It really depends on the design criteria imposed by the agency for whom you are doing the design.  In our office, we keep the resultant always in the kern for public works projects.  However, for private clients, we keep the resultant in the kern for conditions not involving wind or seismic and allow the resultant to go outside the kern for conditions involving wind and/or seismic.  Howver, we do require that the length of the footing under compression be at least 75% of the length of the footing, keeping in mind that the maximum aloowable soil pressure is not exceeded.

The toe pressure is the governing limit for footing when the entire footing is not in contact with the soil.  If the toe pressure is below the allowable then a negative heel pressure is acceptable.

Could there be a concern about pumping?  In the case of a design storm, you'll have rain and cyclic wind loading of the footing.  It seems if your soil conditions are fairly poor, you could start pumping saturated soil from under your footing.

The negative heel pressure will cause the heel to lift off the soil, redistributing the soil bearing stresses.  You'll wind up with a triangular distribution, with the maximum at the toe and a zero stress at some point between the toe and heel.  You need to figure out these values and compare it to your allowable soil bearing.

You can't just leave the negative heel pressure as is and compare the toe pressure to your allowable, because you're actually getting zero (not negative) pressure at the heel and an increase at the toe.

nutte-  My Braja Das reference has a straight forward approach to the increased toe pressure when the distribution is not across the entire footing... I certainly want to keep this increased max below allowable.  It just never says this is o.k.-

My question comes out having a footing proportioned to give me an acceptable FS for overturn, and a toe pressure well below allowable, but the heel pressure is negative (zero).  By the time I increase or adjust the footing size sufficiently to get the heel pressure to zero, the toe pressure is VERY low, and my FS overturn is high.

Part of my dilema I guess is that the various load combinations can come from wind/seismic direction reversals.  One case has a lateral load left and a down force at the top of the wall, the other has a lateral load right and an uplift.  By the time I proportion the footing to account for both cases, and hold the kern, the footings are huge relative to what would be needed for just static vertical loads... clients seem to have a hard time appreciating this  ;)

I don't have a problem designing a footing this way, provided all of the stability checks come out OK, especially if it results from wind or seismic loads.

I just designed a whole lot of shear wall footings with enormous OT moments and little dead load.  I had no choice but to have the resultant outside of the kern (0 bearing pressure at the heel) without making the footings ridiculous.  No one in my office saw this to be a problem as long the allowable bearing pressure at the toe is not exceeded.

Make sure to provide reinforcement at the top of the footing if you do this.

See "Engineer's Notebook", page 17 of Structures magazine, June 2004.

One thing that can be helpful is to calculate the amount of uplift.  When you visualize this, you tend to think of the footing raising up 6" and plopping back down, which would obviously be undesirable.  But when you check uplift amounts, it is usually very small, say 1/16" or so.

You also get into a situation where the slab moments, shears, and bearing are non-linear with the load.  Applying load factors and then figuring moments will give different results from figuring moments and then applying load factors.

The overturning and the bearing pressure should be 2 separate checks.

To check factor of safety against overturning take resisting dead load moment about the toe of the footing and compare these to the overturning moment. bearing pressure does not come into it.

For bearing pressure you should never include a negative soil pressure in your calculations. Even if it is a cohesive soil, what if it was a drought? You should have a positive triangular pressure distribution with the centre of the triangle under the resultant load and zero pressure past the end of the triangle.

I am assuming we are all talking about the ever so popular way of designing cantilever retaining walls.  The equations for overturning are based on simple statics and the equations for bearing pressures are derived based on the assuption that zero is the minimum.  If you get a negative bearing pressure, then the equation for bearing is not even valid.  You will need to devise another approach, maybe one that uses piles or tie backs.

The check for overturning and allowble bearing pressure must both work, not just one or the other.   

JStephen-  thank you, the article was "spot on: ...

I don't like to be too liberal in footings subject to large overturning - what if the footing doesn't overturn exactly about the toe?

Every inch the "real" pivot point creeps in wreaks havoc with the FS you thought you had.

JLNJ,

The type of overturning that you describe can only happen if the bearing capacity of the soil is not sufficient.

When the bearing capacity is sufficient then the footing can only overturn by 'falling over' -like a cup pushed over on a desk- and this requires that it turns around the toe.

" If you get a negative bearing pressure, then the equation for bearing is not even valid."
True, but it's easy enough to obtain a maximum bearing pressure for a triangular distribution.

I also believe it's acceptable to have some uplift for the wind and siesmic load cases.

When checking for overturning, take the factor of safety as ultimate bearing/ bearing at toe, rather than figuring rotation about a point (especially if the FOS is low).

