RPM overflow

[MaxYoshida]

Thanks for the amazing library. Looks like a great potential!

I wired arduino to HC-05, which is paired to a good quality ELM327. I tested some parts of the code. Those are coolant temperature, RPM, vehicle speed, throttle, fuel level, engine load.

They all worked very nice except fuel level which always showed 0, but I think it could be an issue in my car itself as fuel level didn't show as well using Freematics OBDII and their library.

My issue is in RPM, it shows "ovf" and some weird random numbers between 54 - 110 while I am idle (about 720 RPM in reality). Increasing RPM also didn't work for me.

Any ideas why this happens?

Thanks in advance

[PowerBroker2]

Can you run your sketch in debug mode and post the both the printout and the sketch itself? That will give me the best idea as to what is going on.

[MaxYoshida]

Thank you for your response.

Sure. I attached the response I get while debug is on.

Code is basically same as hardware example provided with the library. I just changed baud rate to 9600. As I am using Arduino mega, I attached HC-05 to Serial3. Anyway I attached the code I used to get this debug. You will notice that I commented functions that works fine.

Other things I tried with the RPM: Using same ELM327 module, I connected it to 2 android apps and I see the RPM correct 100%.
I tried Freematics OBDII reader with the library to obtain RPM, works perfect.
I tried this arduino code: https://www.instructables.com/Arduino-OBD2-ELM327-I2C-LCD-HC05-Bluetooth/ (After modifying PID number and using correct equation from wiki which is ((256 * A +B) / 4)), it works fine as well.

Hope these help out in providing a better idea about the issue.

res.txt
code.txt

[MaxYoshida]

I believe I found the problem.. Couldn't easily find it before because it was basically 2 issues combined together.

Issue 1: RPM requires a delay of 400mS at least between readings - At least on Opel Zafira A model 2002 vehicle, not sure about other cars. Adding this as a timer check in the code fixed this problem (or using delay(400); works as well).

Issue 2: This is in the library itself. Inside rpm() function, I see the code:

return conditionResponse(findResponse(), 2, 1.0 / 4.0);

I modified it to:

return conditionResponse(findResponse(), 1, 256.0 / 4.0);

And everything worked. The wiki page indicates that RPM equation is ((256*A)+B)/4
Of course with what I did, I ignored B, which makes reading error by 64 RPM. (Maximum B parameter value effect, which is 256/4)

Would you kindly consider fixing this in the library, and maybe adding a timer for RPM to make sure it is read every X mS where user can change that in code? Would really appreciate this as it would help others.

If I am mistaken in anything I said, or I got anything incorrectly, kindly inform me!

[PowerBroker2]

Up front I want to say I do not give any guarantees that this lib will work or be compatible with cars made before 2008, since that's when OBDII essentially became a standardized interface for virtually all cars.

Issue 1: RPM requires a delay of 400mS at least between readings - At least on Opel Zafira A model 2002 vehicle, not sure about other cars. Adding this as a timer check in the code fixed this problem (or using delay(400); works as well).

I'm not aware of delay requirements like this that aren't handled by the ELM327 chip itself, but you might be correct here. Easy fix at the sketch level for sure.

Issue 2: This is in the library itself. Inside rpm() function, I see the code:

return conditionResponse(findResponse(), 2, 1.0 / 4.0);

I modified it to:

return conditionResponse(findResponse(), 1, 256.0 / 4.0);

The library's RPM function from what I can tell is not at fault, but either the fault of your car or ELM327. Here's why: From the debug statements I can tell you're sending the correct query 010C. The ELM327 is then responding with 410C0CD41E, which is not an appropriate OBDII response. A correct response would be 410CXXXX where the X chars are the RPM data in hex (see here). Your ELM327's response has 2 extra hex chars. From the correction code you posted, it seems the last two chars in the response are extraneous.

Again, as far as I can tell, it isn't an issue with the library, but with the car/ELM327. What I can do, however, is make conditionResponse() a public member and let you write your own RPM function at the sketch level (do not edit the library!).

