The CP2102N has some additional configuration options that are not available on other CP210x devices. These are displayed in the [Advanced Serial Configuration] tab in Xpress Configurator.

These settings are Windows only and application specific. These settings apply to the SERIAL_COMMPROP structure described here: https://docs.microsoft.com/en-us/windows-hardware/drivers/ddi/content/ntddser/ns-ntddser-_serial_commprop

The [Allowed Baudrates] options set the SettableBaud element in this structure, while the [Maximum Baudrate] sets the MaxBaud element in the structure.

The SERIAL_COMMPROP structure describes the capabilities and properties of the serial port, and is returned by the IOCTL_SERIAL_GET_PROPERTIES request, described here: https://docs.microsoft.com/en-us/windows-hardware/drivers/ddi/content/ntddser/ni-ntddser-ioctl_serial_get_properties

These settings for this SERIAL_COMMPROP structure have no direct impact on the device's behavior. For example, if the  [Maximum Baudrate] is set as 300, an application can still request a higher baudrate, and the CP2102N will happily oblige. It's then up to the application to read the SERIAL_COMMPROP structure and then prevent the user from requesting baudrates that are set to be not supported. 

Question
Using CP2105 in Linux embedded platform, when host sends requests to CP2105, but got below error:
"cp210x ttyUSB1: failed set request 0x7 status: -32"

How to interpret the error message?

Answer
Table 6 Interface Commands in AN571 lists all API commands that CP210x device supports. Viewing from this Table, request of 0x7 represents SET_MHS command, which is used to set modem handshaking (DTR and RTS lines) states.

The negative error is returned by the underlying call to usb_control_msg() from the cp210x.c driver.

If this function is not successful, a negative error number is returned.

Looking at Linux USB Error Codes and errno-base.h, we can see that 32 is EPIPE, endpoint stalled.

How does CP210x handle the SET_MHS request from host?

For CP210x SET_MHS (0x07) command, as AN571 stated, the DTR and RTS values can be set only if the current handshaking state of the interface allows direct control of the modem control lines.

When the device receives the SET_MHS command intending to change the state of DTR or RTS pin, CP210x device will then check current Flow Control setting in the device and make a different response.

If the hardware Flow Control is currently enabled, then the handshaking lines have been managed by the CP210x device rather than host, the device will not response the current SET_MHS request, and a STALL would be returned.
 

Question: How can I determine the device revision number for my CP2112 HID USB to SMBus bridge device?  Is this information included in the device top marking?

Answer:  It is not possible to determine the CP2112 device version is from the top marking alone, as this information is not explicitly encoded in the package marking.  To be certain of the device revision, you can query the device over USB and it can return the device version.  To do this, you can issue a feature request over USB (Report ID 0x05) and the device will return the part number (0x0C) and the device version.  More information on this can be found in application note AN495: CP2112 Interface Specification, section 4.5, page 9:

https://www.silabs.com/documents/public/application-notes/an495-cp2112-interface-specification.pdf

You can also retrieve the version number from the device over USB using the Silicon Labs HID USB-to-SMBus API, using the function HidSmbus_GetPartNumber().  For more information on doing this, please refer to application note AN496: HID USB-to-SMBus API Specification, section 2.29, page 22:

https://www.silabs.com/documents/public/application-notes/an496-hid-usb-to-smbus-api-specification.pdf

The above device interrogation techniques are the most reliable way to determine device version number.

Question: 

What is the right way to determine the revision (A01/A02/...) of my CP2615?

1. In Section 3.6.2.3 "Special Operations", the CP2615 Datasheet depicts (in Table 3.13 "Read CP2615 Firmware Revision") an operation available in Configuration Mode to read the Firmware Version Number.

2. In Section 3.2 "iop_AccessoryInfo", AN1139: CP2615 I/O Protocol depicts a "PartID" described as "Encoded part number of the chip (0x1400 for CP2615-A01, 0x1500 for CP2615-A02)."

Answer:

Reading the version via I2C in config mode requires an external I2C Master, e.g. the special-PID CP2112 on the CP2615 EVB. This method is suitable for a prototyping environment (it's how Xpress Configurator determines the attached device), or in a customer's production-line environment (e.g. the external I2C master can verify the proper device version before programming the config). For this method, the CP2615 must be powered up, but doesn't have to be connected to a USB host.

Reading the version via the iop protocol is suitable for an end-user situation, e.g. an app running on the phone or USB host would query the version number before issuing any version-dependent commands (e.g. downloading user profiles, which is only supported on the -A02).

