Apple-1 replica gerber files

Can you upload a photo of the finished product once you get them?

I might be interested, as I have parts to nearly complete another board but have been saving them for a decent replication of the motherboard.

Also, as you going to get the cassette interface boards as well?  Thanks again.

Kills me that I missed this. Will you be doing another run? Or...if I use the same site, can I just upload the gerber files and order them as listed, or do I need to change anything? Also, have these gerber files ever been tested?

If you place anpther order, get them with the connector plated in gold.  doesn't matter if ot doubles the price, as 2 x cheap is still cheap enough. :-)

mi2k wrote:

It was noted by another AF member that a few difeerences from the original A1 boards exist (besides the non-gold plated edge connector):

-The original has a matte finish, this one has a sligtly brighter green.

-The silkscreen doesn't cover the plated areas without soldermask.

-The DIP and breadboard solder pads should be wider although it only really matters in the b

Not $190.. I am selling the motherboard at $145 including shipping worldwide, as economic air service started to resume. Adding the ACI board would be $175 shipped tracked. They are much more costly to make as they take more labor and I have to absorb all the raw material costs which are ordered to my requirement from sources the fab doesn't buy from. I run out of non-NTI motherboards now and the boards being sold are NTI version and the ACIs are based on SCC version. So if you look for first experience with Apple 1 and for testing, the open gerber option should be good enough and most cost effective as JCLPCB would fulfill the orders using standard material and processing; and if you are going to spend a lot of time and effort to populate the board and also would like to display the board, I believe Mimeo (available in non-NTI version) or Newton (currently only available in NTI version) is a better option. Afterall, the components are more costly than the bare boards.

Thanks to Tom Owad (or whoever originated this Gerber file), anyone can now get Apple-1 clone PCBs almost "for free".

It is a sad fact that nowadays, with the "free sh*t" ideology running rampant and corporations only willing to pay for the lowest bidder - even if it's about designing mission critical software and hardware - everybody, even the "brainers" like us, tend to try to get everything "cheap". Preferably, for free, of course. But cheap it must be !

Let me take you on a journey into the world of "cheap":

The company I used to work for has used JLCPCB often, mostly for one-off lab setups, and so far they always delivered.JLCPCB has real engineers take a look at your Gerbers and at least for us these guys caught quite a bunch of possible problems before it was too late. Whether they expend the same amount of effort on their bottom-priced super-cheap two-layer PCB runs can be doubted, as even China has to obey some rules of economic viability. And, the PCBs we built for lab evaluations went into the trash bin after the work was done and all the sought after results had been obtained.

The iron rules of commerce dictate that you might get what you paid for - unless the vendor rips you off - but under normal conditions you won't get something more valuable than you did pay, because any vendor would go out of business when doing so repeatedly. If something is too cheap, somewhere there is a lurking penalty: it won't last, or never does what it is intended for. I have one of these cheap electric hand mixers "Made in China" that never worked right: when beating bread dough the dough always climbs up the kneading hooks until there is a huge nasty mess hanging on the machine and clogging it up. So this kitchen implement is useless for the intended purpose I bought it for and all the $15 or so I paid for it is completely wasted. For $29 there are better ones that do work as intended by the long gone designers of the originals.

To quote the artist and philosoper John Ruskin (one of my favorites, as it's timeless truth):

"It's unwise to pay too much, but it's worse to pay too little.When  you pay too much, you lose a little money - that's all.When you pay too little, you sometimes lose everything, because the thing you bought was incapable of doing the thing it was bought to do.The common law of business balance prohibits paying a little and getting a lot - it can't be done. If you deal with the lowest bidder, it is well to add something for the risk you run, and if you do that, you will have enough to pay for something better.”

As for the PCBs going into Apple-1 builds, I always used Newton PCBs for their quality, authentic looks, and their robustness.

However, I do not agree with NewtonMike's argumentation in this thread that "components are more costly than the bare boards", which obviously is true, but in my opinion, is no factor needing to be considered  at all.  You have already made the decision to build an expensive and obscure toy from a past time long gone. So why wreck that plan by being too cheap ?

IMHO, the most valuable investment that goes into your Apple-1 build is your own time. And the older you get, the more precious this "quality residual lifetime" becomes. This is why I refuse to work in a job anymore.

