Reborn 
I have a few questions about the demonstration code for Lightweight Mesh.
Here are two functions:
HAL_UartBytesReceived and
HAL_UartWriteByte.
I not quite sure about the purpose of these two function. I don't know the UART communication is used for RS232 or for the transceiver itself.
Thanks if anyone can gives me some hints
alexru 
UART is not used for transceiver at all. Those functions are used to receive and send application data over the UART, as it follows from their name.
Reborn UART is not used for transceiver at all. Those functions are used to receive and send application data over the UART, as it follows from their name.
It seems like the UART function is used for control the FIFO. Am I right ?
alexru FIFO buffers are used for UART communications, yes. But they still have nothing to do with the radio.
Reborn FIFO buffers are used for UART communications, yes. But they still have nothing to do with the radio.
Thanks Alexru. By the way, I saw your video about the introduction of Lightweight Mesh. Are you the one who created the WSNDemo code for Lightweight Mesh?
alexru I've created everything in LwMesh.
Reborn I've created everything in LwMesh.
The PDF file of "getting started" says the LwMesh SDK contains precompiled binary.
I searched the file entirely but found nothing related with precompiled binary.
Do you know where can I get these precompiled binary?
alexru They were not included in the package because it is very easy to create your own. In version v1.1.0 information about precompiled binaries was removed.
Just open a project for your platform from the application you like in Atmel Studio 6 and build your own binaries.
Reborn They were not included in the package because it is very easy to create your own. In version v1.1.0 information about precompiled binaries was removed.Just open a project for your platform from the application you like in Atmel Studio 6 and build your own binaries.
I think the default configuration for WSNDemo is coordinator type because the APP_ADDR is set to 0.
But when I connect my PC to the board and use WSNMonitor, noting displayed on the main panel. Does it mean I need to make further modifications in order to get my devices displayed on the WSNMonitor?
alexru What board are you using? Have you checked Fuse settings?
It should just work if the hardware is set up correctly.
Reborn What board are you using? Have you checked Fuse settings?It should just work if the hardware is set up correctly.
I'm using STK-600 + ATMEGA128RFA1, I change the fuse bit based on the documentation provided by the SDK
alexru STK600 requires to connect UART port pins to the on-board level shifter. Did you do that?
Reborn STK600 requires to connect UART port pins to the on-board level shifter. Did you do that?
Yeah, I'm using RS232. Sorry, I thought the fuse bit setting between BitCloud and Lightweight Mesh should be identical. But they are different. I change the fuse bit setting based on LwMesh SDK, it worked.
alexru The only [important] difference is CKDIV8 bit and in the latest release of LwMesh it is no longer important, stack clears this bit. Download the latest version.
Hi, I'm trying to use WSNMonitor to learn about LW Mesh. I have programmed the coordinator with the correct fuses (provided by the SDK) and running the WSNDemo app. However, when I click on connect (on WSNMonitor), with the correct COM selected (and also correct speed 38400), I get the message Can't Connect, and sometimes, even this message doesn't appear. Please, do you have any other idea of what is happening? Thank You :D
alexru Operating system? Hardware platform?
Thanks Alex. I am using my own hardware, that has an Atmega1281 and a AT86RF231, and the communication with the computer is made using Atmega8u2 connected to the Atmega1281 by UART. The computer recognizes the board as an Arduino Uno because it was made using the arduino schematics. My friend who developed the board hardware could read data through UART, using this board, so I guess the hardware is ok. I know that using the own hardware is not the ideal scenario, but it's for a university project. Thank you
alexru So it looks like WSNMonitor can't connect to the Xmega CDC. I have no idea how to solve this. It might be drivers that OS picked up for this device, or how Java interacts with those drivers.
Try to open this port in a terminal and see if you get any output that way.
Yes, I tried to do this, but I got nothing. I don't know how WSNMonitor works, but I have to find a way to read the data using the UART. I will keep working on that, thanks Alex
alexru If you get nothing in the terminal, are you even sure that code in Xmega works correctly?
Before running WSNMonitor, make sure that you get something on a simple terminal.
A few things to check on the mega side:
1. HAL_UART_CHANNEL is set to a correct value.
2. Fuses are set correctly.
3. MCU and radio are connected correctly.
Yes, sorry for delat to answer, I was travelling. HAL_UART_CHANNEL was wrong but I fixed it before posting here. Fuses I have setted accordingly to the guide of LWMESH:
Extended: FE
High: 9D
Low: C2
And about the connection between the mcu and the radio, I found the code to read a register from the radio (PART_NUM) and I have done (if value==3) where value is the value read from the register, the led is on. So.. the led went on.. so I think the connection is okay.. but it's still not working with WSNMonitor..
