Hi there,

It appears that the IridiumSBD I2C and SparkFun u-blox libraries are conflicting when using a SparkFun Qwiic Micro and RockBLOCK 9603 over serial.

Components used:

• SparkFun Qwiic Micro
• RockBLOCK 9603

Steps to reproduce:
This can be observed by running any of the Serial_Examples__Rockblock with u-blox library included:
#include <SparkFun_Ublox_Arduino_Library.h>

After adding the library, for instance using Example5_Sleep, results in error 5, no modem detected:

Output:

Starting modem...
17:00:50.165 -> Calling internalBegin
17:00:50.165 -> Powering on modem...
17:00:50.653 -> >> AT

17:00:50.653 -> AT
Waiting for response OK
17:00:50.653 -> 
17:00:50.653 -> << >> AT

17:01:20.638 -> AT
Waiting for response OK
17:01:20.638 -> 
17:01:20.638 -> << >> AT

17:01:55.728 -> AT
Waiting for response OK
17:01:55.728 -> 
17:01:55.728 -> << >> AT

17:02:25.826 -> AT
Waiting for response OK
17:02:25.826 -> 
17:02:25.826 -> << >> AT

17:02:51.716 -> AT
Waiting for response OK
17:02:51.716 -> 
17:02:51.716 -> << >> AT

17:03:20.631 -> AT
Waiting for response OK
17:03:20.631 -> 
17:03:20.631 -> << >> AT

17:03:50.650 -> AT
Waiting for response OK
17:03:50.650 -> 
17:03:50.650 -> << >> AT

17:04:20.630 -> AT
Waiting for response OK
17:04:20.630 -> 
17:04:20.630 -> << No modem detected.
17:04:50.639 -> Powering off modem...
17:04:50.639 -> Begin failed: error 5
17:04:50.639 -> No modem detected: check wiring.

Potential cause:

I believe this is because the u-blox library defines Serial as SerialUSB:
https://github.com/sparkfun/SparkFun_Ublox_Arduino_Library/blob/97bd455b0e7a05b92bfe5a528c28cc47f2e2d289/src/SparkFun_Ublox_Arduino_Library.h#L55-L67

However, on the Qwiic Micro, we use Serial as the port for the RockBLOCK:

#define IridiumSerial Serial
IridiumSBD modem(IridiumSerial, SLEEP_PIN);
IridiumSerial.begin(19200);

It doesn't seem to matter what order the definitions are made, the u-blox library seems to have final say as to what Serial will be. I wonder if a possible solution is to have a check in the u-blox library for the Qwiic Micro, and not declare Serial as SerialUSB.

Cheers,
Adam

The getSystemTime() function uses mktime() to do the Iridium epoch offset, but does not set the tm struct's tm_isdst member, which may lead to incorrect time conversions (ie a random/junk DST offset).

It looks like the conversion code should either set epoch_start.tm_isdst = 0 or use mk_gmtime(&epoch_start) to get the epoch start timestamp.

I am looking for some help re: clearing the MO buffer.

I understand that when one of the sendSBDText, sendSBDBinary, sendReceiveSBDText, sendReceiveSBDBinary are called, the library does consecutive retries within ISBD_DEFAULT_SENDRECEIVE_TIME seconds. Either there is a timeout, successful transmission or fatal error. This timeout can be adjusted by the adjustSendReceiveTimeout() function.

I have then seen in the send/receive example that after the send/receive functions terminate you always clear the MO buffer. This would make sense in the case when the transmission was successful, but I don't understand why you do it also in the case that it failed. Can't I just call again the send/receive functions without clearing the MO buffer? Is it because when calling them, I have to pass the buffer to be sent?

Hi @PaulZC,

I discovered today that v2.x of the Apollo3 Core is not compatible with the IridiumSBD I2C library, as it no longer includes WString.h. The core does include the following helpful note:

// including WString.h is deprecated, for all future projects use Arduino.h instead

I'm not overly familiar with WString.h but it appears to be used extensively within the library. Should we try to have it included in the v2.x core or is Arduino.h capable of replacing its functionality?

Cheers,
Adam

Hi @PaulZC,

I've been dealing with the dreaded "ISBD_PROTOCOL_ERROR" error recently with a RockBLOCK 9603, and I'm curious, is there a way to dig deeper into the IridiumSBD library to actually determine what's happening when presented with this error? Looking at the code, it appears to be somewhat of a catch-all error return code for a wide range of functions.

Hmm, I do notice that it is most commonly associated with !waitForATResponse. I wonder if it my issue coudl be AT timeout-related. The RockBLOCK is powered directly by a 3.7 V LiPo and the supercapacitors are constantly charged, so everything should be snappy.

