Sources for the 2519

I was curious as to current supplies so I did a few searches. I did come up with one hit.

www.sayal.com which indicates 109 in stock, but you have to request a price. I believe they are located in Canada.

Years ago, before I realized the N and the B were the same part in different packages, I had a hard time finding the "B" part and spent a fortune on a tube of them. I was quoted over $100 each from a couple of places, but did find some at somewhat more reasonable price. Per piece, that is the the most expensive lot of parts, I've ever bought. I hope we aren't returning to those days.

regards,
Mike W.

Totally off topic comment:

Not being *that into* the Apple 1 I was aware that it used shift registers for bulk character storage, but I'd never really thought about how it (or the older TV Typewriters that its video section was based on) managed to pull the same *line* of characters X many times for actually refreshing the screen. This thread made me look up what the part actually did and it makes perfect sense now. (Although I guess I'm still getting a little bit of a headache imagining the circuitry that syncs up the data transfer from the dynamic 2504 bulk storage to the 6519 line memory. One of these days I need to break out the TV Typewriter Cookbook again and actually *READ* it carefully.)

A few years ago, I asked Woz about the TVT influence on the Apple 1 video system. His response can be found at the bottom of my Mimeo page.

http://www.willegal.net/appleii/apple1.htm

An interview printed in Byte Magazine in 1984 doesn't exactly contradict the more recent email I received, but definately puts things in better context.

http://apple2history.org/museum/articles/byte8412/

regards.
Mike W.

Perhaps it was incorrect to say that the Apple I was "based on" the TV Typewriter design, certainly; what I really meant by that was they use similar design principles... and, well, that second reference does imply that Woz may have at least read Don's article before rolling his own condensed/"clever" version.

I'm sort of curious how the price of the Apple I's video system compared to, say, the Processor Tech VDM-1 video card for S-100 systems, an early example of a memory-mapped video system using SRAM instead of shift register memory. (It looks like the VDM-1 was first advertised for sale in November 1975 for $225 assembled. The PolyMorphic Systems VTI is another roughly contemporary board using SRAM.) In retrospect it sort of seems like the Apple I was a slightly backwards-looking design for its 1976 introduction date* but perhaps the components were so inexpensive compared to the SRAMs it was a worthwhile tradeoff?

(*Edit: Granted, it appears that most of the computer was designed in 1975, but... I've skimmed several articles, and one thing I haven't found is the date when Woz first started showing off "the thing that became the commercial Apple I" at the Homebrew meetings. The 6502 didn't officially go on sale until September 1975; apparently MOS was shipping semi-working samples to a few companies before then. Does anyone have a reference saying precisely when the design was set in stone? Wikipedia has an "April 11, 1976" release date that conflicts with a statement that semi-conflicts with "It was demonstrated in July 1976 at the Homebrew Computer Club in Palo Alto, California." That demonstration was the "final" one, not Woz's prototype, I assume?)

Someone bought sayal.com out - I just inquired about the remaining stock. Sold Out.

Date codes on chips from surviving units from the first batch indicate a probable build date of around April or May '76. There is an image of a prototype PCB, with some minor differences from production units, so the design could only have been set in stone with the first production run.

Based on 2504 versus 2102 pricing in a 1978 Kilobaud magazine, cost of the video was probably similar to a SRAM based approach. The thing is that WOZ already had the terminal working, so he didn't need to design or even build something new, he just needed to interface his existing terminal to a microprocessor with memory.

When you consider that the Cromemco Dazzler for S-100 systems was released well before the Apple 1, and the SOL-20 was first publicised in mid-76, you will understand why the guys at Apple were in a such a rush to get the Apple II into production.

regards,
Mike W.

One thing to keep in mind was the parts closet at HP. Woz may not have initially paid anything for the 2504v chips when he designed both his tv terminal and the Apple-1.

Also the Sol-20 prototype was publicly shown in a July 1976 magazine, months after the Apple-1 was designed and began sales. The Sol-20 didn't ship till the end of 1976 and in limited quantities only. The dazzler couldn't really be used as a terminal so I think for early 1976, the Apple-1 was state of the art, the Sol-20 wound up costing a lot more than the Apple-1 when it finally arrived, though it was complete with keyboard and case. By mid-1977 everything changed. I'd also say the proliferation of DRAM in 1977 caused a major drop in 2102 prices by 1978 so I'm not sure your price chart is good for comparison.

