Thanks for the article.China will hold a event such as the FSAE this year.And it is a very helpful for our work.
Thanks for the article.China will hold a event such as the FSAE this year.And it is a very helpful for our work.
Do anything you can,left on regret behind.
--Fsae-China--M-Dreamer
Great post Geoff. Pretty much summed it up perfectly. I can't wait to read more. I'm sure a few of our alumni could probably post up some of their similar writings.
Although a bit long, I enjoyed every word of it. These are the type of posts we just don't get anymore, its sad
-- A brilliant design that is not finished on time is useless. A simple design that is well developed is a much better practice of engineering
- CS
Very good, very solid, I'm making this mandatory reading for all our team leaders! I'm taking a systems engineering class this semester, and a lot of the stuff here sounded pretty familiar (especially the "V" diagram...that's a key figure in the text). The funny thing is, that class absolutely bores the hell out of me. Not because it's unimportant, but because after so many years in Formula, I've already learned a lot of these things the hard way! One important take-away from that class though is the "ideal" amount of systems engineering effort to ensure the project is done on time, within budget, objectives are met, etc. After several empirical studies it turned out to be about 15-20% of the overall effort. Sure makes for a lot of work for us "level 3" folks!
Breaking it into the separate levels is very important. Going beyond that, you can really classify the stages a team goes through by the highest effective "levels" that they are proficient in. I would say that until around 2002 ISU was very much a "level 1" team, maybe level 2 on a good year. We usually got a car bolted together, but were never really competitive. I've seen some documentation that focused on some of those disconnected "wants," for example "this year we're going to dominate competition because we're using an aluminum chassis," or some "innovative" gewgaw. From 2003-2007 we got some good folks in charge, developed at least some sort of rudimentary vision for what the car was supposed to "be," developed some basic points strategy and became a level 3 team, which paid off in some fairly consistent top 25 finishes, including 7th in 2006. Unfortunately we were still lacking in our level 4 strengths, and suffered poor membership, got our shop closed for 2 months for "safety" reasons, and were basically left more or less starting from scratch personnel-wise in 2008. I've been working hard on the level 4 stuff, but it is so true, it's not the kind of thing they teach in classes!!! The best I've found for that is to seek out classes on project management (at our uni it's called OSCM 320, in the business college), systems engineering, etc. Finding people with leadership and management skills who are willing to commit themselves to the team is a harder task, and for that I'm open to advice. Best I know so far is that when you find someone who seems to have some sort of knack for it, seek them out! Don't wait for them to come to you! I'm a military man myself, and I have found that a lot of the military leadership training I've received has been very useful, so that's always a good option. Hell, being in Formula is an awful lot like being in the Army!
Another very good point. I also happen to be taking a class that goes over DFMA this semester, I'm finding it to be another one that is particularly easy after spending some time in Formula. That's actually one of my biggest pet peeves, a lot of folks want to just get in there and start drawing parts up with no clear conception of its functional requirements, and end up with a doodad bolted to a gizmo bolted to another whatsit welded onto the frame. And usually all of those bolts need safety wire . So much easier to just design one part that attaches to the frame, even if it does take a couple more iterations in CAD. Minimizing parts costs perhaps a tad more forethought in design, but saves CONSIDERABLE time in manufacturing, assembly, maintenance, not to mention the "free" benefits in weight, reliability, and cost. Not to mention results in a much cleaner, prettier car. This definitely ties into that "ideal" 15-20% spent on systems engineering.Originally posted by murpia:
Excellent post Geoff.
Just one thing I'd like to expand on:
In between Level 1 and Level 2 is what I believe to be the single most effective way to build a good racecar: minimising the parts count. Parts you don't have weigh nothing and cost nothing and cannot break.
I'm not suggesting another level to the Systems Engineering approach but it's important when cascading out the design from Level 2 to Level 1 that you don't have too many parallel paths.
I don't have much time to discuss further today but I wanted to get the comment in ASAP. I have in the back of my mind a 'thesis' on the Minimal Parts Count Racecar. One day I'll have the time to write it.
Regards, Ian
P.S., I like the Cornell leadership structure, I've been toying with a similar idea for a couple years now. So that's you guys' secret after all these years .
Dr. Adam Witthauer
Iowa State University 2002-2013 alum
Mad Scientist, Gonzo Racewerks Unincorporated, Intl.
My name is Justin Jang and I approve of this message...
Thanks for putting all of this together Geoff, now I don't have to go about putting my thoughts together and on paper/web, and can just reference students to this.
"Man, I need to practice more!" - Kenny Wallace
"Try not to have a good time... this is supposed to be educational." - Charles M Schulz
-OptimumG 2005-2006
-Turner Motorsports 2008-2009
-Black Swan Racing 2010 & 2011 Team and Driver's Champions
-HPD Race Engineer 2011-2014
-Currently Freelance Data/Race Engineer
Great reading! Looking forward to more posts concerning these areas!
