To elaborate on the spreadsheet from last night... in my mind two items which have a major impact on the shape of your torque curve are cam geometry and exhaust manifold geometry. As per series rules I can't play with cam profiles, but exhaust is free.
With regard to header geometry there are multiple schools of thought as to calculating what effect the lengths, diameters, etc will have on performance. Helmholtz resonators are mentioned though I believe are generally out of favor for high-rpm application. I don't really have much to back that up, but it's a simple model. Sam Zimmerman did some real good research into Development and Validation of an Impedance Transform Model for High Speed Engines. Good read, and it's something I'll be taking a look at later down the road. Hopefully I still have the PDF around. It is a good bit more involved analytically and shows both resonant (peak) and anti-resonant points for a manifold. Mid-way between the two is the "wave reflection" theory. This is Caldwell's method that he has a lot of experience with.
The idea is as follows: At the end of the power stroke the combustion gasses in the cylinder are still under fairly high pressure. EVO (Exhaust Valve Opening) occurs at some angle before BDC so the cylinder has time to "blow down" and the engine isn't working against so much pressure during the exhaust stroke. At EVO you have a rapid pressure pulse that's start to travel down the exhaust primary at the speed of sound in the exhaust gas at EGT. When the pulse meets a junction such as a merge collector, you can treat the change in tube cross-section as a change in the medium in which the wave is travelling. As painful as Physics 3 was, you'll note that when you have a change in medium a portion of the wave is transmitted (to go further down into the exhaust secondaries) and a portion is reflected with negative sign back up the primaries. Now you have a negative pressure pulse traveling toward the still-open exhaust valve. The idea is that you tune your header lengths so that the negative wave reaches the exhaust valve right before it closes. The extra pressure differential across the port helps scavenge out more exhaust gas to give you a boost in volumetric efficiency (a tuning peak).
My spreadsheet takes all that into consideration and gives you tuning peaks given the length of primaries, secondary (if they exist) and tertiary (if they exist) tubes.
The other rule of thumb Caldwell had was to aim for about 300 ft/s flowrate in the exhaust at high-rpm peak power. "It seemed to work the best." I can imagine you'd want the exhaust to be open and free-flowing, but with enough velocity to get the gasses out while they're hot and inviscid and carrying some momentum.
Some people try to use the same methodology for intake tuning ("ramcharging" the intake) but I've never been able to conceptualize where you would get a strong negative pressure pulse at IVO, and from practical experience playing with intake runner and exhaust header lengths, the exhaust seemed to have a much greater impact. But for all I know that could just be a fallout of cam geometry...
But what do I know. I'm a tire / suspension guy.
Thursday, October 30, 2008
Wednesday, October 29, 2008
Going to need engine dimensions sooner or later
Got a few things done. Worked a bit more on the frame just to get a rough idea of how the engine bay is going to be arranged. Thanks to Ryan by the way for digging up a couple driver models. Going to have to change the main and front hoop geometry to get the driver's head to pass the requirements in the SCCA GCR.
Unlike the CU Racing FSAE cars, the engine will not be a stressed element of the frame. Sportbike motors aren't designed to be stressed members of open-wheeler frames. As such, the mount points aren't really all that great for having the thing fully stressed. If you try to do it that way you either tend to get stuck with a rigid design where you can't get the engine in or out, or the engine can go in and out easily but you lose all your rear end rigidity from crappy mounts. On this go around the engine bay will be fairly wide and fully triangulated for hub-to-hub rigidity. The engine itself will have some simple mounts inside the bay to hold it in place and restrain it from drive torque.
Right now I'm just using a scaled 600RR model as a placeholder. Going to really need to CMM a 1000RR (or GSXR, whatever) block soon to get some basic dimensions to figure out how much room I need for the bay, intake and exhaust points, etc.
Speaking of exhaust points, threw in a few baseline parameters into my handy wave reflection tuning spreadsheet to get an idea for how long and what diameter the tubes will have to be. Right now it's setup for 4-2-1, though in the end it will likely be 4-1 out one side, or twin 2-1 collectors. I think the last time I talked to Caldwell (the veteran, retired 35+ year race engine builder) his thought was you do get a lot more "streetable" torque band from the 4-2-1, great for FSAE car, but you do ultimately lose out on some peak top end.
Found a couple places that do CNC mandrel bending of tubing. That'd be a great way to do the exhaust. If you haven't built an exhaust manifold before... it's a pain. Real challenging to do in CAD with 3d sketches woven through the frame while matching the right cylinders for the collectors, and keeping equal lengths. Likewise building it is a real bitch getting all the bend and compound angles right. Just building one "on the fly" with no CAD it would be impossible to get the lengths equal... then you have to go in and start to do individual cylinder trimming and all that BS. So yea, having the primaries seamlessly bent from continuous tube would rock.