Thanks
alexru Try to open corresponding COM port in the terminal first and see if you receive anything, then move on to WSNMonitor.
Yes.. I will do that.. if I have success I will let you know.. Thanks Alex
Alex, should WSNDemo application output anything to the UART after the MCU is reset or does it wait for an initial request from WSNMonitor?
I could estabilish a connection with port COM3, but nothing appears on screen.. :(
Alex, what should I get on the terminal when I connect to the COM port? what kind of information ? Thanks
alexru Usually it s a bunch of garbage characters, but if baudrate is set correctly, among those characters you will see a device name, for example "Coordinator" or "Router".
If you have only one device, then you will only see "Coordinator".
The good sign would be to see anything in a first place.
Yes, for the first test i'm trying to see anything.. but it's not working.. I will try to solve this problem. Anyway, let me ask you something. I'm trying to study the code, but I ha ve seen this part:
#if defined(PLATFORM_RCB231)
DDRC = (1 << 4) | (1 << 6) | (1 << 7);
PORTC = (0 << 4) | (1 << 6) | (1 << 7);
#endif
why you configure those bits on port C ? I'm asking because I think maybe it can be important since I'm using a different hardware.
Thank you :D
alexru This is for RCB board with RCB_BB extension boards. RCB_BB uses MAX232-like level shifter that has its pins connected to the MCU, so MCU needs to enable level shifter before use.
It is not likely that it affects you. RCB+RCB_BB is the only combination that I've seen that uses this method.
Humm.. So.. maybe I can comment this lines of code.
Thank you Alex, this LWMESH is a really good job, I'm impressed as long as I go studying the code haha I wish I could do something like that some day.
Alex, I tried to find a code to test only the Uart communication of my board. It worked.. I could write something and then see the echo in the terminal.. but when I use the WSNdemo code I don`t get anything on the HyperTerminal.. maybe I have to change something in the UART configuration of this code.. I will try :D
alexru Are you sure it is an echo from the device and not a local echo from the terminal?
Can you post your test code?
No, I`m not sure, but I think it is right. I just used the source code from this article: http://www.appelsiini.net/2011/s...
The only change that I have done is change F_CPU to 8 instead of 16
alexru You should start from setting HAL_UART_CHANNEL to 0 (default is 1).
Also note baudrate difference (38400 vs. 9600).
Yes, I did this change (HAL_UART_CHANNEL 0) in the config.h file .
About the baud rate, if I put 9600 on WSNDemo code it will generate a problem, no ? On hyper terminal, a connected using 38400 when running WSNDemo.. I will try one more time. Thanks a lot Alex
Alex.. one little thing.. where is defined the value of F_CPU in the code? I searched the information but I couldn't find.. Thanks
alexru It is defined in the project settings or in the Makefile, depending on what system you are using to build the application.
I am using Atmel Studio 6.1 on windows 7. Do you think I have to use the option "Use external makefile" when compiling the code?
alexru No, just look in the existing project settings. There will be a section with compiler definitions.
Thanks. About that problem on UART, I'm running the code with debbuger, step by step, and the code is getting stucked on this part:
void HAL_TimerDelay(uint16_t us)
{
PRAGMA(diag_suppress=Pa082);
OCR4B = TCNT4 + us;
if (OCR4B > OCR4A)
OCR4B -= OCR4A;
halTimerDelayInt = 0;
TIMSK4 |= (1 << OCIE4B);
while (0 == halTimerDelayInt);
TIMSK4 &= ~(1 << OCIE4B);
PRAGMA(diag_default=Pa082);
}
the application does not exit from that while condition.. it seems it stays forever in that condition. So I think that is the reason why WSNMonitor does not recognize the board.. because the code is not running to the end.. or, maybe I'm wrong.. I don't know very much about this
alexru Try a newer version of the stack. I don't know any specific reason why this code won't work for you, but the new one is different and it might just solve the problem.
