WACOM Graphire Drawing Tablet Reverse Engineering

 WACOM ET-0405 is one of the few non-ASIC-based tablets released by WACOM, meaning that their RF receiving signal chain is built with discrete opamps and passives. The pen still has an ASIC. This post contains enough information for you to make your own.

Resources:

Patents:

US4878553, US4999461, US5028745: A general description of WACOM's analog passive-LC tablet design. It contains the block diagram, waveforms, and a general description of the operation.

US5644108, US5679930, US5600105: Detailed description of the ASIC-based pen, including detailed schematics, waveforms

You can find more of those patent stuff online:

FCC:

You can find a huge amount of stuff on the FCC website, including internal photos, schematics, and more. Curiously, the model is licensed under Universal Scientific Industrial Co., Ltd. instead of WACOM Co., Ltd.

Entry: https://fccid.io/IXMET-0405

You can find the original schematic of the mouse, the pen, and the tablet there.

PCB Scans:

I did a 600dpi scan of the PCB myself: 

https://drive.google.com/drive/folders/1vGHHMH9WkavLYoCPT2HWrooUTZOjV1wc

Circuit Analysis:

Signal Chain:

I won't show the schematic here due to copyright reasons. 

IC4 74HC126 serves as the negative power rail generator and RF stimulus.  The RF signal chain starts from IC2B and eventually goes into the ADC input of the microcontroller. My understanding of the signal chain is this:

Tablet Receiver Signal Chain

Theory of Operation:

You can learn how this circuit helps detect the pen position by reading patents (which is, unfortunately, extremely verbose). The first few patents (US4878553) discuss an analog drawing tablet with no ASIC in pens. I think WACOM used those in some of their early models and, curiously, most licensed products (aka WACOM "EWR"). I will talk about the passive EWR technology in the next post.

Starting from the 1990s, WACOM started putting ASIC chips into their pens. This is very powerful because it decouples position detection and pressure detection into two time-isolated channels (i.e. time-domain multiplexing), enabling a much better signal quality in an era where microprocessors are barely able to do complex computations.

The ASIC pen patent went to great lengths to describe an analog transmission system, where the pen samples the data and transmits it back to the tablet using phase-shift modulation. I believe this method has never materialized, as this model uses a digital transmission system.

Captured Data at the Integrator's Output

The aforementioned receiver circuit is basically a very sensitive FSK/PSK detector. From the pen's side the ASIC toggles the capacitance in an LC resonator to slightly shift the resonance frequency, and the tablet is able to detect this minute difference and translate it into a digital readout.

The waveform captured shows 4 distinct segments. I will go over them one by one.

1. Synchronization / Charge-up

The pen needs to know when to transmit and when not to. I think this segment is intended to synchronize and provide power to the pen.

2. Positioning

Next, the tablet started scanning the X and Y coils individually. The peak indicates the location of the pen. It's possible to enhance the resolution by interpolating the curve.

Note that the scanned coils are from a small portion near the pen tip over the entire tablet. The gist is that human hands cannot move instantly, so once the pen position is known, it's unlikely for the user to remove the pen from a small region near the pen tip in 1 frame of time. So WACOM implemented this optimization to lock on to the pen's position and look at the few coils near the pen tip to reduce scanning time. 

3. Digital transmission 

Next, the pen transmitted the pressure and button data by toggling an auxiliary C in the LC tank.The position is already determined in the previous segment, so the MUXes can be configured such that the amplitude is always at the maximum.

4. Positioning

Same as above. I think this is done twice in a frame to improve spatial resolution. Also, the positioning info can be used to maximize the amplitude of the following synchronization segment.

Conclusion:

The design of the RF receiver is quite simple and elegant. In the next ten years, the exact circuit was copied to death by a large number of manufacturers. While WACOM moved on with more sophisticated ASIC solutions. The copycats were able to improve the circuit and programming to obtain 8192-level pressure resolution and tilt sensitivity, without ASIC-based pens! In the next post, I will examine the circuit and firmware design of such a product.