new Lathe!

Finally got a lathe today.  I have been lurking on CL for quite a while.  The guy who had this used it for miniature engineering.  He had an incredible collection of tiny steam engines that he had built.  He had done a lot to this machine to make it work better, and add functionality.  The bad news is that it was totally wedged into the back of a very capable shop with a lot of equipment.  It took us something like 3 hours to get it out and into my car.  I don't know where I will put it yet, but I have a couple days to figure it out while I wait for my headers.

how to drain the stink

I spent the weekend working on the exhaust clearance issue, and figuring out what I want to do for an acetylene torch setup.  I cut the subframe back some more and cleaned it up.  I see three options for the interference issue from last time:

1) modify the subframe:
Pretty substantial. The current header is pointed more or less through the middle of the subframe in structure that I can't just remove. I think I could weld a box tube and some doublers underneath to blow the center section away, but it will be gnarly. Since this subframe is pretty highly stressed, the less I have to chop on it or weld to it the better. As it is I already have to be very careful to make sure I am at least as strong as the original.

2) build a new subframe
Will look cleanest in the end but is the most time consuming.

3) get a different header
Could give me the quick way out if i can find cheap headers.  My first thought is block huggers which outlet straight down near the middle two cylinders.  I think this creates new problems for me (exhuast now outlets where my rack and new subframe are).  Also, block huggers are worst for making power.  My next less intuitive option is to go with a tri-Y:

 

Even though there are more tubes, I lose the flange and the outward cant that is pointing things through the subframe.  To make sure I am not wasting my time, I took some measurements of the image below.  Since I know the distance between the primaries on the engine, I could scale up anything I measured in this image.  I think I may have a collision, but it will be in a different place, and if it is less substantial, I could get away with less subframe modification.  Shipping to me is free, so all it costs me to try is the return trip.  The stakes are big enough, that I figure it is worth a shot.  So I put my bid into the summit fairies. 

 

These are wicked cheap by the way $165.  If they are no good, I may try the block huggers before commiting to the chop and hack campaign.  There is always the kugel option, but that is a long way down at this point: I would have to buy their headers, and their subframe >$1500.  This would also likely mess up my rack location and suck me into their steering arm pieces and the engine frame (>$1700).  So this is sort my whole autonomy at stake here.  Worth a few trips to UPS.

subframe positioning

I decided to try and use the original subframe by cutting the middle out.  to clear the trans:

I plan to clean this up, and weld some plate onto the open section to put the stiffness back.  I think I will also add a piece of square tube to tie the left and right together again.  This tube will connect at the furthest forward point on the subframe (right side in the image above).  It will go through the high area in the sump and hold the rack.  It will be hard to squeeze the exhaust through.  It was hard to get a good picture, but see below.  Probably good that I am figuring this all out now before the engine frame is done and the engine position is totally locked.  

 

planning the subframe

I don't have a lot exciting to share just now.  Now that the engine is placed I am designing some structure.

The structure must:
Hold the engine
Hold the suspension
Hold the rack
Be rigid for each of those things
Attach to the frame and the undercarriage
Dodge the sump
Dodge the exhaust
Be manufacturable by someone one with just cutting disks, drills, mig machine, and possibly a tubing bender
Allow servicability of reasonable items (no good if I have to remove the engine to change the oil filter etc.)
be cheap to make

I have a safety net in all this.  The kugel parts for this part of the job run about $3k All together.  I expect I should do this for ~600-1000 in material.  At that rate I can do it all a second time, and still have some real money left over.

I have some ideas from looking at as many other focus's on the web as I can find.  I will keep you all posted as I start to zero in on a plan.

engine placed

Placing the engine is pretty critical.  Too low, and there won't be enough ground clearance, or room to get the structure and the rack under the engine.  Too high and the hood has to be cut.  Too far forward and the radiator won't fit, and the shift lever won't line up.  Too far back and the exhaust won't make the turn to go under the car, the brake lines don't have room, and the engine may rattle against the chassis while driving.