Thanks, I promise I will try to stop bothering here :) but I'm kind of running out of time. I'm using version LwMesh_1_1_1 . I will try to use a newer version though, if this one is not the newer. I'm trying to understand the code.. some parts I don't understand. For example, I'm reconfigured de LED0 Port.. to fit on my pins.. so LED0 is conected on Port A2 on my mcu, I did:
#if defined(PLATFORM_RCB231)
HAL_GPIO_PIN(LED0, A, 2);
but the led is not turning on (the led is working, I have tested with another simples code). So when I checked the LedOn method:
INLINE void HAL_LedOn(uint8_t i)
{
if (0 == i)
HAL_GPIO_LED0_clr();
.. and this method erases the bit, not activate. Maybe I'm understanding it wrong because I'm a begginer.. Thanks
Ahh! now I see that version 1.2 has been launched.. I will take a look on it! Thanks
alexru I'm using version LwMesh_1_1_1.
.. and this method erases the bit, not activate. Maybe I'm understanding it wrong because I'm a begginer..
If your board uses different LED connection, then you will also have to change '_clr' to '_set' and vise versa.
Oh.. I didn't think about it.. you are totally right! I will do this.. Thanks, and sorry for bothering too much :D
Omg!! The UART now is working!! I can see Coordinator written on hyper terminal, and also the led is working too with the change clr to set!! I can't believe it! haha Thanks Alex :D
Now the coordinator appears on screen. I configured one board with Adress 0x0001, so is a router. It appears on screen too, and a gray line between the router and the coordinator is shown, it works fine. But I configured another board as router too, adress 0x0002, it appears on screen, but no gray line and after about 20 seconds it disappears, but the led indicating that the board is connected keeps on, without blinking... do you have any idea why this second router is disappearing ? Thanks
alexru do you have any idea why this second router is disappearing ?
Run that second router under the debugger. and see what it is doing.
The gray line will only appear on the second report, so it does not look like that second report is happening. Device sends one frame and then just hangs.
Humm ok.. I will try this. but I'm almost sure it's a hardware problem, soldering this small components at home is really dificult.. I will also try to check the solder.. Thanks :D
alexru On the previous version of the stack, I would have assumed hardware problem instantly. This behavior is indication of the interrupt malfunction. On the new stack interrupt line is not used, so it is harder to blame the hardware. But it still might be something like power supply problem.
I will try new batteries and check the solder.. because it doesn't make sense not working, because the first board has exactly the same configuration, just the adress was changed.. and it's working fine
Alex, I don't know if you can help with this.. and I even don't know if it is the correct topic for it, but taking advantage that we were already talking about this, it is worth trying. The situation is.. I have the coordinator, it works fine.. appears on WSNMonitor, and I also have 3 boards that are configures as routers.. board 0x0001 0x0002 and 0x0003.. they have exactly the same hardware (including the coordinator).. So, when I put the battery.. board 3 connects just after plugin the battery... and the led keeps on, indicating connection was done and it also appear on WSNMonitor. The thing is, boards 1 and 2 just keep blinking and don't connect... I know it's a hard question but.. do you have any idea what's going on? the mcu and transceiver are communicating... and I used a spectrum analyzer to see if any RF signal are emitted by the boards.. and it's ok.. they are emitting. The only parts in the hardware are mcu, transceiver, the oscillator, and some capacitors that shows on AT86RF231 datasheet.... Thanks :D
alexru I would say that you need a debugger and have a look at the status code returned in the appNwkDataConf() function. After that it should be possible to track back to the reason for that status code.
Oh.. I have a debugger.. I will try to do that and compare with the same status code returned on the board that works fine.. Thanks
alexru There is no need to compare. On the board that works fine the status is going to be SUCCESS (0). Just let me know what the status is on the board that is not working.
I'm trying to use the search tool to find insert the break point (in Entire Solution) but AtmelStudio returned that no entry was found :(
alexru I don't understand what do you mean.
Sorry I wrote wrong.. I searched the function appNwkDataConf on AtmelStudio search tool, to insert the break point to use the debbug.. but this function couldn't be found on the code, I don't know why
alexru Don't blindly use search, use logic. The function is called appDataConf() and located in the application, as name would suggest.
You always need to go back from the observed behavior. In your case it is "LED does not go steady". Then start looking for something that makes it blink. By doing so you will end up in that function.
