Half problem solved with 74123 ic. But now I have one more minor problem.

17 posts / 0 new
Last post
Offline
Last seen: 1 month 12 hours ago
Joined: Jan 30 2021 - 11:55
Posts: 63
Half problem solved with 74123 ic. But now I have one more minor problem.

Hello!

 

As you read the title I had problem with rolling " \\\\\\\\\\\\\\\\\\\\\\\\\" on reset screen. I read the topic about 74123 or the 27k resistor  next to it that maybe need replacement in order to get 480ns timing for the ram ic's.

 

Anyway, before i "play" with the resistor i replace the 74123 with new one that had already and problem fixed. No more rolling "\\\\\\\\" and reset screen was great!

What happens now, but not often, some times in reset screen before "@" symbol i got lines like that "_ _". Some times only one line, some times more. Some times rolling for sort of time to more than 10 lines, then stop and maybe 3-4 minutes later some more lines appear. But as I says that happens not often.

 

I tried to measure with oscilloscope the timing of 74123 (pin 5, is that correct?) but i saw no stable "ns". I have to say that i am very new to oscilloscope and maybe I did faulty measurements.  

Any help will be more than welcome!

Thank you!

 

UPDATE: 

I take of the 27k resistor and measure it 30,6k. I add a trim pot and drop it to 27k. I don't know if that was the problem, but no line's till now. We will see.

But still interested to get a proper measure for the 480ns timing, so of course i will wait for your reply's!

 

 

 

Offline
Last seen: 20 hours 42 min ago
Joined: Apr 1 2020 - 16:46
Posts: 842
How to fix 74123 Timing woes ...

Read post #10 of this thread:

 

https://www.applefritter.com/content/how-bad-are-mk4096-chips-well-let-me-tell-you

 

Measure the 480ns pulse width the the 74123 pin #13. This pulse always is available.

 

For the other half, the 3.5us, you measure at 74123 pin #5 and change the "10k" resistor adjacent to the 74123. But this pulse occurs only once per character that is output.

So you need some test program to get continous screen output. And for this the DRAM must work. And you need a keyboard unless you have the PROMs which come with my famous IC kits (they have a built-in diagnostics page, no keyboard needed to get an Apple-1 going). Here is a link to a small continous screen output program:

 

https://www.applefritter.com/content/smallest-test-program-ever

 

All the info you need to successfully build and debug an Apple-1 is here in this forum. I wonder why people always ask the same questions again and again. Is the search function on Applefritter that bad ?

 

Offline
Last seen: 1 month 12 hours ago
Joined: Jan 30 2021 - 11:55
Posts: 63
Thank you very much for your

Thank you very much for your answer and sorry that I didn't do a proper search before my posting. I will more careful in the future.

 

Now. I did measure at pin 13 of ic 74123 the pulse width. With 27k i had 310ns. To fix that i adjusted the trim pot I had in place and dropped it to 20,4k to be able to have 480ns timing. Is that ok?  

My question is: The ram pack i installed in my replica is MK4027N-2 and not the original ones. Looking at the data sheets i saw that the MK4027N-2 have 150ns access time & 320ns cycle.  Should i better adjust the timing via the trim pot to something like 300ns or leave it with 20,4k and 480ns timing?

 

 Thank you!

Offline
Last seen: 20 hours 42 min ago
Joined: Apr 1 2020 - 16:46
Posts: 842
In post #3, Zijjel wrote: 

In post #3, Zijjel wrote:

 

"Should i better adjust the timing via the trim pot to something like 300ns or leave it with 20,4k and 480ns timing ?"

 

Uncle Bernie answers:

 

Unless your DRAM does not work, don't change the  480ns timing. All my kits run at 480ns with the MK4027 or equivalent from Mostek, Motorola, or Intersil without any problem ("reliability mods" may be needed).  

If you see DRAM errors with a competent diagnostics program, you may change the DRAM timing a little bit to see if it's a timing issue for  a"bad" DRAM, but later change it back to 480ns.

BTW, the "310ns" pulse widtth  for the 27K resistor is pathological (the 74123 can't complete the full pulse). 20K looks right to me, most 74123 I have encountered need a resistor around that value (19K to 22K) to produce the specified 480ns.