Let me rephrase that:  Take the FOS as the moment that causes ultimate bearing divide by actual moment; also check for allowable bearing.

JStephen,

You should check both as they are two different failure modes (refer my note above).

If your geotech gives you a bearing pressure then it is usually an allowable value and therefore has the factor of safety already included.

The point has been well made that the maximum bearing pressure at the toe should be within the maximum allowable after making the adjustment for no negative pressue at the heel.

Checks should be made separately, for safety against

1) Overturning,
2) Bearing, and
3) Sliding

Yogi Anand, D.Eng, P.E.
Energy Efficient Building Network LLC
ANAND Enterprises LLC
http://www.energyefficientbuild.com

when designing retaining walls to resist the permanent soil loads and dead loads (no wind or seismic) I design the footing with no negative pressure at the heel.  When considering wind and seismic, I will allow some negative pressure because it is a temporary load condition but I keep the resultant within the kern.  

For those of you who "allow negative soil pressure at the heel,"
please explain how this can be justified--adhesion of the soil to the bottom of the footing?  What if the soil is granular?
Negative soil pressure is not acceptable, in my opinion.

DaveAtkins

Only the positive soil pressure is used to react for the vertical loading.  This leaves a triangular pressure distribution under just a fraction of the footing. The resultant of the soil pressure acts at the third point measured from the toe to the zero pressure location.

DaveAtkins-  as is often the case, typing in a thread is different than talking in person.  I can't speak for anyone else, but I meant calculated negative/outside the kern.  I know it's actually zero.  I simply meant that the point of zero bearing pressure is not at the edge of the heel but is back under the footing... part of the heel subject to lifting off the soil/footing interface.  

What I was looking for is 1)if this was o.k. in the short term, 2) under what circumstances, and 3)to what extent.  I hear a lot of "yes, we do that for wind/seismic", and a couple of folks that either gave me their judgement as to how much, or pointed me to published info on it...

My apologies for any confusion.

Well then,
I am okay with a triangular bearing pressure under the footing for ANY load case, as long as the allowable bearing pressure is not exceeded.

DaveAtkins

apsix - my point is that a different expression would be needed for the bearing if you get an uplift at the heel.  One can be derived, but the expression q=P/B(1+6e/B) will not yield the correct bearing pressure based on the assumpitons it was derived under.  Granted, a small amout of uplift will not affect the pressue that much.

Just in case this horse isn't dead yet....
If I'm looking at a service limit state I try to keep bearing pressure on the entire footing, or at least be close. At a strength limit state I'm less worried and generally like the resultant in the middle half or so. Also, if I'm on rock that is stiffer and much harder than my footing, I usually don't mind having a relatively small "triangle."

Recommend you discuss this at the Geotech forum. Ten Structural Eng can't all be right when there opinions are different.

For transportation projects, AASHTO Standard Specification (5.5.5)requires the resultant of the bearing pressure for static loading to be within B/6 of the center of the foundation, for retaining walls on soil.  For seismic loading, the resulting must be within B/3 of the center of the footing.

csd72

It seems to me that the overturning on a compressible material like dirt is not completely analogous to the cup on the table (and that was my point). We don't know the real stress distribution in the soil and the spiked pressure we assume when the resultant falls outside the kern may not be conservative.

I certainly use the same outside-the-kern methodology as others but I often wonder about the actual stress distribution.
   

jmiec - keepint the resultant within the middle third of the footing is analogous to not having negative pressure.

JLNJ,

Try it with a compressible table cloth on the table and it is exactly analogous.

You, like many on this forum are getting two modes of failure mixed up.

Overturning is one mode of failure

Soil bearing capacity is a completely different mode of failure.

As long as the soil is strong enough, it will not yield enough to make any significant difference to the overturning behaviour.

I am in Florida and when the code was changed to include the 0.6D+Wind load case footings starting getting out of control. In this condition I am comfortable with uplift at one edge of the foundation as long as the soil pressure is ok. (This is like a SF of 1.67)

I am currently designing footings for interior columns for a one story manufacturing building. Man the numbers work but these sizes/depths are crazy. Didn't see any footings pull out of the ground when Andrew came thru here.

For spread footings for bridges, the AASHTO LRFD specifications have rules for maximum eccentricity under strength load combinations: For spread footings on soil e < B/4, where B is footing width, for rock e < 3/8 B. Therefore, assuming triangular distribution, on soil you will need compression over 75% of the footing width, on rock over only 37.5%.

(Ref. AASHTO LRFD Specifications, 3rd Edition, Articles 10.6.3.1.5 and 10.6.3.2.5)