Yes.. that's true.. I'm worried about time, I'm afraid I can't do this on time, that I couldn't see the logic.. I will work on it and post here what I find out.. thanks
Sorry Alex, I can't run the debugger because the device is not reading the correct voltage of mcu... it's showing 1.1V but when I put the multimeter it shows 2.6V.. maybe it's a soldering problem, I will try to fix that .. thanks
Now I could run it.. I used watch on this var appState = APP_STATE_SENDING_DONE; and inserted a break point exactly on this line of code.. so I run the debug and before this line, the value of appState is APP_STATE_WAIT_CONF and after this part it becomes APP_STATE_SENDING_DONE .. I don't know if is this what you told me to do.. Thanks
alexru You need to check the value of req->status in the appDataConf().
Hi Alex.. I don't know if this is a good test, but in this part of code:
if (NWK_SUCCESS_STATUS == req->status)
{
if (!appNetworkStatus)
{
HAL_LedOn(LED_NETWORK);
SYS_TimerStop(&appNetworkStatusTimer);
appNetworkStatus = true;
}
}
else
{
if (appNetworkStatus)
{
HAL_LedOff(LED_NETWORK);
SYS_TimerStart(&appNetworkStatusTimer);
appNetworkStatus = false;
}
}
I have inserted a break point on the first if condition, what I think it is "if connection was successful, enter in the first if, if not, then execute the else command" right ?
So, I went step by step debugging, and on the board that works fine, the debug goes into the if condition if (!appNetworkStatus), so.. it's working.... But on the boards that are not connecting.. the debbug goes into the else.. so they are running different..
And when watching the value of req->status:
On the board that works:
req->status = 0
On the board that doesn't works:
req->status = 17;
alexru 17 is NWK_NO_ROUTE_STATUS (0x11). What settings do you have in your configuration file?
The standard configuration.. I have just made this changes:
#define APP_ADDR 0x0001
and I also have uncommented this line
#define NWK_ENABLE_ROUTE_DISCOVERY
alexru and I also have uncommented this line
#define NWK_ENABLE_ROUTE_DISCOVERY
Well, have you uncommented it on all other devices in the network?
All devices in the network must use the same route discovery mechanism.
Yes, I did that.. I first uploaded the code to the coordinator with this lines uncommented, and after that I programmed all boards just changing the address
alexru First of all, check if how it works without NWK_ENABLE_ROUTE_DISCOVERY.
Debugging AODV discovery without a sniffer might be hard.
So I shall comment again this lines on the coordinator and the boards right? I will do this now
Same thing, it didn't work.. I just don't understand why board 3 works fine.. And also the hardware of boards 1 and 2 seems to work fine.. I used a spectrum analyser.. and they are emitting RF signal... is possible to emit RF signal and the transceiver is not working? Thanks Alex
alexru It does not matter if it works or not. What is the status code?
Also, can you show pictures of your boards, antennas, schematics.
When running the debugger with Route Discovery disabled, the code seems to be stucked somewhere, so I couldn't get the status code on the router, just on the coordinator and it was 16... Also.. I uploaded the hardware schematic and board:
Schematic: https://dl.dropboxusercontent.co...
Board: https://dl.dropboxusercontent.co...
Just a note.. on the schematic it's showing Mega128A.. but I have replaced by Atmega1281.. that's because it's and old schematic..
Thanks
alexru When running the debugger with Route Discovery disabled, the code seems to be stucked somewhere, so I couldn't get the status code on the router,
just on the coordinator and it was 16...
As far as schematic goes, I don't think you need C3. And of course, it is always possible that antenna is not matched.
But you need to take care of that loop first.
I wasn't responsible for the antenna development.. my friend did that, and a teacher helped him, I think it's not the ideal antenna but it works since board 3 can connect at a good distance.... Yes.. I will try to find out why board 3 is not stucked and works fine and board 1 and 2 are stucked.. I hope it works :D thanks a lot Alex
alexru Theoretical design of the antenna is just a start. After it is implemented, antenna has to be tuned and matched.For tuning a good design helps, but does not guarantee a success on a first try. Matching depends on the components used, quality of soldering and other external factors that cannot be predicted at design time.
But all this usually does not matter no short distances and won't cause device to freeze.
Huum.. it must be hard to develop a good antenna.. but since I don't need a big distance, there's no problem if they work close to each other.. I just need to find out why some boards works and others don't.. Now I have 3 boards that can connect.. and 2 that can't..
