Arithmetic Processor

37 posts / 0 new
Last post
macnoyd's picture
Online
Last seen: 14 min 45 sec ago
Joined: Oct 15 2012 - 08:59
Posts: 436
Arithmetic Processor

Does anyone here know anything about this board:  Arithmetic Processor AP-1.  It's an oldie.

I've had this for quite a while now and have virtually no information or software on it.

The photo here was taken off the web in my search but I have virtually the same board with a different S/N.

 

[UPDATE]

I found the manual to the board, but there is a program disk called " CCSOFT" that operates it.  If anyone

has this, I'm happy to share the collection (though limited) of what I've found so far.

 

AttachmentSize
Image icon Apple][ Arithmetic Processor99.82 KB
speedyG's picture
Offline
Last seen: 2 weeks 2 days ago
Joined: Nov 16 2011 - 07:45
Posts: 2491
I doubt the card to operate

I doubt the card to operate even with correct software in correct mode without the missing firmware on the PCB ( missing ROMs at lower left side of the picture )....

SpeedyG

macnoyd's picture
Online
Last seen: 14 min 45 sec ago
Joined: Oct 15 2012 - 08:59
Posts: 436
Speedy, the manual says it will ...

Thanks for responding Speedy.  Those ROMS are to hold a small machine language routine for running the card WITHOUT the disk, according to the manual.  So you are half correct.  :-)

But there is no reference as to where or how to get the ROMS in the manual.  They only specify that they are 256 by 8 bytes. (not very big!)  Likely an MMI 6300 series fuse link PROM.  There is also reference to putting RAM in those sockets as well, but for unknown reasons not described in the manual.  (The manual is somewhat vague regarding ROM or RAM useage)  But the software is most important for running the card, and they address it as being Integer and AppleSoft compatible.  Reverse engineering the card will be a significant hassle, as the designer/builder made this card to perform scientific calculations by setting an interrupt at a specified location and jumping to a routine (located on that CCSoft Disk) that executes the calculation(s) and returns the value & status back to the Apple II for display or whatever.  I don't think I would be able to make this card work without that disk.  At least, not very well.

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Dear macnoyd,

Dear macnoyd,

I'm probably able to give you some hints because of... a couple of weeks ago I started to build a Sinclair ZX / ZX81 APU Prototyp card based on an article of an old German computer magazin.

So even that the ROM seems to be missing on your card you are probably able to use the card with some Basic coding as we started here in Germany to do some simple tests.

APU Intel 8231A or AMD AM9511A    are pin compatible

So beside some coding we started with using the original Spectrum ZX ROM without any changes:

Used I/O Addresses to communicate with the APU are as follows, while the APU is using two registers for the calculation(s):  a Data- and a Command Register

D=Data-Adresse             = 31d -> 1Fh    (or  125d -> 7Dh)
C=Command-Adresse    =  95d -> 5Fh   (or 127d -> 7Fh)   

in our case we toggle address bit A6 to switch between Data and Command register  (see my attached schematic snapshoot), so the Apple I/O Addresses needs to be used here, instead.

APU Testprogramm using the Standard Speccy 48K ROM
Arithmetic APU Operation -> 16 Bit Addition: Command 108d -> 6Ch

 

PROGRAM: APU-Test


100 REM TESTPGM 8231A or AM9511A APU
110 LET C=95: LET D=31       REM 95d = Command Register; 31d = Data Register
120 OUT D,10: OUT D,0        REM load value 10 to the APU
130 OUT D,20: OUT D,0        REM load value 20 to the APU
140 OUT C,108                REM use APU command 108d -> 6Ch to do a 16nd Bit Addition
150 LET A=IN D: LET B=IN D   REM get calculation result(s)
160 PRINT A                  REM A should held 0
160 PRINT B                  REM B should held 30   -> which is the result of  10 + 20  (see above
)


 

So I hope that you are able to adopt this to get the Apple card running. Maybe you could upload an image of the backside of your card, so we might be able to re-engineer ? If you could share a BOM that would be great as well.

 

:Update: maybe you can take off all ICs and take a new picture of the Top layer, too?  This might help to see the pcb tracks and to reengineer the card to create a schematic.


Have fun...

