Apple-1 Acrylic Case Prototype

Once comment, you may want to cut out the top area above the large heat sink.  The amount of heat generated by the LM323K is substantial and will cause issues with the plexiglass and if you don't give it a place to go on the top of the case, it will cause heat distorion with the plexiglass  and issues with the actual running board.

... these heat sinks were designed for absolutely free air flow and a "plexiglass ceiling" of only 1 cm distance will already do something to the air flow, although I don't dare a guess how bad it will be. As for the temperature, the air coming off the heat sink will have much the same temperature as the heat sink itself, which for the specified transformers and the TTL "power hog" usual builds will be 40-50 deg Celsius higher than the cool air it gets: assuming a RT of +28C, this would be +78C. Hmmmm.

I would recommend you to do an experiment on a prototype with the correct heat sink and the LM323K mounted and running at the specified input voltage and the typical load current of 1.4 - 1.5A.  Not having a complete Apple-1 build, you could rig this with a load resistor that is rated for 10W and installed on a separate heat sink. Beware of those types which are in a yellow-ish/orange-ish aluminum case ... unless they are properly cooled, they will overheat, and if they overheat, they will act as a little cannon, and shoot their resistor core out with a loud "bang", and the last time this happened to me, that core was stuck in the fiberboard ceiling. Bad karma if it hits a person... this is why it's called a "laboratory" and has "DANGER" signs on the doors  listing all sorts of perils for anyone daring to enter: enter and touch things or look to close at things and come out dead or mutilated, or blind or deaf, or all of the above.

I could give you more exact data on voltages/currents/temperatures if you ask me, the above are just conservative numbers from memory, but I don't have my lab notebook with me right now.

Proper air flow is essential for anything that consumes power and turns it into heat. I have seen too many otherwise good projects fail because the designers did not properly care about the thermal design. This can be the key feature which makes everything possible ... see the supercomputers designed by Seymour Cray. But if you really do the math, you will find out that nowadays, normal but high-performance desktop PCs have a power density in their CPU chips that exceeds the power density in a nuclear power plant reactor core. It is amazing that all this heat can be moved away with relatively cheap, mass produced components. These guys know what they are doing. Ignore the laws of thermodynamics at your own peril. Ah, and with cooling fins, and airflow, you also need the laws of aerodynamics. All this is not trivial.

I agree with UncleBernie regarding the heat dissipation effects of having a plastic surround in close proximity to the top of the heat sink, but if you can either make a bunch of holes or even some slots (might look better) above the heatsink that accomodate airflow, that should remedy the issue.

I personally thought the heat sink got marginally too hot sitting in open air, so I found a fan that would fit atop the four corners of the heat sink itself that pulls air through.  Works well and gives me peace of mind that the +5V relegulator has half a chance at survival.

Besides the heat dissipation from the LM323K, there is one more nuance: how to insert the ACI card?  A sheet of acrylic is covering the connector.  A1 is not very functional without a card.

... and in general the idea is so-so, about the same as trying to sell snow in Siberia, people who collect A1 have the skills sufficient to buy a sheet of acrylic and drill 5 holes in it.

True, but do they want to?

Some of us enjoy the efforts of others, saving us the hassle of doing it. 

Some, maybe not.

macnoyd wrote:

"I personally thought the heat sink got marginally too hot sitting in open air, so I found a fan that would fit atop the four corners of the heat sink itself that pulls air through.  Works well and gives me peace of mind that the +5V relegulator has half a chance at survival".

Uncle Bernie's comments on this:

There seems to be a common misconception about these regulator ICs here on this forum and elsewhere, claiming they could "die" or such. This is not the case. The LM323/LM320/LM340 regulators all include circuitry to protect them from overload and overheating. The basic concept goes back to the legendary late Bob Widlar (who literally "invented" the first key circuit techniques to make analog ICs viable) with his LM309 released by National Semiconductor in 1969. Two other "gurus", Bob Dobkin and Carl Nelson, further developed the concepts (Dobkin gave us the first three terminal adjustable regulators (LM317), and Nelson designed the first three terminal fixed negative regulators (LM320). With their self-protection circuits these regulators are virtually indestructible (as long as used within their specified input voltage range).  The same is true for the 78XX and 79XX series from the original manufacturers (National Semiconductors and Fairchild) and from all the direct copycats who knocked off the circuits and layouts verbatim (to cite a CEO: "to hell with the patents, let's make the money first and then let the lawyers sort it out" - be warned that some ideas are just to good not to be ripped off, standard industry practice, so don't even think of patenting your own  inventions as an individual, just don't do it, they will rip you off !). 

This robustness is the key reason why these regulators are still being manufactured in huge numbers 50 years later, although in most applications  they have been replaced by switchmode regulators (those offer a much higher efficiency). Beware of some dubious East Asian manufacturers or mainland China manufacturers, some of those have modified the circuits and layouts without really understanding the original designs, and these counterfeit regulator ICs are less robust and can be killed under some conditions, such as suddenly applying certain high load currents under input voltages being close to maximum, and if mounted on poor heatsinks. But any of the original regulator ICs will survive that without sustaining any damage.

Don't fear death of these regulators (if they are genuine) !

The takeaway from this is that no Apple-1 builder using any of these regulators from reputable manufacturers must fear that the regulators would die in his Apple-1. They won't. The worst thing that can happen is that they run so hot that the protection circuits will kick in and shut it off. And this means that the regulated voltage on their outputs will collapse to a much lower voltage or even to zero. Once the regulator has cooled down, all will be OK again !

But keep in mind the consequences of self-protecting regulators shutting down when running too hot:

If your regulators have inadequate cooling, this means that the Apple-1 will run for a while but then stop working when the regulator(s) got too hot and did self-protect by reducing their output voltages, and  then, below a certain minimum supply voltage, the  logic ICs could not work properly anymore. Which typically means a program crash or even the screen output disappearing or getting garbled. But after turning the power off and letting it cool down, it will work again. For a while, at least, until a regulator gets too hot again. Rinse and repeat.

If your Apple-1 has these symptoms, do not suspect any IC to be defective yet, before you have checked all power supply voltages still being in spec at the time of any program crash. This may require hooking one or more  multimeters  to the regulated power supply rails in your Apple-1, so you can see when, how, and what happens just before the symptoms manifest themselves. The cheapest digital multimeters for $6.99 to $9.99 will do for that, so there is no excuse.

I have coached many Apple-1 builders (and still do that) and here is a common mistake, especially when using somewhat out-of-spec transformers that increase the input voltages on the regulators even further than normal: the Apple-1 runs at first, but then the above symptoms appear.  And if the "lucky" builder in this situation happens to have a can of cold spray (which is expensive) and - of course - suspects a bad IC, then in a typical case,  soon the cold spray will be empty and still no bad IC has been found. And then they contact me and blame my 100% tested and burned-in IC kits to be  faulty. It then takes me a while to calm them down and tell them the regulator self-protection story. And make them measure the regulator input voltages. And, oops, so far they always have been much higher than if the correct transformers had been used, and all the ICs were good, of course. One particular clown even thought he would not need any heat sink on the LM323K, and the excuse was, he "just had no time to order one and just wanted to try my Apple out right now". He also had no clue how to desolder his poor LM323K again to finally put a heat sink in.  These are the days when I regret to sell my overstock of Apple-1 parts to individuals. I should have taken the tax writeoff and destroyed them.  But on the other hand, happy builders who succeeded with one of my IC kits reward me enough to keep me doing it.