Alex, as you said.. on the boards that doesn't work, the req->status is 17, which I can see on the code that means NWK_NO_ROUTE_STATUS.. it means this board can't find the route? or can't "see" the coordinator.. or something like that? Thanks
I have read this on the manual:
Then node should create a new entry in the Route Discovery Table with parameters from the current data transmission
request. If the Route Discovery Table is full and a new entry cannot be added, then data transmission request should be
confirmed with NWK_NO_ROUTE_STATUS status.
So, maybe it is not creating and entry on the router discovery table, but I don't know why since there is only one board turned on. Maybe I should program the coordinator again with a bigger router discovery table ?
alexru On a high level, it can't find the route. This process consists of sending a request and receiving a response. Any of there can fail.
Without route discovery, device will send a broadcast request and wait for an ACK. This lets you check the system part by part. If C will receive the frame, it means that transmitter is working. If R won't receive an ACK, it means that receiver is broken.
If some error status code is returned, then it will tell exactly how any of them is broken.
Huum.. Thanks, I will try to disable route discovery and try again to debug :D
I tried to debbug with Route Discovery disabled, and the debbuger seems do be stucked in this part of the code:
appMsg.sensors.battery = rand() & 0xffff;
appMsg.sensors.temperature = rand() & 0x7f;
appMsg.sensors.light = rand() & 0xff;
I don't think it means the Tx has a problem, because I think before sending this random datas, the board sends another code to sink or something like that.. maybe I'm wrong
Ah! It was the transceiver... Although it seems to work, I replaced it and now board 1 is working too :D
alexru There is no way to get stuck in this code even in theory. There is no loops here or in called rand() functions.
I see.. but I clicked next step and the debugger does not advance to the next step.. but now that I changed the transceiver board 1 it's working.. I will try the same with board 2 :D
Now they are all working! Thanks Alex... I think they are working fine now. In the end it was a hardware problem. In WSNMonitor each node has this kind of lamp that when you click it turns yellow. Like in this print screen that I have taken: https://dl.dropboxusercontent.co... ..
what is this for? haha Thanks.. now I'm going to read all the SDK manual to learn how to develop my application..
alexru This is lamp sends a command to the node and it blinks. In the current version of WSNDemo it is not supported. It is supported in the next version, which I just submitted for review, so it should hit the web within a week or two.
Huum, since my board has only 1 led I think I cannot see it, but the code I will try to do involves sending a command to a node.. so maybe I can see how it works.. but it is for the next weeks.. firts I have to read the manual as I said.. thanks :D
Hi Alex.. I'm reading about the 2 routing algorithms.. Native and AODV, let just me ask you a opinion... which one you think is better for a small network? about 5 nodes? The only requirement is that the network has to be self organized.. Thanks
alexru For small node count (let's say less than 20) it does not really matter, they both will works almost the same. For larger networks AODV is better.
If it does not make any difference for your application, use AODV, since it is better in theory.
Huum.. yes.. thanks :)
Is there any place where I can find a graphical explanation of AODV, like the one for Native Routing on developer's guide? I'm trying to create one based on what I understood reading the AODV Routing, but I'm not sure my understading is correct. Thanks
Alex, on AODV I'm not sure about something.. is there one Route Discovery Table for each node or just one for the whole process of discovering ? I have read the page many times and I still can't be sure... Thanks
alexru Is there any place where I can find a graphical explanation of AODV, like the one for Native Routing on developer's guide?
Here you go.
Comments to slides about AODV:
Numbers on the links show LQI of a link
We will focus on the node 0, the rest of them behave similarly
Step 1:
Numbers on the arrows -
Route Discovery is a local broadcast frame, each node creates a new Route Request frame with updated LQ information.
Node 0 sends out a Route Discovery Request frame with LQ set to 255, since it is an original frame.
Nodes 1, 3 and 4 receive this frame and update LQ value based on the received LQ and LQI of the received frame.
Step 2:
Nodes 1 and 3 send a Route Discovery Request with updated LQ value
Node 4 sends a Route Discovery Response to the node 0 with a final LQ of the route 0-4.
Node 0 receives the response and updates Route Discovery Table
Step 3:
Node 2 send a Route Discovery Request with updated LQ value
Node 4 sends a Route Discovery Response to the node 3 with a final LQ of the route 0-3-4.
Step 4:
Node 4 sends a Route Discovery Response to the node 2 with a final LQ of the route 0-1-2-4.
Node 3 routes a Route Discovery Response for the route 0-3-4.
Node 0 receives this response and updates its Route Discovery Table since LQ in the response is higher than LQ in the table.