 

The reliability mods are here, post #4 of this thread:

 

https://www.applefritter.com/content/part-path-towards-rock-solid-apple-1-builds

 

In case of DRAM trouble, add the six damping resistors first. Only when this does not help, add the capacitors. So far every Apple-1 using these reliability mods stopped to act up and runs rock solid. But not all builds need the reliability mods. When using high performance bypass capacitors (modern multilayer types, X or Y class ceramics) and when being lucky, the Apple-1 may work OK without the reliability mods. Alas, almost no disc shaped 100nF capacitors use the high performance ceramics, so if you want to make a "1st production run" clone ("BYTE SHOP" version) and use brown disc shaped bypass capacitors, you most likely need the "reliability mods". These greatly mitigate the ringing and the spikes on the power supply lines. The damping resistors mitigate the ringing on the multiplexed address lines for the DRAMs. Some DRAMs  are more sensitive to ringing multiplexed address lines than others. Most of the Mostek MK4027 I had did not need the damping resistors. Motorola MCM4027 needed them more often than not. Intersil always need the damping resistors but it seems they may not need the added  bypass capacitors, but this is based on the first few specimen encountered, so it may change, while with the Mostek and Motorola the experience is based on a much larger sample size close to 1000.

 

A nice can of worms we got here. Every Apple-1 builder should be aware of this. But since yours truly has analyzed the problems and published the remedies, every Apple-1 build can succeed and lead to a robustly working machine. But those who do not follow the path towards rock solid apple-1 builds I have shown you, end up in a world of hurt and embarrassment. "Purists" of course want a machine with no reliability mods, so they can see it crash as often as the originals reportedly did. The  2nd production run ("NTI" version) did use the little blue boxy looking bypass capacitors for a reason. Ironically, on the photo of the prototype Apple-1 used in their ads you can see that this prototype used the good bypass capacitors and not the cheaper disc shaped ones. We can make the conjecture that the prototype probably did work OK, but there was not enough money to buy the higher performance bypass capacitors for the 1st production run, and this is why the Apple-1 is so notorious and so rare and worth $1 Million at auctions  ;-)   ... had it worked OK they would not have bought them back to destroy them. Which is part of the myth and why these originals are so rare and expensive.

 

Offline
Last seen: 1 month 12 hours ago
Joined: Jan 30 2021 - 11:55
Posts: 63
Thank you very much for your

Thank you very much for your great help and congratulations for the way you explain things.

Offline
Last seen: 1 month 12 hours ago
Joined: Jan 30 2021 - 11:55
Posts: 63
.
 ahhhh... Don't want me to relax a bit :)

 

Look at the "0" down right corner. Stuck there. Any advice please?

 

 

Offline
Last seen: 20 hours 42 min ago
Joined: Apr 1 2020 - 16:46
Posts: 842
Don't worry about the power on screen !

In post #6, Zijjel wrote:

 

"Look at the "0" down right corner. Stuck there. Any advice please?"

 

Uncle Bernie answers:

 

This looks good. Terminal section is functional to a high degree. The next thing to find out is if the CLR SCREEN works. While CLR SCREEN button is pressed, screen must go completely blank. When button is released, a blinking cursor must appear in the upper left corner. WARNING: using jumper wires in lieu of a switch has killed many Apple-1 builds (because of miscounting the pin # on the keyboard socket). Instead, I recommend you to build a "Gimmick Switch" (search for it).

 

Never try to start the Wozmon (by RESET) before the screen has been cleared. If there is garbage in the screen memory then there are garbage cursors in the the cursor memory, too, and this leads to weird effects when any software tries to do screen output.

 

Not exactly consumer friendly. Worse, owners of Apple-1 tend to forget the correct start up procedure (RESET and CLR SCREEN both activated, then release CLR SCREEN and watch cursor to blink, then release RESET). And then they think that some ICs went bad.

 

It is important that you can see the blinking cursor before releasing the RESET. Because if there is a bad IC, and this prevents creation of a cursor,  the screen output will hang and the Wozmon will lock up. Wozmon also won't work if the MSB of the keyboard code is not tied "H".

 

Don't laugh - these little pitfalls have befallen so many builders that I could write a book about it.

