ASME-Y drawings standards
Jamie has metrology tools in the
Caution against just using one person as your check. In industry you can release something, then you can do three signature (your immediate superior, then the lead, then the chief engineer or mechanical lead)
Give piece of mind to everyone involved
Look up Industry standards for reviewing things
Checker
+/- 5 pounds is the start of a good plan or goal
If this is a goal, there needs to be a weight tolerance on a by-part basis
Start with your smaller goals and then come to a larger goal that makes sense
How would you determine how much weight you need allowable on a part
5 thou profile tolerance is pretty standard
Take the dimension of the part increase by 5 thou and find the weight
My questions:
The reason the pushrod broke is because we slammed the wheel into the uhaul for gone to engineering
The reason the rod ends were loading in bending is because the design was bad and we were using rod ends in an improper application, the reason they actually bent is that we didn't pressurize the shock before driving to competition
What is "the relevant systems" mean wrt CG
I caution you to use decimals in places that are supposed to be a rough estimate. Are you really sure of your estimate within 5 significant figures of precision
Other ppl's questions:
For CG bias - need to use CAD to validate that 60-40 is a reasonable CG bias front to rear
Do you plan to consider the actual speed you will be going at in determining lateral load transfer and stuff
1g of acceleration not force
Is your 1g a total car of lateral acceleration?
What does it mean to consider weldability
In the past we used aluminum welding, so we want to consider
Its fine to say you don't want to spend time and resources learning how to design for welding aluminum
Challenge us to deep dive a better reason why we don't want to aluminum weld
He said the aluminum welds themselves
The materials lab upstairs can redo heat treating
Waterjetting
Try to come up with cool shit and come up with cool processes for making it
Factor of safety >2.5 - How do you decide what part gets what factor of safety
We don't know how to decide
Recommend: make a list of what you base your factors of safety around and then wholistically go through each part and assign
Where did we get this number from: 2.5 seems really high, used a FOS of 3
Considering fatigue
Need to figure out how understand it and design for it
You say considering the endurance limit, you're never gonna hit the max number of cycles for material
You clearly haven’t gone too deep into it
How would you go about finding cycles for suspension
Road noise isn't going to affect that much
The big loads: accelerations, heavy decels, bumps
How much starting and stopping and turning and things that would actually show up in an amplitude graph
If we are designing for a factor of safety of 2.5, that mean you expect some random peak load at some random time, then you already built in a ridiculous amount of margin for fatigue
A simplified way to characterize cycles is with shock pots
Pissed at UT for never explaining:
The shocks we use: do they have a certain spring rate? We are trying
Do you know what you're actually gonna use your telemetry data for
My questions:
The reason the cardan effect was neglected was
What do you mean by mechanically complex for the upper steering architecture
Why do we want to use Ackermann
Why a hall effect sensor and not just a potentiometer at the pinion
Their questions:
Are we planning to make changes to reduce bump steer
Backlash in the gears might not be as big of a concern in the gearbox as we think. There are ways to design around it.
Look at the track on google earth and look at the geometry. Also look at DoT turn guidelines to see what radius we might see
Look at how deflection and manufacturing issues might affect steering geometry sensitivity
Rack position telemetry might also be cool to get wheel angle telemetry to see how our model compares to real life
How are you looking at tunability. Recommend doing that. Tune your car for the specific road or track you are going on that day
Steering effort: go into what our plan is based on last year to reduce steering effort
How are you going to decrease the steering effort while also making steering faster
These goals oppose each other, how do we measure how much effort is acceptable
Designing around most common corners
20 thou seems excessive for non critical features
It isn't that much harder to get within 5 thou than 20 thou
The brake disk was extremely hevay last year. They are basically harley davison disks that are waaaaay heavier than they need to be.
You said you "want to do" each of the design considerations, we can't do all of them
What is the benefit of deep groove ball bearings over taper roller bearings
The main issue with the caliper mount being machined on the upright was size of stock
You don't actually know how many miles you need to drive
Do you have an idea of how you are going to size your calipers
No
How do you know our calipers are excessive
For frame tabs an mounting, how are you going to go about the spaceclaim process, and have system be able to communicate where they want their tabs to be while also maintaining oversight wrt to weldability and conforming to VR3 standards
Make a tab sketch
Increase in second moment of the area
What is the purpose of experimenting with composites
Other questions:
You should not need to use FEA. If you can't do a substantial amount of analysis with hand-calcs, you cannot really trust your FEA
20 hours of welding practice sounds good on paper, but where does 20 hours come from on a per team member basis.