More about transformer choices:

If you use the 10Vac and 28Vac CT transformers as specified in the Apple-1 manual (or equivalents) then the Apple-1 already runs at regulator voltage headroom that is much too generous: in my builds I measured +10.5V input voltage on the +5V regulator (which would need only 7.5V) and +/- 22...23V on the +/- 12 V regulators (which would only need +14.5V and -13.5V). Linear regulators reduce the excess input voltage to the output voltage by turning the excess power into heat. So, for instance, Pheat = (10.5V-5V) * 1.5A = 8.25W, W being Watts, but when you increase the regulator input voltage to 12V (14.3%  more) then Pheat = (12V-5V) * 1.5A = 10.5W, which is 27.3% more heat. Which means, the regulator temperature gradient over room temperature will not only increase by 14.3%, but by 27.3%, by twice as much as expected from the possible naive (and wrong) assumption:  a little more input voltage indeed will make the regulator run much hotter than just the little voltage increase would suggest to the unenlightened. As a counterexample, if we run the same +5V regulator from the minimum input voltage of +7.5V, the Pheat is greatly reduced, to just 2.5V * 1.5A = 3.75W, which is less than half the heat of the 10.5V case, and the LM323K will stay cool enough to touch it with bare fingers. This would be a nice optimisation !

Why is the Apple-1 power supply as suboptimum as it is  ?

It somehow eludes me why Woz decided to run the regulators in his Apple-1 with such excessively high headrooms. The resulting power supply system is thermally close to its limits and the TO220 regulators having no heatsinks can only make it worse than it could have been. Maybe the People's Republic of California already had a third-world level electrical power grid even 45 years ago, with regular brownouts (and with "regular" I mean several times per day, so you could not reasonably run any useful programs on the Apple-1). Or, Woz just happened to have these suboptimum transformers laying around for his "hobby project" and once pressured by Steve Jobs into "mass production" he did not have the time anymore to shop for and qualify / test more optimum transformer candidates which  would not run these regulators that hot. (And, BTW, about the "third-world level electrical grid": soon coming to a state / city / town near you, guaranteed ! Solar panels and windmills do not a robust and dependable source of electricity make. And the "decision makers" aka "politicians" rarely do have a degree in electrical engineering (or any STEM field in general), but even having that education would not help them much, as the "going green" virus happens to be an incurable, highly contagious disease that attacks, among other organs, also the brain, and impairs the process of rational thinking badly and irreversibly, so they shut down perfectly good coal fired power plants and replace them with dubious non-solutions which over their lifetime will never generate as much net energy as it did cost to manufacture, transport and install them ... a dead end in any case. Don't believe me ? Ask any Texan about having no electricity and no natural gas coming in to heat the house, and that in a state that sits on vast reserves of natural gas !).

Transformer optimisation:

Note that in the Apple-1 manual, there is a note that the +5V side could use a 8Vac transformer, which would be "cooler" (literally). I think that the +/-12V side could be run with a 24Vac CT transformer (or maybe even a 22Vac CT transformer, if such a thing would exist, but alas, catalog transformers tend to jump from 20Vac CT to 24Vac CT). Actually, with the feeble load currents on the regulated +/-12V rails  of much below 100mA each (in my builds I have measured 50mA on the +12V rail for 8K RAM running heavily, it drops to 20mA if the RAM is idle), it is even more of a mystery why using the CT configuration at all, a single secondary winding would do easily, but maybe the two Steves had plans to make an expansion backplane that would have had many slots for daughter cards consuming substantial power each,  and each daughter card would have had its own +/-12V regulators, S-100 bus style.  In this context, having a stronger and CT transformer for the +/-12V side would have made sense. Today, we know that such an expansion backplane for the Apple-1 was never offered by Apple, and the Apple-II had enough slots (lesson learned) and a switching power supply (another lesson learned).

No more F.U.D. about regulators in the Apple-1 !

All this lenghty discussion hopefully was helpful to reduce fear, uncertainty and doubt (F.U.D.) about the Apple-1 power supply in general and the transformers in particular. If there is more interest (comments !) then I would perhaps write up some of the theoretical background such that builders could optimize their power supplies. There is, for instance, plenty of opportunity to reduce either transformer output voltages, or filter capacitor capacitances, but a delicate balance must be kept between them, so you can't reduce the transformer output voltages and the capacitances to their minimums at the same time. Approaching the "sweet spot" for this is a bit beyond simplified hand calculations and would need the use of SPICE (the circuit simulator), however, and I don't know if anyone on this list would ever want to venture into that professional realm.

I test every LM323K before I install and I have had a few real vintage ones fail to input voltage as output.  I do agree the latest/modern versions are pretty bulletproof.   If you don't care about originality then they even make switching supply versions that fit in the package and produce much less heat.

Corey986 wrote:

"I test every LM323K before I install and I have had a few real vintage ones fail to input voltage as output.  I do agree the latest/modern versions are pretty bulletproof.   If you don't care about originality then they even make switching supply versions that fit in the package and produce much less heat."

Uncle Bernie's comment:

So do I. All vintage ICs must be tested, no exception. Actually, what I do is a complete burn-in in a real Apple-1 clone build for four weeks, before I declare the IC set to be fit for sale to other builders. As a semiconductor industry insider, I know many failure mechanisms that can affect vintage ICs, even if they never have been used. And yes, since the beginning of my burn-in activity 3 months ago, I already have little piles of "bad" ICs which were weeded out by my test / burn-in process. I do make some exceptions, for instance, I don't test the more recently made LM311 and 74LSxx ICs in the ACI, nor the TO-220 regulators, but I do test the LM323K and sort them in voltage bins corresponding to my burn-in rigs. The test rig is like this:

LM323Test.JPG

Note that this is a brief test, and no burn-in, which I don't deem necessary for these parts, their cans are hermetic. The test is done at full load but without heatsink, so it must be done quickly. I use salvaged, vintage LM323K, mostly made by National Semiconductor in the late 1970s and early 1980s, and I would have expected to find some bad ones, but so far none was bad, perhaps the vendor who salvaged them did a test already and weeded the bad ones out. They were not exactly cheap, so there was a margin for doing such a test.  And so far none of those I sold did fail. Which is good. Because they are tricky to desolder from the PCB without damage to the PCB. At least with hobby tools. With a professional desoldering station using hollow tips and an electric vacuum pump there is no such problem and the desoldering of a LM323K can be done very quickly.

However, I was told some horror stories about newer LM323K bought off Ebay and / or Alibaba. Not that they would not work, but their output voltages vary widely. With those vintage ones made by National Semiconductors in the heydays of linear regulators all I have measured were in a +/-2% tolerance band, which is much better than their datasheet spec. I was told by Apple-1 builders I coach (which not always use my IC kits) that these dubious newer LM323K may have much larger tolerances, and one number I remember was 4.7V. (Ouch, this is 5V-6% and the TTLs will not be too happy with it).