Unlike the CU Racing FSAE cars, the engine will not be a stressed element of the frame. Sportbike motors aren't designed to be stressed members of open-wheeler frames. As such, the mount points aren't really all that great for having the thing fully stressed. If you try to do it that way you either tend to get stuck with a rigid design where you can't get the engine in or out, or the engine can go in and out easily but you lose all your rear end rigidity from crappy mounts. On this go around the engine bay will be fairly wide and fully triangulated for hub-to-hub rigidity. The engine itself will have some simple mounts inside the bay to hold it in place and restrain it from drive torque.
Right now I'm just using a scaled 600RR model as a placeholder. Going to really need to CMM a 1000RR (or GSXR, whatever) block soon to get some basic dimensions to figure out how much room I need for the bay, intake and exhaust points, etc.
Speaking of exhaust points, threw in a few baseline parameters into my handy wave reflection tuning spreadsheet to get an idea for how long and what diameter the tubes will have to be. Right now it's setup for 4-2-1, though in the end it will likely be 4-1 out one side, or twin 2-1 collectors. I think the last time I talked to Caldwell (the veteran, retired 35+ year race engine builder) his thought was you do get a lot more "streetable" torque band from the 4-2-1, great for FSAE car, but you do ultimately lose out on some peak top end.
Found a couple places that do CNC mandrel bending of tubing. That'd be a great way to do the exhaust. If you haven't built an exhaust manifold before... it's a pain. Real challenging to do in CAD with 3d sketches woven through the frame while matching the right cylinders for the collectors, and keeping equal lengths. Likewise building it is a real bitch getting all the bend and compound angles right. Just building one "on the fly" with no CAD it would be impossible to get the lengths equal... then you have to go in and start to do individual cylinder trimming and all that BS. So yea, having the primaries seamlessly bent from continuous tube would rock.
Tuesday, October 28, 2008
I am so completely cutting frame tubes like this...
I knew tubes could be CNC laser cut, but I didn't know it was like this! Found this gem while at work. What this machine does in 1 second could easily take 5-15 minutes by hand, depending on joint complexity.
Feel free to get up and dance if the music really inspires you.
On a related note, if you aren't using the "weldments" feature in Solidworks to design your frame... you seriously should. It saves so much time and file space its unbelievable.
Roll & Pitch Stiffness
Didn't bring RCVD back home from work today, so I'll have to do this by hand again. I figure roll and pitch gradients (in deg/g) should be ballparked by the following:
Obviously that assumes my roll center is just about on the ground and on chassis centerline, likewise pitch axis about under the cg, w is really w_sprung, etc etc. For now I'm looking at ballpark values.
As an aside, anyone from work would probably be giving me a lot of shit right now about equations and documentation and all that. But you do have to have some level of analytical work and knowing what you're doing, i.e. not just welding a pile of crap together.
In any event, even with the relatively soft wheelrates of last night's 2Hz front and 2.2Hz rear sprung natural frequencies, that still gives me 1.5 deg/g in roll and 0.6 deg/g in pitch. Not too shabby! Lot more rigid than I had expected. Should let me run a pretty low ride height without bottoming the car, even if I had something outrageous like 3g of braking. Going up to 3Hz front and 3.3Hz rear yields 0.7 deg/g in roll and 0.3 deg/g in pitch.
Don't want to go too stiff in roll from just spring, so I do still have some quick tuning options from ARB's, which will hopefully be cockpit-adjustable.
Enough of spreadsheets. Hopefully in this week I can get more of the uprights fleshed out and get suspension CAD started, see what tricks I can pull with the kinematics.
Obviously that assumes my roll center is just about on the ground and on chassis centerline, likewise pitch axis about under the cg, w is really w_sprung, etc etc. For now I'm looking at ballpark values.
As an aside, anyone from work would probably be giving me a lot of shit right now about equations and documentation and all that. But you do have to have some level of analytical work and knowing what you're doing, i.e. not just welding a pile of crap together.
In any event, even with the relatively soft wheelrates of last night's 2Hz front and 2.2Hz rear sprung natural frequencies, that still gives me 1.5 deg/g in roll and 0.6 deg/g in pitch. Not too shabby! Lot more rigid than I had expected. Should let me run a pretty low ride height without bottoming the car, even if I had something outrageous like 3g of braking. Going up to 3Hz front and 3.3Hz rear yields 0.7 deg/g in roll and 0.3 deg/g in pitch.
