 13, 2012 7:21 pm
Location: Bristol
Contact:

Re: Building Rolands new Atom design...

Post by hoglet » Thu Feb 12, 2015 8:20 pm

This picture has started me thinking..... :shock:

Maybe it would be possible for the GODIL to be rotated 180 degrees.

This should be possible without PCB changes, as all the pins on the GODIL are configurable. So it would just be a matter of changing the pin assignments in the UCF file, reprogramming the GODIL, and moving the power jumpers to the other side.

It would then mean the GODIL test connector wires would be much closer to where they connect into the PCB.

Dave
Last edited by hoglet on Thu Feb 12, 2015 8:38 pm, edited 1 time in total.

User avatar
Multiwizard
Posts: 1882
Joined: Wed Jan 11, 2012 9:03 pm
Contact:

Re: Building Rolands new Atom design...

Post by Multiwizard » Thu Feb 12, 2015 8:36 pm

hoglet wrote:This picture has started me thinking..... :shock:

Maybe it would be possible for the GODIL to be rotated 180 degrees.
Thanks Dave, ok then, till now this was my best option... :lol:

User avatar
roland
Posts: 3858
Joined: Thu Aug 29, 2013 9:29 pm
Location: Born (NL)
Contact:

Re: Building Rolands new Atom design...

Post by roland » Fri Feb 13, 2015 9:43 am

hoglet wrote: Maybe it would be possible for the GODIL to be rotated 180 degrees.
This is a great idea, but IMHO a few months too late. If we thought about that in December I could have left a little more space at the connector on the Atom main board which would make it easier to connect the Godil and that connector. Or even better, work out a way that the test connector could be connected with long pins and jumpers, directly to the board. Mk3 :?:

My biggest objection at the moment is that there will be different versions. If Godil fitted this way, you need this version, if fitted that way, you need that version of the bit file. It might be very confusing and if things go wrong you might destroy an expensive Godil.

However, if we want to change it, we have to do it now :!: Only Kees and Wim have received their Godil yet and I can assist them to update their Godil when we meet.

@Wim: how do you think about this, considering a way to construct the connecton when the Godil is oriented in the opposite direction?
FPGAtom: 512 KB RAM, Real Time Clock and 64 colours
MAN WOMAN :shock:

User avatar
1024MAK
Posts: 9987
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Building Rolands new Atom design...

Post by 1024MAK » Fri Feb 13, 2015 10:20 am

I'm easy, as in not fussed. Assuming there is enough clearance if rotated 180 degrees, as long as I know which version I have so that it can be fitted the correct way round, either is fine with me.

Mark

User avatar
Multiwizard
Posts: 1882
Joined: Wed Jan 11, 2012 9:03 pm
Contact:

Re: Building Rolands new Atom design...

Post by Multiwizard » Fri Feb 13, 2015 11:29 am

Hi,

I can not imagine by turning the GODIL 180 degrees on the PCB it would make that much difference, probably it will a tiny tiny bit but not really worth it.

So for me I am thinking of leaving my Atom this way... :D \:D/

But of course for other people it will become much easier to connect a GODIL on to the PCB.


Greetings, Wim... :-)

User avatar
Multiwizard
Posts: 1882
Joined: Wed Jan 11, 2012 9:03 pm
Contact:

Re: Building Rolands new Atom design...

Post by Multiwizard » Fri Feb 13, 2015 11:42 am

KarateEd wrote:Time to build a video showing all the functionality of your new Atom, Wim.
It's not exactly where you asked for but it may be useful for some people... :D

It shows how pretty easy it is to make, of course I am not talking about the keyboard part the way I did it... :cry:

https://www.youtube.com/watch?v=X76YQfl ... e=youtu.be

_________________________________________________________________

I had to cut the Atom PCB at 10.1 Centimeters. (at least in my case)

I did use ordinary weather-strip for bridging the Atom keyboard to the MK1 board. (use only plastic ones)

I have made the slit a bit more narrow by heating it up with a hairdryer and push it down with a metal ruler.

