So, I've been collecting and fabricating part for a few months now. I Just want to get some feedback on how this looks so far and for any advice on what I should be checking before I put in the expensive parts. Instead of the transformers I’m using 2 buck converters. The red one puts out about 8 Volts and the green one generates +15 and -15. This seems to be enough to get through the rectifier bridge voltage drop and still have enough left to let the regulators do their thing. It also lets me run in from a single 12 volt supply. So far, I’m just testing voltages and signals with less costly 7400 logic installed. I’m not going to plug in the $200.00 + shift registers until I’m sure they won’t got pop. So far, I think I have good signs of life as it’s generating the NTSC signal and I’m seeing a screen full of @ signs. I’m assuming that’s what it should do when there are no 2504’s or 2519 memories. Anyway, any advice before I do the final smoke test with the crazy expensive parts would be appreciated.
The 2504 needs +5v and -5v to work, the 2519 needs +5v and -12v. Make sure you have the correct voltage on the board, as well as +12v for RAM operation and you can install all the chips at all.
I don't understand this urge to pull the cat by the balls.
I agree with what 'Macintosh_nik' wrote in his post #2.
Check the presence and correctness of all the various weird supply voltages on these 'rare' Signetics IC sockets (2504,2513, 2519) and also check the supply voltages on the DS0025 socket to be correct.
If you have an oscilloscope, you may check the PHI3, PHI4 clock phases to swing from +5V to -12V and that the -12V phases do not overlap. Any overlap could kill the 2504's. The concern here is possibly bad DS0025 and the abundance of counterfeit DS0025.
There also have been counterfeit 2504 in the grey Signetics package and the Chinese counterfeiters have re-stamped early 1970's Signetics 555 timers because they had the right looks of the package. You may touch the "2504" with your fingertips while turning on the power and if they get hot, immediately turn off the power because a 555 timer in a 2504 place will go up in smoke and melt down the socket, although this takes a few seconds. Be quick enough to avoid blisters on your fingers ... don't wait for the pain reflex, pull the fingers off the ICs earlier. You do have ~1-2 seconds for that.
Other than that, let it run for a while because there are bad lots of 2519N out there (1977 and 1978 date codes) which die within the first few days of operation.
All these perils were not present in my famous kits because I took care of them. One is left, see here:
https://www.applefritter.com/content/few-my-famous-apple-1-ic-kits-now-available
Good luck and tell us how it goes !
- Uncle Bernie
... looks really great, especially with the authentic IC sockets and the (supposed) "Goldie" LM323K. :-)
Well, not happy yet. with all the shift registers in all i get is a screen full of - signs. Pulling out a random 2504 does change the character on the screen so that seems to indicate the 2519 is ok as its getting and seriallizing the data(I have 2 and both gave the same behaviour). I'm fairly sure the 2504s are legit. I got them from octopnv on ebay. same guy that makes and sells the 2504 testers. So far I'm 99% sure its something to do with the clocks. when I assert the clr signal the screen goes black but the - signs come back as soon as I let it go. seems like the chips just aren't getting written to. Anyway going to put it on the scope tonight to see that the DS0025 is outputing. BTw yea, nothing gets hot and no magic smoke. Its probably just a bad contact on some of the jelly bean logic.
Awesome project! I love the repro BEL caps and the 6820 PIA you've got in there. Your PCB looks like it's made from the designs I made in 2022! Neato!
Best,
Logan
I have dealt with this seller from Lithuania, only the best recommendations. If you were buying 2504 at Unicorn, most likely this Ron or Rob would have tossed you 1 non-working piece. They always do that. But octopnv is a different matter. Everything here is always exactly like the drugstore.
Can we see a pic of what's on the screen now? That would make it a lot easier.
Well I finally figured it out, the DS0025 was bad, at least the driver between pins 2 and 7 is. It was confusing me as I had 2 of them and they are damaged in the same way. As I didn't want to wait for a replacement to finish testing I did this to use the good driver on each chip (for now).
ds0025rig.jpg
And it works! on to th CPU section.
goodscreen.jpg
You post #8 tells me that you did NOT check the PHI3, PHI4 phases before plugging in the rare Signetics ICs, as I did recommend in my post #3.