When being programmed with a new configuration, the CP2114-B01 and CP2114-B02 devices can sometimes write corrupted configuration data when SYSCLK is currently being driven by the internal 49.152 MHz oscillator and the size of the new configuration data block is greater than 62 bytes.

The problem does not occur in other SYSCLK modes. To prevent the possibility of corruption when programming a new configuration, the CP2114 should be operating in one of these three SYSCLK modes:

• SYSCLK is driven by the internal oscillator at 48 MHz (e.g. factory-programmed config[0]).

• SYSCLK is driven by an external oscillator at 48 MHz.

• SYSCLK is driven by an external oscillator at 49.152 MHz.

WHQL Certified drivers are selected over non-certified drivers, and higher version numbers are selected over lower version numbers.  Installing a newer driver will not uninstall an older version, but the older version will no longer be used.

For more information, see this knowledge base article from Microsoft: https://docs.microsoft.com/en-us/windows-hardware/drivers/install/how-setup-ranks-drivers--windows-vista-and-later-

What does it mean when cp210xsmt.exe tool produces the message "ERROR: syntax: invalid hex number size with CP210x .configuration file"?

One of the text fields in the configuration file is longer than expected, for example:

DeviceMode {  { 00000000 }  { 00000000 }  } 

Instead of:

DeviceMode {  { 00 }  { 00 }  } 

Question

How can I achieve the Datasheet sleep current numbers for CPT007B, CPT112S and CPT212B using the Eval Boards?

Note - CPT213B does not go into sleep mode to avoid violating the IEC-60730 safety specs.

Answer

The average current drawn is going to be a function (mainly) of the sleep mode scan time (~250 us) and to some extent the sensors included as wake up sources. Scan can take more time if there are more sensors. If all sensors are designated as wake up sources and drive strength is set to 8x, the typ sleep current value shown in the Datasheet can be reflected in an eval board. However, if an eval board is tested for current consumption numbers, one may see a sleep current of the order of a few mA. This is because there are a few additional items to keep in mind - 

• Buzzer circuit

The buzzer circuitry on the CPT boards does not allow it to go to sleep mode. On energy profiler, one will not seeing the sleep scan happen every 250 ms and the current will be of the order of 3 mA. If the buzzer is enabled in the profile but the pin configuration is set to open drain, the current will be of the order of 3.5 uA. Even when the buzzer was disabled the average current stayed at 3-4 uA. If the buzzer's pin configuration is not programmed at all, it defaults to push-pull and if the buzzer circuitry is connected, it will draw more current and not allow the part to go into sleep mode. So the solution is to remove the buzzer from the board, i.e., the transistor and resistors around it. Here is a screenshot from the eval board schematic - 

• Device Revision in CPT007B and CPT112S 

Despite the change made in #1. shown above, some CPT007B and CPT112S boards might still show a current of about 3 uA. This is a known issue. The attached errata throws more light about this issue. The rev A02 of CPT007B and CPT112S shows the correct current numbers.

Problem

There is a bug in the CP2114 Customization Utility that does not preserve changes made by the user in the DAC Configuration window. Example of unexpected behavior below.

1. Connect a CP2114 (B01 or B02) in a PC and open the CP21xxCustomizationUtility.exe

2. Navigate to the DAC Configuration. Read first empty index (3 in this case). Note that DacConfig in this example is currently 00. 

3. Make a change to DacConfig, eg 00 to 11. Close window.

4. Unexpectedly, DacConfig is not 11.

Workaround

To edit the DAC configuration:

1. In the CP2114 Customization Utility, save the configuration to a file using the "Save to file" button.
2. Open the exported file in a text editor.
3. Edit the DAC configuration. Your DAC configuration should be located after the "#DAC config" comment in the file.

01 # [44] PbTerminalTypeLsb
03 # [45] PbTerminalTypeMsb
02 # [46] RecTerminalTypeLsb
06 # [47] RecTerminalTypeMsb
00 # [48] Reserved1
00 # [49] Reserved2
# DAC config
F9
6E
00
47
04
64
FA
64
 o
 o
 o

1. Import the edited file.

The CP2102-GM datasheet specifies three options for pin packaging for the QFN28 device (taken from CP2102 datasheet, page 12:

Although Silicon Labs CP2102 devices are generally shipped in the leftmost "Option 1"/"Irregular Corner" option, the other options are included because Silicon Labs reserves the right to ship these devices in any of those package varieties.

Regardless of the Pin 1 marking option used in a given device on the bottom of the device, the PCB landing pad configuration is the same for all CP2102 devices, and is given in the datasheet on page 13: