« Last post by mcote on May 27, 2016, 09:03:16 AM »
I'm using Torque and its apparently collecting about 80 data points per second. I know this because I have viewed the CSV file it generates and there are about 80 columns of data, with an additional row each second.
How does it accomplish this?
I've done a bit of programming with OBD-II dongles and there are serious limitations about how fast you can query data. I know you can do "PID stuffing" which allows you to query up to 6 PIDs in one request. And I know you can do other tricks like include the expected number of messages and turn off spacing and headers, which speeds things up a tiny bit. Even with all these speed tricks, there is no way I can get anywhere close to 80 data points per seecond.
I suspect part of the reason is that Torque calculates some of the data itself, rather than query it. For example, you get the speed in both MPH and KM/H. Obviously Torque only gets one from OBD and calculates the other. But even still, I think Torque is getting OBD data very fast.
Does anyone know how it works, under the hood? Or even care to take a guess?
What I'd really like to know is, does Torque use monitor mode (ATMA) and just listen for the data that it's interested in? I think this would be extremely fast and you wouldnt even have to query any specific data. However I havnt figured out a reliable way to do this, so I'm still stuck querying the specific PIDs that I'm interested in.
Can anyone confirm or deny whether Torque uses ATMA mode to gather data, or does it use PID query mode? Or is it possible to do both?