Cheers,
Corey

By mid-1977 everything changed. I'd also say the proliferation of DRAM in 1977 caused a major drop in 2102 prices by 1978 so I'm not sure your price chart is good for comparison.

Interestingly, looking in a January 1976 issue of Byte Magazine it looks like 2504s cost quite a lot more than 2102 SRAMs; $9.00 vs. $2.95 in this particular listing. The same issue has several ads for VDM-1 video boards at $169 for the kit, which used 8 of them, but of course the thing to remember about an S-100 board is it included a lot of parts that would be extraneous when building a single-board system with video and CPU combined. So unless Woz's *production* 2504's fell off a truck, er, Apple cart, it doesn't really look like cost drove the decision.

Interestingly, there's a review of the SWTP CT-1024 terminal, which was based on the second-generation "TVT II" design by Don Lancaster. It uses 2102 SRAMs and its complete design was published in a series of articles in Radio Electronics magazine starting in February 1975. (I quickly skimmed the October '75 issue of Byte and it has an ad for this same kit.) Wikipedia(I know, not necessarily a reliable source, but I do think I recall reading a statement like this in the TV Typewriter Cookbook) notes that supposedly one of the stated goals of the TVT-II/CT-1024 was to elminate reliance on ICs used in the 1973 TVT-I design that were going out of production. Considering the paucity of ads for 2504/2519s in these late '75-early '76 Byte magazines one has to wonder if the shift registers Woz used in the Apple I were on that list. *If* that's the case, well, it's sort of hard to argue the Apple I was "state of the art".

(Granted "State of the Art" is a really tricky concept in a field that was evolving as fast as personal computers were in the 1974-1977 timeframe; it's probably not quite fair to say it was "obsolete" but by April 1976 it looks like 2504s weren't something anyone used for new designs anymore. Maybe the two Steves *did* get a huge pile of them for pennies from a reseller that didn't want to get stuck with surplus nobody wanted?)

We do know "cheap" parts were what drove the design. That is why it used the 6502, dynamic ram and stuff. I think you are onto something about the cost of the 2519. They must have been free during the design phase (Woz had already used them in his own TVT design) and Jobs did make some subs for parts that could have been for cost (certain value resistors), so we know he could cut a deal on parts though I wouldn't put it past him to cut a deal for surplus 2519 and then to tell Woz, not to use them in the Apple II since they only have a limited supply. I guess we will never know.

However in early 1976, I can't think of another pre-assembled single board machine which had TV video and a keyboard interface. So that makes it state of the art.

Cheers,
Corey

However in early 1976, I can't think of another pre-assembled single board machine which had TV video and a keyboard interface. So that makes it state of the art.

Perhaps, although I suppose that does point out that "state of the art" can mean different things; you can be state of the art "conceptually" without being so technically.

(And, I dunno, it's sort of splitting hairs to argue the Apple I was "fully assembled, considering it was just a circuit board that still required the buyer to interface a keyboard and supply a case... and technically the cassette interface was a seperate board. It fit into an interesting in-between space between a "computer trainer"-calibre machine like a KIM-1 and something like PolyMorphic System's "Micro Altair" which included a graphics-capable video interface and a similar PROM-based OS for around $800 assembled. It was a progressive design, other than the video system, but I'm pretty sure "revolutionary" wasn't the word for it. The Apple II almost deserves that word, but the 1...)

Also, re: "early 1976", it doesn't look like any Apple I's were actually sold until July, (do you have a reference contrary to that?) and while it generated excitement when people saw such a small "thing" generating video by itself the fact that they were only able to sell around 200 of them over the one-year life of the product sort of implies it wasn't quite what the market was looking for at the time.

It's somewhat interesting to compare the Apple I with the TRS-80 Model I. The TRS-80 was prototyped in January-February 1977, and in a lot of ways it parallels the Apple I; it's a single-board computer with DRAM for main memory, integrated terminal functions, and designed to be as absolutely cheap as possible to produce. The major points at which they differ (besides the CPU) are memory mapped SRAM for video instead of a built-in "terminal"; the Model I actually goes a step further than Apple did for cheapness and used a simple matrix keyboard instead of one with an ASCII encoder. (Also, coincidentally no doubt, their motherboards are almost the same size and shape.) The one thing the TRS-80 could do with its memory mapped video that the Apple I couldn't with its built-in "terminal" was be surprisingly effective at displaying (crude, but usable) graphics and *fast* action video games. That's something that's well-neigh impossible to do well with shift register memory. One sort of has to wonder if the Apple-1 might have been more successful if it'd had the graphics performance (even character graphics) of the TRS-80 and PET; it would have been technically feasible to include it, and if Woz has used 2102's the Apple I basically would have been a Commodore PET a year early.