Step 5:
Node 2 routes a Route Discovery Response for the route 0-1-2-4.
Step 6:
Node 2 routes a Route Discovery Response for the route 0-1-2-4.
Node 0 receives this response and does not update its Route Discovery Table since LQ in the response is lower than LQ in the table.
Route discovery is complete. Route 0-3-4 was selected as the most optimal.
Information from the Route Discovery Response is transferred into the Routing Table.
alexru is there one Route Discovery Table for each node or just one for the whole process of discovering ?
Huum, I think now I can understand everything.. Thanks
alexru Also, AODV is not specific to LwMesh. It is a well-described protocol - see RFC 3561. Frame formats are obviously different, but the idea is the same.
So... it means in AODV once the best route is selected it never changes right? So, let's say the nodes change their physical location, is better to use native routing algorithm right ?
alexru So... it means in AODV once the best route is selected it never changes right?
So, let's say the nodes change their physical location, is better to use native routing algorithm right ?
Also, AODV is not specific to LwMesh. It is a well-described protocol - see RFC 3561. Frame formats are obviously different, but the idea is the same.
Ah.. I see, I studied a little bit of RFC in the past.. I will take a look again
Alex, what really is an EndPoint on the stack context?
alexru If you mean "endpoint" in data request parameters, then it is a logical address of the application inside the device. Device might run multiple logical applications running in parallel. One might handle main traffic, another one might handle network discovery and commissioning, yet another one might be handling over the air updates.
That's what Endpoint means in network header, the fields Source Endpoint and Destination Endpoint? And about this definition of endpoint con config file?
#define APP_ENDPOINT 1
Sorry, but I just have to be sure what I'm talking about.. Thanks Alex
alexru Yes, it is. Most applications are so simple that they only need one endpoint. For more complex applications you will need more.
Note that EP=0 is reserved for stack commands and should not be used in the application.
Huum I see.. I will try to understand better this concept of Endpoint, it's new to me.. thanks again
alexru Well, if you like RFCs, endpoints are like ports in TCP or UDP.
Huum.. with this comparison it became much easier to understand.. thanks :D
Alex, if I develop my own application using the LwMesh SDK, let's say my application is for example, when I press a button on one of my nodes, a led blinks on my coordinator. If a write the code to do this task, Will I still be able to see the nodes on WSNMonitor ? Or WSNmonitor only works for WSNDemo application ? Thanks
alexru WSNmonitor only works with WSNDemo. It displays stuff based on the information supplied by WSNDemo via a serial port. If you can mimic that information, it will display your network as well.
Hum, Ok... I have to simulate a security system with the nodes, I mean, I have soldered a button on the nodes and my idea is when a press the button, let's say, on sensor 0x003 then led blue blinks on my coordinador ( I will solder pins and use a protoboard), when button is pressed on node 0x002, led orange blinks.. and goes on. But I would like to still be able to see the network topology on WsnMonitor, but for now is just an idea, I don't even know if it's possible to do this system. Thanks :D
alexru Of course it is possible. You need a very simple modification to the WSNDemo program. In the data indication on the coordinator check the source address of the sending device and light an LED accordingly. Everything else will work as it worked before.
Yes, I was think exactly this, doing a simple modification.. Although for my graduation project I just have to show that is possible to monitor an enviroment using wireless sensors, I would like to know how to do more complex application because I'm really enjoying this. Thanks
Hi Alex, just a theorical question, to check my thoughts.. why AODV is better with a greater number of nodes? That's because once the route is estabilished it does never change (unless necessary due to node problem of course) so the traffic is not "flooded" with routing process every time ? Thanks
alexru No, both algorithms remember discovered routes. Run-time optimizations in the native algorithm will lead to longer routes. Even with AODV it is not unusual to see 2-3 hops for devices located in one room. With native algorithm it gets even worse.
Over a long time AODV will stabilize on shorter routes and ideally no route discoveries will be done. Native on the other hand will always tend to "optimize" towards longer routes.
