Solicited CAN Messages, Anything special required?

So I finally rolled past 100K on the odometer when I can put more effort into my EV conversion. I whipped out the board I designed a couple years ago to sniff the CAN messages and it still does that just fine. However I tried to get the advanced solicited messages, and I got nothing.

I can see my TX being sent and I can see it being received back through the transceiver. So it would appear it is on the bus. And by "see" I mean looking at an oscilloscope I see a waveform produced when I click send.

Maybe I don't understand the method used in the Prius? If you send a CAN message with an SID of 0x7E3, and 2 data bytes of 0x21 and 0xD0 I would expect another message at a later time from SID 0x7EB with the battery module voltages. I am thinking the message is sent but I am not seeing a response from 0x7EB even with a receive filter set to only 0x7xx addresses.

Is there something special that has to be done?

 

Yes I am dealing with it myself however I am using part of a peripheral library for the microcontroller. I am pretty well versed in how CAN itself works from the physical layer all the way to the protocol itself. Like I said, I have it working for reading messages. That would not be possible without getting the correct time quantas and protocol analyze routines. Also when I send messages, I see it being transmitted.

I see the start of frame, the SID, the data byte count, the data, and the end of frame as well as the required time wait. I get nothing back the way I would expect.

That is why I am wondering if I am interpreting the Prius incorrectly. My interpretation is I send out a message to a SID with a couple data bytes set (and of course in the correct SOF/SID/dN/d[0-n]/EOF/I format) and then some time later but within a few hundred milliseconds I should see a message broadcast from another SID with the response.

Is this incorrect?

 

Actually I am filtering it. The board I designed a while back "sniffs" all messages and passes them over a USB connection. I have it working so that I can actually empty the CAN RX buffer meaning I am getting the data over the USB a little faster then the CAN messages are arriving, so I am not missing any message (like overflows). But for testing to make sure I put in the filter for only messages greater than 0x700, and in loopback I could see my message sent properly but when I actually sent it, I got no response.

I am just hooking into the OBDII port under the dash. I would think if I had some basic error I wouldn't get anything. But I can graph everything passively available on the bus at realtime speeds. Drive around with the laptop and see the HV voltage and current for the pack, door status, everything.

The Prius just seems to be ignoring my messages...

 

Well its not ignoring me, something on the bus is actively ACKnowledging the CAN message so it is not errored. I can see it now on another scope and the signal looks great.

 

Also using the scope to trigger on all 0x7xx addresses shows my message but nothing else ever returned...

 

Looking at one side of the differential line (CANH) I see it transmitted and it is clear that something external (or multiple nodes) are pulling it to acknowledge an error free reception. Just nothing is happening in return.

 

As an update I can change things in the car, but I still can't get these "solicited messages".

I can send a 0x529 frame which contains the info to display on the MFD. I can make it display whatever amount of bars I want on the SOC by clicking a button on my laptop. So I know the CAN communication is working.

There has to be something simple and fundamentally incorrect with my method of getting these other data points.

 

Excellent thankyou.

I did a quick once over, and it would appear this:

Code:

  CANSEND &H7E2,0,8,txData(0),txOk

does the dirty work of sending. I am guessing the inputs are CANSEND (SID,RTR(?),DataLength,&Data-index-0, status)?

From this I gather that you are actually sending 8 bytes of data + all the framing.

I was interpreting it as send 2 bytes of data, and those are the two bytes. Your code sends 8 bytes of data with 3 data bytes pop'd the others 0'd. Is this correct?

I will have to go out to car and try this in a few minutes. Good thing I left my scope hooked up and everything!

 

Correct. That I knew. And the higher the SID the lower the priority. The more 0's, the more dominant the message, the higher priority on the bus. That's why investigative commands are so high.

That's the missing piece of the puzzle. CAN allows, and the Prius does, send non 8byte data lengths. The data length of the packets is all over the place depending on the SID.

So when I saw 0x02 0x21 0x0D, I figured that meant send 0x21 and 0x0D for a total data length of 0x02. But really it is send

SOF | SID = 0x7E | DLC = 0x8 | 0x02 0x21 0xD0 0x00 0x00 0x00 0x00 0x00 | CRC = 0x214C | ACK | EOF



and then some time later [5.716 milliseconds] I get:



It is interesting to note that there is a lot of bus idle time between. Perhaps the Battery ECU actually does a conversion after the command instead of sending the last known sample?

 

CAN requires a byte to say how bytes follow. If you did not specify it, then your library did somewhere. I am coding this at a much lower level so it is not done for me.