As some of you already know, one of my two Apple-1's is being auctioned on eBay this week:
If you're interested in bidding, please note that all bidders must go through a simple verification & approval process by eBay, so please do that before the weekend. If you don't, your last minute bid will likely be rejected.
If you have any questions about the computer or the auction I'm happy to answer them. The computer is currently in Cupertino, California, for a private event...but if you're interested in seeing it then please post here. I've organized a "happy hour" at BJ's Bar & Restaurant in Cupertino on Wednesday, June 8th, from 5-8pm where you can come see it. BJ's is located in the parking lot of Apple's Infinite Loop campus.
I'm a bit surprised that three days after the auction has ended still no peep and no comment on the outcome.
This is history in the making, folks !
For posterity, I have made a screen shot from the Ebay auction before it disappears into the big bit graveyard / Nirvana:
Frankly, I expected a better outcome above $400k, but with the current economical situation worldwide (the wheels coming off all the clown cars by fault of the drivers at the steering wheels) there is a highly likely and obvious reason why this auction could not have reached the hammer price seen in previous auctions. So do not construe this event to be a collapse of original Apple-1 values. We might see Apple-1 prices in the trillions of US$ in few years from now. Or in the Quadrillions. Depends only on how many more USD will be created out of thin air ;-)
If we take the official CPI (which, alas, is a lie), the original Apple-1 of 1976 for $666.66 retail price would now cost $3424.64 (calculated with https://www.usinflationcalculator.com/), but IMHO, the real inflation was much worse, and the price should be over $10000 now.
You can see this as a futile attempt to put some reason into Apple-1 prices, but I'm still off by a factor of 34. And this factor, Ladies and Gentlemen, would have been your speculation gain above inflation if you had bought a few Apple-1 originals, back in the year 1976. I'm too lazy to compare this with the rollercoaster ride of the AAPL stock since they went public. But I think it wasn't a bad investment either. Maybe some reader with more spare time to (apple) fritter away can enlighten us about that, so we can compare.
It also would be interesting to see your opinion about the speculation with Apple-1 clones. Shocked ? No, that's the ugly truth. I always thought that Apple-1 clones are being built by true enthusiasts who want the hands-on-experience but don't have the spare change to buy an original. Or by collectors or museums who are also strapped of cash. But it the last few months it came to my attention that more and more speculators and greedy persons come on the scene. Greed pervades the whole Apple-1 biosphere down to the hellish subterranian realm of the IC brokers, who also smell it, the money, and they now want $200 for a single Signetics 2513N character generator. Sigh ! This greed takes the fun out of our hobby, I was told from several sides now, and sadly, I must concur.
Comments invited !
- Uncle Bernie
I thought it would go higher as well but as some have mentioned it's an NTI version and not the original. The other one at RR Auction right now is an original but it has a lot of chips replaced but it's already at $230K with 8 days left.
As of last year the following article said an investment of $1000 in Apple woul be $1.2 million after 40 years. Better timing in investing would lead to $3.1 million in return so a little bit better than buying 1.5 Apple 1 computers.
The prices of some parts for Apple 1 replicas are truly becoming rediculous. I've seen prices up to $45 USD for a single MK4027-2 DRAM!
In post #4, "justinmc" wrote:
"but as some have mentioned it's an NTI version and not the original."
Uncle Bernie has to correct this fallacy before people start to regurgitate it:
Even the NTI version is an original Apple-1 !!!
So whoever argues that "an NTI version and not the original" is gaslighting you (or is misinformed).
As I am confronted with this fear, uncertainty and doubt over which version to build with my kit builders all the time, I think it's approporiate to clarify what really happend back in the day in terms of Apple-1 production:
1. First, there was the Apple-1 prototype. The one you can see in all the ads. (The hand wired thing Woz had built before the PCB was made is NOT the prototype of the Apple-1, but a nameless predecessor, Woz calls it "my computer").
2. Then came the 1st production run of about 50 Apple-1 sold to BYTE SHOP, which was the deal Steve Jobs had found to make the whole startup viable. These are called "BYTE SHOP" or "non-NTI" versions.