The keyboard and MK1 board are not permanent connected with each-other.

When the Atom is in his normal position the keyboard lies resting on the MK1 board.

User avatar
KarateEd
Posts: 4597
Joined: Fri Sep 20, 2013 10:15 pm
Location: Squamish, BC, Canada
Contact:

Re: Building Rolands new Atom design...

Post by KarateEd » Fri Feb 13, 2015 2:37 pm

Now how about one more showing your new Atom running various programs to show off it's working correctly?

I know it's work to do videos so it's just an ask if you are able and have time.

Love the video btw, very cool.

Ed......:-)
Ed...... :-)

3 working Beebs, 1 RetroClinic Master, 1 normal Master, 1 A3010, 1 Pi2 RISC OS, 2 broken Beeb Motherboards, 1 Omnibus A7000+ server, 1 A7000+ Desktop, 1 PET, 1 C64, 1 C128, 1 Amiga 500 and 1 Roamer.

User avatar
Multiwizard
Posts: 1882
Joined: Wed Jan 11, 2012 9:03 pm
Contact:

Re: Building Rolands new Atom design...

Post by Multiwizard » Fri Feb 13, 2015 7:11 pm

KarateEd wrote:Now how about one more showing your new Atom running various programs to show off it's working correctly?
Hi Ed,

I intend to keep this Topic Building Rolands new Atom design...

And I really like to pick up some of my other Atom projects.

1. Making the Digitizer working in my Gdos-rack.

2. Building further on my CP/M-rack


If you ever decide to visit the Netherlands feel free to visit my attic, then I can show you some of the things a Atom is able to... 8)

For me this new Atom is really awesome, and worth every penny... :D


Greetings, Wim... :-)

User avatar
roland
Posts: 3858
Joined: Thu Aug 29, 2013 9:29 pm
Location: Born (NL)
Contact:

Re: Building Rolands new Atom design...

Post by roland » Fri Feb 13, 2015 7:54 pm

The main problem is that we do not have a program that demonstrates all the features of the new Atom. And for the past 35 years we have build magnificent expansions for the Atom but we almost never wrote any serious software for them. That's why I'm now still struggling with a plain 6809 board showing only a welcome banner :roll:

But I agree with Ed that a demo program would be nice. It would be an added value to my website and maybe speed up the sale of the remaining boards :mrgreen:
FPGAtom: 512 KB RAM, Real Time Clock and 64 colours
MAN WOMAN :shock:

PhilYoung
Posts: 204
Joined: Sun Jun 12, 2011 5:55 pm
Contact:

Re: Building Rolands new Atom design...

Post by PhilYoung » Sat Feb 14, 2015 12:26 pm

Another success for Rolands design:
DSCF0039_rotated.jpg
Keyboard PS2 interface is in progress, my assistant is dealing with that:
DSCF0033_resized.jpg
Cheers,

Phil Young

User avatar
Multiwizard
Posts: 1882
Joined: Wed Jan 11, 2012 9:03 pm
Contact:

Re: Building Rolands new Atom design...

Post by Multiwizard » Sat Feb 14, 2015 12:39 pm

PhilYoung wrote:Another success for Rolands design:
Nice to see another new Atom starting up... :D =D>

Looks good... :)

User avatar
1024MAK
Posts: 9987
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Building Rolands new Atom design...

Post by 1024MAK » Sat Feb 14, 2015 12:58 pm

PhilYoung wrote:Keyboard PS2 interface is in progress, my assistant is dealing with that:

Image

Cheers,

Phil Young
That's a prrreeefect assistant :D :lol:

Mark

User avatar
1024MAK
Posts: 9987
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Building Rolands new Atom design...

Post by 1024MAK » Sat Feb 14, 2015 1:22 pm

PhilYoung wrote:Another success for Rolands design
Ooooh, with that distracting pussy picture, I forgot to say: Well done there Phil =D> =D> =D>

BTW how did you get on getting hold of the DIN socket?

Hopefully Roland has a "pretested" one for me...

By the looks of it, everything else, connector wise is either okay or can be "persuaded" to fit :wink:

More on my board later (resistor fitting in progress)...