Also, the engine must be between 1° and 3° downhill towards the back, dead level on roll axis, aligned with the center line of the car on the yaw axis.

As complicated as that sounds, pretty much I shot for as far back as I could get without hitting stuff, and as high as I could and still maintain pitch angle.  As we saw last time this meant chopping a bracket inside the tunnel.  Once I had the proper angle on the engine, I found that the transmission collided with both sides of the tunnel well before getting to that bracket.  I decided to massage the tunnel in a couple key spots with a BFH.  It worked beautifully.  Car steel is very soft to facilitate stamping and forming.  Even in the spots where there was a doubler, it wasn't hard to knock some good dents.  In stubborn spots I put the back of a ball peen hammer in the precise spot I wanted, then hit the front of the ball peen hammer with the BFH.

 

This all allowed the transmission to go so high that its entire top hit the entire top of the tunnel.  I think it is safe to say that any higher would be impractical.  I lowered it a tiny bit from here and called that good.
 Next was the angle.  I measured before starting the projects and found that the rear fender is 0.25 inches higher than the front fender at stock ride height. Since I know the wheel base, I could measure the current fender heights and figure out what angle the car is at compared to how it should be, and adjust engine angle to account for this.  I started out with the angle finder flat on top of the lower intake manifold.  I had previously noticed that the sump seemed at a different angle so I checked it across three sump bolts and saw it was vastly different than the top of the intake.

I figured that the tail shaft of the transmission is really what matters so I checked it and got the same measurement as for the sump.  

The engine didn't hang level on the roll axis.  I hung some weight from the header to torque the engine to level.

For the yaw axis I held a board against the inside of the bumper and measured the distance from the left and right edges of the crank pulley to the board.

I was psyched that engine did not shift at all when I released it onto the stand.  The leveling feet that capture the screw heads worked really well.

Just for a sanity check I scoured the internet for someone else with this engine trans combo.  I found some pictures and ground clearance info for another guy's set up, and they were identical to my placement as far as I could tell.

engine and transmission height

Today I wanted to look at getting the engine height right.  Before I scattered the focus I measured the distance between the top of the tire and the inside lip of the fender so that now with the car up in the air I can see where the ground is and see what my engine clearance is.  My guess was right on the engine being low.  The drain plug would just barely drag along the ground. So the engine needs to come up about 5 inches for stock ride height.  7 or 8 inches is probably better.

I saw the shifter was not hard against the tunnel, but when I jacked the transmission, there was a lot of resistance like something was hitting.  There is a bracket in the tunnel that was holding the old tunnel inhabitant, the exhaust.  I previously cut the exhaust hanger off, but the bracket is still in the way.  You can't see in the picture but there were a series of studs that held on heat shielding with tinnerman style nuts.  I zapped those off while I had the cutting disc in hand and everything out of the way.  I also made some sharpie lines from the reference hole I drilled for the shifter so that with the trans back in I can evaluate the shifter position.

This bracket was colliding with the transmission tube along the top that holds the shifter shaft.

engine + tranny on jig

I wish I could tell you that from yesterday's post I picked up the transmission, slid it into the clutch and bolted it up.

It was more like:

1. Lift the engine
2. Disassemble the stand and remove
3. Circular saw the stand in half since it blocked the jack from lifting the transmission
4. Take out the trans put it on a furniture dolly and slide it back in.
5. Lower the engine and shim
6. Slide on the trans and bolt up
7. Lift engine and jack trans
8. Reassemble the stand
9. Lower the engine and shim

Completed that in one hour which was exactly my time budget, and progress goal for today.
BTW as you can see, the trans fits in there really nice.  I have plenty of room to dogleg the shifter under the center console if I need it, but it won't need much.  The shifter is only about 2 inches forward of where the focus shifter was.  The engine feels really low right now.  I think I have some fiddling to do yet.

engine positioning jig

My next step is to position the engine in space without using the engine mounts so I can build a subframe.  I made a little stand to do help this that should keep out of the way:

 

 
In the first image you can see that the closest nut has the threads drilled out.  This diameter allows it to slip over the bolt head on the sump bolts.  I made four of these and they all attached to a piece of plywood.  I can thread them up or down to get the engine the right height and angle.  Since I made them independent of each other, I can move them to different oil pan positions if they end up in the way.