Torsten (Berlin - Germany)

macnoyd's picture
Online
Last seen: 14 min 45 sec ago
Joined: Oct 15 2012 - 08:59
Posts: 436
Tokabln,

Tokabln,

Thank you very much for your response.

Maybe late today (if I'm lucky) I will desolder all the parts on the board and take a snapshot of the front and back of the card.

Stay tuned...

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Hi macnoyd,

Hi macnoyd,

glad to share you the information's, while I don't believe that you need to desolder all ICs, just removing them from their sockets might be enough... to see the trace routing.

 

So it's up to you

 

Have fun...

 

 

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Hi macnoyd,

Hi macnoyd,

 

I just found and downloaded the AP-1 manual and took a look...

 

so command 6C, as used above at the shared code, is mentioned at page 1-7, so I do believe that you should be able to address the card via BASIC as we did with our ZX Spectrum computers. To understand how this works, please use the shared PDFs.

 

And I do found a company using the same name as referenced in the manual, so maybe it's worth to contact them.

Update: the folowing link is wrong... this company stays in Saudia Arab and not in California, so it has nothing to do with the Apple II products company

https://www.californiacomputers.net/

 

Kind regards

Torsten

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Hi macnoyd,

Hi macnoyd,

 

as I thought myself... CCS did something similar as some folks of our Sinclair community did... patching the ZX Spectrum / ZX81 ROM. CCS "patched" Applesoft to use the APU instead of the 65(C)02 when using some specific mathematical instructions which are supported by the APU.

 

"Apple APU. The Model 7811 B Arithmetic Processor Unit is intended to increase the execution speed of Applesoft II programs and the number of math functions available. The system employs an AMD9511 APU and plugs into one of the Apple expansion slots. The CCSoft interpreter is loaded from the diskette provided, and the system is ready to, go. The CCSoft interpreter is identical to Applesoft, except that CCSoft sends arithmetic functions to the APU for fast execution. Additional functions include AsiN(x), ACOS(X), LOGIO(X), SINIH(x), COSH(x), TANH(X), INVERSE(x) and PI. "

 

Address:
California Computer Systems,

250 Caribbean Dr.,

Sunnyvale, CA 94086

(Tel: 408-734-5811)

 

So you need someone (or yourself) who's able to implement such patches into Applesoft.

 

Source: https://www.americanradiohistory.com/Archive-Poptronics/80s/1980/Poptronics-1980-06.pdf

(page 40)

macnoyd's picture
Online
Last seen: 14 min 45 sec ago
Joined: Oct 15 2012 - 08:59
Posts: 436
Wow! Really good information!

Wow! Really good information!  Very much appreciated!  You spent a bit of time on this.  Thank you again!

 

While you were finding all this, I decided to go ahead and desolder the board and photograph it to share with you and others.

While the IC's are all marked on the componernt side of the PCB, the resistor & cap values are not, so I will divulge them here so everyone knows.

All discreet resistors are 220 ohms except R6, which is a 2.7K ohm.

All capacitors are 0.1uf decoupling, except C4, which is a 270pf silver mica type.

Q1 & Q2 are FPN2907, and the 8-lead resistor pack is 2.7k.  Pretty much covers it.

Photos:  Larger versions are available upon request.

 

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Hi macnoyd,

Hi macnoyd,

 

that's great... many thanks for that. In between I found the schematic of the board which is attached to the Documentation. So yesterday I started to redraw the schematic using KiCAD. Your pictures are more than welcome as this will help to place the traces as done by CSS. It will take some time to have this finished. There's also one IC which is an Octal Transmitter/Receiver and which seems to be very rare... so maybe another one needs to be used here. All other components should be available... while for the Transistors a BC556 should be used instead of the one used by CCS.

 

I also found additional information about the company... if these are right they seemed to start with S100 Bus System components. Will share the link when I'm at my desktop computer.

 

Stay tuned

Torsten

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
A California Computer Systems

A California Computer Systems history

I hope that this is the right company who has developed and build these cards... because of most stuff I found is related to S100 Bus systems

 

http://www.retrotechnology.com/herbs_stuff/CCS_mason.html

 

Online
Last seen: 3 min 7 sec ago
Joined: Aug 18 2017 - 16:53
Posts: 33
There were some more Applebus
macnoyd's picture
Online
Last seen: 14 min 45 sec ago
Joined: Oct 15 2012 - 08:59
Posts: 436
Tokabln, looking at the schematic ...