3. Then came the 2nd production run, the so-called "NTI" machines. As far as I could find out with my research all the "NTI" Apple-1 were sold by Apple via their various magazine ads. And not all 2nd production run boards were populated with ICs, there were some leftovers, which were wave soldered, with all the components, but no ICs in the sockets. When one was ordered, Apple would plug in the ICs and then test and ship it. At this time they already were busy working on the Apple II.
The most obvious differences between the "BYTE SHOP" and the "NTI" versions (and the root of the notoriety of the Apple-1):
If you look closely at the pictures available on the web, the prototype had small cuboidal bypass capacitors. Because "cuboidal" is a term from evil mathematics, a science which fell out of fashion, and nowadays 2 + 3 = 4 (so-called "woke" math, no kid left behind) we could also use the term "brick-shaped" or, "boxy", which I prefer, as such a small capacitor ain't no brick.
THE DIFFERENT TYPES OF BYPASS CAPACITORS
The "BYTE SHOP" version had disc shaped bypass capacitors. These had a darker brown color. Brand was "BEL". These were quite common in early arcade machines.
The "NTI" version again had the "boxy" capacitors as seen in the prototype. These had a turquoise color (blue-green).
THE BIG BLUNDER AND THE ROOT OF ALL THE PROBLEMS
The two Steves had made a fatal mistake you never, ever do when you start up a production of any series produced item: they did change a key component for the 1st production run, and probably without sufficient testing for consequences of that change in the "Bill of Materials". This is a practice that is strictly prohibited in all industries. You just don't do that. Maybe they ran out of money and had no other choice, who knows.
For the Apple-1, the consequences were fatal: users complained about their Apple-1 suffering from program crashes all to often. According to my research and my lab experiments, the main culprit is the lower electrical performance of the disc shaped bypass capacitors.
From the available body of evidence we can make two conjectures:
- the prototype worked robustly and was OK, and did not crash randomly. Otherwise Woz could not have developed the Apple BASIC interpreter by typing it in by hand from his "hand assembly" code done with pencil and paper. On a machine crashing often it would have been impossible to develop that BASIC interpreter.
- they knew what they had done but kept their mouth shut. And silently and ruefully changed the bypass capacitors in the 2nd production run back to the higher performance "boxy" ones.
THE NTI VERSION IS THE BETTER ONE
So, according to this "Uncle Bernie's theory of the genesis, birth, rise, and fall of the Apple-1", the "NTI" version is the better machine and it should work fine, if no components have deteriorated over time (note that ceramic capacitors also may deteriorate, they may lose capacitance over the years, which can be restored by baking them. This depends on the particular ceramic material used, and is not generally applicable to all ceramic capacitors).
BUT MOST PEOPLE PREFER THE BYTE SHOP VERSION
Alas, for some weird motive which defies logic and reason, people prefer the "BYTE SHOP" version even when building clones. Only about every 5th IC kit I sold was ordered with the little blue boxy capacitors. This is why I stopped offering this option recently. Whoever wants to build a "NTI" version with boxy capacitors can order them from the same distributors where the three "big blue" Sprague capacitors and the heat sink and the transformers can be ordered. It makes no sense for me to provide these components with my kits - sales tax and Ebay fees would increase the costs for them by 30-35% over a direct purchase from a distributor. And on top of that, I can't ship heavy transformers anywhere near the cheap shipping costs the distributors charge.
But back to the BYTE SHOP vs NTI topic.
If I had to build a clone or buy an original, I would go for an NTI, as it is the better and more reliable machine. I have succeeded in buildinga NTI clone with no "reliability mods" whatsoever, and it worked perfectly (actually, it worked in the "gang of magnificent seven Apple-1 clones" which do the burn-in process for my ICs).
Alas, when I ran out of Mostek DRAMs and had to switch to Intersil DRAMs, it needed the six damping resistors, otherwise the burn-in program would throw the occasional DRAM error. But it never needed any additional bypass capacitor. It still runs the burn-in and it must not crash or throw any error for a full month running 24/7.