KTH - Royal Institute of Technology
Senior Designer - Monocoque / Body 09/10
Junior Designer - Engine + etc. 08/09
Some further thoughts/rebuttals to some of the other comments...
its not all about "reducing part count" or DFM. There are areas where that is important, but they are not a "cure all". I'm tentative to put a label like "systems engineering" on it too... just because most catch phrases don't go far enough into the actual heart and soul of the issue. (Mostly untrained) Managers will throw around these types of catch phrases without really understanding what it is about, and so I will not try to summarize Geoff's post in a few words.
It may surprise a lot of the new SAE students and the "club racers" but professional racing varies quite a bit from those ranks. Instead of trying to "shave every ounce" pro racing is more about finding out what is most important and optimizing that.
There are a few items that I would like to point out in the Level 2 area. There are a handful of "parameters" that affect the overall performance of the car that are essentially free (as in, it has little to no financial cost to you and you don't have to put it in the cost report). These include the WD, track, wheelbase, gear ratio, runner lengths, fuel map, spark timing, suspension geometry/settings... among others. One of the first things I did for my formula hybrid team was to take simulation runs on these parameters... the result? We could work our asses off and spend our entire budget on a carbon/titanium/unobtainium car trying to save 75 lbs... or we could choose the right gear ratio for our car.
On a side note, the level 1 and level 2 are pure engineering. Level 4 is mostly management, and level 3 is strategy... which has a class in the college of business, so I guess that is management, but you'd better have someone who understands the engineering behind it.
"Man, I need to practice more!" - Kenny Wallace
"Try not to have a good time... this is supposed to be educational." - Charles M Schulz
-OptimumG 2005-2006
-Turner Motorsports 2008-2009
-Black Swan Racing 2010 & 2011 Team and Driver's Champions
-HPD Race Engineer 2011-2014
-Currently Freelance Data/Race Engineer
Thanks for the positive feedback everyone. I hope you've found the above to be useful and not too demeaning. I'm trying to write such that FSAE novices aren't too overwhelmed, but also so that the old hands don't find it too trivial. Tough line to tread....
Apologies that the first post was a bit lengthy. I just wanted to get a bit of a backbone in place, onto which I can build a bit of a "thesis" over time. I know from prior experience if you try to cut corners on these boards then there are any number of mis-interpretations that can get thrown back at you, and the thread gets hi-jacked into all sorts of unexpected and often trivial directions.
Over the next few weeks I'll tack on a few related pieces to try to summarize my FSAE experience. I've been primarily writing this for my own team so that they understand the principles on which our original design was based, but there is no harm in sharing it around. FSAE is primarily a learning exercise after all.
Agreed with Jack - firstly and foremostly don't forget to have fun. We can overanalyze these things to death, and I've probably done that above. Get the relevant research out of the way, and then get stuck into the fun bits.
Murpia, I like the idea of the "minimal parts count racecar", I think it has quite a place in this event. Z was pushing this very concept on these forums around 4 years ago with his "brown go-kart", but the idea was lost on most. If I had design leadership over a number of years this was the direction I would like to have gone myself, systematically cutting back on complexity until a minimalist baseline was reached. But most of us only get a year or two in leadership before we have to move on and let others have their go (and quite rightly I'd add).
VFR, I was quite a fan of the Cornell system, and was fortunate to get some very helpful guidance from Michael Jones very early in the piece. I only wish we had some good Systems Engineering guidance over this side of the world, I've struggled to find it.
Cheers all, and I'll add another piece soon if anyone is interested.
Geoff Pearson
RMIT FSAE 02-04
Monash FSAE 05
RMIT FSAE 06-07
Design it. Build it. Break it.
Anyway, following on from my first post….
Potential vs. Execution
The point-scoring merit of a particular vehicle can be roughly broken down as follows:
Total Points = Track Speed points + Fuel Economy points + Static Event points
I have already stressed that design decisions made to increase track speed, may not necessarily have a positive effect on fuel economy or static event points (think of what a supercharger can do to fuel economy, or what a carbon tub might do to overall vehicle cost). So when I hear arguments that “design option X” makes the car faster, the first question that comes to mind is whether the designer has considered, or even cares, about the rest of the project (or whether they might even be deliberately ignoring the other aspects in order to “sell” their pet design). Any proposed design needs to be fairly assessed across all the above criteria.
Anyway, the argument I wish to make with all the above is that “Track Speed points” can be broken down even further into three interrelated factors:
• Design Potential: How the car will perform according to the calculations / lapsims that we’ve discussed earlier
• Vehicle Completeness: How effectively we deliver the designed vehicle to achieve its full potential.