Tokabln, looking at the schematic and comparing it to the actual board, you're going to see many differences, so you might want to re-draw the schematic from the board.  (just saying...)  I found a lot of differences when I drew the schematic from a WildCard board.  Other boards as well.  I will eventually re-draw this one but I want to finish up a previous project first.

It would be really nice if someone had that CCSoft disk they would share... (calling out all AF members... )  :-)

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Beside the first version

Beside the first version using ABACOM SPRINT using the uploaded PCB pictures as template. Hope I didn't mist any trace.

 

Schematic needs to be redrawn using KiCAD...

 

 

UPDATE: both pictures are showing the final version now:

 

Top view

 

 

Bottom view

 

 

 

 

 

 

Online
Last seen: 3 min 7 sec ago
Joined: Aug 18 2017 - 16:53
Posts: 33
tokabln wrote:Schematic needs

tokabln wrote:

Schematic needs to be redrawn using KiCAD..

Are you planning to produce some boards? If yes, what about a more "modern" version with replacement of the DP8304 to 74LS245 or similiar and a RAM chip which is available now?

 

Regards

Ralf

 

macnoyd's picture
Online
Last seen: 14 min 45 sec ago
Joined: Oct 15 2012 - 08:59
Posts: 436
Amazing!!

Amazing!!

I can't believe how fast you put that together.

I just did a visual trace and it looks like you've got everything there.

If I were to suggest anything at all, I would set the RAM jumper and INT jumper to accept a 2.54mm (0.1) spacing molex pins rather than a solder bridge.

That way, you can use those convenient molex jumpers.

 

JOB WELL DONE!!!!  WOW!!!

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Hi Ralf,

Hi Ralf,

 

well I'm actually not sure if I'm willing to setup a batch of boards. First task was, to preserve what was done in the past by CCS.

Next step might be to change the DP... to a more modern Transceiver/Receiver stuff. I'm also not sure yet how to address this card, as only Addressline A0 is used here to toggle between the Command-/Dataregister. But that's it,  so it's very different to what we did with the ZX Spectrum. Here we used a fixed address to load/unload the mathematical stuff, toggling between Data- and Command Register as shown above within the shared Basic code.

 

So I believe we need to understand the philosophie behind what was done for the Apple, which I haven't done yet... because I'm not very familiar with the Apple Bus.

 

@macnoyd... yes you are right,  INT / RAM should be using a small Jumper instead of a wired brigde. I will change this and will made the GERBER files available. It doesn't make any sense to produce boards and to ship them to worldwide afterwards as the costs are to high than. So if someone like to (re-)build the original card, he can do that and maybe have some spare for folks within his country.

 

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
CCS APU AP-1 - Clone - Gerber, BOM & Pictures

So I just finalized the SPRINT layout and build a ZIP file containing

 

  • GERBER files for manufacturing
  • BOM
  • Pictures of the board

 

If someone is interested please send me a PM with your eMail Adress and I will share the file, because I can't upload athe ZIP here.

Those who are interested in, please double check the included board pictures with the original pcb pictures shared by macnoyd, just to prevent mistakes. Even that macnoyd has done a precheck, it's not my fault if you are getting wrong pcbs.

 

@macnoyd: I will send it to you in a few minutes using your shared eMail address.

 

 

 

Online
Last seen: 3 min 7 sec ago
Joined: Aug 18 2017 - 16:53
Posts: 33
tokabln wrote:First task was,

tokabln wrote:

First task was, to preserve what was done in the past by CCS.

Thanks. Gut gemacht :-)

tokabln wrote:

I'm also not sure yet how to address this card, as only Addressline A0 is used here to toggle between the Command-/Dataregister. But that's it,  so it's very different to what we did with the ZX Spectrum.

The /DEVSEL line addresses the FPU. /IOSEL addresses the SRAMs. Means in slot 1 the addresses of the FPU are $C09n.

 

SRAM: there are two SRAMs 256*4 in two 16pin sockets. I can only find one chip that fits: Intel 2112. Very rare?