How do you determine an "effective weld" from the number of hours
What metric will you use to measure if a member is ready to weld the frame
Timelines often get rushed.
Something Jonathan did when welding the combustion frame was have people who were interested and had no experience just go crazy on the frame and he would fix it
Don't make this too significant a barrier of entry to learning to weld. Don't want to discourage people
Why make tubes diameter larger and walls thinner?
Is it for bending? A space frame should have minimal bending
The types of load cases we sim for on our regs induce bending to test your joints
Terminology: ultimate yield is not a word
How are you going to ensure that you and your members gain a foundational understanding of solid mechanics such that they can properly utilize FEA, and have confidence in its results
Frame jig - you don't really address
What is "redundant" cross tubing? If it is redundant why is it there
Something you need to consider that you didn't list on there: how are you actually going to test it. You keep bringing up this topic in an abstract sense but I've seen no actual testing plan or specific details of any kind. We already had the discussion that we are not going to put substantial resources into this project because why are we trying to have composites on the frame when we clearly don't understand well enough how to design a regular space frame, which is unambiguously easier
Composite simulations - do you really think it is realistic to know how to sim composites when we barely understand regular structural FEA?
Setting metrics: you are going to get to the point where you need to make compromises. Thinking about this early on will save you a lot of pain
Do you have an estimate of longitudinal CG?
On the timeline: have roll cage design or position finalized
With aeroshell, their timeline is going to be really long. End of october if they want their molds in on time
The biggest driving thing isn't overall dimension, it is the roll cage
Need to give aeroshell the bounding box
What drove 30% weight reduction
Winning solar car weight is 120 lbs.
Be careful with your metric. Using written reference is fine, but think about what you are willing to compromise. Frame isn't the number on priority for weight reduction.
Stop being so vague
How are the jigs made
Egress
What changes do y'all imagine to make egress easier
What does this actually look like
A lot of teams had an open roll cage design
Our was caged in from all directions
This makes packaging of the aeroshell easier
Think about bars they can hold onto, and footholds
For VR3 timelines, you need to play way more in advance, at least on paper
Do be careful about relying on a single manufacturer
Keep CR in the background looking at other options
Minimum lead time is 3 weeks
Ansys: what is stopping you from using beam elements to iterate?
Nothing
FEA was taking 8 hours long. This is ridiculous amount of time. You should be able to iterate with beam elements in seconds. There is no requirement for you to simulate the entire frame as solid elements. You can use beam elements
Great we are trying to validate to hand calcs first
Didn't mention anything about tube material, only tube geometry
Last year used 4130 chromoly steel
Learn about weldability
Timeline
Has this timeline been discussed with the other teams
Your frame welding timeline is in direct conflict with other teams
You keep saying bounding boxes and packaging, how are you going to actually track it
Highly recommend you turn construction lines into solid bodies
What is the point of carbon fiber on the seat? Expanding foam is good and fine, the fsae team did it, but why not do regular upholstery
Wouldn't go with carbon fiber for ergo stuff. Seems like overkill. Could get favorable weight reduction by using some waterjetted aluminum piece or something
CF panels are strong in tension but this case its not the same. You're trying to bend it. Do you really want a composite panel?
Seen great success with 1/8" waterjetted metal
If you want to use a composite panel, why CF? Would be easier to use fiberglass. Rigidity without weight
Fiberglass is cheap and easy to manufacture
Carbon fiber steering wheel
Not really seeing high loads- maybe 3D print something
They said not to do that, especially because of the quick disconnect
Discern what is a structural component and what isn't
Forged carbon fiber will take longer than you think it will
Pedalbox is going to be a timesink
Ergo jig
Great direction to go in
However, can't get reliable information based on height
Short torso, longer torso, etc
Better way to approach: set some standards, measure from top of head to hip-bone, etc
FSAE has a rule for encompassing drivers, solar doesn't
Egress
Have you thought about how you will model egress?
Moreseo happening in the ergo jig?
