This is my first post here. I'm trying to build an Apple 1 replica (I bought the pcb from retroplace a year ago, and started collecting all the parts until I get everything to actually start building).
My electronic knowledge is far from a professional level but I love learning while doing.
So far, I soldered everything, checked the power supply section. The LM323K I had was a restamped transistor (I guess, since it was giving only 0.67V). So I changed it and now I have all the voltages correct everywhere needed.
I proceeded with populating the terminal section, powered it up, and got a nice boot screen. At least I thought, until I noticed one of the _ char was a +
I should mention I switched all the 2509v for mm1404an, because I wasn't able to find 2509v for sale.
When I clear the screen, I get the blinking @
I don't want to go into populating the computer section until I figure why there is a + where there should be a _
I tested removing D2, C3, D4a, D4b, D5a, D5b, D14a, D14b and C11b :
-> 40 x 24 fixed "█" symbols
put back the 2513:
-> 40 x 24 fixed "@" symbols
put back the 2519:
-> 40 x 24 flashing "@" symbols
put back all the mm1404an:
-> flashing "_@" symbols (and can be cleared) but with still that + sign in col 33 line 17
Any help to figure it out would be appreciated. Here is a picture of it:
PS: English isn't my first language. Please excuse any weird sentances and mistakes.
It seems okay, the starting grid sometimes may not have all the "right characters", this is normal.
Set jumpers: "6502" two, "NO DMA", Y-F, Z-D, R-C, W-E, X-0. And also in the KEYBOARD connector a temporary jumper between pins 15 and 16.
Then put all other microcircuits and do Reset (on KEYBOARD connectors connect pins 1 and 9) and see what happens...
Thanks for the details. I went ahead and set jumpers, and put all remaining ICs. I powered it on and... nothing happened on screen, but soon I felt a bad burning smell. So I unplugged it very quickly and touched every component. The only one that was painful hot was the LM320 MP-5
So I'm guessing a bad IC is making a short somewhere and I didn't identify it by touch. While I'm trying to identify the issue, any help to troubleshoot would be appreciated.
Thanks in advance
In post #4, Macintosh_nik wrote:
"Did you buy it on aliexpress? There's a lot of junk out there... I hope your DRAM chips are not from there? It doesn't work, it's quite possible that your processor or PIA is just a relabeled fake, I've run into that."
Uncle Bernie comments:
MUHAHAHAHAHA ! (I nearly spit my morning coffee over the keyboard when reading the above).
"Mostek" DRAMs with a date code hinting they were made in the year 1999 ! Hilarious ! These Chinese counterfeiters don't even do their "homework":
Mostek ceased to exist in 1985 as a brand (it was swallowed by STM) and you can bet they NEVER produced any DRAMs in the year 1999, let alone the long obsolete MK4027 from the mid 1970s !
Same thing with the laser engraved "Rockwell" 6502 and 6520 with date codes from the 2000's ... Hilarious !
Here is a thread about Chinese counterfeit ICs every prospective Apple-1 builder should know:
Buyer beware !
Oh - and the counterfeit MM1404 or Signetics 2504/1404 - especially those in grey mold compound packages - more often than not are based on re-stamped 555 timers, which, when put into the Apple-1 in the 2504 sockets, will only get hot, melt the socket (if they can, depends on the plastic material) and - of course - they also might kill the -5V regulator if this is yet another fake. The real 78xx and 79xx bipolar regulators self protect (they have overcurrent protection and thermal shutdown) and they are very hard to kill, although this is possible by applying enough overvoltage and other mean tricks. But there are *lots* of Asian knock-off three leg regulators which lack the protection circuits, and these are mere junk, designed by idiots who don't know what they are doing. Of course, if you simplify the regulator circuit and take the protection out then you can have a smaller die size and you can sell the part cheaper. But I think the main reason for removing the protection circuits was that these copycats never figured out how to control the process to keep the current limiting set point in the proper range. This is NOT trimmed, it's done by turning knobs in the wafer fab.
- Uncle Bernie
Yeaah... I bought these dram, with a 2000 timecode... (0038)
I removed all the computer section chips, and tested again, in case it caused damages. Everything still works fine with just the terminal section.
