General Category > LisaList2
Lisa Tank Problem
compu_85:
Some more observations:
If we assume that serial numbers start at 1 for a given day and count up, and then add together the highest serial numbers recorded for each day so far, the total is 25,404. I treated serial numbers starting with B and A as if they were different factories, so some says got 2 counts. If we include the 810 count from the 20th day of 1982 the count jumps 26,214... I would say it's an outlier, but we have 5 systems from that day! So I don't know if the other estimate is as trustworthy as we were thinking, since simply counting the serial numbers we've collected is more than double.
Also, it looks like Apple was building 2/10s at the same time as Lisa 1s? I have to wonder if there's some bottom plate swapping corrupting the numbers.
This work is making the 50k number seem more likely.
-J
fri0701:
I've been trying to lay down some "invariants" I see across the recorded serial numbers, and I wanted to hear your thoughts:
1. AppleNet numbers are unique, likely sequential with no gaps, and one was assigned to each machine. I know that we have a conflict with #00104218, but that's the only one we've seen, so I'm tempted to count it as a one-off mistake for now. Since this number is used for copy protection (and perhaps was planned as a global ID number for AppleNet), surely this was meant to be unique.
2. Taking an AppleNet number and subtracting the last digits of the S/N (which I'll refer to as the "batch count" here) gives you a "batch start" number. Each day has one and only one of these "batch start" numbers.
There are some cases where a "batch start" number is shared across multiple days. It is also clear that, although there was a general upward trend, these "batch start" numbers (and therefore, the AppleNet numbers in general) did not monotonically increase as time went on.
My working theory is that on any given day, the factory chose a batch of AppleNet numbers to produce, denoted by a "batch start" number. They didn't necessarily start at the beginning of that range, but picked an open section within it and produced what they could that day. The fact that the same "batch start" number is shared between some days greatly reduces the estimate of the number of Lisas produced from just counting up the "batch count" numbers per day.
As for the total number of Lisas, if we do hold rule #1 above, the number of Lisas produced would be around 12,000 via the tank problem.
Hopefully we can keep refining these theories and numbers as we log more machines! I'd love to hear anyone's thoughts on this.
blusnowkitty:
--- Quote from: compu_85 on February 21, 2023, 03:40:56 pm ---I treated serial numbers starting with B and A as if they were different factories, so some says got 2 counts.
--- End quote ---
I can't prove it, but I'm almost positive that B08Bs were made in California, while As were made in Carrolton, TX. I've heard as much said before that manufacturing started off near HQ and later moved to the Apple II plant in Texas, and this seems to track with the trends we see in the SN tracker. B08Bs account for 1982 and early 1983, by mid-1983 we see a mix of B08Bs and As, and then by late 1983 it's all As.
I'm also inclined to believe fri's post about Applenets being batched instead of sequential.
fri0701:
"1983: Apple opens a highly automated plant in Fremont, Calif., to manufacture the first Macintosh computers. It also said it would shut down a pilot factory in Cupertino, Calif., which made the Lisa computer, and move that production to Carrollton and Cork." https://www.wsj.com/articles/BL-DGB-25630
"October and November saw a series of announcements and releases for the Lisa. On October 4, Apple started shipping versions of the Lisa to Australia, New Zealand, France, Germany and Britain. These shipments came from the new international Lisa plant in Cork, Ireland." https://mirrors.apple2.org.za/www.bitsavers.org/pdf/apple/lisa/development_history/articles/Dines_-_The_Lisa_A_Case_History_1985.pdf
So it looks like we're dealing with 3 plants: Cupertino, CA; Carrolton, TX; Cork, Ireland.
The "Made in USA / Ireland" sticker on the Lisa's case is a good clue. The earliest Lisa I've seen that's labeled "Ireland" is from June 14, 1983. It looks like both A-style and B-style serial numbers were produced there.
The earliest A-style serial number I've seen is March 3rd, 1983. The latest B-style is September 20th, 1983.
Interestingly, in reference to my last post, we also have evidence of Lisas marked "USA" and "Ireland" made on the same day sharing the same "batch start" number.
--- Quote from: blusnowkitty on February 21, 2023, 06:04:29 pm ---I can't prove it, but I'm almost positive that B08Bs were made in California, while As were made in Carrolton, TX.
--- End quote ---
This is an interesting theory... since both styles were made in the Ireland plant, however, I'm more inclined to keep with the idea that the change was to do with the switch from the original Lisa to the Lisa 2 series.
sigma7:
--- Quote from: compu_85 on February 21, 2023, 10:58:15 am ---Here's another question: does anyone else have a power supply with a suspiciously serial number looking sticker on it?
--- End quote ---
Of the 4 power supplies in sight, 2 have those white labels:
a 1.2A is labelled *71090* with the barcode underneath
a 1.8A is labelled *301488* without the barcode
Both of those came to me from Dafax (and also have the round green sticker with the OK stamp). Another 1.8A has the Dafax green sticker but no white sticker.
The 1.8A have the DataPower aluminum foil label with the embossed serial number, so I wonder if Sun Remarketing added the white stickers for inventory management.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version