I have two possible explanations for the lousy tolerances of these newer regulators:

a) Their manufacturers have lost the art (also called "secret sauce") to make them with tighter tolerances. Back in the day they  did control the processes in their wafer fabs to yield a certain bandgap voltage in the finished products, so they would have tighter output voltage tolerances without needing any trim. Using no trim a well designed bandgap reference can be manufactured with +/-3% tolerance and still excellent yields. By proper process control this can be tightened a bit more, yielding lots of parts within a +/-2% tolerance band, and then during EOL test you can bin them into 1%, 2% and 3% bins, and sell the "better ones" to your favorite customer(s) which will praise your awesome quality and low price. "Low price" as in "not using any trim". This one was a well-kept secret in the industry. Any competitor needing trim to get good yields in the same tight tolerance band (better than the datasheet !) already has lost. Fast forward 30-40 years and all the process engineers and test engineers knowing the magic spells have long retired or are deceased, the know-how is lost, wafer fab has been closed and scrapped, and now everything is done by East Asian contract manufacturers, where you can't turn the right knob at the ion implanter to optimize a certain lot of wafers, and the dire consequence is that we will never again get the quality level and the tight tolerances on analog ICs we once had. Digital ICs are a completely different story as the target objectives for such a process are different from the analog world.

b) The regulators which are peddled on Ebay or Alibaba (not exactly being an official distributor like Farnell, Element14, Digikey, Mouser, etc.) actually are scrap that has been discarded my their manufacturer for not meeting spec. These bad parts from the reject bin then go into recycling, to recover the valuable metals in them, which happens in mainland China. And guess what ... the Chinese are known to fish these parts out, relabel them, and sell them, instead of melting them down for the metals. This counterfeiting actually is done by a well organized mafia, and it has gotten so bad that they did extend their tentacles into the semiconductor industry itself. Within the East Asian contract manufacturers who do the packaging and test of the ICs there have been gang members who stole rejects (or sometimes even parts from good bins) right at the source, to funnel them into the non-legit sales channels of these counterfeiters. And even regular customers being cheapskates (or having overly aggressive purchasing departments) bought these parts from the counterfeiters and then complain about the parts not meeting spec to the original semiconductor manufacturer. The quality department opens the IC package expecting some strange die from third parties (a complete counterfeit) but much to their surprise (and to upset the management) there is a "genuine" die in it, which came out of the wafer fab of the company !  How could such a bad part not meeting spec ever slip through the final test and make it to the customer ? Big embarassment ! Except it didn't make it through the final test. Said mafia stole these parts from the reject bin and these parts then made it through dubious and dark sales channels back to the USA or to Europe. You can image how much investigation and detective work went into exposing all of this. The solution that has been adopted is to have no "reject" rails on the testers, but a hose, which is rigidly attached to the tester's "reject" bin port, and goes through a hole drilled into the concrete floor to a high security room in the floor below, where specially entrusted employees, under the supervision of security cameras,  collect the rejects in steel pans, and then they are destroyed / mutilated using propane torches and such, so nobody could ever sell them as an IC anymore. And yes, the charred remains again are being sold by the barrel for recycling / reclaiming the metals in it, but now nobody can fish out good-looking ICs anymore to sell them as such. Ladies and Gentlemen, this is the world we live in. Globalism at its finest ! Ever thought who cooked the precursor chemicals for the expensive brand-name medication (or food supplement) you consume ? You better don't ask. And the product counterfeiting extends into every field of human endavor - counterfeit brake disks have been found in 747 passenger jets, and also the cheap aftermarket brake pads and shock absorbers your local car stealership sold you for the price of the original spare parts may have come from dubious sources, and may not be what they pretend to be. Actually, they might cripple or kill you. 

Caveat Emptor ! (Buyer beware !).

Very, very interesting you write Uncle Bernie.  This is not written in a Facebook group.

As for the parts for A1 from China: either you will get what you were looking for or your money will be returned very quickly.When I was assembling my replica, I bought microcircuits for 2 complete sets: one on eBay, all signetics with vertical markings and the second, almost all from China.  Except for the non-working LM323k and blue capacitors that never came, everything else worked.  There were both the 80s and 90s, as well as completely new ones.  The 74th series is almost completely there, memory, processors, etc.  2513 and 2919 are just not there.

Macintosh_nik wrote:

"Very, very interesting you write Uncle Bernie.  This is not written in a Facebook group."

Uncle Bernie answers:

Yeah, Macintosh_nik, after a professional life spent in the semiconductor industry, there are lots of interesting anecdotes to tell, which, however, in most cases, never get told, and hence, get lost in time.

The "secret sauce" for many great analog ICs has been lost. They can't make them anymore. Unlike digital ICs, the analog IC realm is much richer but also much more difficult and dangerous to navigate. Think of "The Zone" in Tarkovski's movie "Stalker" (one of my favourite movies). To navigate the analog IC realm safely, you need to know the secret insider tricks and techniques. The whole field reeks of magic practised in medieval alchemist's dens but once you have understood - after long pondering  over the often obscure hints from the "gurus" -  on which laws of physics, and which mathematics, and which quirks in the wafer fab this "alchemy" is founded, you become enlightened, and can go on to design your own great analog ICs, like the gurus themselves. Although they had a head start, as they were there, in the 1960s, when all the analog IC design began, and so it's hard to catch up with them. They always will know more. Now with these "gurus" either being deceased (Bob Widlar, Bob Pease), or retired (Carl Nelson, Bob Dobkin) - just to name a few of the more prominent ones - and their disciples not knowing everything that once was and how the whole building was erected from the foundations up (which happened beginning with the mid 1960s), the analog IC industry nowadays is in trouble: too much of the "secret knowledge / secret sauce" has been lost. And today's IC design engineers more often than not are just messing around and produce mediocre designs, especially those from academia, who do not need to design ICs that actually can be mass produced, and the result is that today most of the once sacred pages of the "IEEE Journal of Solid-State Circuits", or "Red Rag", are full of crap. (I stopped my subscription, after 35 years, in disgust about the decline seen in the last 10-15 years, since more and more "papers" therein came from Asian hucksters and swindlers who just seem to make their "results" up, or take ancient circuits, add a few transistors that are not really needed, and then claim "novelty". It's absolutely disgusting !).

The same decline you can see in all other industries, though. Cars ... crap, engines die too soon for obscure reasons (google Theta II engine), Airplanes ... fall out of the sky (google 737MAX), Tools ... they just break and are so faulty that even a master grade craftsman can't produce any good work with them, Pharmaceuticals and vaccines ... plagued by side-effects that cause worse health damages (including death)  than the ailment causes that they are supposed to help with ... and so on and so on.

It's called "Idiocracy", "decline of civilisation", "decline of IQ", although plain old greed and corruption certainly plays a role, too. And complexity also plays a role ... I have seen expert opinions that with today's technology we could not fly humans to the moon anymore and bring them safely back to Earth. That was done 50 years ago using relatively robust and simple technology. The Russian N1 moon rocket was far too ambitious and complex ... and failed. But its very advanced engine technology now, 50 years later, still is the "gold standard" to get satellites into orbit (even in the West, as the Russians sell these engines which came out of the N1 project). There has been no significant progress in liquid propellant based rocket engines ever since. Our cars, if they work, and their "modern" engines don't turn into scrap metal prematurely, don't give us any significantly better mpg than the engines of the 1960s (American gas guzzlers excluded, I mean European cars).

This is called "progress", or lack thereof.

As for the analog IC companies, there once were absolutely great ones with stellar products that had such outstanding performance parameters that everybody else in the industry wondered "how the hell do they do that ? Must be magic ! Or a pact with the devil perhaps ?". Company names like Burr-Brown, Harris Semiconductors, Linear Technology come into mind.

 All these companies had stellar analog ICs. And now all of them are gone. Swallowed by the bigger ones. And the irony is that after these new owners have "consolidated" the new acquisition, they find out that they can't make some of the "stellar" products anymore, and quietly remove them from the released product lists. Customers get a "last call" and then it's over. Sometimes the ineptitude to make those technical marvels is not spotted early enough, and no such warning call is given. Much to the anger, pain and suffering from the customers who need these parts in their products.

This is terrible. And embarrassing. But it's a story that had to be told for posterity. And no, I would never put my works on sites like Facebook along with all the mindless gossip and who has eaten at which restaurant and all the cat photos / cat movies. I do not want to sink that deep. My RQLT is far too valuable to waste it on those sites.

And, finally, to close the circle again to the Apple-1: I think that the yearning for these old, early microcomputers of the 1970s is largely based on the very problems I have pointed out above in this post, the complexity, and terrible nature of their "modern" equivalents. Sure, you can get a Raspberry Pi for $35 or so, and run Linux on it, and run an Apple-1 emulator on the Linux, but the (mostly hidden) complexity is such that you could not fathom all the hardware and software involved even if you would have a 1000 year lifespan. It's impossible to fully understand that $35 gadget and the 10's of millions of lines of code needed to make it run.