• Driver Skill: How effectively the driver delivers the full performance of the vehicle
I’d combine the latter two factors under the term “execution”.
Now I’ve seen plenty of examples where all the focus was on the “design potential” side of things, but the final execution didn’t deliver on that potential. Examples include cars that don’t run properly on competition day, or drivers who obviously lack experience in the car. In our own team s I’ve seen examples where months were spent labouring over a couple of kgs (effectively a couple of points), but on competition day a swag of points gets lost because the acceleration event driver hasn’t driven the car before.
For most teams, the design potential of the vehicle may be worth around 0-20 points relative to your competition (and in many cases the decisions we are labouring over are maybe worth single points if that). Failure to complete the vehicle properly is a penalty of up to hundreds of points relative to your competition, and well trained drivers could be worth up towards 100 points. (I’m being a little vague as it depends on the team and where they are at – but certainly the penalties for poor execution are much more serious than the gains most of us are aiming for in our design stages).
The critical issue is time. We have a limited allocation of time that we have to share across the above three factors, and each deserves attention. Unfortunately the design stage has to be completed first, which is when we are at our least knowledgeable. Therefore we over-allocate time to the design side of things, maybe commit ourselves to a design which overstretches our manufacturing resources, and effectively short-change ourselves of time allocated to the more important latter stages of the project.
The overall point I’m making, with this and the earlier post, is that the parameters we often use to measure the worth of our vehicles is misguided. It is very easy to be convinced that a kilogram or a kilowatt is important, as we worry about the respective spec sheets of our’s and our competitions’ vehicles. To assess the true worth of (for example) a kilogram we need to understand:
• How that kilogram converts to overall competition points (in terms of potential performance)
• How the time cost of that kilogram might effect our final execution of the whole project
Enough for now, I'll get back to this next weekend.
Cheers,
Geoff Pearson
RMIT FSAE 02-04
Monash FSAE 05
RMIT FSAE 06-07
Design it. Build it. Break it.
Thanks Geoff,Originally posted by Big Bird:
Murpia, I like the idea of the "minimal parts count racecar", I think it has quite a place in this event. Z was pushing this very concept on these forums around 4 years ago with his "brown go-kart", but the idea was lost on most. If I had design leadership over a number of years this was the direction I would like to have gone myself, systematically cutting back on complexity until a minimalist baseline was reached.
The idea has been in my mind for a few years now but what inspired me to post it was a recent anecdote I heard.
Apparently one of the big customer racecar manufacturers won the contract to replace a grid of 'open' cars with a 'one make' car. They went through their current design and took out as much cost and complexity as they could, reduced the parts count, integrated the design, utilised symmetry as much as possible, made the car easy to work on, etc.
They didn't need to chase performance beyond an acceptable level for the class, so they chased profit instead. But, they actually ended up with a lighter car, with better installed stiffness of the suspension. It was an aero formula, so outright performance was about the same, but in a non-aero formula that approach would have resulted in a faster car!
As a Design Judge it's these kind of design philosophies that get me interested these days. Any team can come along (and many do) and say 'we targeted lighter weight, better power, more grip, etc.' but very few then go along to add 'and we achieved it through... (insert clever thought process of some kind).'
Regards, Ian
Disagree slightly on the minimal parts bit, though it depends on where you're looking at it.
While it's nice to have lots of features integrated together in one unit (ex: tripod housing integrated into hub), in my experience it tends to add complexity to the manufacturing process.
In a pro shop, doing larger quantities, taking advantage of castings or tooling that's already set up, I'd suspect the minimal part approach works well. May even take less total time on a per part basis if you can take advantage of custom tooling. In a one-off amateur level series like this, especially with very limited fabrication equipment and time, in a lot of places I'd rather use two simple easy-to-make parts in place of one complicated one. There's certainly the opportunity for losing some rigidity, but the manufacturing timeline is critical.
From personal experience, my senior year I machined somewhere around 70-80 parts at a total of ~180 hours, including programming, setup, and run time. A lot of those parts took the "simple and modular" approach. At ~20 hours a week of machining just in itself (on top of welding and everything else), that added up to more than a couple months.
If I had consolidated those parts, stock material cost would have gone up substantially. Since my machine time was free, stock was the only realized cost. Would have also meant more intricate, longer setups, I suspect more run time, and higher chance for scrap.
There's definitely a balance between number of parts, and manufacturability... and for me, manufacturability was a huge design driver in getting the car done as quickly as possible.
Just my 2 cents.
Colorado FSAE | '05 - '07
Goodyear Tire & Rubber | '07 - '11
NASCAR Engineer | '11 - ??