(Another interesting aspect of the "TRS-80 generation", including the Commodore PET, was their memory mapped video systems eleminated the need for hardware cursors or scrolling; it was all done in software. The TRS-80's video system was essentially a clone of the video system Polymorphic used in the "Micro Altair", but that board included niceties like a hardware blinking cursor which were ruthlessly dispensed with to make the system cheaper. Honestly I find it sort of funny that when Woz rips out a bunch of chips and replaces them with some software he's praised as a genius but when a bean counter at Tandy or Commodore forced the engineers to do the same it's generally... less fondly remembered.)

Just one comment to this.
I try not to compare the Apple-1 which really was designed over the course of 1975 (not 76) with machines from 1977, many which were rushed to market.

Woz was showing the Apple-1 off a good year and a half before the TRS80 model-1 prototype was shown. People in the Home Brew Computer Club had Apple-1s even before April 1st 1976 when Apple was founded, they were hand wired versions. The PCB was an idea to speed up the builds. The byte-shop sale was an unexpected result of Steve Job's salesmanship that forced Ron Wayne to get nervous and drop out of the partnership.

By the time the TRS-80 model-1 was being shown, Woz was building a color computer with memory mapped graphics called the Apple II. Woz started the Apple II as an experiment in color graphics. He quickly turned it into an entire computer. I have had emails with him about why he didn't do this or that with the Apple-1. He said it took him a year to get the Apple-1 design worked out, in 3-4 months he had the Apple II, hardware wise, mostly worked out so it didn't make sense to retro-fit anything to the Apple-1. It made more sense to work on the software in the Apple II to maximize what it's potential was. Remember a big thing the Apple II did was change the paradigm which was hardware/software as a "package" to deliver functionality. Using a combination of hardware and software the Apple II in 1977 jumped over everything else in it's price range. Yes it was expensive compared to a TRS80, but it delivered much more, and was much much less than a comparable S100 system which had the same features.

Cheers,
Corey

I try not to compare the Apple-1 which really was designed over the course of 1975 (not 76) with machines from 1977, many which were rushed to market.

Just offhandedly I've been digging a little more into this subject and to be honest I'm finding it sort of difficult to find a new terminal design that used shift registers instead of SRAM newer than 1973 or so. (The 1974 DEC VT-50 used RAM, as did the early '75 ADM-3, and I've already mentioned that the "current" TV Typewriter design used SRAM by January '75.) Undoubtedly there were still people building older designs using shift registers by mid-late '75 (like Woz was), and, again, perhaps the Two Steves were able to get one heck of a deal on obsolete parts to use in their first commercial venture, but... yeah. Strictly speaking there's nothing "wrong" with Woz's design given the Apple I was priced substantially under an S-100 system with a more "modern" terminal board, I just find it sort of an interesting anachronism that the board combines a technology that was clearly on its way out with a brand-new CPU and DRAM for main storage.

Woz was showing the Apple-1 off a good year and a half before the TRS80 model-1 prototype was shown.

The reason I brought up the TRS-80 was simply as an example of a "thrifty" design with a video system based on what *was* "state of the art" in late '75-early '76. (IE, terminal boards like the VDM-1/PolyMorphic/etc. Also, timeline-wise it's a somewhat dirty pool comparison; the TRS-80 was a semi-secret corporate project so of course it wasn't shown publicly in its larval stage. Realistically speaking there's probably only a year between their finalized designs, perhaps less; Don French was homebrewing the design which evolved into the TRS-80 in mid '76, shortly after the Z-80 came out.) But, as you say here:

He said it took him a year to get the Apple-1 design worked out, in 3-4 months he had the Apple II, hardware wise, mostly worked out so it didn't make sense to retro-fit anything to the Apple-1.

