Schematic / PCB history

PCB flavours

The main PCB has following production numbers:

Manufacturers

There had been two large Falcon manufacturers: ATMC and EFA. ATMC had used two sub-contractors, “Golden Horse” (B3) and “ECS Tamsui” (B5), EFA just one, “Ta Yuan” (Y4). EFA had been developing all Falcons since PCB rev. B. List of all Atari manufacturers here.

Serial numbers

This is one big mess. According to this diagram in the Falcon030 service manual:

falcon_serial.png

it should be pretty straightforward but it is not. There are several discrepancies:

  1. Number of digits: ECO statistics document uses serial numbers with the control code having “0” prefix, few Falcons developed by B3/B5 and all of the Y4 Falcons follow this scheme, too
  2. Colour encoding: Early falcons from B3 and B5 follow the rules and include the number indicating colour encoding in their serial number. All of the Y4 Falcons except the 12/1993 Peritel batch however skip this number and rely solely on the PCB production number (ironically, the Peritel batch has got the PCB production number wrong).
  3. Sequential numbering: It’s more or less clear that every factory used its own numbers and there were several numbering resets during the production.

Manufacturing years

From all the serial numbers I’ve seen so far one can conclude:

  1. First batch happened in 08/1992 in B5. Most likely just a few units to be send to early developers and for presentation at Atari Messe in Düsseldorf (held on 23rd August!), where the Falcon has been officially presented (as the Falcon, until then it was known as the Sparrow).
  2. Second batch happened during 10-12/1992 in B5, this was also the only batch producing Peritel models
  3. Third batch happend in 02/1993 in B3, B5 and during 12/1992-02/1993 in Y4
  4. Fourth batch happend in 04/1993 in B3 (s/n numbering reset) and during 03-09/1993 in Y4 (s/n numbering reset)
  5. Fifth batch happend in 06-07/1993 in B3 (s/n numbering reset)
  6. Sixth (Peritel only?) batch happend in 12/1993 in Y4 (s/n numbering reset)
  7. Seventh batch happend in 01/1994 in Y4 (s/n numbering reset)
  8. Eighth batch happened in 09/1994 in Y4 for C-Lab MK II (and maybe MK X) models (s/n numbering reset). Interestingly, one can find PCBs with the old production number (Atari’s) as well as the new one (C-Lab’s) from this very same month.
  9. Nineth and the last batch in 12/1995 in Y4 for C-Lab MK X models (s/n numbering reset)

If we can trust the ECO statistics document, the last approval activity for Falcon stuff had happend around 06-07/1993.

Sometime in 09/1994 (judging from MK II serial numbers) C-Lab had ordered by EFA another batch for the MK II/X models which were delivered around 08/1995 according to this article in ST-Computer.

Another rather confusing fact is that revisions are not always related to manufacturing years. So it’s totally OK to see rev. D/C from 01/1994 and rev. J from 07/1993.

Number of Atari Falcons manufactured

In this intereview in ST-Computer, Bob Gleadow, Managing Director at Atari Deutschland GmbH, claimed that in 10/1993 Atari had sold around 13,000 - 14,000 units and had 4164 units in stock.

Rodolphe Czuba recalls that the number of Falcons sold must be between 10,000 - 12,000 from Atari and 1,000 - 2,000 from C-Lab. The complete batch done was 20,000 units but only 12,000 were assembled and put to market. The remaining 8,000 units were sent (in various state of completition) from Sunnyvalle, USA to Germany (C-Lab) in two big cargo flights.

Highest s/n values per factory:

That gives us 22,184 known original Falcon030 units and 1,688 known C-Lab units. Therefore it’s safe to assume the number isn’t much higher than 20,000.

Still the question whether C-Lab had ordered something from EKA remains – the ST Computer article mentions MKII models still from the Atari batch but quoting C-Lab representatives about a C-Lab batch ordered from EFA directly, for later MKII and MKX models, so who knows? Altough it’s hard to understand why they would order another batch if they had 8,000 semi-completed units in stock?

Anyway, it has been verified on multiple occasions that C-Lab did reuse older boards for their “new” products – i.e. it’s normal to find a MKX with motherboard from 05/1993. So while Bob Gleadow’s numbers would put us somewhere between 17,000 - 18,000 built units (which is quite close to ~16,000 from serial numbers in 10/1993) I think the real count is closer to Rodolphe Czuba’s statement and many MKI/MKII/MKX simply reused already existing PCBs.

PCB revisions

Schematic C303007-001 rev. Schematic rev. date PCB rev. PCB approval date Description PCB rev. changes Notes
A 23/07/1992     Initial release    
B 25/08/1992     Revised per ECO #1488 Image of ECO #1488    
C 25/08/1992     Revised per ECO #1492    
D 25/08/1992     Revised per ECO #1497    
E       Revised per ECO #1498 Image of ECO #1498    
F 10/09/1992 A   Revised per ECO #1505 NTSC A, PAL-I/B A, PERITEL A Production release; PERITEL models had socketed CODEC and no RF modulator
G 06/01/1993 B 26/01/1993 Revised per ECO #1600 Image of ECO #1600   There’s no PERITEL version manufactured anymore
H 20/01/1993 C 16/03/1993 Revised per ECO #1604 Image of ECO #1604 1/2 Image of ECO #1604 2/2    
    D 16/03/1993 Revised per ECO #1613   Common D/C release, manufactured for Atari between 03/1993 - 09/1993, later sold by C-Lab as MKI models
    E 02/03/1993 Revised per ECO #1621    
    F 30/04/1993 Revised per ECO #1625   Corrected error made on ECO #1604
J 11/06/1993 G 04/08/1993 Revised per ECO #1642 NTSC G Improved video quality for NTSC machines
K 06/08/1993 H 04/08/1993 Revised per ECO #1643 Image of ECO #1643 NTSC H, PAL-I/B G Attempt to fix sound crackle during DMA playback
L 06/08/1993 J 04/08/1993 Revised per ECO #1641 Image of ECO #1641 NTSC J, PAL-I/B H  
          PAL J Strangely, there are records of Falcons from 07/1993 with this revision. Maybe done by an Atari dealer in post production to D/C PCBs?
          PAL K C-Lab Falcon MKII, manufactured in 09/1994
Schematic 500047-001 rev. Schematic rev. date Description Notes
1 06/01/1993   Initial release (spin-off from rev.G?)
A 01/03/1993   Engineering release
B 13/05/1993 Revised per ECO #1644 Production release