Don't want to go too stiff in roll from just spring, so I do still have some quick tuning options from ARB's, which will hopefully be cockpit-adjustable.
Enough of spreadsheets. Hopefully in this week I can get more of the uprights fleshed out and get suspension CAD started, see what tricks I can pull with the kinematics.
Monday, October 27, 2008
Hey Jeff Gordon... drink your damn Pepsi.
It's the little things that irritate me. And the big things. And Ohio drivers. But anyway.If you've ever flipped on a Nascar race... before tires start going down and cars fly into the wall, before 43 state-of-the-art carbureted, 2-valve, pushrod V8's fire up, they have the driver interviews.
In every one of them Jeff Gordon is standing there with his Pepsi just visible on the TV screen. Every time it looks like he uncapped it about 5 seconds before hand and hasn't taken a sip. He talks, does his spiel, and then right when he finishes he takes one sip.
Hey! Make it at least look like you're enjoying your sponsor's product! Drink some ahead of time! Why wouldn't you? It's probably hot out there anyway, especially in a Nomex suit. Otherwise ship some back to me.
Probably a good thing I don't get to the track much (well, ever). I'd be completely overloaded on Red Bull from the #84 and #83 teams.
The Friction Egg
Hideously disfigured egg, anyway. Regarding tire grip potential you could really simplify it as a friction circle. A lot of the time it's a bit asymmetric, maybe skewed a bit higher toward drive/brake than cornering.
With RWD vehicle grip you get an egg-looking shape, exaggerated by downforce. Heap of braking grip coming in at top speed into a turn. Bleed off some speed and downforce as you transition to max cornering, then to WOT (assuming you're entering the turn on the brakes). The drive end of the egg is a lot shorter than brake given that you have less speed and downforce at that point, you're only using 2 tires instead of 4, and in the higher gears you probably are significantly power limited.
This will be handy to get corner loads going through a turn. When I figure my pitch and roll stiffness I can combine the two and get roll and pitch angles through the turn which will be an input to OptimumK for dynamic camber and steer angles. Niiiiiice.
And yea yea yea I know it's not very good without a real corner sim and knowing downforce vs speed, length of braking zone, all that stuff. But it's not bad to get in the ballpark.
With RWD vehicle grip you get an egg-looking shape, exaggerated by downforce. Heap of braking grip coming in at top speed into a turn. Bleed off some speed and downforce as you transition to max cornering, then to WOT (assuming you're entering the turn on the brakes). The drive end of the egg is a lot shorter than brake given that you have less speed and downforce at that point, you're only using 2 tires instead of 4, and in the higher gears you probably are significantly power limited.
Wrote up a quick Matlab script that takes some simple vehicle parameters, Max Braking G's, Max Lateral G's, and Max Drive G's and does a simple cornering loading analysis. These plots just use some guess parameters...
This will be handy to get corner loads going through a turn. When I figure my pitch and roll stiffness I can combine the two and get roll and pitch angles through the turn which will be an input to OptimumK for dynamic camber and steer angles. Niiiiiice.
And yea yea yea I know it's not very good without a real corner sim and knowing downforce vs speed, length of braking zone, all that stuff. But it's not bad to get in the ballpark.
Sunday, October 26, 2008
Wheelrates, Spreadsheets, Bruschetta
In reverse order. I made some excellent bruschetta tonight. Not bad for a first shot. Threw in some roasted red peppers. I was kinda weak, went with the bruschetta spread in a jar, but it's got all the essentials. Plus the roasted red peppers, plus fresh mozz, on ciabatta bread and toasted at 450F. Not bad. Plus I endorse Smithwick's as the official adult beverage of suspension design. Or I guess Coors Light if you're still doing FSAE.
Been throwing together some basic spreadsheets... weight transfer numbers... sprung natural frequencies, wheel rates, all that fun crap. Referenced Matt G's tech tip's at the OptimumG site. Good stuff. Think he's still working in Colorado... smart guy.
One of the kinda screwups we did at CU Racing was throwing a ridiculous amount of spring at the car, for no real reason. No downforce and on bias tires where you don't really get any advantage from "locking" the suspension with spring. 120 lb/in wheel rate on a 460lb car! The car does not necessarily respond quicker with more spring. Either the tire or sprung mass is a limiting factor. Think about putting F1 tires on a minivan, or minivan tires on an F1 car. The handling would not be good. The two have to be matched up.