I pinned a strap under one of them so I can lightly pull the engine down on the stand.  This way I can still lift the engine and slide it around to the final spot without lifting it off the stand.

ackerman III (the payoff)

Now that we have some mathematical expressions to describe ackerman in terms of steering geometry I was able to look at how different steering system ideas affect ackerman. First, here is the tool I made for excel:





This has all the variables I defined yesterday. I use the convention to put all user inputs as blue cells. So this describes a car turning left with rear steer. The whole idea of this is that the angle alpha would be different magnitude if the rack moves left 0.5 inch vs right 0.5 inch. So the second group of rows are negative rack position for the other side of the car.

The corresponding left and right side instant center differences are used to calculate ackerman %. Just for reference I threw in what ackerman percentage corresponds to parallel steering. I think that this is sort of the limit of remotely acceptable.

I also made a few plots to help visualize what is going on:

  

Left: shows how two variables are related.  Right: gives a visual representation of of the steering components and their positions based for each row on the left side.

The plots above go with the first spreadsheet and are close to my stock Ford Focus steering geometry. So as the rack moves up (top point in plot) you can see the angle of the wheel changes (segment from 0,0 to about 4.6,0). When we look at the spreadsheet we can see that the factory set up is for slight anti-ackerman (slightly negative). This is what we would predict for a street car. Now we can look at what happens with each of the solutions being considered:

Option 1:  Moving the rear steer rack forward

Several days ago I posted an image showing that the rack on the stock setup can not stay where it is. What happens if we still let it rear steer the knuckle, but move it way forward like this:







Notice that rack movement is 0.6 inch for all rows, bot that turn radius increases for the negative rack positions. So we can concluded that the steering gets less sensitive as the rack moves forward.  We can also see that things deteriorate nonlinearly in terms of ackerman.  By the time we are 8 inches ahead of the axle we are way more anti than parallel steering and although not shown this gets worse for tighter turn radii.

Option 2:  Swap left and right knuckles and front steer.

The shown geometry actually describes this already. Only the ackerman calc changes, and left and right change for forward steer.  As a result we end up with slightly pro ackerman, but nothing that dramatic.  Which was a surprise to me.
Here are plots looking at what happens if we vary front steer rack position

Above we can see that the further forward we move the rack the tamer everything gets.  This is the equivalent on a rear steer car of moving the rack further back.  Only here we diminish pro-ackerman instead of diminishing anti ackerman.  To see how it behaves over the range of steering I plotted + 8 inch offset:

This looks as close to perfect as anyone would need.

Conclusions:

It appears that moving the rack to the opposite side of the axle for either front or rear steer is a bad move.  Things start to head downhill quickly.  The surprise win is that if I swap left and right knuckles, I don't need to correct for ackerman if I can locate a front steer rack far enough forward.  If this works out spacewise, then it will be a low effort and low cost solution to a major hurdle in this build.

ackerman II (calculation from steering geometry)

I had a big elaborate way to calculate ackerman using intersecting circles and a bunch of algebra that I programmed into excel.  Then today I saw the diagram a little differently and came up with a much simpler way to calculate this.  I spent 30 minutes at lunch while eating to replace about 4 hours of previous effort, and it works much better now.  Not sure whether that makes me happy or sad.

The goal here is to be able to find the steering angle as a function of the steering system geometry, and rack movement.  Then if I steer the rack say half an inch +y, I can calculate the wheel angle on each side of the car.  If I know both those angles and I assume the outside angle steers the car, I can figure out how much anti or pro ackerman I have.

Defining variables:

Here is a bird's eye diagram of a rear steer system on the driver's side to help that make sense:

We will make expressions from known dimensions to describe the position of:
The lower ball joint (point A)
The intersection of the hub face and perpendicular line to the tie rod axis (point B)
the the tie rod axis of rotation (point C)
the ball joint on the end of the rack (point D)
Note that we can calculate Dy0 which is the rack y position for alpha = 0.  Dx is mu for all rack positions.  We can then say that Dy = Dy0 + rack position we choose.