I think this proves my point that the "NTI" machines are the more robust ones. Still, they are not perfect, as seen with their sensitivity to the particular type and manufacturer of the DRAMs. It is caused by the marginal power supply bypassing in the original PCB design. First generation 4kx1 DRAMs such as the Mostek MK4096 used in the originals were made with slower process technologies and may have been less demanding for the power supply and ground grids, and more robust against poor power supply bypassing or ringing signals. It's essentially the same story again and again and again: plug a modern, allegedly "drop-in" compatible digital IC that was built with a much faster CMOS process technology into a socket of an older machine which used slow NMOS or PMOS process technologies and works with them, and all hell breaks loose.
INFLUENCE OF THE PCB PROCESS ON APPLE-1
There is another factor, which is the PCB itself. The PCB process back in the day was different. They used HASL (Hot Air Solder Leveling) before applying the silked screened solder mask. Thanks to the tin/lead cover the traces were much thicker all over the PCB. The tin/lead alloy has a much poorer specific electrical conductivity than copper, but the latter is very thin compared to the former. If you examine the cross section of such a PCB trace, a thick "dome" of tin/lead sits on a thin copper base. This construction also causes the "wrinkle" effect in wave soldered PCBs, as it acts as a capillary when molten, which is the most important "proof of authenticity" for the originals.
The modern PCB processes do not use that older methods and the solder mask is not silk screened anymore (silk screen printing requires skilled "artisan" type workers) but it's a photomask process nowadays. The tin is applied after the solder mask, so all traces under the solder mask do not have any tin, only the pads in the solder mask opening have that tin cover. Lead is taboo nowadays thanks to RoHS legislation. The copper-only traces themselves are as thin as it gets. Which is a problem because the electrical performance of the thinner copper-only traces is worse than the thickly tin/lead covered traces of the PCBs made 45 years ago, which worsens the performance of the already quite weak and pathetic power and ground grid caused by the layout even more.
MODERN PCBs ARE WORSE FOR THE APPLE-1
This means that no Apple-1 clone built on any of the modern PCBs can reach the robustness of the originals when it comes to the electrical performance of the power and ground grid and the other traces, when using no further mods. With the power and ground grid layout being marginal in the originals already, the modern PCBs are losers. Which is the reason why builders face problems and why I recommend my "reliability mods".
I have explored this topic very thoroughly over the past 2 years and use of higher performance bypass capacitors helps a lot. The added bypass capacitors of the "reliability mods" also bring remedy. You can further improve things by ordering PCBs with 2 Oz copper instead of 1 Oz copper. This does not yield twice the thickness of the traces in the finished PCB, but it brings improvements which can be measured. I have seen much less voltage drop across the board, and the "pollution" level on the power and ground lines as seen on an oscilloscope also is smaller. These improvements brought about by the higher copper weight are not huge, but like with any incremental improvement process you have to scrape the gains together whereever a little gain can be had. A few percent here, and few percent there, and everything sums up. Each such tiny step moves the Apple-1 further away from the "cliff of unreliability". If you want the great and crude picture, as I see it, the "BYTE SHOP" versions stands right at the cliff. The "NTI" version stands one step back from the cliff. Any build using modern PCBs and the faster 2nd generation 4kx1 DRAMs balances on the edge of the cliff (and some fall into the cliff). Put a DS0026 in (in lieu of the slower DS0025) and all fall. Apply all the known remedies and all the "reliability mods" then you are somewhere in the "NTI" league. Exactly where I can't tell. Somebody having a NTI original should put a DS0026 in and find out if it still works or not (full DRAM test, please). Then I could tell you. With my "reliability mods" a clone runs just fine even with a DS0026.
ARE MODERN PCBs REALLY THAT BAD ?
You might ask why we still can build electronics despite the modern PCB process being so much worse, but keep in mind when making PCBs for Apple-1 clones we plug a 45 year old, hand made, suboptimum two layer layout into a modern PCB process, a move which was not intended by the original designers. 45 years ago you could get away with a marginal layout that spits RF all over the spectrum. Today, all PCBs are designed using CAD tools and the power grid is carefully optimized and simulated. Many more than 2 layers are used to allow for power and ground planes. Otherwise non of our modern electronics would work.