Ah.. I understand, thanks for the awesome explanation :D
Hi Alex, I'm trying to develop that code that I told yo, but for now is just a sketch. I am trying to understand how to receive and process data. Let me tell you I have done and if you can, please tell me if I'm on the correct way.
on appInit(), I added this line:
#if APP_COORDINATOR
NWK_OpenEndpoint(APP_ENDPOINT_COORD, appDataIndCoord);
#endif //where APP_ENDPOINT_COORD defined as 2
and under the appDataInd, I created appDataIndCoord function, as follows:
static bool appDataIndCoord(NWK_DataInd_t *ind)
{
NWK_DataInd_t *msg = ind;
DDRC |= (1<<0)|(1<<1)|(1<<2)|(1<<3)|(1<<4)|(1<<5);
if(msg->srcAddr == 0x0001){
PORTC |= (1<<5);
_delay_ms(1000);
PORTC |= ~(1<<5);
}
return true;
}
This is the correct idea? or I understand wrong? Thanks a lot
Sorry to ask, but, since my hardware is "homemade" I am afraid that if I reprogram many times to test, It will broke down.. so I'm trying to make as few mistakes as possible
alexru 'ind' already has type NWK_DataInd_t, there is no need for 'msg', just use ind->srcAddr.
Delaying for a second like this will work for a test, but should not be done in a real application.
I don't understand how you can break the hardware with reprogramming. Guaranteed Flash endurance is 10000 erase/write cycles. And it is much higher at a room temperature. I can not imagine anyone exceeding this limit. Even if you will reprogram your board every second, it will take about a week to reach that limit.
Yes, but I don't know why, when I try to program let's say for 3 or 4 times, Atmel Studio starts to read the wrong voltage, or says it was unable to enter in the programming mode.
I would ask you exactly this, about this delay.. I suppose it might cause some problems because it stops the processor just to blink the LED... so I was thinking to use interruption, but I don't know how to work with interuption yet.. I will have to learn.. thanks :D
But the main idea is correct right? To open an endpoint with number 2 and write a new corresponding appDataInd for that EndPoint ?
alexru I don't see why you would need to open another endpoint for this application, but there is no harm in doing that. Overall the idea is correct.
It is better to fix the hardware or get the one that works. You won't learn much by guessing, only by doing.
Yes I think so, I have plans to do that, after finishing this project, but for now I have a deadline, so I have to hurry, and import this kind of hardware to my country is quite expensive, besides, I don't have time to wait them to arrive :( Thanks.. I will keep trying here
Alex, on the router nodes..I should handle the button that I want to press to blink onde led on coordinator as a interruption? Otherwise I was thinking to put the code "if button == 1" etc inside appSendData() but somethings tells me that it's not a good idea, because it will only work if the code is processing that part of code.. no? Thanks
alexru I did not understand any of this.
Haha.. Sorry, I was excited when writting... the question is.. the place there I have seen where Nodes sends data, is the function appSendData(), I need to send another data, different from Light Temperature, etc.. I want to send one frame containing let's say number 1 (just to make coordinator check the srcAdress and flash de LED corresponding to node 1). But this frame has to be sended only when I press the button on my node. Should I write this code inside appSendData() or a new function? haha sorry, I'm trying now
Please, Let me just say what I've done and if possible give your opinion. I have created another appSendData() function called appSendDataButton() and the only change was the Endpoint, I have setted as 2 (because on appDataIndCoord I am using another end point.. 2)
nwkDataReq.dstEndpoint = APP_ENDPOINT_COORD;
nwkDataReq.srcEndpoint = APP_ENDPOINT_COORD;
and then on the switch case that controls the state machine, I have added:
case APP_STATE_SEND:
appSendData();
#if APP_ROUTER
DDRA |= (0<<3); //set PortA Pin 3 as input
PORTA |= (1<<3); //set pull up resistor
if((PINA & (1<<3))==0) //where button is connected
appSendDataButton();
#endif
} break;
I think it's not the best way to implement this, since this way I have to press the button exactly when the code is on the if clause.. but even though is not working, but I don't know if it's problem on the code that sends or the code on de coordinator that receives... Thanks
alexru Check how button read is implemented in the SimpleRemote application.
This code, if run, will cause nwkDataReq to be used twice before waiting for the confirmation. If at any point in time you want to send two requests in parallel, then you need to have two different NWK_DataReq_t structures. Alternatively you can introduce a flag that indicates if structure is busy and won't allow double use.
Why are you looking at the button state only in the SENDING state? It will make more sense to look in a WAIT state.
"DDRA |= (0<<3);" does absolutely nothing, but pin is an input by default, so it should work. I'd recommend using macros from halGpio.h to do GPIO operations.
Huum, yeah, I didn't understand very well about APP_STATE_WAIT_CONF, I mean, I didn't find yet exactly where it changes appState to a new state...