Note the shown coordinate system, where:

K1 and K2 describe the hub face and the steering arm offset (for ackerman)

R1 = the tie rod length

R2 = the fixed length from the point of rotation of the knuckle (A)to the point of rotation on the steering arm (C)

R3 = the distance from the point of rotation (A) on the knuckle to the end of the rack (D)

mu = how far back the rack is set

Finding the angles

delta = Asin(mu/R3)

law of cosines to get gamma:

gamma = Acos[(R1^2 - R2^2 - R3^2) /(-2*R2*R3)]

beta = atan(k2/k1)

Finally:

alpha = 90deg - delta - gamma - beta

Once we have these angles it is simple trig to show where each A,B,C,D are which will help us graphically picture what is going on as we will see in the next post.

Calculating Ackerman

I find lots of things talking about ackerman, but I haven't found much describing how to measure it so I had to make up a way:

We can assume that the outside wheel does the steering.  Since we know the wheelbase (B), and just found the angle the right wheel is at, we can find the distance (x).  We can do the same for the left wheel.  Then we can take (Rx-Lx)/Rx to express ackerman as a percentage.  If Lx is further out than Rx (as shown) we have anti ackerman. If it is further in board then it is pro-ackerman.

Next time I will show all this in excel as well as the ackerman dependance on a few of these variables.

ackerman

I decided to focus in the cheapest two front suspension options.
1)  Move the rack far enough forward to get out of the way of the stuff, but still rear steer
2)  Swap the left and right knuckles

The issue that makes or breaks either of these is options is ackerman.  When driving in a circle, the inside tire is going around a smaller circle than the outside circle.  Ackerman steering refers to the system of linkages in the car that corrects for this (sends the inside wheel on a smaller circle).  This is done with the steering arm shown in dark brown below.

For perfect ackerman steering the lines extended from the axis of rotation of each wheel intersect at the line coming from the rear axle (so that all wheels are turning about the same instant center).  When going straight ahead. this means that a line extended from the arms intersects at the rear axle.

We don't want perfect ackerman though.  When a wheel turns the rubber in the contact patch is getting twisted such that the car is travelling straighter than the tire is pointed (called slip angle).  This exact angle depends on lots of things, but since we know that the outside tire is doing most of the turning, it will have a bigger slip angle.  This means that while both tires are turning more than the car is turning, the outside tire has to do this more than the inside tire.  So we need a little less ackerman than perfect ackerman (we will call it anti-ackerman)

Carroll Smith writes in engineer to win that while what I described above is the prevailing wisdom up to the time he wrote the book, he found that using more ackerman than perfect (pro-ackerman) actually led to some really nice handling characteristics.  Particularly that the car was looser (more prone to oversteer).  I would like to speculate on why:

With anti ackerman, if you hit a bump or temporarily lose traction with the outside wheel, the inside wheel suddenly gets more load and has more influence on the direction of the car.  With anti-ackerman, the inside wheel is trying to send the car into a larger radius turn (understeer).  With pro-ackerman sudden loading of the inside wheel will pull the front of the car into the turn and initiate oversteer.  This achieves a similar effect as a static toe out setup without some of the other flakiness that goes with it.

So next we need to describe a way of measuring this.  Then we can look at how ackerman changes when the rack moves forward or backward, or if the steering arms points in a different direction.  This is already getting long, so I think I will break this here and look at that in another post.

Here is a pdf document that describes some basic suspension terminology.

