New Supplier of Apple one motherboards ?

... which appears to be nicer than the other "cheaper" PCBs, because it seems to have a gold-flushed edge connector, which is a desirable feature none of the cheaper PCBs based on the Open-Source Gerbers and usually made by JLCPCB was ever seen with.

Mimeo is all but extinct (the PCB company who made them went out of business, or so I was told), and so the only established, high-quality alternative are the Newton PCBs, which are a bit more expensive than this one, and soon to run out, unless Mike Newton makes a new batch, which is a tough decision facing those cheaper "Replica-1" PCBs based on the Open-Source Gerbers.

I would guess an Apple-1 builder could go with any of the various PCBs offered, but watch out for price gougers who sell the rock-bottom "Replica-1" with no gold on the edge connector for around $100, which IMHO is a rip-off.

I've build three Apple-1 based on Newton (NTI and non-NTI) so far, and three more with the "Replica-1" PCBs offered by "pinguso", see my post #4 in this thread:

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

"Pinguso" is not me, but he offered a PCB set (main board and ACI) based on the Open-Source Gerbers and made by JLCPCB on Ebay for a really great, fair, price. Keep in mind that no Apple-1 build is complete without an ACI. You don't need a cassette recorder, though, which are hard to find in functional condition, but rigging MP3 playback / recoding cables to the ACI is easy, and then you can load all the MP3 files available with Apple-1 software into your Apple-1.

Which begs the question if you will ever need / use the 44-pin edge connector (the ACI daughter card goes into the 44-pin slot, not on the edge connector). If you are not going to use the edge connector,  you don't need to pay for the gold. If you intend to use it, it does work without the gold, too, at least for a while, and then you need to clean it with a pencil eraser now and then to keep it in functional condition. HASL solder layers don't make good, durable edge connectors. Gold over Nickel is a must for reliability. The question is, how good is the gold plating on the PCB you have found. If it is too thin, it is probably not worth the premium price.

Volunteers to the front ! --- Meaning, buy one, and find out, and post your findings here ! 

I just want to make one comment....

  The mimeo isn't extinct, just waiting for covid restrictions to lift enough that I can walk into a PCB house here in the states and explain all the little things on the board that are non-standard to their PCB process.  It's much easier that way than going back and forth on the phone and email.   When a board run is a few thousand dollars, I don't want to have any mistakes or misunderstandings.  There are a few PCB houses in the north east USA that I could drive to when I am allowed.  So my new plan is to have a stock this summer of Mimeo and ACI boards. 

Hopefully you can be patient.  

Cheers,

Corey

Great news Corey! I have been planning to get a Mimeo board for a long time to supplement my Newton NTI. I even have a few duplicate components to start out with. :-)

This sounds good !

I was trying to get my hands on a real, thick, high-quality MIMEO PCB since I had been lured into the Apple-1 scene, and certainly would buy (1) from the new lot, and if had all the desirable features, I'd probably take a few more.

Please send me a message via the Applefritter message system once you have the new boards.

- Bernie

I just got word from another Apple-1 builder that the "golden" look of the above PCB the post #1 is about might be based on an optical illusion, because the PCB traces will reflect any ambient light, and if that ambient light is yellow-ish, it appears to be gold flushed, but it isn't.

If this is the case, and I can't know for sure, the PCB in question might  be just another run-off-the-mill PCB made by JLCPCB, based on the open-source Gerbers. If this is the truth, then the price asked for this PCB is way too high, and the deal offered by "pinguso" is much better, as it includes the motherboard and the ACI card, IIRC for about half the price of that offer seen in #1. You just need to be patient because as far as I see on my radar, "pinguso" only orders small batches of boards, which take weeks to arrive, and they sell out quickly.

Much the same supply side issue as with the 100% tested and burned-in IC sets I offer on Ebay. I can only make two of them per month (two Apple-1 clones running 24/7 for 4 weeks to complete the burn-in), and people just need to be patient and please do not engage in bidding wars, unless you need them soooo baaaaadly. There are more of my IC sets to come, and beginning with mid January 2021 I can offer one of these sets per week, as I am gearing up to have four Apple-1 clones running the burn-in.

We Apple-1 enthusiasts have to work together to thwart the price gougers. I'm always trying to find cheaper parts sources, and I have deep enough pockets so I can buy plenty, if I find some, to keep the "Spice flowing" (from "Dune") but the real challenge is to find the 2504 / 2513 / 2519 (or equivalents) at a good price. They are slowly running out. I was told just a few days ago that there a no 2519 in stock at any IC broker in North America anymore. And these Chinese forgers will gladly stamp "2519N" on any ordinary 16-pin TTL IC made by Signetics back in the day, which used the same grey mold compound.