I want to enjoy my residual quality lifetime before some random medical event like a stroke etc. takes the joy out of my life.

I don't sell my precious lifetime anymore for fraudulent fiat "money" the international banksters routinely conjure up out of thin air at the costs of a few electrons worth of electricity. The more they "print", the more worthless it becomes. Beware that with all the known "more real assets" the owners of this slave planet also defraud us as they manipulate and skim off ALL the markets (except maybe the local farmer's market, cash only, who knows). Try to sell a gold coin if you don't believe me. And don't even get me started on real estate. The property tax alone will steal it from you quicker than you think, but slice by slice.

But back to the PCBs. I just have finished my third Apple-1 build based on a Newton PCB and I am fully satisfied. No problems with these quality PCBs at all, even when I had to desolder the occasional bad component, the fragile carbon composite resistors being the worst offenders. I would not use them if it wasn't for the authentic looks, else all my composite resistors would go where they belong: into the trash can. I also had the "luck" to get a new +5V regulator delivering a measly 4.85V - lesson learned: do not buy those. Instead, get a bunch of used LM323 made in the 1980s by Linear Technology or National Semiconductor (same know-how of the same "gurus") and select for the correct voltage. Still cheaper than the disappointing new one from a reputable distributor. Which, BTW, still was in spec (+/- 5%). What a POS. How dare they ! But now with all the competition gone they can't make them anymore as they used to do back in the day. There was a constant fight over making ICs that are better than the datasheet spec and customers who did 100% incoming screening knew which manufacturer had the "best" ICs, and placed their orders accordingly. The trick was to produce the quality at low cost. Nowadays, with dwindling demand for those nearly 50 year old regulator IC designs, they just can't do this tight process control anymore. What you nowadays get (for shocking prices nevertheless) just barely meets spec.

So my semiconductor industry insider advice is: avoid "newly made" specimen of any ancient IC designs like the plague. There often was "secret sauce" involved and the recipes for that have long been forgotten since they shut down the obsolete and contaminated old wafer fabs and now run the occasional wafer lot in more modern fabs, using more modern tools and processes that attempt to mimic the original process from the 1960s and 1970s. Theoretically, these modern tools should allow impeccable control of all geometries and implants to a precision the process control engineers of the old fabs could only dream of. And even the cheapest silicon wafers today are at least 1000 x better than those used in the 1970s/1980s. Still, with the "secret sauce" lost, and the small quantities made nowadays, they just can't achieve these good results anymore. So, a more modern process to make something that came out of the 1960s or 1970s does not mean the newly made widget has better performance. More often than not, the opposite is true. Here is another example: 

A few weeks ago I finished an OSI Superboard II build based on the Klyball PCB, which most likely was made by JLCPCB. And I am NOT happy with the quality of this PCB. I only paid a bit over $40 for it, including theft (aka sales tax) and shipping. I should have paid more for a PCB built on better base material where the traces do not come off so easily. Not having any solder mask on that PCB for reasons of authentic looks may have contributed to the flimsiness of the traces. I dare a guess if they still would use the melt-on tinning process of the 1970s and 1980s all of their flimsy traces would come off at once. Alas, this long forgotten PCB process step would be essential to make fake "original" Apple-1. When I compare my genuine OSI PCB from 1980 with the new build ... oh the horror ! But I need the replica to prevent wear and damage to the precious original.

So, when it comes to PCBs, you get what you pay for. You may overpay a bit here and there for the high quality ones but compared to the value of your life time hours that go into a build it's a negligible expense in any case. Remember the disaster with those cheap Apple-1 PCBs that came out of Russia a few years ago ? They only sold one batch, and never sold anything again. The quality was that bad, or so I was told. I wonder how many of those Russian PCBs became builds that work.

The same rules apply to your choice of all the other parts. I've just put a complete, tested IC set up for auction on Ebay as an experiment into Apple-1 builder psychology. Do they take the bait and are willing to pay a bit more for a proven and tested IC set that is known to work, or will they opt for junk ICs with tarnished leads and of unknown functionality some greedy surplus dealer took out of a moldy warehouse that is so cheap to rent because it has a certain ... scent ?

Let's find out how this experiment goes ! I'll keep you current !

In closing, I sincerely hope despite of the excessive length of this comment you got some insider insights you can't get elsewhere. You are welcome !