In any event, starting at a guess of 43/57 static mass split F/R, and 2.0Hz front sprung mass natural frequency, the wheel rates come out to be pretty low. Much lower than I'd be able to get away with using a Cane Creek damper and anywhere close to a 1:1 motion ratio.
- 2.0 / 2.2 Hz split F/R :: 88 / 141 lb/in wheel rates
- 3.0 / 3.3 Hz split F/R :: 198 / 318 lb/in wheel rates
- 4.0 / 4.4 Hz split F/R :: 352 / 565 lb/in wheel rates
Crikey
Posted up on Apex Speed regarding typical G-loading of these cars. Initially I had expected somewhere around 2.0g. I'd say the best FSAE cars can get up around there on a hot day on clean asphalt, and they have really low footprint pressures and rubber about as sticky as you'll find (though not necessarily up to peak temp).
The claims are anywhere from 2.0 to 2.7g. As with anything motorsport-related I take that with a grain of salt. On initial aggressive cornering or braking you'll get an artificial "blip" of acceleration, from the chassis yaw or pitch acceleration multiplied by the distance of the accelerometer to the CG. If your accelerometer is at the true CG it's not a big deal, but if you're using say a MoTeC dash for your lat g measurement it can be significant. Generally filtering the trace over a short window (a moving average over say 0.125s) produces a more realistic measurement and will knock the peaks down a few tenths. Still that's in the range of probably 2.5g for a fast corner and a little less for a slow corner.
That brings me back to 50/50 on the differential option, Salisbury vs ATB. With a CGH near axle height that will certainly exceed the max torque bias ratio of the Quaife ATB out of slow corners and cost a heap of grip.
It will really come down to CG height. If I can design the car light enough, say in the 600-some pound range, I can throw a heap of ballast in the bottom of the frame and knock the CGH down significantly. Diff will have to be revisted much later in the design cycle.
In other news, I'd avoid "Red Bull Cola." After having become an addict of Red Bull after 3 FSAE years, I was expecting RBC to be something great. It tastes like maybe a 50/50 blend of Coke and Dr Pepper. Lame. If they took the flavor of Red Bull energy drink and just put that into a normal drink, that would sell great.
The claims are anywhere from 2.0 to 2.7g. As with anything motorsport-related I take that with a grain of salt. On initial aggressive cornering or braking you'll get an artificial "blip" of acceleration, from the chassis yaw or pitch acceleration multiplied by the distance of the accelerometer to the CG. If your accelerometer is at the true CG it's not a big deal, but if you're using say a MoTeC dash for your lat g measurement it can be significant. Generally filtering the trace over a short window (a moving average over say 0.125s) produces a more realistic measurement and will knock the peaks down a few tenths. Still that's in the range of probably 2.5g for a fast corner and a little less for a slow corner.
That brings me back to 50/50 on the differential option, Salisbury vs ATB. With a CGH near axle height that will certainly exceed the max torque bias ratio of the Quaife ATB out of slow corners and cost a heap of grip.
It will really come down to CG height. If I can design the car light enough, say in the 600-some pound range, I can throw a heap of ballast in the bottom of the frame and knock the CGH down significantly. Diff will have to be revisted much later in the design cycle.
In other news, I'd avoid "Red Bull Cola." After having become an addict of Red Bull after 3 FSAE years, I was expecting RBC to be something great. It tastes like maybe a 50/50 blend of Coke and Dr Pepper. Lame. If they took the flavor of Red Bull energy drink and just put that into a normal drink, that would sell great.
Friday, October 24, 2008
Uprights
After a couple years of learning by doing and screwing up, I think I may finally have got the beginnings of a decent upright and wheel assembly. Top hat not shown, nor a lot of other things.
Uses two pretty robust DGBB's for the live spindle at the moment. I'll have to evaluate if they can sustain enough lateral loading. Radial loading generally winds up being the most severe in this application from forces at the contact patch being translated as a force and moment at the bearing.
For example, 700 lbf lateral force at the contact patch becomes roughly a 700 ft-lb moment at the bearings. Split that on two bearings with say an inch between centers and now you're talking about 8,400 lb (over 4 tons!) of radial load on the bearings. Now that I think about it I'll probably have to tweak the bearing spacing in the current design and make the upright a good bit fatter, and/or run even more stout of bearings. Maybe ACBB's. Still just a mock up now.
Billet aluminum design at the moment. Sending it out to be cast would be outrageously expensive for tooling. In theory I could have some cast at Dave's place, since he has his own damn aluminum foundry, but AL356 and AL357 (high strength casting alloys) don't have nearly the same mechanical properties as AL2024, Al7075, AL7x49, AL7068, etc. Plus there's porosity, etc. In any event the design is pretty simple and shouldn't be hard to setup on a CNC. Symmetric left to right so programming costs are reduced.