Cheers,
Adam

In the file "IridiumSBD.cpp", currently at line 1403, there is this code:

while (wireport->available())
{
  wireport->read(); // Mop up any unexpected bytes
}

That code can be removed, since the Wire library uses packages of data and not a stream of data. There is no need to remove unused data.

Explanation: Common-mistakes#4

In the code above that, is this:

wireport->requestFrom((uint8_t)deviceaddress, 2); // Request two bytes
if (wireport->available() >= 2)

The peripheral can not send less or more. There is no need to allow that it is higher than 2.
If two bytes are requested, then test for two bytes:

if (wireport->available() == 2)

Due to supply-chain shortage, some components in RockBLOCK 9603 3.F have been replaced with alternatives. This does not adversely affect the normal operation of the device.

Because of these component changes, 3.F now includes a transistor circuit on the On/Off control pin to ensure correct power up at all supply voltages. The On/Off control pin should not be clamped below the RockBLOCK supply voltage and should be disconnected/left floating to turn on.

Potential issue:
This change has introduced an issue when interfacing directly to a 3V3 microcontroller pin while supplying the RB with 5V. The new On/Off circuitry now provides a high impedance (1Meg resistor) to the super cap voltage, equal to the supplied voltage. This may exceed the maximum IO voltage of the micro.

If the RB was supplied by 5V, the connecting device should also be 5V tolerant. If not, it is recommended to use an interfacing circuit, eg MOSFET, transistor, relay, etc. to connect to a microcontroller.

More info:

https://docs.rockblock.rock7.com/docs/power-supply#onoff-control

I compiled with PlatformIO (still Arduino platform) and got a warning that boolean is deprecated.

Perhaps makes sense to replace with bool in the IridiumSBD::isConnected() function

Hi @PaulZC,

I've found myself on several occasions needing to decrease the value of ISBD_STARTUP_MAX_TIME from 240 seconds to something shorter for testing or other purposes. This usually involves going in and modifying the library directly, but I was thinking that it could be useful to have a function that could set this timeout value directly.

Please let me know what you think! Happy to whip up a PR.

Cheers,
Adam

I am trying to use the I2C Iridium modem with an Itsybitsy M4.

I am getting compile errors for the library attributed to SAMD51 problems. Help?

Arduino: 1.8.13 (Mac OS X), Board: "Adafruit ItsyBitsy M4 (SAMD51), Enabled, 120 MHz (standard), Small (-Os) (standard), 50 MHz (standard), Arduino, Off"

.../Documents/Arduino/libraries/SparkFun_IridiumSBD_I2C_Arduino_Library-master/src/IridiumSBD.cpp: In member function 'int IridiumSBD::getSystemTime(tm&)':
.../Documents/Arduino/libraries/SparkFun_IridiumSBD_I2C_Arduino_Library-master/src/IridiumSBD.cpp:237:14: error: aggregate 'tm epoch_start' has incomplete type and cannot be defined
  237 |    struct tm epoch_start;
      |              ^~~~~~~~~~~
...Documents/Arduino/libraries/SparkFun_IridiumSBD_I2C_Arduino_Library-master/src/IridiumSBD.cpp:257:24: error: 'mktime' was not declared in this scope; did you mean 'mktemp'?
  257 |    time_t epoch_time = mktime(&epoch_start);
      |                        ^~~~~~
      |                        mktemp
.../Documents/Arduino/libraries/SparkFun_IridiumSBD_I2C_Arduino_Library-master/src/IridiumSBD.cpp:259:16: error: 'localtime' was not declared in this scope
  259 |    memcpy(&tm, localtime(&now), sizeof tm);
      |                ^~~~~~~~~
.../Documents/Arduino/libraries/SparkFun_IridiumSBD_I2C_Arduino_Library-master/src/IridiumSBD.cpp:259:40: error: invalid application of 'sizeof' to incomplete type 'tm'
  259 |    memcpy(&tm, localtime(&now), sizeof tm);
      |                                        ^~
exit status 1
Error compiling for board Adafruit ItsyBitsy M4 (SAMD51).


This report would have more information with
"Show verbose output during compilation"
option enabled in File -> Preferences.

To receive data, the method is send a blank MO message, with this method(blank message), it will not consuming the credits,
but sometimes with same syntax, it send non-empty message, so it consuming the credits,
i dont know why it sometimes send blank message, and sometimes send non-blank message

I suspect that I have a mismatch in my Time.h library for this driver.

Can someone help point me to the right library or an updated version of the Iridium Driver?

K