I've been avoiding chiming into this thread so that I don't seem like I'm just putting down cheaper board sets from the Mimeo.  Mike Newton is right, the board is one of the smallest costs in the entire replica build.   Worse than that, if your "cheaper" board has a defect, you have bigger problems made worse if you spend hundreds of dollars or euros on just a set of accurate vintage NOS TI sockets.  This doesn't count the accuracy of the PCBs.  There are a lot of "accuracy" mistakes on this "open" and free set of gerber files.

The Mimeo boards that Mike Willegal used to sell and that I have sold all are "bed of nails" in-circuit tested and were produced by a PCB house that does work for the US military and prototype boards for the aerospace industry.  This extra testing step and the use of the highest quality materals eliminates the potential issues that might cause you grief when troubleshooting the board on power up.    

Just so everyone is aware, the vendor Mike and I used for the Mimeo went out of business due to Covid.  It pushed them to the point of closure and I'm out of PCBs.   I am now on the hunt for a good PCB house in the states that will not only use materials of similar quality and price, but will do the in-circuit testing of the boards before delivery.   I have tried asian PCB houses for other smaller and cheaper projects and was not impressed with the quality and wouldn't risk a replica PCB built using vintage components with one.  I think Mike Newton is lucky that he works in the PCB industry so the Newton boards are much better quality than typical PCBs manufactured to the lowest cost boards you generally see coming out of Asia.  

 I should also mention with all the replica boards, I can spot one as not original in about 2 seconds without even flipping the board over to look for any branding in the corner.    I will not share how publicly since I do this for muesums and auction houses who rely on my knowledge to prevent fraud.   I say this not to promote my services or belittle anyone elses work, but to say in all this, please do not delude yourself that you can spend a ton of money and make a perfect replica.  So if making your own PCB and using the cheapest parts makes you happy, great.  If searching for the most accurate vintage components you can, then great.  I would just suggest not using your own PCB to do the second option as the risk is very high, forgetting the fact that the Mimeo and Newton are more accurate but not perfect PCB sets.

Cheers,

Corey

I believe that one way to spot a real board is the rippled effect under the soldermask on the back of the PCB. From my observation, the original boards were HASL  plated BEFORE soldermask was applied (Hot Air Solder Leveling - basically the bare copper is dipped in molten solder post etching, then as the board is removed the excess solder is blown off by an airblade leaving just a thin layer of solder over the tracks)  - when the board went through the wave soldering machine, the HASL layer on the underside of the board was heated up to the point where it melted - giving the tracks a rippled look.

They don't do it that way anymore, so it would be  difficult to replicate that look  - you'd have to ask them to HASL the board before soldermask is applied - highly automated production houses like JLCPCB probably wouldn't be able to do tis easily. There would be bit of experimenting required to get it right.

Slightly off topic, but I'm chasing a few bits for my Apple 1 build - if anyone has parts to trade, please let me know. 

Mike Newton tried this process in one run of Newton boards.  It did have some issues.  I'll leave it to him if he wants to chime in on details...  btw.  Some original boards have this "ripple" effect very bad and some don't where you need to see the board in person to see the ripple.

mi2k wrote:

"I have built one of these to try out substitute parts to the now scarce Shift Registers and PROMs,  did not hesitate to cut traces and add jumpers to do so ( wouldn't have done this to a Newton or Mimeo board ). "

Just a week or so ago I did cut traces and added jumpers using a real Newton board ... but I can afford to do this. The engine of my car exploded two days ago at a bit over only 86000 miles. These are the worse costs ...

BTW, with my criticsm above in this thread I did not intend to put down the cheap JLCPCB boards... they have a purpose ... as you point out in your last post ... but as far as I am concerned I would not use one to put in quality sockets and my precious "residual quality lifetime" in for a build ... but perhaps, for my current experiment, I probably would have preferred to cut traces on one of these "cheap" boards. The Newton boards I have are just too valuable for this type of throwaway experiments. But they have the quality you pay for. Worth every penny. Only the nasty shipping costs hurt. Why don't you organize a group purchase at Newton Mike to save on shipping ?

Other than that said, I think your concerns about the Shift Registers and the PROMs are far too pessimistic:

When I was shopping for the elusive Signetics 2504 I found out that the drop-in replacement National Semiconductor MM1404 is still available at IC brokers by the thousands. No problem if you can meet their minimum purchase order...