So even here, Buyer beware ! Some people got burned (literally) by forged MM1404 or 2504 coming out of China. These are re-labeled 555 timers or 741 opamps and will just get hot when put into an Apple-1.

This situation also is the reason why anyone engaging in providing PCBs and such must play it safe and not order too many at a time. Once the critical ICs are gone for good, and have turned into unobtanium, all the other parts stock becomes worthless.

Plan and hedge accordingly.

UncleBernie wrote:

but watch out for price gougers who sell the rock-bottom "Replica-1" with no gold on the edge connector for around $100, which IMHO is a rip-off.

I agree. There's noting wrong with cheap replica PCBs if the limitations are stated and if they're sold at a cheap price, but $100 ain't it.

However, when ordering PCBs from inexpensive Chinese fabs, even if you do select hard-gold-over-nickel edge fingers, the default is that you don't get nearly as much gold as edge connectors need. For example, since I want to do some hardware hacking, instead of doing it with the Mimeo 1, I just had ten new boards fabbed with the free gerbers, and no gold. From PCBway, they cost me $18.30 each, plus $65 for express shipping, so $24.80 each in total. I had them five days after placing the order. I didn't try to make any special arrangements for the silkscreen over non-soldermask, so they came out about the same as the JLCPCB boards.  If I selected gold edge fingers, they would have been $47.80 each, not including the shipping. However, that's only for 10 microinches of gold over 120 microinches of nickel. That little gold will wear off after 25-50 mating cycles. The gold is effectively just cosmetic.

As an example of proper edge connectors, the PCIe specification requires a minimum of 0.7 microns of gold (28 microinches) over a minimum of1.2 microns of nickel (47 microinches). That should be good for over 500 mating cycles. The nearest match PCBway offers is 30 microinches of gold over 120 microinches of nickel. If I select that, the per-board cost is $65.80. If I had those made for resale, using more accurate gerber files and with the proper silkscreen (even over the non-soldermask areas), I'd feel justified in selling them for somewhere between $100 and $200 each.

I've drawn an Apple 1 schematic in Eagle, and started a PCB layout, but I'm not sure whether I'll finish the layout. I mostly wanted a schematic that is organized more logically. Cramming everything exept the linear power supply into two D-size sheets was impressive, but in my opinion it makes the circuitry hard to understand. I drew it on twelve B-size sheets. I used Eagle 7.7 because I've owned a full license since back in the Eagle 4.11 days, but I'm not willing to rent software so I can't use a newer version of Eagle. In principle I'd like to use Kicad since it's open-source, but every time I try to use it I get frustrated and go back to Eagle.

I also tried importing the available gerber files into Eagle, but the import was a disaster. There were thousands of errors, and the resulting PCB file just looked like random vectors.

Due to the above controversy about different Apple-1 PCBs from different vendors, I could not resist to snap a few photos in a weak attempt to point out differences and features to watch out for. With "weak attempt" I mean that I know all too well to which length a professional photographer would go to get the pictures right, that is, looking like the real thing, but I have neither professional photo studio with some $10000 in lighting equipment, not a professional grade camera, only a very old one, Toshiba PDR-M4 I bought back in 1998, when it was on final sale for "cheap". All I could do to get the best results with these handicaps is to take the photos outside, in natural sunlight, and still, most photos show a blueish tint, but note that I put a white paper in the background, so if you happen to have the photoshopping software, which I don't have, you can download these photos and adjust the RGB gain such that the paper looks plain white. Then, all the other colors should be correct, too. The correct information is there in the digital photos. BTW, the background paper is from the one and only patent they got on the NMOS 6502. Quite an interesting read. Essentially, it claims protection for the on-the-fly BCD correction logic, and nothing else. This is why some non-licensed 6502 knockoffs do not support the BCD arithmetic, and it might be the reason why some CMOS versions "need" an additional cycle to do it, all to dodge the wrath of the patent holder.

But now to the photos:

it's a stack of three different Apple-1 PCBs made by different processes. The lower one is a Newton-NTI. The middle one is one of Uncle Bernie's specials (not for sale), but still worthwhile to show, as they were made by JLCPCB's "ENIG" process based on the Open Source Gerbers. The top one is a standard process JLCPCB made "Replica-1" PCB I bought from "Pinguso", who is a honest Ebay seller with the best prices, and not one of these pesky price gougers who just want to rip us off.

Here is the photo with the least amount of blueish tint:

A1_PCB_4.JPG

The rest of the photos follows at the end of this text.