Perhaps it's a little unreasonable to have expected him to keep up with the advances in the field that happened in 1975; things were moving very quickly and it *was* his first computer. (Undoubtedly he needed the experience to be able to work out the Apple II in "3-4 months".) And I'm not debating whether or not the Apple II was "revolutionary"; it was the first machine that used the shared-memory DRAM architecture used on pretty much every 6502-based home computer built for the next decade. (Closest comparison would probably be something Atari was playing with in their labs but since the 400/800 didn't debut until '79 so the Apple II totally wins.) Again, I was just musing about the seeming anachronism of the *Apple-1's* video system, and Woz's explanation is perfectly reasonable: building the Apple 1 was an exercise in learning how to build the *computer* part and he mostly just reused his old terminal design unchanged. (Other that deriving some DRAM refresh timings off it.) That does mean, according to his own words, that the video system wasn't "state of the art". Maybe the whole package was by some measures, but not the video system.

Remember a big thing the Apple II did was change the paradigm which was hardware/software as a "package" to deliver functionality. Using a combination of hardware and software the Apple II in 1977 jumped over everything else in it's price range. Yes it was expensive compared to a TRS80, but it delivered much more, and was much much less than a comparable S100 system which had the same features.

And... how did the TRS-80 *not* do this? Yes, it didn't have color graphics. It did however pack nearly the same functionality as a $1,500 SOL-20 into a box that could be had for $399 minus video display and cassette recorder and a lot of that price reduction was due to the clever use of software to replace hardware that you didn't strictly need in an "integrated" computer vs. a collection of semi-independent boards on a bus. I know there's this grand narrative that tries to make the Two Steves out as unique geniuses operating in a vast wilderness of ineptitude and wastefulness, but... frankly it's a little insulting sometimes how it minimizes the contributions made by the other pioneers in the field. Bleah.

(I do imagine the brass at Radio Shack would have loved to have had Woz working for them. His lobotomized disk-drive idea is exactly the sort of penny-pinching Tandy was after...)

Anyway, sorry for the thread hijacking. The thing I will really give the Apple 1 credit for is being an early demonstration of the idea that perhaps the "Box of Slots" paradigm wasn't necessary for building an "interactive" machine. Looking back the hobby was pretty fixated on the S-100 bus for "real computers", with single-board machines generally consigned to being "trainers". But building an S-100 system necessitated a lot of redundancy and the bus wasn't particularly well designed, among other things its limitations were in part responsible for the bad reputation DRAM had in the homebrew computer field. (Early DRAM boards for the Altair predating the Apple I tried to use signals over the S-100 bus for refresh and the idea never really worked. It's sort of amazing that DMA graphics boards like the Dazzler were able to be made to work *at all*.) The Apple 1 showcased the "economy of scale" that could be achieved by cramming CPU, RAM, and video onto the same board and letting some of the circuitry do double-duty. All I'm saying is if there's criticism to be has for it it's just the fact that it wasn't *quite* there in terms of being a killer product, and I'm pretty sure some of the blame *has* to fall on its 15-seconds-to-fill-the-screen video system. (An SRAM-based Apple I would have been able to play a killer game of Pong, particularly if they'd incorporated a semigraphics mode...)

Not meaning to diminish the Apple 1 in any way (after all I just completed assembling and bringing up one of Mike'e W's excellent reproduction PCBs and most definitely have pride in ownership), if it was not for Apple's ultimate incredible success and dramatic story, I believe that the Apple I would be relegated to a place in history (and value) comparable to the Altair, Scelbi, KIM-1, TRS-80 etc, each of which brought their certain key innovations to the table.

(An SRAM-based Apple I would have been able to play a killer game of Pong, particularly if they'd incorporated a semigraphics mode...)

After studying the Apple I circuits, one thing that seems apparent is that with a very minimal amount of additional logic, there could be a special mode with a VBI interrupt and a software routine that could be set up to create an instant screen update by copying a defined DRAM area to the video shift registers on demand. The principle is partly analogous to the way the clear screen functions. This could then eliminate the screen update time and create simple character based animations. Too bad it was not done.

(Note: Deleted the duplicate post. Hope that's all right.)

The Apple I deserves its place in history. It was a quantum leap in 1975-76 in computer innovation.
Woz's genius comes from taking concepts that existed previously in larger computer systems and compacting them all into a "small package" of 61 chips.

