I was invited to help debug an original Apple 1 in a few days and I thought I would see if the folks in this group would like to share some ideas. The machine is number 33.It is stuck with the whole screened filled with the pattern "@0" and my belief is that this pattern is simply the powerup contents of the video shift registers and that somewhere in the CPU data path up to the PIA video output has an issue.My plan is to run a range of tests and tools using my MCL65+ in place of the 6502. It can simply emulate the 6502 to see if the original ceramic 6502 is broken, can emulate the motherboard's RAM and ROM, and can perform hard read/write/dump tests which it can report to us via the USB serial interface. I welcome any ideas or suggestions to debug this rare and historic machine. Thanks,-Ted
Anonymous
User login
Please support the defense of Ukraine.
Direct or via Unclutter App
Active forum topics
Recent content
Navigation
No Ads.
No Trackers.
No Social Media.
All Content Locally Hosted.
Built on Free Software.
We have complied with zero government requests for information.
Don't forget to take some pictures and post them here if they let you!
Absolutely! I am very excited and honored to get the chance to work on one of these historic computers.
Can you share any experience/advice on typical failures for these original machines? Are any chips more susceptable to failure? Do chip sockets fail?
I have a script of things to check in mind, but what would your plan of attack be?
Cool, I've never seen the original Apple-1. Can I give one piece of advice?
IMG_20210317_091040.jpg
Yes, the owner says they do get this repeating pattern - but he said that he swapped most or all of the video section chips in an attempt to fix but no luck.
Would we also get this same screen results if the CPU was inoperable? I believe the main interface between the CPU section and the video section is through port-B on the PIA - so that is one of the first places I intend to look at. If there is no activity then I assume the video circuit will just display the pattern held in the uninitialized 2504's.
I am also wondering why the character @ would come out on uninitialized 2504... It is ASCII character 0x40 - So does that mean that five of the 2504's are outputting 0's and only one of them is stuck at a 1? Or maybe the character ROM 2513 is addressed to the the '@' character? at input address 0x00?
In theory, if one of the 2504's was bad then the CLR should not be able to clear it.
In post #5, MicroCoreLabs wrote:
"I am also wondering why the character @ would come out on uninitialized 2504... It is ASCII character 0x40 - So does that mean that five of the 2504's are outputting 0's and only one of them is stuck at a 1? Or maybe the character ROM 2513 is addressed to the the '@' character? at input address 0x00?"
Uncle Bernie answers:
Half of the shift register cells in these 1404/2504 work inverted, actually there are two blocks of 512 bits inside, running from opposite clock phases, so the designers got twice the speed, one bit in and one bit out per clock phase going low (to -12V). They called that "double pumped". Note that PMOS clocks are active when at a negative voltage.
So half of the bits are '0' on power up and half are '1' (even bit / odd bit wise) and together with the cursor logic this makes the @_@_@_@_ ... pattern on power up where the '@' should blink as all of those are 'cursors'. A CLR SCREEN is mandatory to clean up this mess before you give it a RESET. Otherwise the many 'cursors' will cause the cursor state machine to have the electronic equivalent of a nervous breakdown (erratic behaviour of output on screen).
If you pull all the 2504 you can selectively pull the output pins in their sockets "H" to see if all the downstream logic works. Use a ~390 Ohm series resistor on a probe wire to +5V for that. The resistor value is not that critical but helps to prevent blowing things up. You do not want to poke around with anything directly connected to +5V.
- Uncle Bernie
Ah.. DDR.. makes sense now. But if I we are getting stuck at this pattern then is it possible that the CPU simply has not udated it yet? Or another reason could be that the CPU is trying to update it but the cursor 2504 is broken and stuck outputting zeros.
I can probably find out by probing the CB2 output from the PIA.
The CLR signal from the keyboard looks asynchronous to the CPU, so maybe would temporarily clear the bits to the 2519 and I should see a uniform pattern on the screen?
If there is no activity then I assume the video circuit will just display the pattern held in the uninitialized 2504's.
The CPU does not initialize the shift registers: only the hardware can do that when CLEAR SCREEN is pressed. The Apple I does not have the ability to clear the screen from software, which is why there is a CLEAR SCREEN key directly connected to the terminal hardware. (The terminal interface is a true "glass teletype": the only thing it is capable of doing is to "print a single character at the current cursor position and advance the cursor" when the PIA writes to its port.)
If CLEAR SCREEN is not used after powerup (conveniently pressed simultaneously with the RESET key), the shift registers will contain garbage. The RESET key, itself, has no connection to the terminal section. It's interesting that the original manual doesn't state this, but at the time it must have seemed too obvious to remark.
The owner reports that nothing happens - maybe a flicker - when the CLEARS SCREEN is pressed. The screen remains full of the @0 pattern. He also reported that he swapped all of the video ICs.. So I should follow up and confirm this. Perhaps the OR gate at C9 is bad and the CLR signal is not generated... What else could cause this hehavior?
This is the display.
It appears shifted to the right with potential overlapping/wraparound at the edge rows. The user confirms that he swapped all of the video chips and pressing CLEAR does not change the display.
Holding the CLEAR buton should pass a '1' through the OR gate C9 to the MUX C4 and C14 and force zeros on all six outputs which would both cause the 2504's to be cleared a well as feeding 0 to the 2519 which should clear the screen as long as it is pressed. It seems to me that there could be a fault in the path from the keyboard key to the output of the MUX C4/C14...
Screen Shot 2023-04-20 at 7.36.02 PM.png
Another question: Are these multi-voltage chips sensitive to powerup sequence? What if filtering caps on one rail are bad and delay the rise of a voltatge rail so that the IC only recieves voltage they are skewed by milliseconds?
Is tampering with an original Apple 1 endangering it's collectible value?
Tampering - yes, fixing it - no. It's always more valuable if it works.
Hi,
You wrote it would be machine number 33.
Do you mean it has a serial number 01-0033 written on the back? In this case it is maybe not listed in the Apple-1 Registry.
Would be great to add it with some pictures and information. Most important is to preserve the history.
Thank you
It's up to the owner if they want to diclose to the world that they posess such a valueable asset.
Because of that I am asking you if you would forward my request to the owner. You already wrote here that you want to publish pictures here at applefritter. Now it is no secret anymore that a #33 exist but before I add it to the Registry without any further information and pictures it is better to hope and wait for pictures.
Maybe the owner is unaware about the Registry? It is the mission of the Registry to preserve information of all Apple-1.
He can contact me anytime. I have preserved already information about Apple-1 that are NOT listed because the owners didn't want to publish it.
Of most Apple-1 only pictures and information are bpublished but nothing about the owner.
Did I?
They are.