wood truss load responsibility

huangyhg

wood truss load responsibility
i'm providing some engineering consulting services on a structural failure of an agricultural building where the roof collapsed under snow loading.  the roof was comprised of wood trusses spaced at 4'-0" o.c. and spanning about 75 feet.  i analyzed the trusses, using the truss manufacturer's spec sheets for capacities, and found pretty quickly that they failed due to an unbalanced snow loading on the leeward side of the roof.  that's not really the issue.
most wood trusses are called out by the designer of a project.  this is usually an engineer.  but in this case, there was no engineer.  the "farmer" was coordinating the activities of the work, called up a local lumber yard and ordered the trusses.  the lumber yard conveyed the order to a truss manufacturer who simply plugged in 20 psf for the roof snow load and sent the underdesigned trusses on their way.
the truss plate institutue specs and all other info i can gather dictates that the designer is responsible for calling out or determining the proper live loads. but in this case, the farmer didn't have the background/know-how to do so.  he had no idea what a snow load even was.  the truss designer knew that these trusses were headed to a ground snow load area of 45 psf.  
wouldn't it be the truss designer's responsibility to at least question his own decision to use 20 psf?  
check out our whitepaper library.
wtca
trussengineering.com
have a full article regarding truss design responsibilites.
lots of questions to get answers for...
did the building require a building permit?  due to size or span, etc?  if it did, why wasn't one obtained?  did the building inspector know that construction was ongoing?
was a contractor involved?  he may have known something about permits? loading, etc?  was the loading used common for the area?  most truss suppliers, here will querie a loading if they don't think it's correct.
why did the owner not get permit or professional help? ignorance or to save money?
were shop drawings (even 8-1/2x11) prepared by the truss supplier?  often the shop drawings have a note regarding spans greater than 60' requiring an engineer.  are there any attached or referenced notes to the shop drawings?  did the shop drawings or transmittal require approval of the loading? most truss suppliers' software accommodates unbalanced loading, reduced loadings (for ag buildings, etc.), slippery roofs...
did the truss actually fail due to load?  the 4' span would indicate purlins (truss supplier's name for even 1x4's @ 16 on the flat) and metal roofing.  was a 'slippery' roof considered in the analysis?  was the weight of snow actually determined? or had there just been a lot of snow? i use a 4" dia. piece of septic system pipe, with a sharpened and serrated end to obtain snow cores for weighing.
how did the owner install the trusses?  were they damaged in the process of installation? 75' span trusses are something not normally handled by forklifts.
did you take a look at the horizontal bracing system.  many of the failures i've seen are due to incorrect horizontal bracing or damage during installation.
truss fabricators prefer to be responsible for what they do best... make trusses.  the drawings can specify that they be responsible for lateral loading, etc.; basically everything above the top plate.  is there any correspondence from the owner to the truss supplier to the effect that he was doing this himself? or did the owner simply order trusses, say, from a lumberyard, to span 75'?
i've got to run... and will add to this, later.
dik
it is possible that the web members were not braced properly.
this can be a classic example of the great big loop hole between the ubc and the tpi code.  the ubc refers the designer to the tpi which in turn says that the engineer of record is responsible for the web bracing.  usually the truss company's calcs will indicate which webs must be braced and where.  then the design engineer has to find the compressive force in these webs and determine how the keep the web from buckling.  this can either be 2x4's connecting the webs to a hard point or adding another   
dik:  thanks for your response and multitude of questions. to answer a few:
1.  the project was out on a farm. there was no jurisdictional authority requiring a permit or inspector.  the county maintains that they do not require any code compliance on farms.
2.  there was a hired erector but the farmer was acting as the coordinating entity in putting the pieces together.   the farmer had a 4 year construction engineering degree (b.s.) but no design degree or experience.
3.  8 1/2 x 11 truss "shop drawings" were indeed prepared by the truss designer....using 20 psf live load in a pg = 45 psf area.  from his testimony, no unbalanced loading was used in the design software.
4.  the failure, from my analysis, was definitely due to truss failure.  i inspected all the trusses and found very good quality materials, nailing plates, bracing, bridging, etc.  the actual snow load was definitely a drifting on the leeward side (6" deep on windward side and 30" on leeward).  the metal roofing, the open terrain, the agricultural type, etc. were all included in my analysis.
the main issue is who is responsible for "catching" the 20 psf design in a 45 psf ground snow area (the unbalanced snow load on leeward side was 30 psf).  the truss designer was asked "if you were asked to provide 20 psf trusses in a far northern area, near the canadian border, would you question it?".  he answered yes.  in this case...at 45 psf, he didn't question it but just provided 20 psf.
since the tpi and wtca only seem to deal with the common "ideal" situation where there is an engineer/architect dictating loads, this situation seems out of the ordinary.  
most truss engineering software automatically tests for
an unbalanced load case. if the software has the capability and the option has been suppressed, blame might shift toward the truss fabricator. trusses more commonly fail due to improper erection & handling. with the actual load applied, does the csi exceed a factor of safety of 2.5 for wood and 4 for steel? have you determined the failure mode?
the csi factors in the original design were on the order of 0.87 to 0.93 but that was under a uniform ll of 20 psf.  for the unbalanced ubc snow and under the unbalanced actual snow the factor was on the order of 1.6 to 1.95.  (with 1.66 to 2.0 being the theoretical failure point).
i've passed this on to a classmate that works in the truss industry for some additional info.  looks like it's mostly a matter for the lawyers (i do a lot of forensic stuff for lawyers/insurance).