I'm not sure if the DS0025 in your photo of post #8 has laser lettering. If so, it's a Chinese fake. Laser lettering on ICs was invented several decades after National Semiconductor stopped making the DS0025. Same story with the 6502 ... NONE of the genuine NMOS 6502 made by MOS TECHNOLOGY, SYNERTEK or ROCKWELL ever had laser marking. All these with laser marking are fake ! (Relabled CMOS versions or a mix of oddball CMOS/NMOS salvaged from various electronic trash).
There are LOTS of 8-pin MOS gate drivers which were made in the past 50 years, on various process technologies. There are CMOS ones which can drive several Amperes (briefly) into the gates of large MOSFETs, and these are ubiquitious in applications like automotive etc. and still being made, except today they are in SMD packages.
These drivers typically follow the pinout of the DS0025 (legacy inheritance ... they tried to get into the DS0025 sockets, and so opted for the same pinout, and then the follow up product had to have the same pinout again, until now).
Some of these more modern ICs have complementary drivers inside (one inverting, one non-inverting). Which could explain your problem with one channel not working.
Others have no pull down resistors on the inputs (the DS0025 has them) but may work by happenstance / randomly picked up charges.
Most modern CMOS versions are NOT specified for the 17V supply differential they "see" in the Apple-1. It's above their abs,max voltage.
I'd recommend you to take a ohmmeter to your "DS0025" outside of the socket. Each input (pins #2, #4) should have a pull down resistor of about 250 Ohms to V- (pin #3). A huge tolerance of +/-50% is OK. But the resistor must be there. If it isn't, or if it is much greater than 500 Ohms, the "DS0025" is counterfeit.
You can also take a Q-tip drenched with Acetone (nail polish remover) to the lettering and gently rub. If black paint comes off, it's a counterfeit.
Tell us what you find and if it's counterfeit, warn us from the seller ! (I'd bet seller is in Shenzen, China ;-)
- Uncle Bernie
P.S.: I got burned by counterfeit DS0025 myself, but managed to get my money back. But not unless invoking clawback by the credit card company. This happened several times, with different sellers. One of them where I bought only 10 to probe the waters refunded me the money and told me to keep the ICs (because he did not want to pay the return postage as Ebay demands). I have worked out a circuit mod for the Apple-1 to make most CMOS "DS0025" fakes run fine in the Apple-1, BUT since the counterfeiters never tell us what was the basis for the fake, we can't tell if they are specified for the 17 Volts supply voltage seen in the Apple-1. Some known CMOS types are specified for that. If anyone is interested in the information, tell me. I'll post something in an extra thread. This could at least take the pressure off the builders who are desperate for DS0025 but only get fakes. They could buy a known, genuine CMOS substitute for a cheaper price and apply the mod.
Actually I did. Ph3 was dead, ph4 was fine . Same with the inputs to th ds0025. 3 dead, 4 fine. I traced ph3 back and couldn't find any errors so I pulled the ds0025 and tested the inputs again. Both were perfect. Pin 2 on both chips was shorted to gnd. Thats why I redirected ph3 to pin 4 of the second ds0025. This is what they look like now. Good 15 volt swing and no overlap. And no they are not laser etched and they pass the acetone test
IMG_0593.jpeg
Ok, I have the CPU section running but when I reset the cpu, it just starts printing the character set, over and over. Did I buy the wrong proms?
Based on what 'mc6809' wrote in post #10, we can't rule out counterfeit DS0025 completely, but it is more likely, based on the information now at hand, that this is a systematic failure in these two examples (and possibly, more ICs from this lot are affected).
There never should be a short to ground on the input pins of these ICs (although the pulldown resistor of nominal 250 Ohms must be there).
Nondestructive decapsulation in an industrial QA lab could tell the story ... it involves squirting hot concentrated sulfuric acid though a fine nozzle on the IC to expose the bond wires and the die. (Don't try that a home, it's a machine which does this dangerous step, and the fumes get sucked away and get neutralized).
Except for the counterfeit ones I was quite lucky with the functionality of the (real) DS0025 made by National Semiconductor, less than 1% were bad / unfit to work in the Apple-1.
(I still have some left over - send me a PM if you want one).