Haven't planned a set method of testing egress
It is hard. Esther never figured it out. Try and find something that isn't just "I think somebody could jump out of that"
If this is intended to race ASC, ergo needs to think about race strategy, check boxes
Measuring people front
Be more granular with how you measure
In your jig, have something where people sit down
You can have the steel backing help the roll cage structurally
Sensing braking
Be careful with measuring the angle of the pedal. Don't pick something that is separated too far from the actual action of the brake calipers closing. Keep it more closely linked, ie fluid pressure, force, etc
Controls actually has all the sensors. Taking over the brake line sensor, since they will need for electrical braking etc.
How measure braking force for drivers
Calculate how much force we get through the pushrod of the master cylinder from pedal ratio
They figure out how much pressure this goes into the brake lines
No, like how much force the driver is actually applying to the brake pedal
Figure out how hard a driver can step
The takeaway: think of a way you can get to that metric and multiply for adrenaline
If you didn't have plans in measuring how strong your drivers are when braking, how do you size your components
Expanding foam is kind of hard. Need to consider. If you add too much padding it is useless
What is the value of the driver model
This is purely for packaging
How does your dynamics team feel about CG with your driver
Plan for ballast
Again, CFD is impossible without an understanding of the thermodynamic and heat transfer principles behind it
Do you have any specific plans wrt characterizing the thermal conditions of the battery? If you want to do hand-calcs of thermal FEA, you need inputs. If our cells have poor documentation that is a problem, if we don't know how much heat they are generating in varying conditions that is a problem. It is good and fine to say we are going to figure it out, but the how is the important part. It is your responsibility to learn
How are you going toe valuate the viability of the heat pipe and cold plate prototypes
You mention a lot of testing for the composite panels. You need to make a structured plan for how you are actually going to conduct this testing, the logistics of where and when, the variables you will be changing, how many tests, and how you will document the results. Testing without organization is a waste of time
Average RMS current draw
Dependent of the weight of the vehicle, haven’t fully tested
8 amp nominal low voltage system
30 amp nominal on motor
40 amp pull nominal
Don't know what the cruising current draw is for the motor
What about charging
Able to chart at 6 amps during MPPTs
Assume maybe an amp extra?
5 amps per cells, do you even need that much cooling?
Average ambient outdoor temperature
32 degrees Celsius
Kevin: pick one thing you want to have that safety factor on. If you have a bunch of safety factors stacking up
If you want to set the cell temp at 45 C, have maximum operating envelope on every part along the way so you don't derate your pack
If these cells are so poorly documented that they harm our engineering process, maybe we need to find better cells
Dr Subramanian's lab - mech e professor with a lab electric used to characterize their battery
Are we testing water ingress
Yes, without the cells in there
Other teams have seen 5g as development level for bump loading, will it kill the design if you bump that mechanical to 5g
Have we considered a new electrical architecture for the battery? Having current collecting board
Really only think we need 3 point bend test and shear test, and then can extrapolate everything else in a spreadsheet - Kevin
You have a PCB for every battery module? That is a lot
When we are rushing to manufacture the wires can break off easily
The fuses are on the BPS side. If you have live wires that are not intrinsically fused, put the fused on that board. Also put it on the BPS
There is a 1D vibe table at ferguson at pickle
The majority of failures are stuff that are poorly soldered and poorly crimped. Do quality control
Only get automotive rated connectors. If you get JST connectors, bullet connectors, they pop off. You need positively locking connectors
When did we actually see signal interference at comp?
I'm kind of confused what the problem with the E stop is? Why do we need to look at other solutions?
You should create a standardized inspection scheme for verifying that connections are done correctly, won't snap off etc. You want piece of mind, and even experienced people make mistakes. You want redundancy.
Why don't we just buy more crimpers. Like literally rn
Other:
Panel mount connectors
Electrical standardizing can and 12v to use panel mounts
Are there plans to look into that
We started trying this last year, but we never really implemented it
Kevin: Recommend that the board in the enclosure not mount to the panel mounted connector. Generally for 2 to 4 pin connectors it is hard to find panel mounted connectors.
Want to spend more time learning about what we can use, see what's available to us
Are you shielding high voltage wires?
