1
Preventing Electrical Damage in ThingMagic Nano Modules
Dan Ratner
7/21/2016
Since the release of the Nano module, we have had a few reports of Nano modules ceasing to work
properly after weeks or months of successful operation in customer-designed readers. The two must
common symptoms are:
The module fails to initialize and will not communicate with the host
The module will initialize, but presents a firmware CRC error (Error number 0200h) to the host
indicating that the application firmware is corrupted.
We believe the cause of these failures is memory corruption resulting from the circuitry around the
Nano not following the interface guidelines we give in our User Guide, which are based on requirements
from the supplier of the processor IC we use. There are two areas are of particular concern:
1. Input lines remaining high when the Nano module is turned off
2. Programming pin SWCLK allowed to float if not used
Input Line Requirements
In the Nano Design Guide, we give the following requirements for input lines. (The relevant clause is
underlined for emphasis.)
Input High-level Voltage
1.9 V min to indicate high state; 3.7 V max when module is powered up, no
more than 0.3 V higher than Vout when module is turned off to prevent
damage.
Vout is the DC voltage that the Nano sources on pin 10 and is nominally 3.3V as long as the DC input
voltage (pins 16 and 17) is greater than 1.8V.
This input voltage restriction applies to both the serial interface input line (receive side) and the GPIO
lines. Our recommended carrier board schematic diagram shows a buffer circuit that would enforce this
requirement for the RX input line. We did not take steps to protect the GPIO lines on the carrier board
because they can be used as either inputs or outputs depending on how the Nano is programmed.
We have discovered that the most common “damage” that is caused by violating this requirement is
corruption of the bootloader or application memory. Application memory corruption causes the
modules bootloader to report a firmware CRC error, indicating that the firmware is not operational. If
Bootloader memory is corrupted, the module may not respond at all. The altered memory cannot be
repaired through user-accessible module interfaces, so the module cannot be repaired in the field.
2
Options for the way Nanos are shut down (by DC power or Shutdown line) and by the way the Nano and
input drivers are powered (by an independent supply or the same supply) leads to three solutions,
depending on the use case:
Case 1: If the Nano is powered by the same source that powers the circuitry that drives the
input lines to the Nano (so that everything is powered up or down together), and the Shutdown
line is not used to power off the Nano, there should be no potential for the input lines to remain
high when the Nano is powered off and the Nano should never experience this memory
corruption issue. See Figure 1.
Figure 1
Case 2: If the Nano is powered by a different circuit than the logic that drives the RX and GPI
lines, so that the Nano’s power could go down while the voltage to the input lines remains high,
the memory corruption issue could occur (regardless of whether the Nano is powered down
through dropping the DC supply voltage or using the Shutdown line.) See Figure 2.