Thank you both for pointing me to the right direction.
P.S. : @UncleBernie I saw you have some IC kits on ebay. Do you have some DRAMs in stock to sell ?
Ok so I got new DRAMs, and went ahead with 4K and desoldered the w wire. Now it powers on with the boot screen, I press clear, I get a single @. Then reset, and there is something weird.
(Yes 2 @) and when I type in, it writes between the / and the first @.
I also noticed that if I type more than one line, some characters previously entered change. I ignored it a while and typed in the test program from the manual. The chars display but while running it changes some characters 1 or 2 lines above. I assume it's a bad shift register, so today I'll try replacing them one by one (I have 3 spares)
After looking at it back from the beggining, I spotted a few ICs being in their LS version:
- the 74S257 switched to 74LS257. Looking at both documentations, they have small differences (like propagation delay from data input being more than doubled).
- the 7427 switched to 74ls27, which has a different high/low treshold.
I started collecting the parts for this project long time ago, and obviously didn't pay enough attention. I'll get some more precise IC references (each time with a couple spares, as these are cheap enough) before resuming troubleshooting.
I don't think it's a bad shift register, they have very different symptoms, you can read more here. The 74ls257 works without problems in one of my assemblies, about the 74ls27 I have to look in my notes, I can do it tomorrow. Tell me, do you use a 27k resistor in the area between B2-B3? Try putting 20-22k, it's not hard.
I tried replacing the 27k by a 22k then 20k but no change.
Thanks for that precision.
I have an old analogic oscilloscope, that goes as low as 500nS/div so the readings are not as precise but, after a few measurements on pin 13 of the 74123, i decided to solder a 20k potentiometer, with a 5k resistor in serie. When I turn the potentiometer, it changes the duty cycle, but not the frequency.
This is what it looks like with X: 0.5uS/div and Y: 2V/div
You're getting close. I tried 74ls27 and 74ls257, on my assembly it works. Here's a little video. I still think the solution to your problem is in the area around 74123. Try swapping out the 8t97 and 74ls257, as well as the other 74123 and the 47pH capacitor. The Apple-1 is not exactly an ordinary computer, in the sense that it has never been mass produced and therefore has not undergone the necessary technical checks. Sometimes components that don't work in the Apple-1 are fine in other systems...
Ok, so far:
- I changed the 27k resistor for a 22k.
- I swapped the 74123 for another one
- I changed the 74ls257 for a set of 74s257
Still exact same issue (wasn't able to reset the processor for a moment, due to a bad 74s257, that I've been able to identify)
I'm out of spares for the other chips. They are cheap enough so I ordered a few spare 8t97 and 7427.
I'll keep this thread updated when everything is delivered.
I changed the two 8t97 and the 7427.
I also found a digital oscilloscope and fine tuned the pulse on pin 13 of 74123 to 480 nanoseconds with a 18k ohms resistor.
Still no change.
Maybe I'll start checking all the capacitors. Wondering if it could be a bad prom tho. What do you guys think ?
Where did you get that PROM from? How was it programmed? Those PROMs are hard to find and even harder to find a programmer that can handle them. This is where buying a tested parts kit from Uncle Bernie might have saved you some trouble.
The symptoms of bad PROMs are poorly understood... I tried swapping A1 and A2, as well as putting clean chips. In both cases I just didn't get /, so it's unlikely it's the PROMs. But just in case, where did you get them? I hope not on aliexpress, the probability to buy them there is 0.
Pretty hard to work blind. Can you post a good photo of your assembly here? So you can read the markings on the microcircuits? And the back side too. What do you use as a keyboard? Try to put another 47pF capacitor, I once helped a guy in a Facebook group with a similar problem, it helped him.
You can test your PROMs, but probably not now, but when your build starts interacting with you.