 

Regards

Ralf

 

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Hi Ralf,

Hi Ralf,

 

thanks for the explanation... so that make sense... because CCS talks about using additional RAM or additional ROM within the manual. So maybe I can change the RAM part from 2 x 2112 to another 256 Byte RAM type. Let's see.

 

Also thanks for the card address... I found it in the doc right now... but just one question left... how would this address ($C09n) change when A0 is toggled...

  • are we then talking about $C090 when A0 is low
  • and $C091 when A0 is high?

The reason why I'm asking is, that A0 is toggling between the Data-/Command Register of the FPU. In such case recreating the above ZX Spectrum code shouldn't be a problem, so macnoyd could start some testing when he's resoldering the card ;-)

 

As a DP8304B Octal Bus driver replacement a 74LS244 (not pin compatible) seems to be okay to do that job.

 

Thanks again and maybe we can discuss offline (in German) how to enhance the card. I will send you my eMail Address via PM.

macnoyd's picture
Online
Last seen: 14 min 45 sec ago
Joined: Oct 15 2012 - 08:59
Posts: 436
Fantastic job well done by Tokabln,

Also want to note that the buss driver IC is readily available, so not worth changing unless you do a board re-design.

https://www.ebay.com/itm/282106996683/

The 2112 sram are rare. 

[EDIT] But not too rare ... https://www.ebay.com/itm/323529866297

 

I was going to check the datasheet of MMI 6301 and see if  the pinout is compatible.

[EDIT]  The MMI-6301 (256x4) seems to be compatible right down to CE, so I believe a pair of these would be required for resident Apple ][ code.

I've gotta guess these would be "fun" to program, LS nibble in one & MS nibble in the other.  That and these are fuse link PROMS, no room for error.

But they were (at the time) what was available at reasomable cost.  EPROMS were much more expensive in the late '70's.

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
That's great, just purchased

That's great, just purchased a couple [EDIT:] of the DP8304  ... probably for the German community ;-)

 

 

Online
Last seen: 3 min 7 sec ago
Joined: Aug 18 2017 - 16:53
Posts: 33
macnoyd wrote:The 2112 sram

macnoyd wrote:

The 2112 sram are rare. 

After reading (and understanding?) the manual I think having this RAM is not necessary. It's a nice feature to have 256 bytes of RAM (or ROM) for 6502 code, but now in 2019 it's expensive to get these obsolete chips. Means: no need to populate Q1, R4, R5 and the jumper for RAM.

 

Regards

Ralf

 

 

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Hi Ralf,

Hi Ralf,

 

yes you are right. I will put this hint onto the schematics including that 2112 RAMs are used at position U4 / U5.

 

Have fun

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Hi macnoyd,

Hi macnoyd,

 

as long we do not have any content to program the PROMs, there isn't any demand to find a programmer ;-)

 

Btw: just sent an email to a guy while hoping he's able to help us here... you are on bcc:

macnoyd's picture
Online
Last seen: 14 min 45 sec ago
Joined: Oct 15 2012 - 08:59
Posts: 436
Got it ...

tokabln wrote:

As long we do not have any content to program the PROMs, there isn't any demand to find a programmer ;-)

Btw: just sent an email to a guy while hoping he's able to help us here... you are on bcc:

 

Yes, got it.  Thanks for keeping me in copy.  I hope he responds and has something for us.  That would be great if he does.

 

Regarding the ROM / RAM sockets, I think it would be relatively easy to change this to a single 2K x 8 socket instead of the (2) 256 byte by 4 configuration that currently exists.  At the time, the 2112 and 6301 were good choices, though today, maybe not so much.  Fortunately, there's a 2K x 8 RAM & EPRom that closely matches, pin for pin.   For me though, there's something nostalgic in keeping the design like it is.  Not all would share this opion, but that's OK.

I can draw up a schematic using 2716  EPROM or 6116 SRAM layout if  it's a consideration, but I'd like to do it only after we obtain the CCSoft disk.  Without that disk, we would need to re-engineer the software, and I'm not in a good place to do that right now.  Too many other projects in the works.  The CCSoft disk is what makes this board desirable, in my honest opinion.  I'm hopeful that this disk exists somewhere and is discovered.