For grins I may draft up a welded sheet 4130 design. I could have all the pieces lasercut to almost snap-fit together before welding. My issue with 4130 weldments is that I feel they're hard to do stress analysis on. Even if your welder is consistent, it's hard to say what exactly the alloy percentage, microstructure, and mechanical properties are of your weld zone once the filler metal has been diluted into it. Then you've got distortion with heat treating, and you'll have to do spot machining afterward anyway.
Currently mocking up the front assembly with a 4-piston radial-mount AP caliper and ventilated rotor.
Uses two pretty robust DGBB's for the live spindle at the moment. I'll have to evaluate if they can sustain enough lateral loading. Radial loading generally winds up being the most severe in this application from forces at the contact patch being translated as a force and moment at the bearing.
For example, 700 lbf lateral force at the contact patch becomes roughly a 700 ft-lb moment at the bearings. Split that on two bearings with say an inch between centers and now you're talking about 8,400 lb (over 4 tons!) of radial load on the bearings. Now that I think about it I'll probably have to tweak the bearing spacing in the current design and make the upright a good bit fatter, and/or run even more stout of bearings. Maybe ACBB's. Still just a mock up now.
Billet aluminum design at the moment. Sending it out to be cast would be outrageously expensive for tooling. In theory I could have some cast at Dave's place, since he has his own damn aluminum foundry, but AL356 and AL357 (high strength casting alloys) don't have nearly the same mechanical properties as AL2024, Al7075, AL7x49, AL7068, etc. Plus there's porosity, etc. In any event the design is pretty simple and shouldn't be hard to setup on a CNC. Symmetric left to right so programming costs are reduced.
For grins I may draft up a welded sheet 4130 design. I could have all the pieces lasercut to almost snap-fit together before welding. My issue with 4130 weldments is that I feel they're hard to do stress analysis on. Even if your welder is consistent, it's hard to say what exactly the alloy percentage, microstructure, and mechanical properties are of your weld zone once the filler metal has been diluted into it. Then you've got distortion with heat treating, and you'll have to do spot machining afterward anyway.
Currently mocking up the front assembly with a 4-piston radial-mount AP caliper and ventilated rotor.
Thursday, October 23, 2008
More on the original 312B
I'm sure many of you are familiar with the Shell / Ferrari commercial. If you aren't, you've missed out.
The original 312 is featured from 0:19 to 0:50, and the 312B from 0:50 to 1:12. That YouTube video preview shot is also the 312B.
There is an alternate version of this video as well, with some longer scenes. Enjoy.
The original 312 is featured from 0:19 to 0:50, and the 312B from 0:50 to 1:12. That YouTube video preview shot is also the 312B.
There is an alternate version of this video as well, with some longer scenes. Enjoy.
Brakes and Diff
Heard back from one of the Rally engineers at AP regarding some brake options that would work. Good stuff. Also picked up the brake spreadsheet from Race Tech magazine (good suggestion, Dan). Close to working out what bore sizes I'll need. Getting further along on the brake system packaging is going to be key. Rotor and caliper are going to define some of the hub and upright dimensions, and from there I can get an idea of where I can put a-arms. Since I don't feel like spending money just yet I should be able to get away with using the demo version of OptimumK for some initial looks at kinematics.
Some comments from Billy Wight as well. From what I recall having looked at the original DB-1 spring availability and damping rates I still feel it may be a good option for this application. I'll have to go back over all my notes from '06 and '07. It has a hell of a lot of adjustability. I believe Ohlins also has some data or a damper rate plotter for the TTX-40 on their website to get an idea for where that's at. If I were to buy "proper" racing dampers, even 2-way adjustable R/C dampers are expensive as all hell. We shall see.
Differentials are another good point, and exceeding the max TBR of the ATB. It will be close. Running a 60" track width... to stay within the 4:1 TBR the following options are available
Some comments from Billy Wight as well. From what I recall having looked at the original DB-1 spring availability and damping rates I still feel it may be a good option for this application. I'll have to go back over all my notes from '06 and '07. It has a hell of a lot of adjustability. I believe Ohlins also has some data or a damper rate plotter for the TTX-40 on their website to get an idea for where that's at. If I were to buy "proper" racing dampers, even 2-way adjustable R/C dampers are expensive as all hell. We shall see.