It should be like this, check every character
FF00: D8 58 A0 7F 8C 12 D0 A9
FF08: A7 8D 11 D0 8D 13 D0 C9
FF10: DF F0 13 C9 9B F0 03 C8
FF18: 10 0F A9 DC 20 EF FF A9
FF20: 8D 20 EF FF A0 01 88 30
FF28: F6 AD 11 D0 10 FB AD 10
FF30: D0 99 00 02 20 EF FF C9
FF38: 8D D0 D4 A0 FF A9 00 AA
FF40: 0A 85 2B C8 B9 00 02 C9
FF48: 8D F0 D4 C9 AE 90 F4 F0
FF50: F0 C9 BA F0 EB C9 D2 F0
FF58: 3B 86 28 86 29 84 2A B9
FF60: 00 02 49 B0 C9 0A 90 06
FF68: 69 88 C9 FA 90 11 0A 0A
FF70: 0A 0A A2 04 0A 26 28 26
FF78: 29 CA D0 F8 C8 D0 E0 C4
FF80: 2A F0 97 24 2B 50 10 A5
FF88: 28 81 26 E6 26 D0 B5 E6
FF90: 27 4C 44 FF 6C 24 00 30
FF98: 2B A2 02 B5 27 95 25 95
FFA0: 23 CA D0 F7 D0 14 A9 8D
FFA8: 20 EF FF A5 25 20 DC FF
FFB0: A5 24 20 DC FF A9 BA 20
FFB8: EF FF A9 A0 20 EF FF A1
FFC0: 24 20 DC FF 86 2B A5 24
FFC8: C5 28 A5 25 E5 29 B0 C1
FFD0: E6 24 D0 02 E6 25 A5 24
FFD8: 29 07 10 C8 48 4A 4A 4A
FFE0: 4A 20 E5 FF 68 29 0F 09
FFE8: B0 C9 BA 90 02 69 06 2C
FFF0: 12 D0 30 FB 8D 12 D0 60
FFF8: 00 00 00 0F 00 FF 00 00
I went back to my eBay history, to find out that I bought the proms from you Macintosh_Nik (in February)
So I'll assume it's been properly tested.
I have already ordered a new 47pF capacitor. I don't have any spares. I will test changing it. In the mean time, I will provide a high res picture of front and back as soon as possible, and a video of the official manual test program running. I'm using a modern keyboard attached with a appleII to usb keyboard and a cable I modified, but I already know it's not the issue since it does the same with just a jumper between 15 and 16 in the socket and a short to clear and reset (and a jumper between the PIA and the 6502 to avoid random chars)
Then you don't have to worry about PROMs. If you bought this in February, you are one of my last buyers.
I would also like to make sure that your keyboard adapter is not causing this phenomenon. It would be great to check how ClearScreen works (pins 12 and 16 have to be shorted in the keyboard connector) and Reset without it (pins 1 and 9). It is better to do it with some old button and a couple of wires.
Yes, I tested without keayboard, with a jumper between 15 and 16 and a wire to briefly short between pins to ClearScreen and then Reset, with the exact same output.
Here are large pictures (clickable) of my board:
Is 1ЛБ554 a legal replacement for 7410? I understand this is a Soviet microcircuit, in my records I found that К155ЛА4 = 7410...
It's the same, but yeah, it's one of those I'm still waiting for the real reference to be delivered. I must confess I've had a hard time finding all the exact IC references for months and I took some replacement chips sometimes, which I'm trying to replace now. I'm not surprised that with "almost" the right chips, it "almost" works :D
I desoldered the 47pF and measured it. It reads 21pF. So I guess I'll wait for the new ones to be delivered. You pointed at it several times, I should have verified it earlier. My bad
I've changed the 47pF capacitor and the two 1ЛБ554 for two 5410 I had spare (7410 for military use), and still the same issue.
In the mean time, I'm trying to understand more about the whole thing. I came accross this webpage.
Here is a video I made for everyone to see what happens when I type in.
I typed the program given by the manual, trying to ignore weird things on screen:
0:A9 0 AA 20 EF FF E8 8A 4C 2 O (RET)
I think you probably have power spikes on +12 -12 +5 -5 lines. Did you test components before and after installing on the bare unpopulated board? Specifically the 22 uF capacitors? I would suggest adding the 6 resistors on the bottom of the board as shown in post #4 of this link. You probably won't need the small capacitors shown all over the board.