No they weren't. This is what caused the Emi problem
Shielding is gonna happen this year
How fast is the CAN running at
1.5 kbps
No, the frequency of the CAN
Each system will have its own internal CAN line
YIKES - Kevin
That is a lot of wires. Is that four can lines all separate
Doesn't make sense to have 4 different CAN lines. Should have enough bandwidth to process all of the systems
It's not implemented yet.
We need more discussion. Reduce the number of CAN lines
Figure out how fast and how frequently you are pulling sensors, what is your critical communication, and then see if you can squash all of that into one CAN line because it will really simplify your architecture
Were able to fit telemetry and controls into 1 CAN line
In terms of waterproofing in the context of panel mounts, are we waterproofing the whole enclosure or the board?
The enclosure
In terms of the connector, we will have one board per enclosure
Cable sizing
Are you going to run a lot of different sizes of wires
We have more than we needed last year
We were using the 50 year old pickle wires
Going to invest in new wires
If we are buying industry grade connectors might use 2 gauge for the wires
If your RMS is 30 amps you can run the 12 gauge wires
Low voltage, 22 gauge to 16 gauge
Want to reduce the variance in cable sizes and standardize it
Do we have twisted pair CAN wires
We make them on our own, EPLAN helps with length
Kevin says we should buy them and they have shielded
If you want to make your CAN line faster, shield bold sides of the wire, make it shorter, reduce number of wires
How did you determine that 8 fans were needed
How are you going to validate these hand calcs?
How are you going to learn about and teach members about heat transfer
What testing are you going to do to characterize the thermal conditions of the battery
Are you just measuring the temperate of the cells? That doesn't tell us the heat generation if we don't know the thermal capacitance of the cells?
What if a thermistor is on a cold spot of a hot spot
They epoxy it to the center of the cell
In general 60C is a good target for the center of the cell
You need to set the RMS current number at the golden number that you look at
It is what you try to design around
It drives heat generation for everything in the whole loop
It is what causes the heat generation. If you cant quantify it you need to choose a number and if a car goes above it you have to rerate
Thermal mass of the cells would also be good
Design our cooling system slightly above the power budget and then having tachometer and anemometer at the end of the cooling pipe.
You will get leakage through you lid
Tachometer: built in to the fan and measures the speed of the fan to extrapolate flow rate
Anemometer: goes at the end of the coolant loop to measure how fast air is leaving
You can estimate drop
You should run your fan duty cycle based on how much at the end of the loop
At the battery exhaust
The biggest thing that affects your cooling design is the THERMAL GRADIENT BETWEEN CELLS
The cells first will be cooled more than the ones in the end
Tend or parallelize your airflow paths so your cold air touches the most number of cells in the shortest path, so you have the best thermal gradient
Why does surface finish matter if we will be vinyl wrapping it
What types of reference points do you want to have on the mold
How are you going to develop your knowledge of fluid mechanics
Is there a numerical reason for going to 4 ply. It is flimsy with 6 ply? How do you know you did the 6 Ply
Why was the six ply flimy. Maybe the error is with your manufacturing and not with 6 ply
If you have a lot of resin in between your ply you are cutting a lot of strength
How do you plan to validate that?
We are doing testing. Advait will talk about composites stuff
He would go with 1 or 2
Not 1. Concerned with pin holes, don't want to do bondo work
I would make the structure that holds it strong
Esther Ply
Understand wanting to reduce the pky. Even one layer reduces 5 hours of time
Discern who is focusing on the track race and who is focusing on ASC
The shell is not structural but it still does have to withstand these environmental conditions
The sheet amount of time that goes into such a thick layup, the resin is already blushing by the end
References in the mold
Wheel cutouts can have bosses/indents for the cutouts
What references were you thinking about
Just having a boss
Even just having an accurate center line, would be useful in terms of measurement
What does your reference physically look like on the mold
Say you have the bottom shell mold that is female, and you have a little bump out of the bump on either side and pull a string across
Drill the mold, put a locating dowel pin in, then do the surface prep to seal it, then you have dowel pins/locating and you'll increase your bearing reinforcement, and you have actual locating holes
Worry about setting a locating feature and it being on the actual part
Want to make sure that you can hand it to inexperienced people
Nothing on the mold surface, it is on the boundaries?
For the example of the wheel cutouts,
You don't want references on your mold surface
If you see these reference points changing, you could have laser cut plywood that follows the curve
If you want more information on that style: talk to electric
Why 3 minutes for tilting?
