Capping on cylinders 3

[emanc]

We encountered a problem on sylinder tests when we switched testing laboratory. The 1st lab uses sulfur capping on samples; we get results way above the designed strength. The 2nd lab only uses rubber capping and we get results way below the designed strength. We suspect the rubber capping causes eccentricity in load application thus samples easily breaks. Comp machine of lab 1 have expired calib record while that of lab 2 have been calibrated recently. But we dont't expect too big difference in results just because of difference in calibration frequency. Any thoughts on this?

 

[civilperson]

The sulfur compound with good end molds gives the true strength if the operator complies with ASTM rate of loading, (not jog the pressure to over 50%). To use rubber caps, a majority of cylinders need to squared with a masonry saw unless extreme care is taken in the making and storing of cylinders.

 

[concretemasonry]

Also, take a look at the failure of the broken cylinders. Often this can also be a clue to ant problems.

 

[emanc]

Cylinders in lab 1 tends to shear almost parallel to cylinder length and samples appears like bulging, while in lab 2 cylinders break on top corners of the samples which indicates sample is not being axially crushed. We will apply sulfur capping in lab 2 and see what the results will be, though we are not confident if lab techs in lab 2 can apply sulfur caps properly as they are not used to it.

 

[concretemasonry]

Are the labs and technicians certified?

 

[emanc]

Both labs are certified by dept of public works, im not sure of the technicians. Anyway lab 2 technicians already said they dont know how to apply sulfur caps so we will go back to lab 1, we've been using lab 1 for some time and so far we did not have any problems with the structures they have tested. thanks.

 

[Rjeffery]

emanc (Civil/Environme)

The second lab may not be allowing the cylinders to break at their highest potential. the polimeric caps store energy and release it as the fracture process begins... this gives the appearance of a failure which may or may not be complete failure. Also, check the diameters of the cylinders being broken... there may (should) be a variation.

 

[msucog]

i'll throw in my 2 cents...sort of...what range of values are you talking about? for example, is one lab getting 3100psi while the other is 2900psi for a 3000psi mix. i think the data should be brought in to the discussion. also, you could always try companion cylinders with the two companies out of the same sample at the same time to help narrow down (or perhaps open your eyes to) the range of test result values. there are "ranges" within each company, each lab within that company, and especially between companies. however, there should not be gross differences in the values being obtained. my point: 200psi in the values may be a wide range for you when it's actually "reasonable". aci has a publication related to evaluating the statical side of it all. don't remember what it is off the top of my head.

 

[emanc]

example is: for a 4000 psi design, lab 1 result is 4300 psi while lab 2 gets 3600 psi.

 

[DarthSoilsGuy]

[i'm assuming you're a member of the design team]

What is the concrete supplier's position? are they doing their own testing as well? i can't imagine they are being quiet. Are the failures true failures? (More than 500 psi individual, running avg of 3 below design strength)

when we compare breaks b/n the two labs, is it on the same cylinders batched from the same truck? Or, are we talking about an overall statistical difference comparing the concrete when lab 1 was doing the work alone verses lab 2 doing it alone?

unfortunately, what you know is outweighed by what you don't know. you would have a hard enough time trying to figure out what was going on if you were the lab manager for either lab 1 or lab 2. it would still be tough to really nail down if there 1-Was a problem and 2-What it was, if you were by some weird convenience lab manager of both companies.

So many variables involved, capping/neoprene is probably a red herring. i can spend hours talking about how many ways a cylinder can break low. i can spend minutes talking about how it can break high. now that you know that lab 1 is out-of-date with their calibration, it needs to get addressed (one of the few break high possibilities). the real question is "what to do about the existing low breaks?" From what's posted, there isn't enough to discredit the low breaks.... even if field testing passes.

Official & accredited calibration of the break machine may have to be scheduled way down the road based on the demands of the CCRL or (fill in the name of lab accredidation organization). But, it can be checked quickly with a proving ring and dial gage measuring deflection to confirm existing breaks by lab 1.

good luck.