With the Apple-1, you can understand everything within the machine, down to the last bit, and down to the last transistor. Really ! Even the 6502 processor can be fully understood, the transistor level netlist is available on the web, it came out of the Visual6502 project. And this netlist is verified to work - I even use them in my own Apple-1 emulator.

Only a few people in the world have the original 6502 schematics, but anyone could use these netlists and turn them into human readable schematics with just a few weeks of work. There also are the 6502 schematics made by Beregnyei Balazs, which are great to understand how the 6502 works, but I have hand-checked them against the original schematics and there are a few little bugs. Which, however, does not diminish the tremendous value of  Beregnyei's work for the 6502 community. The original schematics are on two pages and they are quite messy and hard to read even for IC designers, while Beregnyei's work is very clear, nicely organized, and on one huge page. I recommend them for anyone wanting to venture into the inner workings of the 6502. For those not being IC designers already knowing two phase dynamic NMOS logic, get the book "Intoduction To VLSI Systems" authored by Mead / Conway, not to be confused with the crappy later books of the same title smeared together by mere epigones.  You can get the Mead / Conway book from 1979, used in good condition, for about $30, and it is a fine introduction to the exactly same circuit design style as used in the 6502. If you really want to know how these things really work, go for that book and then study Beregnyei's schematics !

I am surprised that in the United States someone has watched the films of Andrei Tarkovsky, his style is so unlike Hollywood, however, and Soviet too.  At home among strangers, a stranger among friends ...

If you haven't watched Solaris yet, if you liked Stalker, it will not leave you indifferent either.  

I have a question for you, Uncle Bernie: Why did the original Apple 1 use expensive Cornell Dubilier and BEL guitar capacitors?  Why did you use these and not the cheaper ones?

Macintosh_nik asks:

"Why did the original Apple 1 use expensive Cornell Dubilier and BEL guitar capacitors?  Why did you use these and not the cheaper ones?"

Uncle Bernie / Radio Eriwan answers:

We don't have a narration on this capacitor topic straight from the horses' mouth (only Woz himself could tell the truth), but we have sound circumstantial evidence about the background of the bypass capacitors in the Apple-1. This is my own opinion how the events may have happened:

1) The hand-wired prototype of the Apple-1 which Woz himself had built did work fine and did not crash its programs very often. Maybe it could run for days without any program crash. Circumstantial evidence for that: before Woz had the ACI, he used to type in his 4K BASIC interpreter by hand and then demonstrated the BASIC. This would have been all but impossible if the machine would have been unstable / crash prone.

2) Once Steve Jobs had talked Woz into "mass" production of the Apple-1, a PCB had to be made. According to the story, there were several prototypes of the PCB. Circumstantial evidence for this: historic photos exist of Apple-1 PCBs with no solder mask.

3) Now, a conjecture of mine: once they had put together the first PCB based Apple-1 prototype, they found out it was not running stable anymore, but was prone to frequent program crashes. They traced the problem down to ringing seen on the regulated power supply lines (which would have needed a fast enough oscilloscope to see). The deadline to deliver the first lot of Apple-1 to Byte Shop was approaching. They had invested all their money into this project and were desperate to find a quick remedy to prevent (or at least mitigate) the program crashes. The obvious choice any competent digital design engineer would pursue in this situation would be to look for "better" bypass capacitors which would dampen the ringing and the spikes on the power supply rails down enough to make the machine run better and more reliably. In this situation, costs for these capacitors would only be a secondary consideration. So when they did find capacitors which improved the situation, they did use them, even if they were somewhat more expensive.

4) We do not know for certain at which point in time the "better" capacitors have been used. We know that one of the first Apple-1 was given to Lisa Loop's institution which taught kids to program in BASIC. And this Apple-1 crashed all too often, frustrating the kids (and the teacher). When Lisa Loop complained, Woz himself tried to fix this Apple-1, and after that, so the story is told in Tom Owad's book, this Apple-1 worked somewhat better, but did still crash occasionally. Woz later gave her one of the first Apple-II that came off the production line, to remedy the problem once and for all.

From this story we can conclude, like Sherlock Holmes would have done, that the two Steves did not find a complete remedy to the problem. Had they found a complete remedy, Woz could have fixed Lisa Loop's Apple-1 to run rock solid and never crash by itself. This did not happen. The complete remedy is here (see post #4 in the following thread:)

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

... but it came 45 years to late. There is no doubt that if the two Steves would have revised the PCB to take these 24 additional capacitors, this Apple-1 based on such a hypothetical Rev.B PCB would have run solid as a rock and never would have crashed on perfectly good programs. There would not have been any nasty phonecalls to Apple (the company) by angry / frustrated / desperate Apple-1 owners distracting Woz from working on the development of the Apple-II and Disk-II. There would have been no  "buy-back" program for the Apple-1 with the target objective to destroy them (just to relieve Woz from answering all these nasty phone calls). The originals would not be so rare nowadays and would probably not fetch 1 Million US$ at auctions. But this is an alternate history timeline. It did not happen.

What the two Steves had created is known in the industry as a "phone call generator". Don't do that.  In one of the semiconductor companies I worked for this was proverbial and cast in stone: "Don't design a phone call generator, or you will get fired and lose your stock options." Bad karma. Don't design a phone call generator ! (I wonder what did happen to those engineers who designed the Theta II car engine --- maybe they had to flee the country other than losing their jobs. If they are still among the living. Don't design mass produced stuff that fails in the field. Very bad karma !).

5) The NTI capacitor mystery. We all know that there were two different "series runs" of the Apple-1: the so-called "non-NTI" or "pre-NTI" first "series", and then, the "NTI" 2nd series. Experts like Corey might want to chime in and correct me, but as far as I can see it, the "NTI" was the same PCB layout, but made by a different PCB company, which put their "NTI" logo on it. However, these "NTI" machines also came with totally different bypass capacitors: little boxy looking blue ones, instead the brown disk shaped ones of the previous, "non-NTI" machines. It is said that these new capacitors improved the problem with the program crashes even better, and it is also said that they were novel multilayer ceramic types. But again, my conjecture is that even these did not bring a complete remedy, otherwise, it would have been logical to only "buy back" the first series (non-NTI) which did not have these improved bypass capacitors, or to do a recall of just these limping non-NTI machines and fix them by removing the bypass capacitors and putting the better ones in. Again, this did not happen, they wanted all Apple-1 back to destroy them, so this circumstantial evidence supports the verdict that even the novel multilayer ceramic capacitors in the NTI-versions of the Apple-1 were no complete, 100% remedy to make the machine run rock-solid.

Here is some recent interesting find of mine. I was able to get some little, blue,  multilayer ceramic capacitors made in 1983 by muRata / Japan. At that time, the Japanese already had the best and most advanced process technology to make such capacitors. The ones I have were specifically designed to provide the best bypass capacitor performance (not all capacitors are good as bypass capacitors, i.e. the low-loss ones optimized for tank / RF filter circuits are no good as a bypass cap, because in bypass capacitors you want a specific kind of losses to dampen the parasitic LC tanks).