Want to think more about. Came up with it a few days ago and said it was a good idea. Need to solidify it more, but seemed like a reasonable number
A good target but not really important
How are you going to develop your knowledge of fluid mechanics
A shell made up of the best performing sections from 3 different shells may not make the best shell. How do you even characterize if a "section" performs well
I'm not sure we really fully understand the cons of an external suspension. Why did they move away from it on texsun? We should investigate
How are you going to make sure that the ansys sims are valid
What is the defined engineering fundamental when it comes to aerodynamics?
For combustion and electric need to worry about downforce, angle of tilt of airfoils
For us, it is really just the equation for CD
How the roughness of the surface impacts the reynolds number
What we see is still a good usage of engineering fundamentals - Wesley
Ansys fluent gets really bad when you start to go transient. Your ability to visualize fluid flow is good, but it becomes a multidimensional thing
Sharp corners bad
Avoid 3D printing in a wind tunnel. The surface will affect the results a lot
Your job is not finding the most optimized shell design that can exist
You will be limited by frame and dynamics
How to make sure sims are accurate: you won't know, just pick the best design
Excel sheet is a good way to keep track of things
Statistical analysis: how do you figure out which combination of things will be the best. There are so many combinations; You will notice trends in the data from Ansys, some of them will be conflicting
Try to think of how you can keep track of that
Unfortunately the best way is just to run everything
There's a reason why F1 still has a wind tunnel. If sims were perfect they would just use those. Doing physical validation is important. You could get a good surface quality 3D print if you go to SLA or SLS, and start doing more surface prep, you can get a close surface prep
We have machines that can measure surface roughness, you can calculate how much your different roughness affected your quality
The biggest way this aeroshell will impact is drag, but WEIGHT.
Energy is your acceleration, the more mass you can drop the more efficient. See aero push dynamics, battery, frame on not having a car that is so big and so long.
He looks inside the car and there is a lot of empty space
Esther's comment on the size
The array, want to have discussions with them. Designing for 4 m^2 isn't the best
Need to have a plan in mind
Start with the frame, see how many cells you can fit inside the frame. Start with the most compact frame you can, and see how many cells fit in that footprint
From there increment out, instead of starting at 4 m^2
Push back against array. Array will want more cells, and Aeroshell needs to push back with reason
Most critical part for shell to care about is the roll cage. Slope and height of the roll cage
Design approach is great. Can look at other cars
This depends on us having our top shell and bottom shell fixed. Who controls this?
Are there going to be versions of this as well or will you split them after
If one team is working on the nose and also the top shell, this will be another factor
Might be Aidan does the top and the bottom and everyone works with it
His pitch: nose separate, canopy separate, tail separate
BUT: these all rely on tangency with the profile of the top and bottom shell
Esther did the whole profile at once, and then split it across the top and the bottom
Maintain tangency
The downside: only one person works on it
Once the frame and dynamics are set, you can start doing the body work
Packaging is our hardest battle
Esther says have ribbing to support the topshell outside the bottom shell
Aero and Solar's next action item is to have a big discussion regarding the roll cage. It will affect everything
In her experience, the biggest headache has been fitting the roll cage
It kept changing every day
Occupant cell affects the angle of your windshield and the location and height of your canopy
Say you have a set length. You know where it is steeper you know where it tapers off
You also want the canopy to be aerodynamics. It is all because of the roll cage
With the discussion
Frame wants it to be structural
Aero wants it to be aerodynamics and manufacturable
There is a lot of wiggle room, more than you think there will be
Packaging is king
Are we doing bulkheads?
IDK
What constitutes "finishing" wrt the composite frame aka sandwich panel testing
For windshield, you could get a fiberglasss surface, heat it, and droop the polycarb over it
The issue with thermoforming polycarb is that it sucks up every bit of texture that is there
The thermoforming tool is 19000 dollars because any degree of granularity creates fogginess on polycarb
Try to design the canopy in a way you don't need thermoforming
Once you cut the canopy all the stiffness comes from the polycarb
Don't use carbon fiber for the canopy
Just use fiberglass for the canopy
Or you could do what
Map out where fiberglass goes on the mold.
