Saturday, June 30, 2012

I'm approaching a technological platform dilemma

Microsoft, Google, and Apple are the new "Big Three" and each offers their own flavor of the consumer multimedia experience. Personally I use Windows at work and for heavy-duty computing at home (graphic design, audio production). I also have an Android tablet and phone. 

The trouble is that I like a unified experience. I don't like having Widget X to do Task A, and Widget Y to do Task B. Seamless availability of content across hardware form factors is a big selling point for me - and clearly for consumers. Apple did very well on this level for many years with very good integration of home computer, mp3 player, phone, tablet, etc. They were the only show in town that really nailed those. 

Moving forward, Android and Windows are at a point to meet or beat that. Windows has been well-entrenched as a home and business computing power for decades. Now they're breaking in with Windows Phone 8 and the Surface tablet. Android has had the smartphone and tablet (the latter becoming increasingly impressive with the Nexus 7), and has dabbled in home computing with netbooks - though those clearly aren't in the same league as a workstation.

For me, buying anything Apple is clearly out of the question. I already have two different platforms I use, I don't need a third - and they will never be compatible with stuff for my job. Then it comes down to Android and Windows. Do I keep things disjoint, or do I try to go all in? This wouldn't even have been a question if not for the Surface tablet to be honest, which I think gives Microsoft an edge in cross-platform architecture. For that matter, I'm fairly certain my next tablet will be the Surface. It seems like the only option out there that isn't just made for viewing content but creating it. I like that - I'm a creative guy.

So then if I have Windows at work, Windows home workstation, Windows tablet... do I stick with an Android phone? I love a lot of the services Google offers (Gmail, Maps, Music) and I like Android as a mobile platform... but what I need to determine is if there's an advantage in unifying the whole experience.

Sunday, June 24, 2012

If I had to design a(nother) road racing car from scratch (Part 1 of ?)

It's fucking late. Napping earlier was a poor choice.

This tune however is never a bad choice!

A thread on F1 Technical got me thinking about this. When I was doing the FSAE thing I'm not even sure what approach we used for coming up with the force element rates every year. Whatever it was certainly wasn't too clever as we arrived at a point through tuning where disconnecting a front bar did absolutely nothing for handling (probably should have taken the damn thing off completely for that matter - weight savings!).

In any event for most racing series this is all probably a moot point since you're entirely likely to have a known baseline and be tuning around it rather than wholesale fundamental setup swaps. But if I were doing it from scratch, I think you need a methodical constraint based approach... otherwise it is too open ended. Intimidating even. If you sit down and think about it logically though and make some design decisions, the spring rates and such solve themselves for you. Same goes for kinematics. Anyway:

Step 1 - Come up with a baseline lateral load transfer distribution as my first constraint.
If I have the same tires front and rear I'd probably put my front percentage a few points higher than my rear weight distribution (since taking nose weight out is going to be in the direction of oversteer.. with few exceptions). If I'm 52% rear biased on static weight, maybe my front load transfer distribution starts at 55%. If I have larger rear tires than fronts I can start to take that back down as otherwise I'm promoting some understeer. If I have a RWD car with a bunch of power I'd go in the direction of higher front LT percentage to keep the inside rear wheel loaded on initial throttle and be able to put some power down at the expense of a little mid corner understeer.

In any event, this will tell me what ratio of roll stiffness (spring plus bar) I have to keep. I won't know the final answer until I get to the track, but this is a starting point.

Step 2 - Ratio of front and rear axle ride rates as my second constraint.
Fundamentally I'd say it's more than likely I don't want to couple pitch with ride so this will likely be fairly proportional to mass distribution. Maybe a little stiffer rear if you subscribe to the flat ride philosophy, and even then that would be dependent on track speed. Having some aero knowledge helps here too. E.g. if the car is massively sensitive to front ride height more so than rear, then perhaps I go with stiff front soft rear if pitch variation is no big deal.

Between these two steps things are starting to get pretty well tacked down. On a 50/50 weight distribution car with same tires all around this would lead you to slightly stiffer rear springs than front, with a small front bar. I think that's reasonable.

Step 3 - Overall ride rate
We need something at this point to lock in some absolutes, and a ride rate would do it. Can approach this a few ways. Unless you have previous working knowledge of where you need to be this is a bit arbitrary - but that's OK when you get to setup packages and being able to sweep this at the track. Let's say we pick 1 Hz as a ride frequency. With that and the total car mass and a guess at tire rate, your wheel and spring rates are immediately defined. Boom! No debate or guess work, the design choices you made set them already. Bar rate(s) are also immediately defined, pending Step 4.

Incidentally at this stage some baseline stabs at damping rates fall into place. Can even just assume they're linear.

Step 4 - Do we need more roll stiffness?
This is where it gets a little fuzzy to me. With the springs you have in the car, does that give you a good enough roll stiffness rate? Who is to say what's "good enough" ? Probably dependent on your kinematics, how much dynamic camber you want to target, and how much static camber you want to get away with. That should pretty much be sufficient to lock this one down in a logical rather than total guesswork manner. CG height should drive this as well - higher it is the more bar you will need to control it.

What sucks is if you decide you do need more roll stiffness, it implies needing to add a rear bar - otherwise front bar only would change your LLTD found in Step 1. On an independent rear suspension, adding a rear bar adds an element of direct and immediate load transfer in single wheel bump. Not directionally good for getting power down off corners with a RWD car. Fundamentally I think if I can get the roll stiffness I want with my springs, so much the better. Plus, no rear bar (or hell front bar) = less pieces to fail and less weight.

Step 5 - Game plan some packages
Crucial step. You are probably not going to nail it out of the box with this baseline setup, so you need to know how to adjust. The most obvious items would be +/- "a bit" on lateral load transfer distribution and overall ride and/or roll rate. Figure out what size springs, bars, and dampers you will need to have the adjustment range you want. Buy them and bring them to the track. 

From there you can do some of the detail work and track tuning. Looming questions are still out there, like how much rebound vs compression damping you want or how much high speed versus low speed. Rig testing, to me, would come in at this point. Map out which are the big hitters for ride / mechanical grip metrics, which aren't. How do those damper changes move the needle in comparison to +/- 0.5 Hz of ride rate?

If I ever want to do a F1000 again, or turn my 350z into a track car, this will likely be the approach taken.