Yes, this is where the Apple I was innovative: the packaging. ("Quantum Leap" is probably pushing it.) The closest forerunner to the Apple 1 in this respect was probably the "Sphere 1" computer: the $860 (kit) price included a single board computer (6800 CPU) with 4K of dynamic RAM, a machine language monitor ROM, a keyboard, and a 32x16 video display. (Using SRAM.) Like the Apple 1 the cassette interface was an add-on module; unlike the Apple 1 the video board was also a seperate module. (The raw CPU board included 16 general-purpose I/O lines that could be used to bit-bang a teletype. Which, likewise, was an innovative feature at the time.)

Of course, it gets a little tricky putting the Sphere 1 on a timeline because while it was being advertised as early as July 1975 the company ended up being rather late in delivering working product. They were taking orders from before the first issue of BYTE, but it doesn't appear that many made it into the wild much earlier than October '75. But nonetheless the Sphere had HUGE "mindshare" in the latter half of 1975 and if there's any computer the Apple 1 is a "copy" of it's that. (I suspect Woz would disavow it having any influence on him while he was designing the Apple 1 and he's *probably* telling the truth. I've seen him mention HP's line of BASIC-programmable "calculators" which were essentially personal computers as one of his inspirations and considering where he worked and his own admission that he was sort of caught off-guard by the existence of a homebrew computer club it's *possible* that the Sphere flew by him unnoticed. That said, well, in this article he slags off on the work of pretty much every other early pioneer computer company, including Sphere, but maybe he's doing that in the rear-view mirror?.)

Over the lifetime of the company, which folded by the end of '76, they delivered about 1,300 units. Customer satisfaction was pretty low, the Apple 1 was a "better", or at least more reliable, machine than the Sphere but considering Apple sold only 1/6th as many and came (more than) half a year later... I dunno. If you seperate the Apple 1 from the company/people that spawned it can you *really* argue that it was a hugely significant machine in and of itself? If Apple were just another corpse piled in the dustbin of history next to MTI/Processor Tech/PolyMorphic, etc, I'm sort of skeptical anyone would much remember the machine. But clearly that's not the world we're living in.

Notably the SWTPC 6800 (November '75) also lacked a front panel. 8080-based systems with boot ROMs *instead* of front panels (rather than in-addition-to) didn't really seem to catch on until the first few months of '76. Presumably everyone building 8080's felt like they needed to copy the Altair... which circles around to the point about perhaps the focus on fitting everything into the S-100 bus was a little counterproductive. Oh well. In the world of computing there are few curses worse than the need to be "backwards compatible".

Ed Roberts (not Steve Jobs) was the person responsible for creating the demand for personal computers.

Truth be told if you went back in time and deleted any one person from history it's almost certain it would have played out much the same way. There were at least half a dozen inventors getting close to building a working airplane around the time the Wright Brothers actually succeeded with theirs, in large part because the enabling technologies were becoming sufficiently mature to make combining them into a working machine a realistic possibility; the question was who was going to be first and how soon, not whether it was going to happen at all. Likewise It's pretty easy to argue that "Personal Computers" were becoming "inevitable" for the same reason by the by 1970s...

That said: yes, and what I said above is not meant to diminish Ed Roberts' contribution in the slightest. The excitement was in the air *way* before Steve Jobs arrived on the scene, and the credit deserves to be spread far more equally than it seems to be these days. It bothers me *a lot* that a simplistic "Steve Jobs invented the personal computer." narrative is probably what's going to end up in my kid's history books.

2519's now on ebay at $110 (no affiliation): http://www.ebay.com/itm/For-The-Apple-1-Signetics-2519b-Chip-1974-Date-Code-Tested-Good/141460513719?_trksid=p2050601.c100261.m3456&_trkparms=aid%3D111001%26algo%3DREC.SEED%26ao%3D1%26asc%3D20140813125820%26meid%3D4f023763bad0460b98570777017024ae%26pid%3D100261%26prg%3D20140813125820%26rk%3D3%26rkt%3D4%26sd%3D141454380104%26clkid%3D1731164668878960333&_qi=RTM1963931

Apparently that is the going rate, he already sold one at the price....

If you don't care about package or date code.... Just had this pop up on my eBay "followed searches"

http://pages.ebay.com/link/?nav=item.view&alt=web&id=261654147519

It may go for a reasonable amount.