These muRata caps are made from a superior X7R ceramic material. My most recent Apple-1 clone (still waiting to get the big blue filter capacitors) was built with these bypass capacitors, and other than that, it uses a super-rare DS0025 instead of the more ubiquitious DS0026 clock driver. And guess what: so far, when running from laboratory power supplies, no program crashes and no memory errors did occur, despite there are none of my reliability mods in this machine yet.  However, it has some other mods: unlike in the original Apple-1, the TTLs all are LSTTL. There is also a difference in the PCB process, as I specified the thickest copper weight and the ENIG (gold over nickel) process, which may or may not have helped. And despite these encouraging results are preliminary (the build is not yet completed, as it still runs from lab power supplies) it seems - at least at the moment, the jury is still out - that by use of  optimized bypass capacitors, the Apple-1 reliability can be greatly improved, even without adding additional ones (which the DRAM manufacturers still would recommend in any case to keep their DRAMs happy ... the original Apple-1 only has one (!) bypass capacitor on the -5V rail near the DRAMs, and it would need 7 more,  if the later recommendations of DRAM manufacturers were followed ... probably not an oversight by Woz, because in their early datasheets, the DRAM manufacturers did not show the huge current spikes on the power supply pins of their DRAMs, and did not give proper advice how to address that problem. Later DRAM datasheets did, including oscillograms of the current spikes, and recommended PCB layouts with plenty of bypass capacitors for the +12V and -5V rails --- gotten too many phonecalls perhaps ?).

So all this circumstantial evidence and my recent discovery points into the direction that the two Steves were on the right track by experimenting with the bypass capacitors back in the day, to rescue their Apple-1 product, and to save their fledgling company from utter disaster (angry customers wanting their money back always is a disaster).

The part of Macintosh_nik's original question about why "Guitar" capacitors were used has not been answered yet. Alas, I don't have any specimen of those original capacitors desoldered from an orignal Apple-1 (yikes ! 1 Million US$ at risk !) and at the moment I don't have access anymore to a network analyzer to measure it to compare it with other and / or more modern capacitors. But I have a conjecture: the cheaper ceramic capacitors use Hi-K dielectric materials, so less layers are needed for a given capacitance (more layers = more costs). But these dielectric materials also happen to be microphonic. Which is undesirable in HiFi equipment and electric Guitars.  So my conjecture is that "Guitar" or "HiFi" ceramic capacitors are made out of less microphonic dielectric materials, which, as a side effect, might also cause a better performance as bypass capacitors, other than the higher price. But without taking specific samples of such capacitors to a network analyzer, I can't really say if this conjecture is correct.

Thank you for such a detailed answer!  Much is becoming clearer now.  I still have a lot of questions for you, but I remember your RQLT and try to look at first, maybe this forum already has answers.  I don't speak English, I translate through Google translator, it takes a lot of time.

There seems to be a couple more questions:1. What does testing your awesome chipsets for 2 weeks involve?  Are these some kind of tests (I know how to make a memory test, Hello World test, ASCII test) or is the board just turned on around the clock?2. You write about a group of A1 builders that you train, can I somehow become one of them?3. How to write a microcircuit or IC correctly?

Macintosh_nik asks:

"1. What does testing your awesome chipsets for 2 weeks involve?  Are these some kind of tests (I know how to make a memory test, Hello World test, ASCII test) or is the board just turned on around the clock?2. You write about a group of A1 builders that you train, can I somehow become one of them?3. How to write a microcircuit or IC correctly?"

Uncle Bernie answers:

Hi Macintosh_nik,

is google translate so good now that it can translate my convoluted English into Russian ? (Wow. The Russian language is one of the most difficult. Russia never really needed cryptography ... the language alone will hide all the secrets from non-Russians ...  just joking of course ;-) ... but this is based on my experience with manuals for Russian made equipment I used.

To answer your questions:

1) The burn-in of my "famous IC kits" actually takes not only two, but four weeks running 24/7. The PROMs with the WOZMON are replaced by special burn-in PROMs which exercise the terminal section and agressively test the DRAM.

The terminal section test is a sliding pattern over all possible characters, so each character will appear in each possible position on the screen, both horizontally and vertically. Any hiccup that disturbs the pattern persists for a while and is easily spotted. Any DRAM failure will switch the terminal section test off and just continue the DRAM test and give detailed error syndrome messages.

I check the burn-in rigs several times per day and especially each morning and each evening. If an IC dies or starts limping, it is pulled out and gets a red dot (the totally dead ones are thrown away, but if they start limping the effects seen can be funny, and I collect them to make a "chamber of horrors" showing all the symptoms. One of the rules is that if a new IC goes into a board, its "harvest date" is revised such that even the replacement gets a four week burn-in. Most of the "bad" ICs die in the first night, however. A few die in the first week. I keep statistics on that and hence, the four week time.So far none of my IC kits had a failure at the customer. Which is good.

(Of course the whole process is a bit more involved and includes pre-screening of some ICs, and margin tests over supply voltages and temperature, all tricks of the trade from many decades working as an analog IC designer.)

2) "Training" Apple-1 builders: I actually don't "train" them. I "coach" them, which means if they have any question or problem with their build, they can contact me via a special email account, and I will do my best to answer the question and solve the problem. If they are in the USA, I even may do that over the phone. The contact email comes in each IC kit. I don't offer this service to builders not using my IC kits. I have probed these waters by making a few exceptions to that rule, but it takes so much time to hunt down bad ICs that it's not desirable for me.

About half of this is not even about issues with the circuits, but where to buy which parts at good prices, or if substitute parts would work, or which tools to use. IC sockets and bypass capacitors always being a hot topic. So far each builder who contacted me and followed my advice ended up with a successful build that works rock-solid, no program crashes, no trouble. But not everybody buying my IC kits contacts me, though, but some did it only too late, after they had made costly mistakes. So far the small margin I factor into the price of these IC kits was able to pay for my prepaid cell phone cards and the gas for my car which is consumed by this service (I don't want to have internet at home, so I need to drive to a free wifi place.) As long as I don't incur losses, it's OK for me. I do learn a lot from the builders I coach. Every little pitfall is duly recorded to be avoidable the next time. A continuous improvement process. Progress is being made !

3) How to "write" a microcircuit or IC correctly... here, Google translate is lacking. I don't know what you meant with "write". But if you did mean how to "program" a PROM correctly, you need a good, professional grade programming system that back in the day was certified by the PROM manufacturer to correctly implement the prescribed programming algorithm. And then, this ancient programming system must have been rebuilt (weak electrolytics replaced, connectors cleaned, new ZIF sockets for the PROMs) and it must be newly calibrated (all voltages and pulse widths correct). For me it was quite a journey to get my Data I/O System 19 with Unipak up and running, and it was quite expensive (I'm still in the hole money wise with it, and running out of PROM blanks does not help with recovering the money invested ... because I can't sell individual PROM kits anymore due to blank PROMs getting scarce / too expensive).

Just if anyone wonders why I do all this, first, it keeps me busy during the pandemic, otherwise I would use most of my time to fly around to visit places I'm interested in. Using private airplanes, of course, as flying with airlines is too terrible since the TSA managed to turn a 3 hour inner-USA flight into a 6 hour long ordeal (and you can't carry a handgun anymore, you have to check it in even with more hassles, paperwork and loss of time, which rises the question if renting a limo complete with  bodyguards at the destination would provide the same security level with less hassles, but I'm too much of a cheapskate for wanting to pay for such a service). Second, I'm sitting on rare Apple-1 ICs worth more than $20000, and I want to get that money back. Third, I want to know if my system for building Apple-1 works, and if I really have succeeded in fixing its most prevalent problems. Note that in the past, everyone who had sold Apple-1 kits gave up, and this - I think - is because the success rate for builders using those untested ICs likely was far too bad, and so they actually sold a "phone call generator" and produced too many very unhappy and angry customers. Which is no viable business, of course. In short, while I'm trying to get my money for the ICs back, I'm also running an experiment to find out if complete Apple-1 kits could be commercially viable. Which would help to sell off my IC overstock quicker. But we are not there yet. There are still a lot of open questions and unknowns before I can offer complete kits.