- Uncle Bernie
In post #11. 'mc6809' wrote:
" Ok, I have the CPU section running but when I reset the cpu, it just starts printing the character set, over and over. Did I buy the wrong proms ? "
Uncle Bernie comments:
this could be the 'diagnostics' page I put into my 512x4 PROMs which came with my kits. Not sure, though, because there are other people in the world who have PROM programmers. Pull one of the RAMs (power off, of course) and see if you get error messages in hexadecimal. If so, it's my diagnostics page. But where is the Wozmon ? It also should be there. On 512x4 PROMs, the selection is done by lifting pin #14 on both PROMs to enable the diagnostics page (floating TTL pins are "H"). If the pins contact the sockets, they are "L" when the Apple-1 accesses the PROMs, and then it's the Wozmon.
When I stopped making kits (ICs ran out and I'm not willing to pay usurious prices for the 'rare' ICs), I dismantled my burn-in rigs and gave away a few sets of the 'diagnostics' only PROM sets which were in the rigs as a freebee with my last kits. These were meant to be used and then sold - as 'diagnostics' of course - by the builder. Maybe something got mixed up ... show me a photo of the PROMs, please.
- Uncle Bernie
Yes, thats exactly whats happening, if I pull one of the lower ram chips I get B5@E000 repeating. I figured its was a diagnostic mode as E000 is where the basic ram is, I just couldn't figure our how to get to WOZMON. Thanks for the info, I'll try lifting pin 14 on the proms when I get home tonight. seems the person I purchased these from is using your version.
In post #14, 'mc6809' wrote:
" Yes, thats exactly whats happening, if I pull one of the lower ram chips I get B5@E000 repeating. I figured its was a diagnostic mode as E000 is where the basic ram is, I just couldn't figure our how to get to WOZMON. Thanks for the info, I'll try lifting pin 14 on the proms when I get home tonight. seems the person I purchased these from is using your version. "
Uncle Bernie comments:
This indeed looks like the output of my 'diagnostics' page. But you should have pulled a DRAM from ROW B, not from ROW A. If you pull a DRAM from ROW A, a bit is totally missing from the pointer to the memory location being tested, and depending on the bit location, this may then lead to a bogus address reported in the message ... the 6502 does not work very well with a missing bit in the zero page (although I tried to mitigate this issue, but 256 bytes of code can take you only so far...)
If you lift pin #14 then you would get the diagnostics page, so don't do this with the PROMs you have. Send me a photo of the PROMs in which I can read the lettering on them. This should help to diagnose what is going on.
- Uncle Bernie
https://www.ebay.com/itm/315030808178
These are the ones I purchased. Either they are programmed wrong or i mis-understood the listing. they say 6301 so yea, 256x4. No second page with these.
I've looked at the Ebay listing you pointed me to, and these are indeed PROMs programmed with my diagnostics page BUT they are WRONGLY advertised as "Wozmon". (Ouch).
The fun part is, this guy seems to have the MOST PECULIAR Apple-1 build I have ever seen, ALL the ICs are in Textool zif sockets so I suppose he wants to produce Apple-1 IC kits ! At least somebody who steps into my footsteps and makes 100% tested Apple-1 IC sets !
I have no problem with that, but when he screws up the contents of the PROMs he will make a lot of unhappy customers. I doubt that his PROMs are genuine MMI. I do have a few genuine MMI PROMs (256x4) and the logo is different. Try the Q-Tip / Acetone rub on these PROMs to find out the truth. My money is on ex-Eastern Bloc PROMs made by TESLA in Czechoslovakia. Here is the datasheet first page from my TESLA databook:
PDRM4689.JPG
There also are some USSR made PROMs of this size (DIL16, 256 x 4 bit). I don't know if CEMI in Poland ever made PROMs. Generally, the quality of the USSR ICs was much better than other Eastern Bloc semiconductors. This is because they went into Soviet weapons systems. All the 'consumer grade' ICs made in the Eastern Bloc back in the day were of dubious quality at best. The worst I've encountered were "hobbyist grade" ICs made in the GDR, "U880" and the like, they blatantly sold test rejects to their hobbyists (I'm glad I did not live there). I'm still struggling with repairing the GDR made Chess Computers in my collection. Fidel Castro has the same type, except it was put into a nice chess table, and man, if I could get it and move out of Cuba ... unless it was eaten by Termites. Only two of these chess tables with integrated computer exist in the world. But back to your PROM problem.
Please send me a PM stating where you live. Maybe I can help you with the right PROMs and with a DS0025 that works.
- Uncle Bernie