Consider making the area larger
Just do the layer of kevlar and the layer of fiberglass. The stiffness will come from the polycarb. Don't need to worry about reinforcing with CF. It is strong through continuous strands.
Or make it all kevlar
Wesley
Carbon control arms
Trying to make it all yourself out of prepreg
That is a very challenging thing
There are issues with doing it with the melt out core
Making your control arms air foils, the performance gain you will see from that is super super minute
Making a carbon control arm interior that is stiff, and then putting an aerodynamic shell on the outside would be easier
Just buy carbon rods off rock west and then do an outside carbon shell
Worried about sealing the split mold
If it's not a perfect match and you don't have good compression, you will have a bad seal
You will have leakage, you won't get vacuum
Interested in the canopy, and what determined our metrics for the plys. The ply count seems arbitrary
It's regs :shrug:
The issue we had with the PVA was that we would spray the PVA, and you'd do it on Saturday, they workday would end and everyone would leave. Then you'd go that week and do it in April (when it was rainy) and PVA dissolves in humidity.
When you spray for PVA< spray at 7AM, then layup that mold. If you wait too long you will have rpbolems
Last years they did apply the same day as layup and still had the same issue
You have to let the PVA fully cure and then spray on another layer
Or spray a bunch of thick layers and then do it once
Demold R&D is interesting. If you're having molding problems, acetone and hammers
Composite frame and testing
Seems great, seems like it would be fine
However, aluminum core density is what makes or breaks your bending strength. If it is too low of a density you will just crush your core, if it is too high, should have just made an aluminum monocoque
Do remember you need adhesive sheets, don't just hope and pray it will bond. You need some kind of post bond or added adhesive
You would be making the skins independently. Secondary bonding
ASTM standards - > good luck. Their composite standards are OK.
There is no set way to get a panel's strength
You are better off doing the standard composite theory test to determine your individual moduli in tensile compressive mode and a shear mode to correlate back to the simulation
Panel strength doesn't matter: there are so many orientations the panel and carbon can came
You want to test the material not the panel
The properties of prepreg changes over time
You can test wet layup material properties. Its not an easily repeatable process wrt to density which is your biggest player
Your panel might be better or might be worse later
The testing is mainly to see if we need filler etc. Figure out what we need to use
Molding for bottom shell and top shell: gel coat
It is ok.
For bottom shell, gel coat does maintain its thickness, but bad for vertical walls
The high vertical wall you'd have a lot of buildup on the bottom
Imagine you are pouring molasses down the side
Testing demolding
Hot wire cutting suffers with the 2d part of it
Our molds are not 2 dimensional
The biggest worry is that our molds are not 2 dimensional
A bowl is way harder to demold. We are an elongated bowl
Look at milling pucks of foam into bowl shapes
Could even scrape it out with your hands into a bowl
Pushrods
Pretty easy
Adhesive testing in pushrod, matters how your dynamics is set
If your dynamics is well thought off, and your rebound strength doesn't go into a tensile strength
You don't technically need a bond
Buckling testing can be important. FEmap will do your work for you
A 6x safety factor will still be 8x lighter than the steel equivalent
This is not as big of a concern
Resin Infusion
The time it takes for the resin to flow vs the tacking and curing time will be a problem
For this big of a part, how do you get a good seal and succeed at a drop test
Need a 6000 dollar pump. If you want to do this you need to own the pump
Battery layup
Might look at fiberglass instead of kevlar
IDK how much impact resistance you need
Cooling on cells
Cutting holes in the top shell would make it extremely flexible
Youd probably need to increase weight back up
Try seeing at intercooling through your core
Save weight, save time, save hole saws
Steering wheel
Don't let ergo do the composite steering wheel
You will spend a lot of time going back and forth with them
Clayton
Doing big layup over summer?
Why are you trying to get your plug manufacture in may
If you let it sit at pickle over the summer something bad will happen
Actually we just need it by start of school year
Resin infusion
You have to be very sure
You will have at least 6 resin entry points on something this big, with a perfectly sealed vacuum bag, and many pots of resin being mixed at the same time
Resin infusions this big you'll have a team of 6 professional layup people working on it
Esther
Having a mold by May: keep it
Deadlines will always get pushed back
Keep your timelines early
So easy to push stuff back than to quicken it up
Keep your timelines early
Eesha