Offline
Last seen: 1 month 12 hours ago
Joined: Jan 30 2021 - 11:55
Posts: 63
 That's are great news! Thank

 

That's are great news!

 

Thank you one more time for the useful informations. I did the test's you told me with clear & reset. Everything works great.

 

No, I don't use jumper wires for clear & reset. That's can be very dangerous specially when you are excited!  

I build a very easy pcb with button's to be sure.

 

 

 

Looking at you title on your last post: "Don't worry about power on screen!" I want to ask if really don't matter. Even if i see much more garbage on power on screen (specially in fast reboot happens) but clear and reset work ok, means that everything is smoothing and work properly? No need to worry?

 

 

 

 

Offline
Last seen: 20 hours 42 min ago
Joined: Apr 1 2020 - 16:46
Posts: 842
Apple-1 start screen indeed is irrelevant - best ignore it !

In post #8, Zijjel asks:

 

"I want to ask if really don't matter. Even if i see much more garbage on power on screen (specially in fast reboot happens) but clear and reset work ok, means that everything is smoothing and work properly? No need to worry?"

 

Uncle Bernie answers:

 

Yes, indeed, what I said is true: Don't worry about the start screen (as long as there is a character field, of course ! --- if you see vertical zebra stripes the 74166 at D-1 is bad, and if  the whole picture area is filled with '@' (and not only 40 x 24 characters) the 2519N is bad (most of the time). But random "garbage characters" in the regular 40 x 24 character field are OK !

 

I don't know how the myth that a "_@_@_@_ ..." pattern is the "right" power on screen came into being. This is nonsense and has confused many builders. True that this pattern is likely to appear after the Apple-1 has been turned off for a prolonged period of time, but it is not guaranteed at all, and some minor glitches may happen and then first time builders like Zijjel  get spooked by the occasional "wrong" character in the start screen (see post #6 above).  I've explained how the 1k x 1 shift registers work and why that pattern is generated here:

 

https://www.applefritter.com/content/apple-1-start-screen

 

But this is an artifact and it is not guaranteed that all of the shifter cells in all the 1404 "wake up" in exactly the same way. All this happens while the power supply voltages are still ramping up and so some logic gates may trip earlier then others. The Apple-1 has no power-up reset and it just so happens that all the counters, clock pulses, etc., wake up and then - hopefully - after some 100000 clock cycles the whole logic works properly. Depending on how the power supplies come up, you may even see the screen starting to scroll the garbage out to the top. This rarely happens with transformers but if you use switchmode power supplies (which ramp up slower) it is seen now and then. People think this is a bug. No, Woz used one spare TTL gate to implement this feature: any "fake/garbage" cursors in the vertical blank region cause the scrolling so they get moved out of the cursor memory area corresponding to the vertical blank (which could throw the cursor logic state machine off track).  Also note that the Apple-1 video section has no dedicated logic for blanking the  screen outside of the 40 x 24 character field ... the blanking is done within the character data path, so any faulty IC there can cause extraneous characters (or trash) appear outside of the 40 x 24 character field.

 

All these issues (except potentially bad ICs) go away with the CLR SCREEN.

 

If you turn off the Apple-1 with text on the screen and turn it on after a brief period (seconds), you can see the old screen contents again, but in a garbled and moved / scrolled state. Again, this is normal, and CLR SCREEN will fix it.

 

The Apple-1 is the type of  quirky product you could not unleash against the unsuspecting computer user of today. The spooky "terminal section" behaviour patterns alone would ring the service hotline off the hook, even if everything works OK and there are no bad ICs. Then add in the frequent program crashes, the notorious ACI, and the other quirks, and you can see how the 150 or so Apple-1 that had been sold indeed rang the phone at Apple off the hook ... and only Woz understood his creation well enough to answer these calls, greatly distracting him from working on the Apple II and the Disk II. This is why they decided to buy the Apple-1 back to destroy them. Of course nobody got his money back - Steve Jobs was too much of a shrewd businessman already - for each Apple-1 turned in the owner got a discount voucher towards the brand new Apple II, or so the story is usually told.