[PowerBroker2]

Release 2.6.3 will now let you make that custom rpm function like this:

float rpm()
{
  if (myELM327.queryPID(SERVICE_01, ENGINE_RPM))
    return myELM327.conditionResponse(myELM327.findResponse(), 1, 256.0 / 4.0);
  
  return ELM_GENERAL_ERROR;
}

[MaxYoshida]

The response file I sent while debug is on did not respect the time delay (Which I have no idea why it needed to be 400mS here. On Freematics sensor 180mS was enough for RPM reading. Maybe chips handle delays differently?)

I would like to take another debug sample tomorrow and post it. Just to be sure what effect the 400mS delay will have on the ELM327 response.

Again thanks for your support so far. Let's hope what we find here helps someone else.

[MaxYoshida]

Ok, this took few hours, but here is what I found: Freematics OBD is faster as it uses more powerful processor and communicates by cable. My ELM327 communicates by bluetooth and it is slower to receive a response from it through HC-05. This means on the car I used for the test and specific modules, I found that there is a minimum time delay to get a response from the car. Time is as next:
ELM327 + HC-05: 375mS, anything less and the car will start sending nonsense
Freematics OBDII UART: 180mS, anything less there is no valid response
Can bus shield from DFRobot: 305mS, anything less is not a valid response

All tests were made on Arduino mega2560

Regarding the RPM value, I still have no idea what's happening.. But sometimes it doesn't work at all with the library and sometimes it does, while using another code and same hardware (HC-05 + ELM327) it works fine. Maybe there is a missing command for ELM327 or the car that is only required in some cars such as Opel Zafira 2002?

I think that's all I can get in debugging. Its up to you to close the thread if you find that we can't add anything else to it.

[PowerBroker2]

I think it would be interesting to look at the debug printout of a run where you do have that ~375ms delay AND it still doesn't work.

[MaxYoshida]

Sure. I attached it. But it is with 500mS delay.

At this point there is 1 thing I could think of.. there is a missing configuration. If I recall correctly, when I jumped between codes trying to make it work, sometimes RPM did work as is in your library without any modifications to it, sometimes didn't on the same exact code! What makes me unable to think is that other things I tried do work, problem is just in RPM and it never appeared on Android apps or other codes I tried.. At same time, the other code I tried (Attached it) gave me a very weird issue.

When trying that code to read a single PID in a loop, it works just perfect, but when I ask it to read 2 or more PIDS in the loop, none of them gives me correct value! I have absolutely no clue what's going on..

On the other hand, your library and android apps have no issues handling multiple PIDS at same time, but your library most of the time doesn't show RPM correctly at all.. At this point I am stunned and can't think of a logical reason how all of this is possible

500ms.txt
HC_05.txt

[braedenv25]

@MaxYoshida . you said in a previous comment " I tried this arduino code: https://www.instructables.com/Arduino-OBD2-ELM327-I2C-LCD-HC05-Bluetooth/ (After modifying PID number and using correct equation from wiki which is ((256 * A +B) / 4)), it works fine as well. " . Does this mean that using this code you have gotten the RPM to have a constant output in the serial monitor? If so could you attatch the code?

[MaxYoshida]

@MaxYoshida . you said in a previous comment " I tried this arduino code: https://www.instructables.com/Arduino-OBD2-ELM327-I2C-LCD-HC05-Bluetooth/ (After modifying PID number and using correct equation from wiki which is ((256 * A +B) / 4)), it works fine as well. " . Does this mean that using this code you have gotten the RPM to have a constant output in the serial monitor? If so could you attatch the code?

Not a constant output, but a correct output. You know that RPM fluctuates (for example between 780 - 794 while idle on my vehicle). When increasing RPM and viewing it by eye, it seemed correct to me as it follows as expected

I attached the exact code I used in my vehicle

OBD_.txt

[braedenv25]

@MaxYoshida Would it be possible to have a constant almost real-time output like the example @PowerBroker2 had with his HUD project? Also thankyou so much for the working code pertaining to that other project it's a great help.