 

 

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Yes... indeed. The first

Yes... indeed.

 

The first Gerbers are for the nostalgic folks... while I'm actually in contact with RalfK. He has some good ideas and some requests that such a "new" card should offer. Nevertheless I believe that I will produce up to 5 cards (maybe 10) for the German community based on the old design. For my personal card I just ordered some RAM, to have a more historical version ;-)

 

So when I'm done with the CCS original schematic (while I will take care to use the card as reference), I probably will start to do a redesign (probably with Ralf) and will share and discuss with you folks if something else is needed or requested.

 

Torsten

macnoyd's picture
Online
Last seen: 14 min 45 sec ago
Joined: Oct 15 2012 - 08:59
Posts: 436
If you are ordering cards

If you are ordering cards and are getting them with gold immersion on the card edge, I'd like to buy one from you if you have one to spare.  :-)

Please keep us posted when something new comes about.  Great work!

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Folks,

Folks,

 

today I received some interesting hints and links from the German C64 community where I'm a forum member, too...

 

Write your own 9511 software ?
NASA paper - https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19850026198.pdf

Compute 11/1980 magazine
Compute 11/1980 - https://www.commodore.ca/gallery/magazines/compute/Compute-007.pdf

 

Have fun... while I'm still seeking to get more updates... so stay tuned.

 

Torsten

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Just to let you know... the

Just to let you know... the cards seems to be shipped be the manufacturer seeed... so I will keep you posted when the raw pcbs arrived.

 

Torsten

 

Update: I ordered 20 pcs. of raw pcbs (5 are already reserved) and I do have preorders from

 

Preorders as of: 02/26/2019

1 raw pcb by macnoyd

1 raw pcb by stynx (not sure if I'm able to deliver incl. parts)

1 raw pcb by ralfk

1 raw pcb by speedyG

2 raw pcb by myself

 

I too ordered 10 pcs. of the DP8304B for the German community while I will take at least 2 pcs., but not sure when they will arrive.

 

Update: 02/26/2019

added speedyG to the preorder list

 

 

speedyG's picture
Offline
Last seen: 2 weeks 2 days ago
Joined: Nov 16 2011 - 07:45
Posts: 2491
I will add an order for one

I will add an order for one aw old card  too.
SpeedyG

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Folks,

Folks,

 

the raw pcbs arrived today. So I need to build at least one card to see if they will work. Beside two pictures of the card.

 

Will keep you posted.

 

 

 

Have fun...

 

 

 

Online
Last seen: 3 min 7 sec ago
Joined: Aug 18 2017 - 16:53
Posts: 33
Great :-)

Great :-)

macnoyd's picture
Online
Last seen: 14 min 45 sec ago
Joined: Oct 15 2012 - 08:59
Posts: 436
Nice work

Looks pretty authentic.  Well done!

Offline
Last seen: 5 hours 34 min ago
Joined: Jul 5 2018 - 09:44
Posts: 252
That board looks great!  I

That board looks great!  I hope it works as good as it looks!

 

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Folks,

Folks,

 

beside the populated CCS 7811 Clone card (except the RAM/PROM chips).

 

Next step will be to do some testing and to document the building and some technical stuff.

 

Some additional words.

I do have a couple of raw boards available and a few spare parts to serve the above requests.

I'm not able to sell the APU (Intel C8231A) as I do not have them available anymore. 

I can also share the transistors (if the testing will work), the R-pack and probably some TTL which are not typically available and only a small batch of the DP8304 gate.

 

I will setup a list of what's available and what's not in a few days.

 

Have fun...

tokabln's picture
Offline
Last seen: 1 day 19 hours ago
Joined: Dec 30 2015 - 10:48
Posts: 154
Dear folks,

Dear folks,

 

the card is working... beside two test programs I used.

 

The first one is taken out of the CCS Owners manual, while the second once is a small basic program doing a short calculation, like 10 + 20 = 30

 

This means I can offer the card and some (not all) parts to a couple of folks, while I will start with the above list.

I do need to do some calculation and to investigate relevant shipping costs, at least for the US folks.

 

Beside the two progs as well as the CCS owners manual.

 

 

 

Log in or register to post comments