 

Now, with the Apple-1 cloning scene, all these gremlins come back to haunt the unsuspecting builder who does not expect such quirks and shenanigans. The builders of my kits get a 62 page long "Tips and Tricks" pdf which (if they read and follow it) protects them from the most common worries and pitfalls. Every pitfall and difficulty reported by "my" builders went as a "Tip" into to the "Tips & Tricks". Just as a matter of statistics, every kit sold has created one additional page in this document. But now it's very close to be complete. There are still some improvements to be made. The worst thing is the ACI, of course, where the work is still ongoing. I think the current "improved ACI" has most of the bugs and shortcomings fixed, but so far I could not get it up to the same specs as the circuit in the Apple II, so this is work-in-progress. It would be too easy to copy the circuit from the Apple-II. I did that more than a year ago, but it just looks wrong when put on the ACI card.

 

Comments invited ! 

Offline
Last seen: 1 month 12 hours ago
Joined: Jan 30 2021 - 11:55
Posts: 63
I am excited with what I read

I am excited with what I read!

How much i wish i had in my hands this pdf !  Too bad I did not supply the materials from you.  

Thank you very much !

Offline
Last seen: 1 month 12 hours ago
Joined: Jan 30 2021 - 11:55
Posts: 63
Hello! This is today screen

Hello! 

This is today screen after clear & reset.  

This lines and question marks didn't saw immediately. Part of that symbols rolling. Then stop and then rolling again.   

I didn't did any change from yesterday that all worded great.  

I don't know if is something i can do, measure with oscilloscope or change components that may have. Or is more clever to take care with keyboard solution to be able to run a diagnostic program.

 

Anyway, I really don't feel disappointed. Not at all! This problems help me learn!

 

Have a great day!  

 

Edit-1:

I discover that this faulty reset screen happens after a "cold" boot. I mean that Apple-1 was off all the night. Now that components are a bit warm with new reboot progress looks like reset screen work great. No rolling symbols. I don't know if this tell's something.  

 

Edit-2:

No, just did it. Keeps rolling. 

 

 

Offline
Last seen: 20 hours 42 min ago
Joined: Apr 1 2020 - 16:46
Posts: 842
Great example for keyboard input gremlins !

In post #11 above, Zijjel showed a symptom he can't explain. 

IMHO this is a great example for keyboard input gremlins.

When no keyboard is connected, the PIA picks up noise at the STROBE input from the keyboard connector, and random character inputs appear.

This is a known effect.

If you look closely, after the input buffer is full , Wozmon purges it and give you a new '\' prompt.

This happens after all 128 characters (or 3 x 40 character lines plus 8 characters in the 4th line), clearly visible in the photo.

All you need to do to stop this is to solder a 2k resistor from keyboard connector pin #14 (STROBE) to pin #9 (gnd).

Oh, and this proves that Wozmon is running correctly. Time to get a keyboard. You could try my keyboard emulator if you have an old DOS based computer with a printer port. Alas, the embargo against our Russian friends (I mean individuals from  the Apple-1 scene, not Putin or his gang) deprived us from the great keyboard converter from Russia which would plug directly into the Apple-1 keyboard socket and accept a standard PS2 keyboard.

 

Comments invited !

Offline
Last seen: 1 month 12 hours ago
Joined: Jan 30 2021 - 11:55
Posts: 63
Great info once again! I

Great info once again!

 

I confirm that by adding 2k resistor on my small clear & reset pcb now works great! 

Sure i will try the Dos emulator keyboard! It is great solution and helps me learning! I like very much projects like that!

 

Thank you!

 

Offline
Last seen: 20 hours 42 min ago
Joined: Apr 1 2020 - 16:46
Posts: 842
Zijjel gets a reward for willing to try my keyboard emu cable:

In post #13, Zijjel wrote:

 

"Sure i will try the Dos emulator keyboard ! "

 

Uncle Bernie says:

 

look at post #8 of this thread:

 

https://www.applefritter.com/content/apple-1-keyboard-emulator-cable-plans

 

I've posted the most current version of the DOS executable I have been using since a year or so. It's of course experimental, like everything I do in terms of Apple-1 work.  And at some point there (hopefully) will be enough users cheering me up to finally make a "final" version with all the features and DOS/LINUX compatibility.