As you can see in the photos, the standard process PCB (on top) has no gold plating at all on the edge connector. It has the same tin plating every other PCB trace else has. The consequence is, if you ever use that edge connector to plug things in, then it will work for a while, and then it will fail, and you will need to take a pencil eraser to the edge connector to get it shiny again. So it will work again, for a while. Rinse and repeat. Tin or lead/tin plated edge connectors are no good ! Back in the 1980s, during the home computer price wars, some manufacturers stopped ordering gold plated edge connectors to reduce costs, and the consumers using them sooner or later had contact problems ... note the words "using them". If you never intend to use this edge connector, having no gold plating there is OK, and you saved a lot of money. As the ACI daughter cards plug into the green or blue 44-pin socket, at least for me, I don't see any reason why to ever use the edge connector. For me, having gold plating there is more for optical reasons (the "authentic looks"). It boils down how pragmatic you are, and what exactly you intend to do with your Apple-1 clone, and how much money you have to squander, to decide to go for the gold or not.

The middle PCB shows the pathetic "gold plating" you get from the JLCPCB "ENIG" process option. See also the previous post by "brouhaha", which is related. The "gold" is only a few atom layers thick, and, alas, it shows. On their webpage which pretends to show you how the final product looks, it looks as if it were thick, fat, 24 carat gold, like on these pimped-up Desert Eagle pistols. Naaah. What do you expect from a nation where "cheating the customer" is deeply ingrained in their culture. They don't even see it as "cheating". For them, it's "winning". Being a (retired) semiconductor industry insider, I could tell you horror stories ... but the best one is that phone call an ex-colleague got from his mother who lives in China. She had bought a carton of eggs on the local farmer's market and when back at home, she found out the "eggs" where realistic looking plastic shells filled with water. She was furious and called her son ! Now think about it. How much does it cost to mass produce real eggs from real chickens. In China. And much does it cost to make the injection mold forms for these fake eggs. It eludes me how this fraud ever could be profitable. But they do it anyways !

The lower PCB is a Newton-NTI PCB which shows nice gold plating on the edge connector. I don't know how the PCB house Newton Mike uses does it, but back in the day there was a manual process called "tampon galvanizing", which only uses minimum equipment: A tool holds a small sponge, the "tampon", soaked in gold-loaded potassium cyanide solution, and it is connected to a DC voltage source of 3 Volts or so, and the other wire goes to the edge connector pad to be gold plated. Then, the "tampon" is gently rubbed against the edge connector pad, building up layer upon layer of gold. This only works if the base layer is nickel. Tin or Lead / Tin don't work. And it is a tedious, time-consuming, manual process that is all but extinct nowadays. There are of course more automated, batch processes to do that gold-plating of edge connectors but they require much more investment in the production equipment and many more process steps. I wonder why JLCPCB apparently does not offer proper gold-plated edge connectors, only their ENIG process. But I did mention the "tampon galvanizing" process for a reason: it may be able to salvage the ENIG PCBs and add some more gold, such that the connector at least looks better. Alas, I found it impossible to buy potassium cyanide here in the USA, despite the nearby Cripple Creek gold mine uses it by the tons in their leach process, which is done in the open. Another case where pesky lawyers deprive us consumers from the materials needed for our crafts / arts.

You might also notice some differences between the various PCBs. The PCBs based on the Open Source Gerbers have small round solder pads, and the Newton-NTI has larger, oval solder pads. There is also a size difference. Newton measures 15 9/16 inch x 9 3/16 inch (395 mm x 233 mm) and the Replica-1 PCB measures 15.5" x 9" (394 mm x 229 mm). Disclaimer: these are measurements taken with a ruler from Dollar Tree and in my age, the eyes are not the best anymore, so don't blame me if there are slight errors in these numbers. The major takeaway that applies in any case, even in the presence of possibly faulty measurements, is that the physical size of these PCBs is slightly different, and the screw hole locations also are slightly different, so you need to be aware of these mechanical differences when building custom enclosure(s) for your Apple-1 build(s).

As for functionality, I have built several Apple-1 clones based on Newton NTI, Newton non-NTI, and the Replica-1 PCBs, and all of them work fine, at least after my reliability mods were applied, so other than the differences pointed out above, there is no reason to worry.

These are the rest of my PCB photos:

A1_PCB_1.JPG

A1_PCB_2.JPG

A1_PCB_3.JPG

A1_PCB_5.JPG

A1_PCB_6.JPG

A1_PCB_7.JPG

You can't beat the gold IMO, if the board was done right.  Yes it costs more, but I think it's worth it.