item 1:
although the county may have no requirements, does the state?  in which case, is there a stipulated snow load?
item 2:
a different level of technical expertize is expected from someone specializing in a type of work.  what is the technical skill of the truss designer? is he a certified technologist/engineer? again, the greater skills, the greater the expectations.
this may also be a problem to the owner, since he may have the only technical 'credentials'.  
i assume the owner has undertaken the coordination of the work: foundations general framing, etc.  in the process he has put himself at risk; the work undertaken may not even be covered by insurance (i gather that there were no fatalities, only material damage).  he has benefited from the reduced load (cheaper), the lack of professional input (also cheaper) and the lack of a general contractor (also cheaper).
item 3:
the critical question is where did the 20 psf loading come from?
i've not experienced a snow loading that light; in our environs it's typically 35 psf and up. is this a common or an unusual load for the area where the trusses are fabricated , where the lumberyard is located, or where the project is located?
is it a common design load for ag buildings in any of those locations?
our building code has reduced snow loading and no windloading for residential buildings; it's not uncommon for a truss supplier to run a truss not realizing that the building is commercial and, therefore, has a different loading criteria.  if the truss design software had these features then as ctruax noted if the truss supplier switched these off, he may be accountable.  is the supplier equivocating when he says that the software did not use unbalanced loading for the design? or does the software have that feature and was it just disabled?  a lot of the truss software is provided/rented by the truss plate suppliers and have an incredible number of features.
did the lumberyard stipulate this loading to the truss fabricator?  the reason for the question is that it's not common for a truss supplier to 'pick' a load out of the hat.
if the truss designer knew where the trusses were to be erected, and he selected the design load, then there is some negligence even if he provided shop drawings to the owner for approval.
if someone stipulated the design load, whether or not the truss supplier knew the location, then i would suggest that the truss supplier had a reduced duty of care to confirm it.
has the truss supplier/designer provided trusses for the area where they erected, other than this job?  if they (supplier or designer) had provided trusses for 10 projects in the area, they should have known.
what information is on the shop drawing?  did the owner sign it as being either reviewed or approved? again a difference in the degree of acceptance.  was there any acceptance of the 20 psf live load?  was everything handled by phone, or was there a written contract of any kind?
the tpi or tpic (in canada) provides fairly good information/documentation, but lean towards the truss manufacturer as providing a material only. other technical issues should be handled by the designer/professional/or someone else. their guidelines (they actually have printed material) wrt to truss responsibility, fabricating, handling and installation are good and if followed will result in a 'successful' project.
i often provide copies of their erection and bracing information to clients with the usual disclaimers.  i've also received copies of this information attached to shop drawings.
have you determined the loading at the time of failure, not just the code stipulated loadings.  what was the actual extent of snow on the roof?  have you run this on another 'roof truss program'?
will add some material after i hear from my classmate.
dik
dik...whew...you ought to be a lawyer.
item 1 - no state requirment either.
item 2 - the truss designer had a couple of weeks training on the software (from the vendor) and no formal engineering training.  the farmer had the b.s. degree like i said..but no design experience.  i'm sure he didn't know what a snow load even was.
item 3 - the 20 psf came from the default on the software.  probably because south of missouri/kentucky the roof snow doesn't control design, but rather the roof live load of 20 psf.
the truss designer just took the order, "decided" himself to use 20 psf in the software (no one stipulated it) and designed the trusses.  he knew where they went.  he hadn't done a lot of trusses in the further north area before.  the lumber yard was just the middle man.
the owner didn't accept the shop drawings,....he just erected them (which is a form of acceptance..but an ignorant one in this case).
the bottom line is:  we engineers have a duty to protect the public welfare.  anyone else in the construction industry does not have this duty explicitly laid out as a foundation of their ethical behavior.  in this case, the truss designer just used the "standard" ll on the trusses and sent them north to snow country.  his job typically is supplying trusses dictated by someone else.  in this case, the person asking for the trusses didn't know boo about snow load and a bunch of pigs had a shock.
we have a similar situation in ontario where farm buildings are excluded from the ontario building code and the reference is to the national farm building code of canada.  it's possible to construct a 100,000 sq.ft. farm building without any engineering input; at least there are stipulated design loads. there's no question about our duty to ensure public safety; our thing slips through the cracks in the same fashion.
now that we know what happened, the reply is easy... if the truss supplier took it upon himself, without any guidance or direction whatsoever, to assign a live load or use the default out of ignorance, then i would suggest that the truss supplier is fully responsible for replacing their defective product... (it is not fit of purpose) and their responsibility could include any other damages sustained by the owner <g>.
the problem is that rarely are the supreme court justices unanimous...
btw, your item 2 sounds very much like a concern i had about using fem stuff just a while back...
reply from my classmate (wood truss specialist)
paragraph 2 just about sums it up... will get a copy of the *.pdf from the tpic site.  likely one similar is available from the tpi site in the us.
message as follows:
hello dik;
the responsibiltie of various parties in tpic opinion can be found in pdf format at tpic.ca
the responsibility is the supplier to verify that loading is correct for building location. specified ground snow load can be obtained from environment canada or local bldg. inspector.
farm trusses are required to be built to part 4 with unbalanced , wind , & built up snow conditions. farm code allows load sharing up to 4' o/c spacing as well as importance factor of 0.8 for low human occupancy. ontario bldg. code regulates farm bldgs. using 1995 farm code.
vic tanner p.eng
timber systems engineering inc.