 

[msucog]

being a testing firm myself and having gone through these things many time, i'm willing to bet the supplier blames the testing lab (regardless of the facts). either way, the provided numbers are too too close without backup data to even throw reasonable estimates about what is actually going on. in case you're not aware, aci does have provisions for acceptance of results that darth elludes to in his first sentence. if the 1st lab had lots and lots of results at 4300 plus or minus say 200 and then the 2nd lab comes in with lots of results at 3600 plus or minus say 200, then one of the labs is doing something different (whether it's #1 or #2, who knows--most would blindly kick the lab providing lower numbers).

i guess i'm going to make this thread a little more complicated (may be echoing questions already posted)...why did you switch labs in the first place? and did it happen while in the middle of a project or on different/between projects? what time of year are you comparing results (both same time of year---winter/summer)? how far apart in time are the results you're comparing? were the mixes (not just same design strength...but mix designs) the same from the same plant on the same project at the same time?

as darth points out, there's lots of factors that can booger up finding a clear and defined answer to the difference...unfortunately, it's usually easier for everyone to just kick the messenger (the testing firm). however from my experience at my firm, i seldom narrow down the problem to our lab or testing personnel but then again we're more critical and demanding of our people than other firms around here. i can usually determine that the problem lies with the contractor...sometimes the supplier...sometimes it's just the oddball thing that happens with a non-perfect world. to be fair, i do finds faults from within occassionally and typically can find many more faults with other low-ball testing firms. so testing firms are not completely immune to problems but more often than not, it's my opinion that it's the contractor's and/or supplier's problem.

and without changing the subject, i'll just go ahead and throw this out there before it comes up, don't go out to core the area, find a higher strength in the core breaks and then say "i told you your cylinder breaks were too low"...i'm sure there's plenty of threads out there discussing this. in other words, "them cylinders and cores ain't the same animal" (as a coworker of mine once said).

 

[emanc]

msucog-
It so happened that the day we have to break cylinders lab 1 techs are apparently not available so we have to go to lab 2 wherein we dont usually go to break our samples. The following day I required the contractor to break another sample from the same pour to lab 1 as I was not confident with lab 2 method and how samples break.

darth-
I dont belong to design team, I am with the project management team so we are the ones accepting/rejecting the contractor's work on site.

 

[msucog]

huh, it's hard to say without a track record for comparison. sometimes strange things just happen and perhaps the difference is in the capping type/quality. which lab made the cylinders?

 

[BigH]

We had something like this recently. Our lab (calibrated compressive machine - 9 months) was getting 23 MPa. Had a problem and had to go to the plant site to break. They got 29 MPa on same mix. Talked them into calibrating their machine along with special calibration of ours - they were reading 33% too high! Not the same - but shows another light as to what might/can happen.

 

[boffintech]

I bet Rjeffery is on to something. "The second lab may not be allowing the cylinders to break at their highest potential."

I've run hundreds of tests using both sulfur and rubber caps. When using rubber caps, chips often fracture off the top or bottom edges of the cylinder (usually top). They need to keep loading the cylinder to get a full break; there's more psi in there.

 

[emanc]

msucog-
both labs are independent, they are not the ones taking samples but the concrete supplier.

 

[DarthSoilsGuy]

the concrete company made the samples.......
well there goes the possibility of a discrepancy in the sampling methods! Now,.. i can imagine that they are being quiet.

Lab 2 not breaking the cylinders is a possibility. this possibility could be confirmed/rejected by dumpster diving and capping the "low break" cylinders, since the internal crystalline structure should still be intact. getting the cylinders from lab 2 to take back to lab 1 will be awkward. i would not want a testing lab to learn how to cap on my project. I am assuming that there are no reserve/hold cylinders that can be tested.

how many samples did lab 2 test anyway? if 1. Nothing exceeds -500 psi ....and.... 2. The average of any 3 consecutive tests at 28 days of the same mix meets design strength (the average of the 2 28-day cylinders = 1 test) .....then you've met ACI strength requirements.

when you say that you're on the PM team, does that mean that you work for the general contractor and the work in question is by a sub or are you working for a Construction Mgmt group for a CM At Risk job. doesn't matter much, but Lab 1 should have their machine checked and you should avoid the situation where only you and lab 1 know about it.

 

[msucog]

now darth, you wouldn't be implying anything with your statement would you??? i personally don't trust suppliers even though i try to.

my final thought: you got too many variables and likely will only come to the conclusion that there's 8 different fingers pointing 47 different directions. i suggest you look up the acceptance criteria in aci 318 and 301 (i think it's there too). darth pretty much said it. for future pours, hire a testing firm to sample and test the results. if it happens that it's your supplier, then have them break them (but i wouldn't have the people i'm buying from also giving the testing...but that's just me). besides, if it's a special inspections job, there must be an independent special inspector doing (or at least observing--and that's only a maybe as a possibility) the sampling/testing. that independent lab must be paid for by the owner or owner's rep (not the contractor).

 

[uwrockguy]

This is a common issue and to get to the root of it would require many man hours. There have been numerous good points made that I completely agree with but I'll take a stab at clarifying a few things.