 

There are a lot more tools in Uncle Bernie's shed. I have my own Apple-1 emulator and a whole toolchain to automate the assembly and conversion process for Apple-1 program development: assembly language source code in, press a button, and   a binary and a .APL file falls out. The .APL file is the hexmon command format which  my keyboard emulator can feed to the Apple-1 (invoke with F3). F1 and F2 are for clear screen and reset.

 

There are further tools which use the output of this first stage of the tool chain to make synthesized AIFF files for the ACI (both regular speed and turbo loader speed) and there also is a .TUR file format for the keyboard emulator cable which can load large programs in mere seconds, too.

 

I just can't publish these tools because they are in an imperfect state and there are lots of quirks and bugs left. For instance, my Apple-1 emulator uses X11 but on most modern versions of Linux it just hangs and can't output stuff to the X11 pop up window which emulates the Apple-1 screen output. On older versions of Linux it works fine. And with almost no Apple-1 users out there in the world showing interest in my tools it's simply not worth my RQLT to hunt these bugs down and get the tools in a more useful shape. This is a pity of course but this is how it is.

 

Comments invited !

 

Online
Last seen: 35 min 1 sec ago
Joined: May 4 2021 - 06:35
Posts: 153
Hi Zijjel,you can consider

Hi Zijjel,

you can consider also my open source alternative:

 

https://p-l4b.github.io/USBKEY/

 

 

With it you can:

- use a standard USB keyboard;

- tele-type on your Apple-1 from an external terminal;

- send APL files to it via a small bash file.

 

Some small programs (i.e. Mike Willegal's memory test, ASCII test) are embedded in Arduino code and can be loaded quickly.

 

Enjoy! :-)

 

Claudio - P-LAB

 

 

 

Offline
Last seen: 1 month 12 hours ago
Joined: Jan 30 2021 - 11:55
Posts: 63
Thank you both for your

Thank you both for your fantastic job you do in this community. Great work and very helpful! Congratulations!

I will go with Uncle Bernie's Dos emulator for the beginning as it is more quick for me and more retro style! But definitely I will build and your project P-Lab!

 

Thank you!

Offline
Last seen: 20 hours 42 min ago
Joined: Apr 1 2020 - 16:46
Posts: 842
Hey, Uncle Bernie's keyboard emu also is open source !

In post #15, Claudio wrote:

 

"You can consider also my open source alternative"

 

Uncle Bernie clarifies:

 

My keyboard emulator also is open source. The source code is at the top of the keyboard emulator thread. Anyone with a brain could use it, but it needs a C compiler which is able to produce an executable for the ancient DOS operating system. Not everybody wants to expend the time to find one. People told me. They want executables. So I can decided not to waste time with publishing keyboard emulator C source code anymore. But anyone who asks me by the "Send PM" system can have it. I see no reason to hide my bad programming style as I wrote C programs which were so good that even today,  33 years later, nobody was able to write something that even came close. Oh, and "science" still does not understand what I did, 33 years ago. These professors and Ph.D. students are too weak of mind to conceive the revolutionary algorithms I found back then. Which were the foundation of my self-made wealth.

 

So I don't need to "hide" my source code. It works, it is robust, it does the job, and it is cheap to write (in terms of time invested). But you all need to understand my RQLT is running out so I'm not inclined to waste time with publishing stuff (here: source code) nobody ever would read or even use. It's like shouting into the forest and expecting that somebody shouts back. Just a waste of time. And I have no time to waste.

 

And if you consider the "total cost of ownership", nothing, absolutely nothing can beat the cheapness of my keyboard emulator solution, as long as you have some old leftover DOS machine around and did not throw it into the trash. I never did that except for hopeless cases which could not be kept in running condition.

 

Compare this to all the programmers and tools etc. you need to buy / build the p-lab solution. Alone ordering the microcontroller from Digikey or Mouser will cost you more than my whole keyboard cable.

 

The only thing I wonder is why only 3 to 4 people in the world (less than 1% of Apple-1 clone owners) have built my keyboard emulator cable. It's not bad. It works, and gets the job done. I use it all the time because every other method to get software into the Apple-1 is more awkward.

 

Comments invited !

Log in or register to post comments