I agree that Russian is much more complicated than English, it is more literary, English is more musical.  Your "intricate" English is more like Russian, all these beautiful phrases and complex sentences with many commas, etc.  Perhaps Google translates Russian into English well and vice versa because Google itself is half Russian (I'm talking about Sergey Brin).  When the pandemic ends and you will visit where you dreamed, perhaps you should visit Russia, I am sure you will like it here.  You will walk along the Neva embankment in St. Petersburg with a volume of Dostoevsky, you will go to the Tretyakov Gallery in Moscow.  If you find yourself in a bar and at any table you say that you are from America and do not know Russian and ask for help, then you will definitely not be able to pay for a drink that day.  You will always be treated.  Just don't try to drink Russian, it doesn't make any sense.

 Now about the case: on the first question, everything is clear, I remember about your RQLT and I think you are unlikely to share PROM firmware for testing.

On the second question, I just have the least difficulty getting parts, I have been buying on eBay for 10 years through one scheme and have no problems with delivery to Russia.  I am more interested in technical issues, since I am not a typical A1 builder, a year ago I had never soldered anything (only copper pipes - I am a plumber) and did not know how to use a multimeter.  But I am very persistent and learned everything quickly.  Now I'm waiting for the oscilloscope to be delivered to me, in my knowledge there are still "white spots" that I would like to eliminate with your help.

Google still did not cope with the third question, I meant how to correctly name a microcircuit or IC in English?

Dear macintosh_nik,

actually, I have released my burn-in software a while ago, it's in the PROMs for the A1 and A2 positions of all my recent IC kits, and it can be activated quite easily:

PDRM1004.JPG

The reason for doing this is twofold, first, I had to switch to using 512x4 PROMs because my 256x4 PROMs stock gets low, and I being stingy, and hating waste, I did not want to waste half of these PROMs by keeping it empty. Second, it's really useful to allow Apple-1 builders to test if their build really is good, without even needing a keyboard yet. All you need is two wires, one to do a CLR SCREEN, and one for the CPU RESET, to start the burn-in software. IMPORTANT: always remember that Apple-1 screen output does not work properly without a cleared screen to begin with.

As for hobby electronics, anyone with a brain can do that. I did learn digital electronics as a 12 year old boy from a TV show which was a digital electronics course aimed at adults. The book to accompany this TV show was available from the public library for free. I did learn TV analog electronics from another book which showed TV screen pictures with defects, and "if you see this, change the tube of the vertical oscillator" and such. Within 2 years, age 14, I was already able to understand and repair color TVs and reading their schematics (to trace the signals, not to understand in detail how these circuits really worked). Much later, after University, I worked for a big semiconductor company and guess what - I designed analog circuit blocks for TV ICs. The irony, however, was that all these last-generation ICs for CRT (Cathode Ray Tube) based TV chassis already did all the signal processing in digital. But on the chassis, the TV signals still were analog (for legacy reasons). So we designed programmable gain amplifiers, fast video ADC, fast video DAC, and PLL to clock everything, but all the signal processing on chip was done by a huge digital block  designed by the digital team. We (the analog team) only did countless variations of PGA, ADC, DAC and PLL, for year after year after year, on many processes. Which sometimes was very difficult because some were special digital-only CMOS processes with embedded DRAM (trench capacitors) and the MOSFETs available had terrible threshold voltages to do any analog work. So we had to "invent" new analog circuit design techniques to make it work. And we succeeded !

As for the name  of ICs / microcircuits here in the West, the "official" name used in the industry is "Integrated Circuit", or shortly, "IC", pronounced "I see". In the public (journalists !) they also use the word "microchip", and some companies have that word even in their name, but it is wrong, as the "chip" is only one part of an IC. Sure, all the transistors are on that silicon chip (called "die" from "dice" in the industry, not "chip"), but the chip alone is useless for customers and can't be sold as a product (some exceptions), but in general, the "die" must be "die bonded" on a leadframe, then it gets contacted by "bond wires", which connect the "pads" on the "die" with the leadframe, and then the whole thing needs to be encapsulated, mostly in plastic mold compound. An only then, after final test, it is a complete "IC" that can be sold. BTW, the bond wires, which often are gold, are the reason why recycling of bad / dead / junk ICs is so attractive. There is much more gold in a barrel of junk ICs than in the same barrel filled with the best gold ore from the world's best gold mine. To extract the gold, you need to grind down the ICs to dust, and then leach the dust with potassium cyanide solution (or mercury, which is even more toxic). This is why this recycling is done in China now. They don't care about toxicity and contamination of their environment. But they are glad to sell the gold reclaimed from the junk ICs.

... that is, if you build an analogy with the LM323k IC that are sold on aliexpress, can they be rejected by the manufacturer for one reason or another and sent to China for processing, where they are again sold as workers? And here's another question - in the Facebook group, many A1 builders put a fuse in front of the transformers.  For me, just connecting them correctly became a real quest, so I never got to the fuse.  What threatens the lack of a fuse and how much is it needed?  And if it is still needed, then where should it be installed and which one.  We have a voltage of 220v, I used Hammond 186D10 and 186D28 transformers. And yet, why in your posts usually at the end everything is repeated several times?

macintosh_nik asks (in post #20)

 "If you build an analogy with the LM323k IC that are sold on aliexpress, can they be rejected by the manufacturer for one reason or another and sent to China for processing, where they are again sold as workers?  ... What threatens the lack of a fuse and how much is it needed?  And if it is still needed, then where should it be installed and which one."

Uncle Bernie answers:

1) The rejects found to be bad by the manufacturer usually are sold to China as scrap for recycling of metals. But the Chinese have been seen to fish the good looking ones out and re-stamp them / re-sell them which they should not do. It's fraud to sell defective ICs as being good ones. And this scrap is the source from which many Chinese counterfeiters get their ICs. Unfortunately for us, most Apple-1 ICs must be old because they were made 40-50 years ago. Unless you find a reputable source in the USA (local surplus shop) where they were sitting in the shelves for this half of a century, the only places to source them is a) IC brokers and b) the Chinese. And you don't know if the IC brokers source "their" ICs from China. And you don't know if they are genuine, or counterfeit, or defective, or just too old. As long as you only buy a few and don't risk much money  you can do that. But when I bought the many thousands of ICs for 100 Apple-1 I could not risk to use such questionable sources.

2) Transformer fuses. The original Apple-1 manual shows a fuse in the primary circuit of the transformers.  The idea is if there is too much current on the secondary side (due to a short or a bad component, or an IC plugged in backwards) the fuse will blow before things get so hot that fire could break out. If you have no fuses and short a transformer secondary then the transformer will get hot and hotter until it makes smoke signals. No good ! Some transformers do have thermal protection built in but if they overheat the thermal fuse opens to prevent fire and the transformer is trashed.

The problem is that one fuse on the primary, although better than no fuse at all, does not really protect against the fire hazard, especially when two or more transformers are fed with the one fuse. Modern design practice is to have one fuse per secondary winding on each secondary winding. For instance, if you buy TRIAD world series transformers types VPP16-1900, VPP20-1500 or VPP28-1060, the datasheet says that "Fuse must be used on secondary as conditions of acceptability for UL Class2/3 operation." You can use these transformers with no fuse, but then you are not meeting the UL standard (and can't legally sell the power supply as it would be considered to be unsafe). I have attached one of these datasheets for reference: 

https://www.applefritter.com/files/2021/03/10/VPP28-1060.pdf

The other ones you can download at any distributor who sells the transformers (such as Mouser).

