[NTLK] Cadillac prototype

[Justin Ried]

Hello everyone,

Thank you to Grant for the nudge to join this list. I recently acquired an Apple “Cadillac” prototype, which is the “Senior” product concept that was dropped in favor of the “Junior” product concept that eventually became the commercially released Newton products. Cadillac is substantially different than other Newtons, and is a much larger tablet-like form factor. I have found a few examples of Cadillac in museums and private hands, and the NewtonTalk Glossary has been particularly helpful.

My own example has a clear body, and all of its original components seem to be in place. A gallery is available here:
https://drive.google.com/drive/folders/17BIoW0O4ShLOBT40Ce4gyO1e35YuE6oc?usp=share_link

The other clear example that I have seen once belonged to an Apple employee: James R. [Dick] Huston (Apple Employee #25). That example was sold on eBay and appears very similar, but is missing several key components including its pen, the serial board behind the serial port and IR transceiver, and its ROM card. It is otherwise very similar to mine. 

Mine also comes with a PCMCIA card installed in its expansion port (separate from the ROM and RAM slots). That PCMCIA card has a significant impact on the Cadillac’s power usage profile, detailed below. 

I am particularly interested in finding out whether my example can safely be powered on, and if it works as originally intended. I have used a bench power supply and probed Cadillac’s internal test points to get a pretty good understand of voltage rails and their requirements. I have also disassembled the battery pack and found it has seven NiCd cells in series, but the positive lead is intentionally disconnected from the battery pack's own (mostly) unpopulated PCB and covered in heat shrink. This leads me to believe that the charging logic was never fully implemented on Cadillac’s battery pack PCB, and it has never been powered by the batteries. The battery pack’s barrel jack and J2 header connections are wired up to reach the Cadillac’s onboard power connectors when the battery is installed, so it clearly was designed to receive power via the AC adapter, even in this prototype state.

An old webpage on archive.org that shows a working Cadillac prototype notes that it runs on a 7.5 volt 1,000 mA AC adapter. I assumed that would apply to my unit also, and have investigated Cadillac’s voltage requirements using a bench power supply - ultimately finding 7.5 volts not to be sufficient with my own unit.  Details below. 
Apple Newton “Cadillac” prototype — key findings from the last day

The biggest change in our understanding is that Cadillac does not behave as though 7.5 V is necessarily its intended full operating voltage. The previously cited 7.5 V figure comes from a collector-era web page and should not be treated as authoritative without independent confirmation.

The battery pack contains seven NiCd cells wired in series, which I verified physically when the pack was opened. That means the pack is nominally about 8.4 V, and a freshly charged seven-cell NiCd pack can be roughly 9.8–10.15 V immediately after charge. The battery cells themselves are currently disconnected at the positive end, so present bench-supply current is not battery-charging current.

Measured power behavior

With the PCMCIA card installed, Cadillac shows a very clear voltage-dependent transition:

Input	Current	Power	Behavior
7.50 V	~0.124 A	~0.93 W	Stable, silent
~8.31–8.35 V	variable	variable	Rapid ticking begins
~8.35–8.5 V	large current excursions	—	Audible ticking / converter cycling
8.50 V	~0.257 A	~2.18 W	Stable, silent
9.00 V	~0.243 A	~2.19 W	Stable, silent
10.00 V	~0.222 A	~2.22 W	Stable, silent
Above about 8.5 V, current falls as input voltage rises while total input power remains almost constant at about 2.2 W. That is strongly suggestive of a regulated switching-converter load rather than a simple resistive load.

The transition around 8.35–8.5 V is abrupt rather than gradual. Video showed the current jumping substantially within only a few hundredths of a volt, accompanied by audible ticking from inside Cadillac. Once the input reaches about 8.5 V, the ticking stops and the unit settles into the ~2.2 W state.

Starting Cadillac directly at 8.5 V avoids most of that cycling and brings it immediately into the stable higher-power state.

PCMCIA card is clearly significant

The 68-pin PCMCIA-format card has a major effect on Cadillac's behavior.

At 7.5 V, with the card removed, current drops to about 0.051 A / 0.382 W.

At 9.0 V, however, removing the card produces the opposite effect: current can rise dramatically, with readings around 0.44–0.70 A and roughly 4–6 W, accompanied by a faint high-pitched squeal from inside the unit.

With the card reinstalled, Cadillac returns to the much calmer and very repeatable ~2.2 W regulated state between 8.5 and 10 V.

That strongly suggests the card is not merely an optional peripheral. Its presence materially changes system state. It may contain firmware, configuration data, memory, termination, or some other prototype-specific function, but its exact role is still unknown.

Internal hardware observations

The main logic board contains two large Apple-specific LSI ASICs:

U3: LSI L1A7879, marked “PROTO”

U15: LSI L1A7777 / 343S0703B / ©1992 Apple

No public documentation or datasheets for those exact ASICs have been found so far.

The power section contains a UC3846 switching PWM controller and substantial magnetics. The audible ticking/squealing appears to come from the power-conversion circuitry, not the sound subsystem: plugging headphones into Cadillac's headphone jack did not reproduce the ticking in the headphones.

When Cadillac is in the stable ~2.2 W state at 9–10 V, there is slight warmth over the central Apple/LSI logic area, while the power section itself does not feel noticeably warm. That suggests a meaningful portion of the supplied power is reaching the logic circuitry rather than simply being dissipated in the converter.

Power-button behavior

The POWER button itself appears electrically healthy but does not initiate the major power-state transition.

Its signal changes cleanly from about 5 V released to 0 V pressed, and pressing POWER consistently increases total current by only about 1 mA. That remains true even in the higher-power 8.5–10 V operating state.

This matches the possibility that Cadillac is intended to power up automatically when its main supply/battery is present rather than being turned on by the front POWER button.

Earlier internal rail measurements

Before the unit was closed back up, we confirmed that several internal rails were present:

raw/input-related node: about 7.41 V with a 7.5 V bench input

intermediate rail: about 5.52 V

regulated rail: about 5.04 V

another intermediate node around 5.8 V

These remained stable during POWER-button presses.

Current working interpretation

The most plausible model at this point is that Cadillac has a low-power/partial-power state below roughly 8.4 V, a narrow startup/UVLO transition region around 8.35–8.5 V, and a stable regulated operating state from about 8.5 V through at least 10 V.

The fact that the seven-cell NiCd pack is 8.4 V nominal makes that threshold especially interesting. It raises serious doubt that 7.5 V was truly Cadillac's intended adapter voltage.

At the same time, we still do not have proof that 10 V was the original adapter specification. We only know that Cadillac behaves much more like a normally regulated electronic load from about 8.5–10 V than it does at 7.5 V.

The other major unresolved issue is that the LCD remains blank. However, the current draw, ASIC warmth, PCMCIA-dependent behavior, and regulated power consumption all suggest that “blank display” should not be assumed to mean “completely dead.”

The two most valuable questions for Newton experts now are probably:

Does anyone have original Cadillac adapter specifications, photographs, or an original PSU label?

Does anyone recognize the 68-pin PCMCIA card or the custom Apple/LSI ASICs L1A7879 “PROTO” and L1A7777 / 343S0703B?

Those two pieces of information could explain most of what we're seeing.