Mark

PhilYoung
Posts: 204
Joined: Sun Jun 12, 2011 5:55 pm
Contact:

Re: Building Rolands new Atom design...

Post by PhilYoung » Sat Feb 14, 2015 2:04 pm

1024MAK wrote:
PhilYoung wrote:Another success for Rolands design
Ooooh, with that distracting pussy picture, I forgot to say: Well done there Phil =D> =D> =D>

BTW how did you get on getting hold of the DIN socket?

Hopefully Roland has a "pretested" one for me...

By the looks of it, everything else, connector wise is either okay or can be "persuaded" to fit :wink:

More on my board later (resistor fitting in progress)...

Mark
Hi Mark,

The DIN socket came from ebay :

http://www.ebay.co.uk/itm/270630491914,

so hence I have 9 left over, are you having problems finding one ? I could send you one if you like, not sure what the postage would be since it isn't a convenient letter size.

The only difficult socket to fit was the DC in, which I did by folding the +ve pin (which is underneath) across itself at 45 degrees, and then flat against the base of the socket so that I could solder it from under the board. I didn't bother with an audio jack and put in a link to replace it.

It actually didn't work first time. When I had to remove the misaligned CPLD socket it turned out that I had nicked a trace, but once that was repaired it sprang to life.

Cheers,

Phil Young

User avatar
1024MAK
Posts: 9987
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Building Rolands new Atom design...

Post by 1024MAK » Sat Feb 14, 2015 2:25 pm

Hi Phil

Well I already have some (NOS) 7 pin DIN sockets that I hoped would fit:
image.jpg
image.jpg
But as you can see, they have the wrong PCB pin-out. And of the three electronic part suppliers that I checked, they either did not stock them, or had the same pin-out as my my DIN sockets, or a completely different pin-out!

So yes, if you can send me one, that would be helpful. Let me know the cost by PM and I will send the amount by PayPal or any other sensible arrangement of your choice :D.

I have yet to rummage through my remaining selection of various "barrel" 5.5/2.1 and 5.5/2.5 power connectors (I have already tried one common type), but it does look like doing a bit of bending will be called for.

I have two types of stereo dual switched jack sockets that are very close to fitting. I think if just one pin is trimmed a bit, one of them should fit.

Mark

PhilYoung
Posts: 204
Joined: Sun Jun 12, 2011 5:55 pm
Contact:

Re: Building Rolands new Atom design...

Post by PhilYoung » Sat Feb 14, 2015 2:55 pm

1024MAK wrote:Hi Phil

Well I already have some (NOS) 7 pin DIN sockets that I hoped would fit:
image.jpg
image.jpg
But as you can see, they have the wrong PCB pin-out. And of the three electronic part suppliers that I checked, they either did not stock them, or had the same pin-out as my my DIN sockets, or a completely different pin-out!

So yes, if you can send me one, that would be helpful. Let me know the cost by PM and I will send the amount by PayPal or any other sensible arrangement of your choice :D.

I have yet to rummage through my remaining selection of various "barrel" 5.5/2.1 and 5.5/2.5 power connectors (I have already tried one common type), but it does look like doing a bit of bending will be called for.

I have two types of stereo dual switched jack sockets that are very close to fitting. I think if just one pin is trimmed a bit, one of them should fit.

Mark
I've just checked, the ones I bought have the row of three pins on the back edge, and the row of four slightly inboard of that, so the reverse of what you have. They certainly fit, as you can see.

I'll check how much postage would be, and PM you. Won't be until Monday of course. Shouldn't be too much, I thought you were in Holland originally.

As to the DC inlet, I bent and trimmed the pins as described. It's only the underneath +ve pin that was a slight problem. They are cheap enough to experiment a bit if needed,

Cheers,

Phil Young

User avatar
roland
Posts: 3858
Joined: Thu Aug 29, 2013 9:29 pm
Location: Born (NL)
Contact:

Re: Building Rolands new Atom design...