For the Apple-1, with its old ICs, that could die and get hot at any time,  having fuses on the secondary windings may or may not help against fire. They would for sure protect the transformers, if the fuses are properly selected. But the bad IC still may smoke without blowing the fuse. You need "slow" fuses of a current rating high enough, such that they blow before the transformer smokes, but not blow due to the charge current of the big electrolytic filter capacitors. All this is a bit tricky. The voltage regulators in the Apple-1 also self-protect and limit their output current and their temperature to safe levels, so this protection may kick in before the fuse blows, so the fuse will never blow despite one IC is already making smoke signals. I know all this from hands-on experience with my own Apple-1 builds (ouch, hot !). So I know how tricky the fuse issue is. This is why I avoided the topic so far. But you talked me into making a comment on this ...

If you try to find a fuse for the seconday side, the voltage rating is irrelevant. Fuses blow due to current alone. So a 230V/400mA "slow" fuse will blow somewhere between 600mA (after a longer time) and 1.1A (after 1-10 seconds) even if it is used in a 8V circuit. The blow characteristics can be found in the datasheet of the fuse (yes even those humble components have datasheets). But using a 12V automotive fuse on a 230V circuit would be a death trap (lack of insulation against the 230V). Don't do that.

3)  Why my posts often repeat things at the end .... first, to sum up what the most important takeaway of the (long) post is. The long part goes into the details to prove my point. I could be lazy and not explain "why", but then people would say I'm just making it up and there is no proof. Second, a lot of my posts are glued together with cut-and-paste and I don't have the time to remove superflous information for every special case. Sorry for that, but my time is precious and I try to be as efficient with my time as possible - even if that means I copy/paste a lot off stuff that is not as brief as it would be possible if somebody would pay me for my time invested. Third, sometimes my mind wanders off into other topics which somehow got triggered by the actual topic. But this is inevitable with old people like me. Once they start telling stories, they never stop. As long as they are not babbling idiots due to Alzheimer or other forms of age related Dementia, like our beloved President seems to be, this is fun and everybody likes listening to the stories. Much better and less boring than watching TV !

It seems Uncle Bernie and I have annexed this post, I hope this will not cause inconvenience to the author, but rather the opposite, he will be in the top for a long time.

 Thanks Uncle Bernie for the previous answers.  I very much regret that I did not know about this forum before, in a group on Facebook people are extremely reluctant to share information about A1, only beautiful photos and "blurry phrases".  Like it's a Manhattan project, no less. There is another question for you, Uncle Bernie, on the ACI card there are 4 resistors 10 kOhm 1% (RN55D1002F), I seem to have missed that 1% is needed, put the usual 5%.  The board works, I downloaded the basic and the Mastermind game through it, but the LED never turned on.  This may be somehow connected with this, or this LED lives its own life and you shouldn't pay attention to it at all (as I was told on the Russian forum).  In general, I would like more information on this card, how to set the sound level correctly when playing audio files, whether the download quality deteriorates as the entire system warms up, etc.  In general, I am a little disappointed with the use of A1, there is very little software, the one that is somehow does not capture (you need to graduate from Ivy League University to understand how to play this Mastermind - but I figured it out!).  So far, I realized for myself that building A1 is much more interesting than using it.

The LED will only turn on if you are using a real cassette deck.  Most modern digial audio sources (i.e. iPod, Computer) can't drive the signal enough to turn on the LED.  There is a reason why many of us who grew up in the 60's, 70's and 80's are all partialy deaf from listening to music on headphones.

Thanks for the answer, otherwise my hands were already scratching to remove this LED and put another one on.  I really liked your first YouTube video, do you plan to continue?

I do,  I have more stuff I'll be working on soon.  

macintosh_nik wrote (post #22):

1) I very much regret that I did not know about this forum before, in a group on Facebook people are extremely reluctant to share information about A1, only beautiful photos and "blurry phrases".  Like it's a Manhattan project, no less.

2) There is another question for you, Uncle Bernie, on the ACI card there are 4 resistors 10 kOhm 1% (RN55D1002F), I seem to have missed that 1% is needed, put the usual 5%.  The board works, I downloaded the basic and the Mastermind game through it, but the LED never turned on.  This may be somehow connected with this, or this LED lives its own life and you shouldn't pay attention to it at all (as I was told on the Russian forum).  

3) In general, I would like more information on this card, how to set the sound level correctly when playing audio files, whether the download quality deteriorates as the entire system warms up, etc.

4) In general, I am a little disappointed with the use of A1, there is very little software, ...So far, I realized for myself that building A1 is much more interesting than using it.

Uncle Bernie answers (this time in reverse order for a good reason, and note that I have added the corresponding numbers above):

4) Agree that there is not much software. Microchess is fine. The rest is mostly "look daddy my first BASIC program runs" level. Not even amateurs. Bloody beginners. Back in the day, there were magazines which printed pages upon pages of such crappy BASIC programs for the buyer of the magazine to type in. Very rarely there was a "gem" to be found. I always wondered what the point of these magazines was, wasting so much paper for that. Only later I found out that all the "articles" are only filler material, what these magazines really were about were the advertisements paid by small and large businesses. This is where these magazines made the money. Now they all are gone. Most of them withered away in the late 1980s.

As for the Apple-1, it's up to us users to write better and greater software for it. Due to its limitations, it is actually fun to program because you literally have to do (almost) everything by yourself, so you can learn a lot, which is the point. There are no sophisticated and complicated calls to 16 KByte or 32 KByte of operating system in ROM. I always hated to work with such complex firmware because it always was poorly documented and not open source. So more often than not, you were losing time poking in the dark. With the Apple-1, there is no such problem. The firmware written by Woz is open source and the few subroutines it offers are easy to use.

The problem with writing better software for the Apple-1, as I see it, is that there is no good toolchain yet to develop the A1 software on a PC, emulate it there, and then, once debugged, download it into the real Apple-1. Once such a toolchain exists, there (hopefully) will be more good programs for the Apple-1.

Another issue is using a cassette recorder as a mass storage device. This sucked even back in the day. I remember these first microcomputers in 1977/78 (I'm old enough). The KIM-1, the PET2001, the TRS-80, all only had cassette. It sucked. We hated it. But there was no other way. Then the first 5.25" floppy disk drives appeared and the whole situation changed dramatically. This is where the microcomputer / home computer "revolution" really took off. We ditched the cassette recorders and never looked back. Alas, I found it to be very difficult to adapt Woz' original floppy disk system design concept of the Disk-II to the Apple-1. In its original form it critically depends on exact cycle-by-cycle timing within the RWTS code, and the non-transparent DRAM refresh of the Apple-1 whacks that and causes it to fail.  

3) ACI sound level: if you have a cassette recorder which actually can light up the LED at high volume settings, in my experience, this is already too much and the load operations will fail. With one of my recorders I have found that if I turn the volume up until the LED dimly lights, and then turn it slowly down until the LED just stays dark, this is the right setting. But this may or may not apply to any other cassette recorder.

If you have an oscilloscope, you can attach it to the output of the LM311 (pin #7), and this can be done at either the 47K or 3K resistor connected to this pin. Then look at the header tone of the recording which is long and homogenous enough to adjust the volume. Adjust the volume from low to high slowly. At a very low volume, there will be no movement of the oscilloscope trace. As you increase the volume slowly, you will be able to see the signal, but initially it will "break apart" at places, where it is still too weak. Increase the volume further until the "breaking apart" effect disappears and the signal looks good. You will note some expansion and contraction of the signal over the time axis, which is the "wow" effect of the cassette recorder itself, that is, small periodic speed changes. These are caused by the "ping-pong" type speed control circuit of these cheap recorders, and by mechanical slop in the drive belts. If the cassette recorder has aged drive belts which are too loose, the "wow" effect is more pronounced and you should change the belts. I found it is very hard to find good replacement belts. Those I bought off Ebay looked good but under magnification they had micro cracks started at their surface. They will not last long.