Differentials are another good point, and exceeding the max TBR of the ATB. It will be close. Running a 60" track width... to stay within the 4:1 TBR the following options are available
- CG at axle height (~11.25"), TLLTD +10% forward of static axle loads. Potentially lot of U/S
- Drop ride height, stiffen springs, CG at 10", TLLTD +5% forward of static axle loads. Might be possible.
- Compromise, 10.5" CG height, TLLTD +7.5% forward.
Wednesday, October 22, 2008
Aero
Good way to finish off lunch break..
A co-worked stopped by my desk yesterday. Said co-worker happened to be involved in the aero development for a certain high-downforce FSAE team for the past couple years. How convenient. Talked a bit 'bout this and that, and how aerodynamically inefficient the 312B airfoil arrangement is. I'll have to come up with some compromise between a more "modern" platform and the historic one. I'm stubborn that I want to have something that looks like a late 60's F1, but I'm more stubborn on developing hardware superior to the competition. Aren't I an ass? Checked the GCR and there is a spec for maximum front and rear wing width but nothing on plan area...
In any event I know I can get a heap of mechanical grip and pull some trick stuff with the tires, which will hopefully make up for some of the aero deficiencies. Spoke with both the current Hoosier and former Goodyear sportscar guys regarding tire choices. I got a few options which will be ridiculous grip. Right tire is the ace of spades...
A co-worked stopped by my desk yesterday. Said co-worker happened to be involved in the aero development for a certain high-downforce FSAE team for the past couple years. How convenient. Talked a bit 'bout this and that, and how aerodynamically inefficient the 312B airfoil arrangement is. I'll have to come up with some compromise between a more "modern" platform and the historic one. I'm stubborn that I want to have something that looks like a late 60's F1, but I'm more stubborn on developing hardware superior to the competition. Aren't I an ass? Checked the GCR and there is a spec for maximum front and rear wing width but nothing on plan area...
In any event I know I can get a heap of mechanical grip and pull some trick stuff with the tires, which will hopefully make up for some of the aero deficiencies. Spoke with both the current Hoosier and former Goodyear sportscar guys regarding tire choices. I got a few options which will be ridiculous grip. Right tire is the ace of spades...
Tuesday, October 21, 2008
Design Overview
I'm using the overall dimensions of the 312B as a rough starting point. At least I know a guy will fit in it. At this stage of the game I'm doing a very rough fleshing out of roughly where components will go, how much room I have to play with, etc. As I continue I'll step through the design process, both the concepts and hard numbers. I'm open to suggestions, by the way.. particularly regarding aero.
- Weight: 1000 lb, including driver
- Wheels: 13" bead diameter x 10" wide (maximum)
- Wheelbase: ~94"
- Track: ~60"
Engine
There are a number of options available here. At the moment I lean toward the Honda CBR1000RR, just as I've had experience with some of the internals on the 600 F4i from FSAE, including the gearbox. Word on the street is the '08 models boast 180-ish HP at 12,000 rpm, but that's just a wikipedia stat.
Differential
At first I was 50/50, but now I lean 80/20 in factor of an ATB differential over a Salisbury-type. If I was doing FSAE again I would probably lean toward the Salisbury in order to get some locking on coast to tighten the car up entering the corners. Otherwise with the ATB they seem to really shift to oversteer when on the brakes or in lift-off transients. Plus, they're traction-limited in enough corners you can steer the rear around with throttle.
With the F1000 I'm hoping the much longer wheelbase will mitigate the longitudinal weight transfer effects a bit... keeping good rotation while trailbraking but not wanting to spin the car around. And then the ATB wins with power application, and keeps up the differential action at all times so I'm not scrubbing any speed off going tight mid-corner and on track-out.
Exhaust
I much preferred the drivability of a long-header 4-2-1 exhaust to the 4-1 merge on a FSAE car. The '04 and '05 cars (and I believe '03) all suffered from the same issue, in that there would be no torque at all up until probably 7-8k rpm, and then bam the torque would flick on like a lightswitch. If you were out of the powerband you were junk. The 4-2-1 on the '06 and later, with longer primaries, seemed to drop the start of the powerband down into the 5k rpm range. Much better for a "street" course.
Twin 2-1 exhaust pipes on either side of the car would look very cool. I may have to consult with the folks at Burns Stainless to get their input. Would be nice to be able to drive this thing at an SCCA autocross and not just at Mid Ohio or Beaver Run.
Frame
Welded tube spaceframe. Composite tubs are not allowed in this series, nor would I probably be interested in making one. I'll have to check the SCCA rules but I believe 1.25" OD x .095" wall is recommended for roll hoops. At the moment with the placeholder frame I have 1.5" OD x .125" wall. Easier to start extreme overkill and back down.