Post by roland » Sat Feb 14, 2015 3:03 pm

Hi Phil,

Nice job =D> I didn't even know that you already started building your new Atom. How did you connect the keyboard to your Atom?

Today I received the first set of Mk2 boards and they eliminate the problems with cassette, audio and power connector.

Cheers,
Roland
FPGAtom: 512 KB RAM, Real Time Clock and 64 colours
MAN WOMAN :shock:

PhilYoung
Posts: 204
Joined: Sun Jun 12, 2011 5:55 pm
Contact:

Re: Building Rolands new Atom design...

Post by PhilYoung » Sat Feb 14, 2015 4:54 pm

roland wrote:Hi Phil,

Nice job =D> I didn't even know that you already started building your new Atom. How did you connect the keyboard to your Atom?

Today I received the first set of Mk2 boards and they eliminate the problems with cassette, audio and power connector.

Cheers,
Roland
Thanks ! There was no problem building it apart from fitting one of the CPLD sockets the wrong way round and having to laboriously remove it. Also all the resistors are carbon film (light brown) except the 180ohm just below the VGA socket which is metal film (blue). I find this very annoying in an OCD sort of way.

There is no 'real' keyboard though. I've built an arduino shield with a PS2 connector and an 8816 cross-point matrix on it (shamelessly copying from Phill). So the arduino reads the keycodes coming in from the PS2 KB and opens or closes the appropriate switch in the 8x16 matrix which is of course connected to the Dutch Atom KB connector.

I'm trying to decide whether to keep the original Atom keyboard layout or use the normal keyboard layout. It's not yet finished, eventually I'll try and do it as a small self contained board.

Cheers,

Phil Young

User avatar
roland
Posts: 3858
Joined: Thu Aug 29, 2013 9:29 pm
Location: Born (NL)
Contact:

Re: Building Rolands new Atom design...

Post by roland » Sat Feb 14, 2015 7:32 pm

We have been looking at Phill's matrix with the 8816, but I cannot find a 8816. That's why Guido is building a PS/2 to Atomatrix interface with only using one PIC.
FPGAtom: 512 KB RAM, Real Time Clock and 64 colours
MAN WOMAN :shock:

User avatar
1024MAK
Posts: 9987
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Building Rolands new Atom design...

Post by 1024MAK » Sat Feb 14, 2015 7:46 pm

As you can see, progress is being made :D
image.jpg
Mark

User avatar
roland
Posts: 3858
Joined: Thu Aug 29, 2013 9:29 pm
Location: Born (NL)
Contact:

Re: Building Rolands new Atom design...

Post by roland » Sat Feb 14, 2015 7:53 pm

I hope I have your Godil here before you have finished your board [-o<
FPGAtom: 512 KB RAM, Real Time Clock and 64 colours
MAN WOMAN :shock:

PhilYoung
Posts: 204
Joined: Sun Jun 12, 2011 5:55 pm
Contact:

Re: Building Rolands new Atom design...

Post by PhilYoung » Wed Feb 18, 2015 3:01 pm

Hi,

It looks like I spoke too soon. Once I got a keyboard interface working I was able to try running the programs from the Atom Archive, and rapidly discovered that there is a problem with my Dutch Atom and I'm a bit stumped.

What is happening is that writes to (or possibly reads from) RAM are being written to two locations #1000 apart. So for instance:

!#4000=0; !#5000=0; !#4000=RND; P.&!#4000; P.&!#5000

indicates that the same value has been written to both locations (or is being read from the same location). It also works the other way round (write to #5000 and read #4000).

Checking further with a short program running from the upper text space shows that any writes to #2xxx are also written to #3xxx, and any to #3xxx are also written to #2xxx. And then the same pattern working up the memory map, ending with #6xxx and #7xxx duplicating each other. Maybe I've got RAID1 memory.

So far I've tried another RAM chip - actually a 32kB SRAM on an adapter - with the same result. I've re-soldered the socket for the RAM and the memory decoder 9572XL and looked for any dry joints or bridges - no change. Finally I've reprogrammed the CPLD - no change.