I think this sums it up quite well: the original Apple-1 were not as wonky as we initalliy believed based on the failed attempts to build clones using modern PCB processes and faster (and hence, more aggressive) 2nd generation DRAMs. But they indeed were marginal, close to the "cliff of unreliability". And the ""BYTE SHOP" version was closer to that cliff, thanks to the substitution of the proven bypass capacitors of the prototype with (cheaper ?) bypass capacitors of poorer electrical performance. This mistake was corrected in the 2nd production run, or "NTI" machines, and the higher performance bypass capacitors were back. Still, the "BYTE SHOP" version fetches a higher price in auctions than the "NTI" version and is also preferred by builders of clones (by a ratio of 5:1). We know remedies for all of this and can make any Apple-1 work robustly by applying the proper mods and by carefully selecting components and the proper PCB manufacturing recipe. This improved know-how over what the two Steves had back in the day enables the new production of Apple-1 using industrial wave soldering techniques. But without the "wrinkles" ;-)
Comments invited !
- Uncle Bernie
To be more exact I should have said original production run. As you said they are both "Apple 1" but it's like revision 0 & revision 1 although they were not labeled as such.
The original production run is more popular because people don't care as much about reliability since most of these are just for display and not atual use.
In post #6, "justinmc" wrote:
"To be more exact I should have said original production run. As you said they are both "Apple 1" but it's like revision 0 & revision 1 although they were not labeled as such."
Uncle Bernie comments:
I would not call this "revision 0" or "revision 1". There were no changes in the circuit nor in the layout. The different PCB manufacturer slapped their "NTI" long on. And they reverted back to the higher performing bypass capacitors that already had used in the prototype. The color of the 22uF capacitors changed from yellow to blue early in the 2nd production run - but that does not qualify as a "revision" either. At least according to common industry practices. A revision entails more significant changes.
I think the reason for the preference of the BYTE SHOP version is that they were the first ones produced. Along that kind of "logic" you want serial number 001.
- Uncle Bernie
I have to say that Uncle Bernie makes a pretty compelling case for the NTI version being the preferred ones -- a working NTI is better than a broken "original batch" BYTE SHOP version. It isn't like there are thousands of either running around. It's only a few dozen of each version still in working order these days really.
It could be referred to version 1 and version 2 as well, or whatever the proper terminology to denote that they are not identical. The NTI version removed the extra 22mf in the silk screen and an additional hole was added in the top right corner of the board as well as the NTI logo of course and the matte green soldermask changed to glossy. But as you say maybe not enough to officially designate a revision but certainly not 100% the same.
Once the new owner has his Apple-1 in his hands (it's in transit at the moment) I'll share more details on the auction process, my thoughts on how it went, and what advice I would share with others who may be interested in selling their Apple-1's in the near future.
The summary is I'm pleased with the result and even more pleased with where the computer is going!
Moderator please delete! Comment failed to post becasue it contained an Emoji! Oh the horror!
Moderator please delete! Comment failed to post becasue it contained an Emoji! Oh the horror!
Now that the auction is over and the Apple-1 is on its way to it’s new home, I want to share some details about the auction process and my thoughts on how it went and what I would do differently in the future. Warning, this is a long Uncle Bernie-style thorough post with a lot of thought put into it.
Bidding for the “Schlumberger 2” Apple-1 ended on June 2nd at 9:00AM Pacific. The highest bid was $340,100. A total of 28 bidders participated and placed 113 bids over the ten days of the auction. The listing had more than 140,000 page views and was Watched by more than 6,500 eBay members. The auction fees charged by eBay came to just over $8K.
From a financial perspective, the auction was a success. I purchased two Apple-1 boards from the estate of the previous owner with the intention of restoring both and keeping one. The proceeds from this sale covered my cost of both Apple-1 computers and that was what I had hoped to do. For this particular sale, the starting bid was only 99¢ and there was no reserve. Had I elected to set a reserve, it would have been at $250K and my target was to net $300K. My stretch goal was $400K, which we did not achieve…but only due to technical issues experienced by a bidder from Europe. Based on speaking to both the successful bidder and the guy in Europe, had his bid come in on time the closing price would have been $400,100.