First, we need to incorporate some statistics here. Discussing a few points in space is difficult. If these samples were cast for ACI 318 code compliance, there should be two cylinders if cast in 6"x12" molds or three if cast in 4"x8" molds. Either way, a result is the average for that age (a single cylinder test is meaningless under the code, think of it as information only). You need to calculate the average (which is pretty much what everyone does), but you also need to look at the standard devaiation and the coefficient of variation for that test. Consult ACI 214 for assistance but this can be done readily in MS Excel or equivalent software by calling out the correct function. You need to understand how "good" the lab is at defining a result. If the coefficient of variation routinely runs over 4.0%, run away from that lab. A compotent CCRL accredited lab should be cranking out at below 2.0% with occasional instances between 2.0 to 4.0%. Now, once you start developing a history, that is more than 10 results on the same mixture, you can then determine the standard deviation for the population. At this point, you could then define the material variability by evaluating the population's standard deviation and laboratory's within-test standard deviation. ACI 214 spells out how to do this in detail so consult that document for assistance. ACI 214 is a required document for ACI's Concrete Laboratory Testing Technician - Grade II certification. All accredited labs should have at least one on staff (if not, ask why they don't as it is in their best interest). My point is that this individual should be able to assist.

If anyone is still reading this after the dismal science discussion, there is really no statistical difference between using an unbonded cap to a sulfur cap if the corresponding ASTM is being followed (C 617 vs. C 1231). You can easily screw up a sulfur cap just as bad as using an unbonded cap if you deviate from the documents. Sulfur caps require special gear and ventilation for protection while capping and attention to detail is a must as air voids, thick caps, and off axis capping can lead to garbage results. Unbonded caps must be rotated, as they will tear and degrade over time. Per ASTM C 1231, you can't use the pads past 100 uses. This is usually the biggest violation. All accredited labs must keep a record regarding the number of uses, make sure they are doing this. I personally recommend unbonded caps for concretes under 6,000 psi as it is far safer and friendlier to work with. Sulfur mortar is a solid at room temperature and must be heated to roughly 265 deg F in order to cast caps on to the concrete samples. The process is dangerous because your body is mostly water and it will boil that water in your skin if contact is made. It also generates sulfur gas which can cause skin irritation, nauseau, etc. Unbonded caps can be very successful if the testing protocols are followed; if not, garbage in equals garbage out.

One final note, a simple test on the labs part can be done. A compression machine is a simple device. A typical hydraulic machine has a base that is a ram and a top that holds the spherical head. The spherical head must be able to pivot to accept the 0.5% allowable off-axis sample (otherwise you would point load on one side). You should be able to walk up to this head, and be able to move the head with your hands around the spherical joint. If you can't, well...maintainence well need to be done, the head is not functional. Also, don't accept a load verification as compliance, there are many calibration points to consider besides "load". Surface planeness being one of the more important ones.

Good Luck.

 

[msucog]

very good points uwrockguy. in particular aci214. i doubt most contractor (maybe even architects/engineers) will understand what you're trying to say without a lot of explanation. the instances i used 214 to evaluate test results for presentation to the project team, we drew a bunch of completely blank looks. we pretty much had to explain the results and then after a brief pause, we said "so based on aci214, this indicates that the results are consistent and the mix is consistent...whether one or the other is doing it completely wrong every time, the results are consistent. and since it's difficult to get consistent results by doing the process incorrect across several different persons, it is more probable that the results are representative." the issue can be very confusing especially when discussed field strength versus cylinder strength...they're both the same to folks that do not understand what the results represent. all in all, i'd say the having a good understanding of the statistical analysis is the more difficult part of what we do. then, if you're tasked to explain it to people that are only interested in seeing passing results, the whole subject is nearly impossible to get everyone on the same page. as i said, at the end, you will usually have to say "this means that the test results are (good or bad)"--again remembering that this does not mean the product is necessarily good or bad.

as i mentioned previously, i think you've got too many variables over a short period of time to draw any meaningful conclusions without a massive research project in to everyone involved. from my experience, such a project would likely still not provide meaningful conclusions because the result variance is so close. to resolve the issue of "failing tests" you could always just go core the in place concrete and i'd suspect those results would be satisfactory to the strength (but you never know about that oddball result that may show up). if you can use the aci result acceptance criteria to resolve the issue, then i'd suggest doing that and get the matter over and done with. if you've got lots of failing results in the future, then back up and rethink the whole situation. good luck.

(btw, i'd rather dig a ditch by hand all day than cap with sulfur cement. even with a vent hood and all that, the stuff makes you smell like sulfur for days. not to mention the hazards of dealing with the stuff.)