It works fine otherwise, so the display is rock steady and I can load from and save to the MMC card. I've not tried sound yet.

Any ideas what to look at next ? I have a scope but no logic analyser unfortunately.

Cheers,

Phil Young

User avatar
roland
Posts: 3858
Joined: Thu Aug 29, 2013 9:29 pm
Location: Born (NL)
Contact:

Re: Building Rolands new Atom design...

Post by roland » Wed Feb 18, 2015 4:32 pm

Hi Phil,

This is a weird problem. I would think there is a problem with A12. Maybe you can test this signal with this program:

Code: Select all

P=#2000;[ STA #2100;JMP#2000;]
A12 should be low all the time

The same program at #3xxx

Code: Select all

P=#3000;[ STA #3100;JMP#3000;]
should have A12 high all the time.

The combination of those should give pulses on A12:

Code: Select all

P=#3000;[ STA #2100;JMP#3000;]
I think that's a good start to look at the A12 behaviour. If that is not correct you can start with replacing the '244 address buffers. But I don't have any clue why this would not influence reading from the rom or I/O.

Greetings,
Roland
FPGAtom: 512 KB RAM, Real Time Clock and 64 colours
MAN WOMAN :shock:

PhilYoung
Posts: 204
Joined: Sun Jun 12, 2011 5:55 pm
Contact:

Re: Building Rolands new Atom design...

Post by PhilYoung » Wed Feb 18, 2015 5:29 pm

Thanks Roland, I'll try that tomorrow.

It seems likely that A12/A13 on the RAM is the problem in some way. It looks from the schematic that A12 to A16 to the RAM are actually RA12-16 from the memory decoder CPLD. Only A14 on the EPROM comes from the CPLD (RA14) and this seems to be OK.

I'll also check to see the relationship between RA12/13 and A12/13.

I've checked continuity between RA12-16 on the 9572 to the RAM, and all is correct, with no shorts. Might be a faulty CPLD? It re-programmed OK though. But that CPLD has slightly distorted pins which I straightened, so maybe static damage to the IO blocks occured.

Cheers,

Phil Young

User avatar
roland
Posts: 3858
Joined: Thu Aug 29, 2013 9:29 pm
Location: Born (NL)
Contact:

Re: Building Rolands new Atom design...

Post by roland » Wed Feb 18, 2015 7:44 pm

PhilYoung wrote: It seems likely that A12/A13 on the RAM is the problem in some way. It looks from the schematic that A12 to A16 to the RAM are actually RA12-16 from the memory decoder CPLD. Only A14 on the EPROM comes from the CPLD (RA14) and this seems to be OK.
The ram is "logically" divided into 32 blocks of 4 kB. So the address lines A0 ... A11 are driven by the 6502 and the other address lines are controlled by the CPLD. That explaines why other memory is not suffering from this issue. I mixed up with my 6809 board where A13 and A12 are connected to the cpu's address lines. So forget the nonsense about replacing the 244's :oops:
PhilYoung wrote: I'll also check to see the relationship between RA12/13 and A12/13.
For these relationships the table at download/file.php?id=9814&mode=view might be a good starting point. I don't think that there are (m)any changes to the final cpld code.
PhilYoung wrote: But that CPLD has slightly distorted pins which I straightened, so maybe static damage to the IO blocks occured.
How did that happen, the distorted pins? Could be an explanation for the memory not working correctly....

Cheers,
Roland
FPGAtom: 512 KB RAM, Real Time Clock and 64 colours
MAN WOMAN :shock:

PhilYoung
Posts: 204
Joined: Sun Jun 12, 2011 5:55 pm
Contact:

Re: Building Rolands new Atom design...