Having spent many frustrating hours welding the 2007 frame, in a variety of awkward positions, actuating the TIG pedal with anything from my foot to elbow to ass, I'm trying to ensure the joints aren't completely asinine this time! Plus I'll be able to spend extra time making sure I have good fit-up, etc. Good fit up is the difference between a joint taking 10 minutes, and over an hour!
Rigidity and safety are at the top of my list. I'll take a weight knock for both. Rigidity shouldn't really cost much in weight if I design it clever enough, and given that this will be going faster than the 60mph of the Silverdome parking lot, I want something safe.
Wheels
13" bead diameter obviously. I'd really like to get single-piece wheels for rigidity, but that may be tough to get a 13x10" and 13x8" single piece with the dimensions I want. If anything I may have to go with a BBS 3-piece, with Jongbloed as an alternative. Kodiak was impossible to deal with and the Keizer may as well have been made out of cardboard. Camber and toe compliance otherwise are going to cost a heap of grip and perceived response linearity.
Brakes
Disc brakes on all corners. Beyond that I'm pretty open to sizing. I've sent notes out to Wilwood and AP for ideas regarding rotor, caliper bore, and master cylinder sizing. Since this is basically a double-weight FSAE car and since I'm not a brake expert, my initial thought is to run small 4-pistons up front, and 2-pistons rear. I may be able to get away with 2-pistons up front, but then again I don't want to skimp on brakes.
It's likely the front rotors will be ventilated. Rears may be solid disc. I had initially wanted to run inboard brakes on the rear to drop some of the unsprung mass and help with any TLV to get extra grip out of the corners. I may still do that. Cooling ability will be the deciding factor. Easier to get ducts to work when outboard and away from the exhaust...
Tires
Couple options here. I'm tempted to run Goodyear FM tires. They have bias tires in R160 up to R600 or so, and radials in R250 and R430. R250 in itself isn't as grippy as the R160 but the radial construction may pay off in terms of footprint efficiency, wear, and rolling resistance. Goodyear doesn't have an awful lot appropriately sized in 13" bead diameter. One other option would be to mount up the D2692 FSAE slick on a 8" wide wheel front and rear. They aren't particularly responsive, though maybe bumping the inflation up to 18-20psi would stiffen the carcass up. That compound on this type of car would probably be ridiculously sticky. That may be a good one in the bag o' tricks.
The other option is to run one of the zillion Hoosier tires available in the appropriate size, in R25 or R35. They even have a 13" radial. We'll see what Jeff Speer @ Hoosier has to say when he gets back to me.
Suspension
Pushrod-actuated double a-arm front and rear. I have some interesting ideas regarding the kinematics. On the front I'll take a look at running a negative FVSAL, putting the instant centers outboard of the respective wheels. The idea being I can run a lot of static negative camber for the radial tires, but on the brakes the tires will go more positive toward 0 camber and give me a heap of braking grip. On-throttle the tires will go more negative camber which may help get an extra bit out of the outside tire, but I'm not sure. I wonder if flipping the IC's flips the jacking effect... ie you wouldn't have jacking on the sprung mass unless the roll axis was under the ground there. Who knows.
The Cane Creek Double Barrel is probably perfect in this application if I want to use close to a 1:1 motion ratio. Looks lighter and more compact than the Ohlins on the Stohr car, and hopefully isn't as expensive while still giving 4-way adjustability.
Allright that's a lot. Until next time...
Formula 1000
Since graduating I've wanted to build a project car. With the '07 Car there were a lot of compromises that had to be made, and some just flat out bad decisions. It wasn't what it had potential to be. So on one hand I've wanted another crack at it. On another hand, going to town on my 2007 Nissan 350z isn't really an option. Don't have an alternate car to drive, I need something that's not a complete race machine for daily driving, and don't feel like buying another car to work on.
Enter F1000. Single seat open wheeler, 1000lb minimum weight, 1000cc sportbike motor. Some aero allowed. Basically a double scale FSAE car. I could just buy one, for example this excellent Stohr vehicle.
But what fun would that be? Plus, buying a full car is F expensive... and being the overconfident jerk that I am, I figure I can do better. I can do some design for now, which will at the very least keep me busy. From there I can get an estimate on what it will take to build, and make the call as to committing the time and resource to fabrication. Instead of taking the normal formula car approach, I'm taking it back old school for inspiration.