The computer is going to a great home and the successful bidder already owns a “Byte Shop” Apple-1 along with a number of rev. 0 Apple II’s.
The next topic I’d like to cover is the decision to work with eBay versus traditional auction houses. I chose eBay for four reasons (in order of importance):
Traditional auction houses provide a turnkey service. The store the item, insure it, photograph it, market it, handle the financial aspects of the transaction, and ship it. They also usually have a network of perspective buyers and loyal customers that they can tap into to maximize the chances of success. Obviously there is a cost to this “white glove” service: typically they take 10% of the hammer price from the seller and charge the buyer a 20-25% premium. For example if the hammer price is $400,000 then the seller nets $360,000 and the buyer pays $500,000 plus sales tax and shipping (another $20-40K). My view is that for most Apple-1 owners, this is still the best way to bring your Apple-1 to market.
With a standard eBay auction you get none of this since even their payment platform doesn’t support amounts above $20K or something like that. However, the fees are much lower. They charge the seller ~14% on the first $7,500 and 2.3% on the amount above that, plus some nominal listing fees depending on what options you choose (as little as $1).
Despite this I went with eBay. Beyond fees, one major factor was timing. I wanted to have the auction in late May or early June, and the other auction houses I spoke to were proposing September or November. I generally have a pessimistic view on where the market for collectibles is headed over the next 6-12 months, so holding the auction as soon as possible was an important consideration. We ended up postponing the start to early June because another Apple-1 auction started in mid-May. Despite my efforts to avoid being up against another Apple-1 auction, a second board was listed at the end of May and that sale just ended. Having three Apple-1's on the market at basically the same time is less than ideal and is probably one of the reasons that none of them achieved the results seen in previous years. The first didn't meet reserve, mine sold for $340K, and Roger Wagner's just sold for $375K.
Another important factor in my decision to work with eBay was how much involvement I would have in the process. I wanted to be involved in all aspects for two reasons: it's fun and it helped me increase awareness of my collection. My goal is to have a permanent, easily accessible home for my collection and I believe increased awareness will help me achieve that. At least that's the dream!
So what was it like working with eBay? It was great! They have a team dedicated to the sale of collectibles on their marketplace and that team was very enthusiastic and supportive. They were intimately involved in the entire process from creating the listing, to marketing, to helping me with payment options, and even media/PR activities. They invited me to visit eBay and held a media day, and even their CEO came to see the computer and expressed his support for the auction. I couldn't have asked for more from them. At the closing price of $340K, the net proceeds were approximately the same as what I would have gotten with a traditional auction house. If we had hit the figure of $400K, it would have been both financially and timing-wise a win.
I'm not sure I would do much differently with the sale of this particular Apple-1. Its condition was rated as 6.5/10 by Corey, it lacked original manuals or ACI, and its history is a bit murky due to the desire of the previous owner's family desiring to remain anonymous. The only two things I can think of is to be more proactive in supporting potential bidders to bid early and avoid using sniping tools that can fail, and also to end the auction on a weekday when eBay customer support is working and available to help bidders having issues.
My other Apple-1 would likely be rated a 9/10 and should I ever decide to sell it, I probably would skip auctions all together and instead take my time and conduct a private sale. It would be a lot less work and be less stressful, but would probably take longer. That view may change, but that's how I feel at the moment. If you have any questions or thoughts please feel free to share them. I hope this long post is helpful to other Apple-1 owners who are or may consider a sale in the future.
The RR website shows a sale price of $468,750 to me. Where did you see the $375,000?
That figure includes the 25% buyer's premium. The hammer price was $375K, and the consigner gets the hammer price less any seller fees (10% in the case of RR).
Congratulations on the sale! It was very interesting and informative to learn how it all works. If I had that kind of money, I would buy your board and give it to the Apple Museum in Moscow, so far there is only a replica, although it's quite accurate. But someday we'll definitely have an original! My daughter will be married at the end of the summer in Moscow, I'll try to visit the museum and take some pictures.
Thanks for the explanation. They really should make that clear to avoid deceiving people into thinking they can sell for more through their auction service. Congrats on selling yours in a more transparent manner on eBay