I will use macros from halGpio.h it's really better, and I will also take a look at the application that you suggested.
Can I send 2 differents NWK_DataReq_t structures inside appSendData?
alexru APP_STATE_WAIT_CONF denotes that request is pending and application is waiting confirmation. Confirmation function is appDataConf(), which changes state to the next one.
You really need to understand how state machines and even-driven design work. Until you do, you won't be able to write sensible applications. At least you need to have a clear understanding of WSNDemo sate machine.
The state application spends mode of the time in is APP_STATE_WAIT_SEND_TIMER. And that's the state where it makes sense to wait for a button press.
Can I send 2 differents NWK_DataReq_t structures inside appSendData?
I'm really trying to understand the state machine.. I know without understand it I will not be able to do what I want. But some parts I couldn't understand very well yet... like when appDataConf is called, and then the state is changed to APP_STATE_SENDING_DONE in the end. So, ok.. and then? the program should go back to appSendData where appState is changed to APP_STATE_WAIT_CONF later.. and then I can't see where appState is changed again to another state.. But it's not fait to ask explanation about this.. I shall study and find out... Thanks Alex
alexru So, ok.. and then?
Then handler for this state is executed:
case APP_STATE_SENDING_DONE:
{
#if APP_ENDDEVICE
appState = APP_STATE_PREPARE_TO_SLEEP;
#else
SYS_TimerStart(&appDataSendingTimer);
appState = APP_STATE_WAIT_SEND_TIMER;
#endif
} break;
End device goes to sleep, router starts a timer and goes to a timer waiting state.
Then timer callback (appDataSendingTimerHandler) is called and it transitions application into a APP_STATE_SEND state.
That's why I was not understanding.. I didn't know what time callback handler was for (SYS_TimerStart(&appDataSendingTimer);)..
it like when appState = APP_STATE_SENDING_DONE, then after a time appDataSendingTimer is called and the state is changed to APP_STATE_SEND again.. that's it ? Thanks
alexru I didn't know what time callback handler was for
appDataSendingTimer.handler = appDataSendingTimerHandler;
that's it ? Thanks
Yes, I could find it, but I couldn't guess what was it for, I'm not very advanced in microprogramming, I'm a beginner.. Thanks Alex.. I will study more and try to do the modifications
Ah.. it's kinda of working now... it's blinking the led on coordinator.. although I have done it in the most correctly way.
I have done this:
static void appDataSendingTimerHandler(SYS_Timer_t *timer)
{
uint8_t buttonState = HAL_GPIO_BUTTON_read();
if (APP_STATE_WAIT_SEND_TIMER == appState)
if(buttonState==0)
appState = APP_STATE_SEND_BUTTON;
else
appState = APP_STATE_SEND;
else
SYS_TimerStart(&appDataSendingTimer);
(void)timer;
}
and created a new state:
case APP_STATE_SEND_BUTTON:
{
appSendDataButton();
}break;
I Think it's not perfect because, it's not all the times that I press the button that the led flashes on coordinator, so it might be due to delay of transmission, or the code wasn't on the part that read the button when I pressed.
alexru Please use code tags around your code, otherwise it is unreadable.
This is a very strange code. It will only sample button on the timer boundary. Usually button presses should be handled at the moment of pressing, not on some relatively long timer edge.
You also need to put braces properly in the conditions. Right now logic is not correct for what it supposed to do,
Yes.. It's strange haha although is the only when I can think now, but of course I will try to do better.
I don't know the best place to put the code to read the button, I will think more about it...
And, you are right, the logic is wrong, I forgot to put braces.. Thanks :D
Alex, just for understand a bit more.. this function
static void appDataSendingTimerHandler(SYS_Timer_t *timer)
it's kind of keep running in parelallel with the code checking every time if appState = APP_STATE_WAIT_SEND_TIMER ?
alexru No, this function is called after a specified time interval. Nothing in the system runs in parallel.
"If" condition in the function is there to prevent invalid state machine transition. Normally by the time it is called appState will always be APP_STATE_WAIT_SEND_TIMER.
Ah.. I understand now :D.. now is working.. the 4 leds on coordinator are Toggling when I press the button on each node :D Thanks very much Alex...
IT's working, although I might find a better place to read the buttonState.. because sometimes I have to hold the button for it to send the frame.. or maybe is just because I haven't done any process of debounce on the button...
How can we use HAL_UartWriteByte() function to write a float value to the UART