Allow me to introduce the 1970 Ferrari 312B, Formula 1 car. It's something different, and I think it looks tight. There may actually be some benefits to this general design (a tube!) Given the relatively low horsepower of these cars, getting rid of the sidepods will significantly cut down on frontal area and might up the drag-limited top speed. Again, with low engine output power I have no need to house enormous radiators in sidepods, and the oil cooling is taken care of on the stock engine configuration. The scoops you see poking up just under the rear wing on the original are for oil coolers... though on mine I will likely use this area for the coolant radiators (radiator on the original being up in the open nose). F1000 does allow an underbody, which I can probably get away with by using a rigid piece of carbon or s-glass.
Enter F1000. Single seat open wheeler, 1000lb minimum weight, 1000cc sportbike motor. Some aero allowed. Basically a double scale FSAE car. I could just buy one, for example this excellent Stohr vehicle.
STOHR CARS HAVE WON SIX CONSECUTIVE DSR NATIONAL CHAMPIONSHIPS.
IN 2007 WE HAVE FORTY TWO NATIONAL EVENT WINS AND
NINETEEN NEW TRACK RECORDS.
But what fun would that be? Plus, buying a full car is F expensive... and being the overconfident jerk that I am, I figure I can do better. I can do some design for now, which will at the very least keep me busy. From there I can get an estimate on what it will take to build, and make the call as to committing the time and resource to fabrication. Instead of taking the normal formula car approach, I'm taking it back old school for inspiration.
Allow me to introduce the 1970 Ferrari 312B, Formula 1 car. It's something different, and I think it looks tight. There may actually be some benefits to this general design (a tube!) Given the relatively low horsepower of these cars, getting rid of the sidepods will significantly cut down on frontal area and might up the drag-limited top speed. Again, with low engine output power I have no need to house enormous radiators in sidepods, and the oil cooling is taken care of on the stock engine configuration. The scoops you see poking up just under the rear wing on the original are for oil coolers... though on mine I will likely use this area for the coolant radiators (radiator on the original being up in the open nose). F1000 does allow an underbody, which I can probably get away with by using a rigid piece of carbon or s-glass.Who the hell is Jersey Tom?
My name's Tom, originally from the glorious Garden State - New Jersey. I spent from 2003 to 2007 at the University of Colorado at Boulder in Mechanical Engineering. At some point in 2004 I thought it would be fun to join the Formula SAE team, and then followed a very long three years. For those of you not familiar, while you're already trying to crank out an engineering degree in 4 years, you spend 20-70 hours a week and a good bit of cash out of your own pocket building a small open wheeler powered by a sportbike motor. You blow off class and homework and GPA building a vehicle which in all probability will fall apart at competition. But its fun, and when you do drive it the first time it's borderline better than sex. Definitely better than sex with a hooker. Plus there's no STDs, errant children, or awkward mornings and phone calls. Other than "Hey hotshot, still think you can machine those uprights in a day? Ours just snapped."
On the 2005 team I mostly wrenched on crap. Since I was working in a machine shop at the time and had some experience there, I got to do a bit of fabrication for the car. Wound up spending a good bit of time getting to know the ins and outs of the chassis, and went to competition. Tough break for our can-do-it-all team captain, problem in enduro and a low finish (after the 2004 team's best-place 23rd).
As a junior on the 2006 team I decided to get more involved and did a design-related independent study. The previous upright design took one or two guys about a week each to CNC machine on campus. Having them outsourced took a couple days each and $1500 for a pair. I thought this was ridiculous. I designed a new, modular upright with "DFM at all costs" as a goal. Pulled some weight out and reduced to fabrication time to probably 3 hours each. There were some wheel bearing "issues" and a lot of joint stiffness compliance, but it just barely held through competition and we held a school-best 22nd place finish.
2007 rolled around and I found myself as one of the co-captains. We started with a budget of -$5000 (that is a negative sign) and 9 senior design members, down from 11 in '06 and 13 in '05. Once again I designed the wheel assembly, and once again had compliance issues. I also CNC machined the vast majority of complex parts on the car, welded the frame, and did some of the DAQ wiring. This had been 3 people's work a few years previous. Bad decision. Needless to say I was pretty busy and did a pretty poor job as a project manager. Got the car done, got to competition, but had a driveshaft "issue" in enduro. Still, I learned a lot, and the judges liked some of the tire data work I had done.Afterwards I traded in beautiful, sunny, no-humidity Boulder, CO for grey, rainy Akron, OH. Why? Great question. I'm still not sure that was a wise decision. But in any event I'm here, working as a tire/vehicle dynamics engineer, and I feel like building another racecar. Which brings me to my next topic...
Subscribe to:
Posts (Atom)