Post by PhilYoung » Thu Feb 19, 2015 11:55 am

Roland, some results below.
roland wrote:Hi Phil,

This is a weird problem. I would think there is a problem with A12. Maybe you can test this signal with this program:

Code: Select all

P=#2000;[ STA #2100;JMP#2000;]
A12 should be low all the time
A12 at 6502 and 27256 both low as expected
RA12 at 628128 high (incorrect)
roland wrote: The same program at #3xxx

Code: Select all

P=#3000;[ STA #3100;JMP#3000;]
should have A12 high all the time.
A12 at 6502 and 27256 both high as expected
RA12 at 628128 high (correct in this case, but possibly stuck on high)
roland wrote:
The combination of those should give pulses on A12:

Code: Select all

P=#3000;[ STA #2100;JMP#3000;]
A12 at 6502 and 27256 pulsing (both negative 1us duration, 7us period)
RA12 at 62128 high (incorrect)
roland wrote:
I think that's a good start to look at the A12 behaviour. If that is not correct you can start with replacing the '244 address buffers. But I don't have any clue why this would not influence reading from the rom or I/O.

Greetings,
Roland
So that is fairly conclusive, the RA12 line is misbehaving. I'll do some more tests but it looks like RA12 is being pulled high by another signal, or is possibly just stuck high. So it's either a faulty 9572, or a problem with my soldering (possibly when I had to remove the PLCC socket).

So I'll order a couple of those. Whilst I wait for them to arrive I'll try programming the CPLD to just divide the clock down and output it to various pins including the RA12 one so I can determine if that pin working at all. If not, remove the CPLD and check if there is any connection between that pin and any others - if so it's a soldering problem.

I think the damage was due to dropping the thing when unpacking. I didn't notice until I tried fitting it, and noticed the slightly misaligned pins.

Thanks for the help,

Phil Young

User avatar
hoglet
Posts: 9237
Joined: Sat Oct 13, 2012 7:21 pm
Location: Bristol
Contact:

Re: Building Rolands new Atom design...

Post by hoglet » Thu Feb 19, 2015 6:01 pm

Phil,

I have had problems with PLCC sockets, with the sprung pin getting bent slightly and not making contact with the IC.

I would try checking continuity between the pin on the PLCC IC (using a fine needed to touch just the IC pin) and the pin on the RAM chip.

I think there will be a break in this path somewhere.

Dave

PhilYoung
Posts: 204
Joined: Sun Jun 12, 2011 5:55 pm
Contact:

Re: Building Rolands new Atom design...

Post by PhilYoung » Thu Feb 19, 2015 6:30 pm

hoglet wrote:Phil,

I have had problems with PLCC sockets, with the sprung pin getting bent slightly and not making contact with the IC.

I would try checking continuity between the pin on the PLCC IC (using a fine needed to touch just the IC pin) and the pin on the RAM chip.

I think there will be a break in this path somewhere.

Dave
I'll give that a try tomorrow, thanks for the suggestion. I could also try (gently) pressing down between the socket pin and the chip contact with a suitable wire and see if that fixes the problem by directly bridging the two,

Edit: I've just tried that now, no luck. The 9572 pin is definitely connected to the SRAM and not just the socket contact. I popped out the CPLD and checked there was nothing in the socket and it all looks good. Also that pin isn't connected to an adjacent one, I suppose I should check against all the other pins as well.

I don't have a spare 9572, but I do have a 9536, so I might try editting down the VHDL to become a standard Atom so it will fit and see if that works. If so that will prove the original is faulty.

Cheers,

Phil Young

PhilYoung
Posts: 204
Joined: Sun Jun 12, 2011 5:55 pm
Contact:

Re: Building Rolands new Atom design...

Post by PhilYoung » Sun Feb 22, 2015 12:45 pm

Fixed it. And of course it turned out to be something quite mundane.

Reprogramming the CPLD to output the clock increasingly divided down on the RA12-RA16 pins worked, simplifying the VHDL to just echo A12-A15 to the outputs didn't. I then remembered to check that the inputs address lines were actually active. A12 wasn't, stuck on high - and hence so was RA12 causing the original problem.

Checking the board revealed this:
Dutch_Atom_Highlight.jpg
There is a missing bit of track between the 27256 pin 2 (A12) and the CPLD pin 22. I repaired this, and it all now works.

Case closed, thanks for the suggestions,

Cheers,

Phil Young

Post Reply

Return to “8-bit acorn hardware”