If you find my recommendation to use an oscilloscope to adjust the volume exaggerated, well, this is how I got my own ACI to work. By just fiddling with the volume control blindly, I never found the sweet spot where I could load the 4K BASIC without errors. The oscilloscope solved the problem in minutes. Then I built an improved version of the ACI, which works better. But I'm still not satisfied enough with it to publish instructions for this mod. It's actually the same circuit based on the 741 opamp as used in the Apple-II, and is more robust against the volume setting,  but there still are some issues I had no time to explore and weed out yet. I do not want to publish half-baked mods.

2) If you do a mathematical analysis of the ACI circuit around the LM311, you will find that 1% resistors are not needed, and 5% resistors will do fine. You have confirmed my theoretical find in practice so there can be no dispute ;-)

What I think has happened is that Woz was afraid to specify common carbon composition resistors there, which had (or still have, if you can find them) 10% or 20% tolerance levels. At low volume settings, a particular bad and unlucky combination of 20% tolerance resistors may not work. At higher volume settings, even that would work, but at these higher volumes, the noise, hiss and buzz from the cheap recorders could interfere with proper program loads.

As for the "Russian" method, just ignoring the red light, in this case it is right. The LED circuit of the ACI is terrible and ineffective. I think they should have hooked it to the output of the LM311, with a small coupling capacitor, so if the output does not change, no light, and when the output changes, the LED lights up. Then the LED as a volume setting helper would probably work. I didn't try this out yet. But in the current form, it's useless.

1) The Farcebook problem: I never use such "social media websites", for several good reasons:

a) it's a time thief with very little value to be found and extracted,

b) they install spyware on your computer which tracks you, secretly turns on the microphone / camera on your computer/smartphone to listen to all of your conversations and to observe you, and

c) based on this data collection (mostly done by AI bots) they build a personality profile of you which knows more about you than you would want, and sell that data to third parties (if it's free, YOU are the "product" they sell out).

d) Furthermore, they claim the copyrights to everything you post on their site. They can sell every one of these "cat movies" or, in our case, A1 photos, to make money (if they find a buyer). And if you, as the originator, then put the same photo you made in a book you write, they can sue you for violation of "their" copyright. 

They are just disgusting Orwellian eavesdroppers, peeping Toms, and snitches of information to the Gestapo and to rip-off artists. They also are IP / copyright thieves. There is a reason why the Sugermountain man reportedly needs 20+ armed bodyguards. It's obviously not that many because he is afraid of getting kidnapped and held for ransom, as any rich man fearing that needs no more than 2-3 bodyguards against that threat, but needing 20+ bodyguards suggests that he fears to get shot by a sniper, contract killer, or worse. An angry FB "customer" perhaps, who got spied upon, his data stolen and sold, and then may have sustained damages from that, such as a wrecked private life or wrecked business, he is mad like hell, but he can't get justice in a court of law because FB user agreement everybody klicks away (thereby "signing" and agreeing to it) was written by very good lawyers and is watertight, so he could get justice only by putting a bullet into the man who is responsible for his misery, which he would think is Mr. Sugarmountain, but actually, it's himself, and in all fairness, this disgruntled FB "customer" should better shoot himself to get justice, because it was his own utter stupidity of enabling them to spy on him and to steal and sell his data and personal profile. A smart man stays away from bad people, spies and snitches and thiefs (and criminals in general) and so needs no bullets to mete out "street justice".

More and more people wisen up against the inherent evil of these harmless looking "social media" sites which are all but harmless, and refuse to install their apps, and this is why now almost all smartphones and tablets and notebooks come with these spying / snitching apps preinstalled. Which are hard to remove even for experts. All the "turn off microphone" sliders etc. appear to work but don't turn anything off.

This, my friend, is not the "Manhattan Project", it's called a "Panopticon", and these people are not free anymore, they are in a prison of their own choosing, under surveillance by the prison guards 24/7, and against their will they are turned into whores who are pimped out to third parties so Mr. Sugarmountain can get rich and richer.

Yet another problem is with the poor quality and information content, and I think this is rooted in the fact that people posting there a) don't know much about the subject, or, b) don't want to tell their "secrets" because they use FB only as a front for a sales pitch ... "Want to know more, visit my shop, and pay for the information you seek". I think that in most cases, it's a) which applies. These are hobbyists dabbling around and barely got their Apple-1 working without understanding it, and now they brag how smart they are, but actually they know nothing and can't tell nothing of value on the subject matter. And this they seek to hide behind the shroud of "secrecy": look, I'm much smarter than you but it's my secret. Stupid impostors.

I think this explains it all. The world would be a better place without FB or Instacrap and all these other evil sites who are just honeytraps to lure people in and then spy on them and sell their information out to third parties, and to give a platform to Narcissists and idiots of all sorts.For me, it is painful to watch how even my own extended family members of the younger generations are so addicted to these sites that their face is glued to the screen of their smartphone all the time, even during banquets, they put the smartphone on the table in front of them and always look into it, being so much sucked into whatever crap there is to be seen on that screen, that they are unable to participate in any conversation or discussion, much like some severe case of Autism or brain damage, mental cripples or such, as if straight out of a nursing home. And worse, they continue to do that even when at their workplace. They are useless. I disinherited all of them in my Last Will.      

Geez UncleBernie ...

How is anyone supposed to respond to that?

Macnoyd wrote:

" How is anyone supposed to respond to that?"

Uncle Bernie answers:

No response necessary, dear Macnoyd. I don't want to start a discussion about the perils of social media websites and the internet in general.

It is perfecly fine with me if people take my technical advice and then, perhaps also heed my warnings about FB and all, which I intentionally placed last, as it is not exactly Apple-1 related.

In my opinion (which is quite strong) this type of internet tech may have been invented by well-meaning people having noble motives. Such as the inventors of radio and TV once must have had. But all of this turned into a monstrosity, evil to the core, serving evil purposes, and turning all too many people into a sort of drug addicts, impairing their functioning in normal life. This is very concerning. And the damages caused by it are astronomical.

As to the eavesdropping, you might think I'm paranoid. This is not the case. You just need to talk to any expert who works to keep the IP and trade secrets of corporations and/or the classified material of the Government safe from leaking. For instance, you cannot have any smartphone in any conference room where critical business is discussed. It really has gotten so bad.  Much of this is based on the voice recognition technology embedded in any smartphone OS. Or in little helpers like "Alexa". And the business behind harvesting all of this data involves billions of US$ of revenue per year.

When I had read "1984" as a kid I thought that the totalitarian government must have furnished the "Televisor" in each room for free. I was mistaken. Neither the "Goebbels Snout = Volksempfaenger" of the Nazis nor the modern surveillance devices were furnished by the State. People buy them with their own money, out of their free will, thereby  enslaving themselves and losing their privacy and freedom.

Do you have an opaque tape over your notebook camera as recommended by Edward Snowden ? Do you ?

As for the internet addiction, it's terrible. I'm a victim myself. Forget "self discipline" and "self control". These are illusions. For me, once I visit youtube, the whole day is gone. There are so many valuable and well made videos there which can answer everything I ever wanted to know, and also everything I don't want to know. Once I watched one video about a topic, another dozen pop up which are even better !

My solution was to take a sledgehammer to my internet pod (after cancelling the contract) and now I have to drive to a free wifi place to get internet access. The time I can spend is limited by the notebook battery. A smart arrangement with just enough built-in friction and resistance to keep the bad habit at bay. I full understand why my nieces and nephews are so addicted to FB and Instacrap.  Maybe it's genetic. But they seem to be unable to see the damage and time theft and then engineer a solution.

This said, without radio / TV / FB / Instacrap / youtube, the world would be a better place. All of this is very, very dangerous mind control technology which ensnares people and steals much of their life time. Add the eavesdropping / profiling to that, and you have a dystopia. This is the world we live in.