Here is a great article about ackerman that references a few articles and texts.

rack vs tranny

The tranny on a front engine rwd car is in the same space as the rack on a rear steer car.  Which is why I am having to change some things.  I put the subframe back in the Focus loosely so I could measure some steering geometry and took a few pictures:

Here is a good look at the knuckle:

Here is a look at the engine bay with the subframe back in:

Here it is closer.  Notice how the steering rack totally blocks the tunnel:

Here is a look at the passenger side knuckle in the car with the tie rod hooked up to it loosely.  Steering arm on the left, brake caliper mounts on the right (you can see it from the other side in the above picture.  I was talking with Seth earlier today and we conjectured that you can tell if a car is front steer or rear steer by which side the caliper is on (should always be opposite).  Go ahead, impress your your spouse or GF now.  Did you know that on almost all cars the gas fill is opposite side from the tail pipe?  I have yet to find a counter example of that either.

Front suspension options

As I work out where the engine will be located, I am also trying to come up with a strategy for the front suspension.  Here is what the front suspension is like (looking in from the front) on many front wheel drive cars including my focus:

The rack sits behind the engine and transmission.  The steering arm comes off the back of the knuckle (known as rear steer).  The problem with this is that on a rear wheel drive car, the transmission and engine need to be right where the rear steer rack is.  Here is how a similar setup works for a front engine, rear wheel drive car:

Notice how the the steering arm comes off the front of the knuckle here.  This puts the rack much further forward.  On a mustang the rack is under the sump which has a bump in it to make room for the rack:

Here are some solutions that won't work:

 

Use the rack from the mustang and just swap the left and right knuckles on the focus to make them front steer.

I really liked this idea for a few minutes until I realized that the ackerman would be reversed that the car would be undrivable (outside wheel trying to turn a tighter turn than the inside wheel).

 

Stay with rear steer, just put the rack below everything

This is a no go because the rack will be way below the scrub line of the car which means if you run something over the rack will get hit which is wicked bad news.  Even if this wasn't an issue, having the rack really low will cause horrible bump steer.

 

 

Here are some solutions that might work:

 

Just move the rear steer rack way forward, but still let it rear steer the knuckle.

The rack can be moved forward a bit. In Carroll Smith's book "Engineer to Win" he describes just that. It will cause the rack ratio to get quicker at large steering angles, which is fine. It will also increase ackerman.  (I think I will discuss the implication and magnitude of this in a future post) I have to assume there must be a reason nobody moves the rack forward to solve this. There are a couple I can think of: The tie rod forces increase as the rack moves forward, and the steering ratio gets slower (at small steering angle) and becomes nonlinear. Also there is a very limit range of motion in the rack ball joints.

 

Use the mustang knuckle on the focus strut and control arms

If it fits or can be made to fit.  Also will have to think hard about the effect on handling.  The focus frame rails are much wider than the mustang.

Swap left and right knuckles and build something to move the tie rod outboard to the correct location for ackerman steering.

Does caster and king pin inclination / scrub radius get messed up?

Take the whole mustang suspension over to the focus

The mustang does not use coil overs like the focus.  The springs rest on perches:

So the perches have to go too, or make new ones.  Again the Focus rails are much wider., but I think the track width is the same. So the perch would be interupted by the frame rails leading to some crazy fab.

 

Find some car with a front steer knuckle, or whole suspension that fits or nearly fits.

Junkyard scavanger hunt.  Seems unlikely, but might be worth a quick check.

 

Buy a mustang II front suspension.

These are used by hot rodders everywhere because they are self contained and weld right to frame rails of various widths.  The black beam in the image below represents the frame rails.  would have to cut the fender where it meats the frame rail.  Would have to work out the geometry on thisto ensure it would fit on my wide frame rail spacing and height.

 

 

Here are some solutions that will work:

 

Make the focus knuckle into front steer

Kugel makes a piece that bolts on that moves the steering arm knuckle forward of the knuckle.  This is less than ideal, and the price is especially not ideal (~$700).  This is how all the v8 focus builds I have found handle this.  This is the fail safe if I don't find something better.

My plan is to collect info on the maybe items to sort them into will or won't work.  Then evaluate the will work items to find the [cheap quick right] balance.

For reference, here is an article about how different front suspensions and power steering work:

play time

but not play time for me.  After traveling for work last week, I spent the weekend assembling a play house.  It turned out great, but it made a mess out of the shop and I am ready to get back to the car.