I was 99% sure that replacing the bad capacitor would solve the problem. Try swapping the C11B for any other 1404 from the top row. Will it change the starting grid?
I did test a lot of components but I don't think I tested all of them, can't remember having tested or not the 22uF capacitors. Just found out that I made a mistake with the 5300uF capacitor and got a 5000uF instead. I don't think it really matters.
One thing I remember is I tested all the resistors and they were all at their max tolerance value (5% and sometimes more, max 10%, new old stock Allen-Bradley). I could replace them with more reliable ones if worth it (I'll take new ones, not nos Allen-Bradley).
I added the resistors from UncleBernie reliability mod, no change, added all the small capacitors, no change.
Yes, I was very excited finding that bad 47uF capacitor and replacing it, thinking "finaly, it will work"!
I swapped the C11b for the D5a, and I have the exact same starting grid.
We all use old Allen Bradley resistors, they all show 5-10℅ more from the rating, it's not that important. I used a big blue 3500uf capacitor and had no problems with that, so your 5000uf is a very unlikely version. Your video part works fine, your processor part works too, but something is interfering with it...
With the new 47ph capacitor did you change the value of the 27k resistor?
Is your timing still close to 480ns?
Could you have mixed up the small yellow 0.01uf(4pc) and 0.001uf(1pc) capacitors?
The timing is now 424ns, I'll adjust the resistor.
I double checked and the 0.001uF is at its place, next to the 74123, and the 0.01uF are all at the right places too.
So: the video part works, the processor part probably also since you can enter commands. DRAM and PROM checked, you tried other 7410, 74123, 74s257 and 8t97, timing is close to 480ns. That leaves only 74154, PIA and the processor out of the chips. I have never had a non-working 74154 or PIA. The 6502 processor may behave unpredictably if it's from eBay, but I've only met ones that didn't work at all, no Reset. I'm running out of ideas, I need to think...
Have you given the digital oscilloscope back yet? I have screenshots of some measurements from a working board.
So: the video part works, the processor part probably also since you can enter commands. DRAM and PROM checked, you tried other 7410, 74123, 74s257 and 8t97, timing is close to 480ns. That leaves only 74154, PIA and the processor out of the chips. I have never had a non-working 74154 or PIA. The 6502 processor may behave unpredictably if it's from aliexpress, but I've only met ones that didn't work at all, no Reset. I'm running out of ideas, I need to think...
Have you given the digital oscilloscope back yet? I have screenshots of some measurements from a working board.
Thank you for all the screenshots.
I tested with an other 74154 and 6502 without any change (even tried with a 6502A I had around).
Here are the measurements I managed to take with the oscilloscope ( I get to keep it !)
The first thing I notice is there is a lot of noise on D-Z and Y-F jumpers (and random spikes on Y-F). Also, on your screenshots your 74123 pin 15 shows a 2MHz frequency, mine shows a 1MHz. But maybe that's just an auto-detect fail.
Seing that much noise on some points I wondered if my cheap transformers could be the culprit. I have some triad transformers waiting to become the "original" PSU but for now I made everything using transformers that work directly on the 230V 50Hz french power line.
I saw on the board that you use a 74HC32. Now: I am not an expert at all but when I tried HC versions of the TTLs, it did not work for me. If you have a spare 7432 you could try that if you not did that already.
HC? I didn't notice that. But yeah... they won't work in a lot of applications unless the circuit is designed for CMOS. HCT would have a better chance of working, but still, even 74LS would be a better choice or possibly 74H or 74S if that doesn't work.
Also, it might be worth looking at Uncle Bernie's valuable insight in this thread about Apple-1 memory and reliability mods...
Finaly, it works! Thank you retroplace for spotting the 74HC32. I looked at the board so much I didn't see it. With a proper standard 7432 it works as expected. Thank you. And thanks to everyone that helped here, especially to Macintosh_Nik for spending so much time helping. I'm sooo happy it finally works (did I already say that?)
Glad to hear you got it working.
Welcome to the wonderful world of Apple-1!