These 1K x 1 shift registers were sort of an industry standard back in the day. Actually, together with early 1970s calculator ICs, the very first mass produced PMOS devices which really made money. Lots of companies made them.  Synertek had the SY1404 and SY1404A. Beware of other 1K x 1 dynamic shift registers like the much cheaper AM2808 --- those are not double pumped and would need twice the clock speed to work in the Apple-1.

The real issue is the Signetics 2519. There was never a second source for them. At least I did not find one. There is a guy on Ebay selling them for $29 if I recall correctly. The might be the guy who cleaned out the IC broker I bought the mine from. I could get further 50 pcs for $21 ea but I don't need any at the moment. Do you need 50 ? No ? Buy one from the $29 guy.

The 2519 could be replaced by 6 x CD4557B variable length shift registers. Using them you could get rid of the DS0025, too. I only got DS0026 but then you REALLY need to add some bypass capacitors. Preferably hidden in the sockets as seen elsewhere in this forum. I did ask a local company how much it would cost to mount six 4557 die into a cerdip: several hundred dollar each. Not attractive for us. Or should I say "not cheap enough" ? Actually, their price is OK given the hassle, the substrate they must make, and the manual die bond and manual ultrasonic bonding. They use these methods mainly to make substitute drop-ins for the military with, like us, often uses obsolete and long forgotten ICs. Heck, if an Apple-1 original worth $500000 needs a 2519 I can make some hundreds with a really custom die for a fraction of that sum.

As for the PROMs, I could provide all you will ever need, but they are AM27S21APC with date code 026EY5L, whatever that means, and not the original MMI. And I have email correspondence with Steve Wozniak which implies he is OK with me burning his software into them for other hobbyists. I don't have his written permission on paper, though.

BTW, MMI had terrible yield problems in their wafer fab around the mid 1970s and their PROMs/PALs ain't as good and reliable as those from National Semiconductor and AMD.  It eludes me how AMD could make such excellent fuse links but for NSC the reason is obvious: it's the same recipe as used for the thin film resistors of precision analog circuits, for which NSC and LTC are still famous. Their arch-enemy Analog Devices apparently never found the "secret sauce"  to make the best thin film resistors although they tried for 50 years and TSMC offers them only since a few years and the jury is still out whether theirs will rot away (or change values) as with all the other fools who tried to make those. They are in the group of the most nasty components you can try to make in a wafer fab. Making BJTs is a piece of cake compared to quality thin film resistors.

So I would suggest we don't need any PCB modifications yet. We could do some to improve the power grid and the bypass capacitors but according to my investigations this is not necessary if enough added bypass capacitors are hidden in IC sockets. I am more concerned about the reflections / ringing on the multiplexed address bus of the DRAMs. Alas, not much can be improved here in terms of total trace length which in the original layout is longer than necessary, and just a bit above the critical length for Schottky drivers. If we use faster 6502, such as 6502A or 6502B, and modify the DRAM timing a little bit, we could use less critical 74LS257. I am still figuring out the details. But so far nobody has responded to my post on this forum featuring a test program that tries to find out whether the ringing observed with the 74S257 drivers really causes DRAM failures.

Comments invited.

Uncle Bernie

Corey986 wrote:

Mike Newton tried this process in one run of Newton boards.  It did have some issues.  I'll leave it to him if he wants to chime in on details...  btw.  Some original boards have this "ripple" effect very bad and some don't where you need to see the board in person to see the ripple.

All the Newton NTI and non-NTI boards have HASL before soldermask like the originals and yes this becomes non-standard now and yield is low. The material is special-ordered 1.8mm thick old style white laminate and old-style soldermask is also ordered and processed to my requirement. So it is really difficult to locate a fab who are willing to take my order which is basically tailor-made. And, no, I have never ordered from JLCPCB, as they don't even entertain any of my requirements for that low price range. So, I don't think the JLCPCB boards can be compared to Mimeo or Newton boards.

It all boils down to whether you just need a functional replica or cosmetically accurate replica. I believe most of the interests can be fulfilled by such functional replicas as the DIY-from-open gerber (shared from a Russian hobbyist I believe, at least not Mimeo or Newton gerber) or Replica-1 like boards,  but still there are some interests in cosmetically accurate replicas. If I haven't made Newton boards, I would still go for the Mimeo boards only.   

