In the June 2016 version of "Corvette Magazine", there was an excellent article by Hib Halverson titled, "Suspension of Belief". Hib Halverson is a legend in Corvette circles, so I knew when I saw his name as the author that it would be a great article.
I must first say that I have been reading Corvette magazines for literally decades now and without question "Corvette Magazine" is the absolute best. The articles are extremely well written. The photography is second to none. The physical quality of the magazine itself is top notch.
I am thrilled that a question I sent in made into the Letters to the Editor in the March 2017 edition of the Corvette Magazine, and even more thrilled to be called a "deep geek" by Hib Halverson!
Below is my question and at the end I include a image of the cover of the March 2017 edition as well as my question:
"I really enjoyed Hib Halverson's Suspension of Belief article in the June Edition. It was the best article that I have ever read on this topic. I have had MagneRide in my 98, 11 GS and now in my 16 Stingray and absolutely love it.
Whenever I have read articles on this topic, it seems there are theory response times and implementation times. In Mr. Halverson's great article and in many other articles, the theory time seem to be 1ms as stated "The calculations are performed once per millisecond, so at 60mph the system is calculating the optimum damping for every inch of travel." Whereas the implementation time, at least for the Stingray and the 3rd generation MagneRide, is 5ms as stated in the article, "I was covering 176 feet every second, and MR took about 5ms to respond. That means the car moved around 10 inches while MR "decided" what to do and shocks began to change." My question is, where is the latency (delay) with the sensors, microprocessors and other systems that are feeding the variables into the MR system so it could execute at its fastest possible speed of a change in 1ms? In other words, what will have to change to have for a MagneRide the ability to change at its fastest speed? I realize 5ms is ungodly fast and MagneRide is a true engineering feat, as well as this is a geek's question, but Mr. Halverson's excellent article made me think of this."
Hib Halverson's Response:
"First...thank you for the kind words on the MR story.
As for your question, I didn't discuss latency in the article because of length constraints. Admittedly, to "deep-geeks" such as yourself, the absence of that discussion does beg questions about it. While I understand the latency in the system, I decided to send your question to my pal, Darin Dellinger at BWI, to get an answer straight from one of the top engineers working on MR. Here's what he said:
"This is a very good question! Today's MagneRide electronic control module (ECM) works on a 1 mSecond cycle time, as it has from the beginning. As you mention, reading the input, processing the information, and driving a physical output into an electrical load are all individual actions that consume time. A fast processor internal to the ECM is a must, as are responsive circuits on both input and output. All of these things and a few more have changed over time to improve the overall system response time, as well as other attributes. In the race to faster overall processing, all of these things must considered and then implemented with an eye on value.
I'll add that you sort of answered your own question with regard to from where stems the current generation MR's latency. I say that because you seem to get the idea that doing the computations is one thing and how quickly the entire system can change states is another.
To get the overall system response time from about 5-mS down to 1-mS would be a huge engineering challenge. In fact, I think just to reduce it by 50% to 2.5-mS would be a tough mountain to climb.
Only time will tell if additional decreases in latency are possible."
Below is the cover and my question in the magazine EVERY Corvette owner should subscribe to:
No comments:
Post a Comment