As there are multiple sources of low cost bare boards made from the open gerber, I have no plan to order new batch and will sell the remaining boards in bare form or assembled while stock lasts. 

Corey986 wrote:

Mike Newton tried this process in one run of Newton boards.  It did have some issues.  I'll leave it to him if he wants to chime in on details...  btw.  Some original boards have this "ripple" effect very bad and some don't where you need to see the board in person to see the ripple.

Oh yeah, the "ripple" effect, at some point, as I would like to recreate the ripple effect which can't be made by hand-soldering, I tried to outsource the assembly job and asked the assemble house to set up their wave soldering machine using some Newton board. In a few runs, the contact temperature was so high and contact time was so long that, some soldermask really peeled off from the back, especially at the LM323K section. I am pretty sure the ripple effect as seen in the original Apple 1s can be replicated if the wave soldering setting is just right, but eventually I gave up the idea as doing the custom declaration for import to / export from China can be a nightmare and I don't think there is a demand on assembled boards processed like that.    

Unfortunately I do not have an original Apple-1 so I need to use a substitute to make my points, but I have a board made in Silicon Valley in 1979 using wave soldering and showing the ripple effect:

Ripple_Bot.JPG

You can see a pronounced "ripple" effect and if you look closly it seems it starts at solder pads and then, due to capillary effect, the solder gets sucked in under the solder mask. I can't explain why it can travel so far from a solder pad and it must "lift off" the solder mask to do so without breaking it up.  Note the bottom row of vias near the edge connector. This area was protected by a glued-on heat shield that after wave soldering was stripped off. The vias down there were covered, too, and so they are not solder filled. You can see that NO "ripple effect" is seen on the trace runs beginning from these non-filled vias. All the solder that was sucked in and making the "ripple effect" emanated from the higher up via that are solder filled. It did not get sucked in far enough to reach the non-filled vias.

Here is the top side:

Ripple_Top.JPG

You can see there is almost no "ripple" effect seen here, other than on a few places near pads with component legs going in. The "nicest" one got a red arrow.

Still, although there is no pronounced "ripple" effect seen on the top side, the traces still look a bit round-ish below the solder mask.

Conclusions:

The 1970s PCB process used a thicker layer of HASL than today's processes can do and IIRC there was an additional step to reflow this thicker layer so as to get the round-ish shape of solder on the copper trace.

When they then silk screen on the solder mask, this round-ish shape of solder below the solder mask forms a capillary that will contain liquid solder during the wave soldering process and since it is a capillary, it can suck in more solder from any via or pad hole. This creates the "ripple" effect. With a thinner solder layer there is less capillary effect and less suction.

To recreate this effect today we would need the thicker HASL layer and the reflow process to make it round-ish.

Then we need a solder screen material that does not crack and peel off when the capillary effect sucks in even more solder. It must be elastic and rubber-ish when hot and then get back to solid. This material may not be available anymore today.

All of this process of course needs tight control of preheat, wave temperature and height, postheat for cool down and then proper speed, too. Unless a factory does this all the time for mass production and constantly monitors and adjusts the process parameters listed, there will be too many defects and boards need to be re-worked manually or, in bad cases, scrapped.

There is "secret sauce" involved in all these manufacturing processes and all the recipes for that "secret sauce" got lost when these factories and workshops were closed down.

Today we use SMD technology involving silk screened lead-free solder paste and reflow processes. Due to lack of lead, the outcome is crap that does not last for decades, as tin whiskers will grow and make shorts, and the brittleness of the solder joints cause them to crack under thermal cycles.

I consider such products fraudulent as their quick decay is built-in. The aircraft / avionics industry and the military still are allowed to use solder alloys containing lead and for special applications even containing cadmium. Nasty stuff, but since you don't eat electronic boards, not dangerous for you. 

I refuse to buy "modern" consumer electronics using lead-free processes and keep a private stash of good 1990s tin/lead solder at hand. I wash my hands with Delead soap after exiting the lab and so far I have no lead poisoning. There is even a drug you can use to remove heavy metal from your body. The wood processing industry uses it by the truckload, and it is cheap. As a pharmaceutical pill, not such much, last time I paid ~$50 for three pills, which is a complete treatment, for once a year. I do that every year as I love to eat tuna.

You can actually see this ripple effect on older boards when soldering by hand also.   When working on old processor tech boards (circa 1977/78)  if you give it too much heat or let the heat travel on some of the larger traces, you can see the ripple happen in real time.   It's not isolated to wave soldered boards.   It is just more common with them back in the day.

I wanted to share my experiences with the "open source" Gerber file pcbs. Over the last six months or so, I have been producing 150 board with JLCPCB while selling around 75 to this day. So far, I had not a single complain about the quality of the PCBs apart from one customer that received damaged boards due to the packaging being destroyed.

All boards are electronically tested by JLCPCB (you have to pay a significant amount more to have them do this) and while I would agree that there are definitely manufacturers out there with better quality, I have to come to the conclusion that JLCPCB does a really good job at a very competitive price.

I have yet to build an Apple-1 replica myself from those boards, but I started collecting the parts to do so. Let's see how that goes …

Best

Armin

how many buyers of your 70 boards have even started soldering?I think the number will be less than 5% ... if at all so many.  Because the hobbyists will certainly have their own PCB made in China for little money.But I agree with you, the Russian board Gerber PCB is generally technically flawless.It's a shame that JLCPCB only offers ENIG (in gold, ... poor quality) or in normal soldering silver.I already have three different qualities here and have now ordered the most expensive variant with extra expansion gold contacts.But for the picture frame alone, or as a gift, the variant (free Gerber Russian variant) ( same as well as from Armin) will always be sufficient.And if you start Armin yourself now, you should also buy the components from Bernie,  otherwise it won't work ;)

But you will surely have a friend solder it anyway.  :) 

AppleMicha wrote:

It's a shame that JLCPCB only offers ENIG (in gold, ... poor quality)

ENIG (electroless nickel, immersion gold) is a solderability treatment. Its only purpose is to prevent oxidation of exposed copper while the bare board is in storage. It then melts away when components are soldered on. It was never intended to provide a hard-wearing surface for edge fingers.

The best treatment for boards that will be brush electroplated with nickel and gold is probably OSP (organic solderability preservative), since it can be cleanly removed using common degreasers. This re-exposes bare copper that can then be activated and plated using normal methods. Briefly brush electroplating the fingers with 15 µ-inch of nickel and hard gold will provide a robust mating surface that will not tarnish, fret, or scrape through over thousands of mating cycles.

OSP is less protective than ENIG on component pads to be soldered, so it is not usually specified.

While I agree that only a handful of people that bought the board actually will complete it, I know already of one person who successfully did. So that gives me a good feeling about selling them.

And yes: You can bet I would buy the parts from Bernie. And I definitely would solder it myself.

Hi NewtonMike

Is your board still available? Where to buy?

Thomas

I bought a board from Armin and have just finished the soldering. Together with a Chip-Set from Bernie the system works well, nearly from the first power-on (there was a small problem with a IC socket).  In my eyes the pcb is in a goot quality, yes, it's too glossy but for me it's ok. 

Thomas 

Hello Thomas,..have you bought a board within a day and soldered it within a day?

(because of the time between the request from the NewtonMike Board and the order from Armin as you write here).

What Armin sells is nothing else that anyone can do for about less 20 euros. (the Gerber file is also free).

Yes, if you have manufactured 5 boards at JLCPCB (cost 59$+25$ Shippng to germany),

you have to find 4 buyers, which was not a problem so far, and in the end you had your own board for free.My first Apple1 is also built with the board (which a nice Russian drew, not Armin).Of course, it works with the components from Bernie.

I ordered the (expensive) board from NewtonMike anyway, because it is just a "tad" nicer.But everyone has to think for themselves ... and I will build a second Apple1,

and this time with even more original components, such as resistors, capacitors and diodes.

Greetings Mikel

Our cars are in the same garage. I also soldered the first replica just to work on the Russian board, now I'm thinking about the second one is more authentic. Assembled all the components except for the board. I never thought I would need two replicas, but it's true. Appetite comes with food.

yes ,.. :)  thats rigth :)i "need" gold-Finger´s :D  

Thank you "Newton" !!

DSCN9455.JPG

Yeah, I really wanted a Newton board, but you got the last one I believe...

Best,

Logan