Sources of error in Sand Cone Method? 3

[lviescas]

I have to use the Sand Cone Method for density testing on a job where having a Nuclear Densometer on site all the time is not practical. I have only been doing Sand Cone tests for 6 months, and I understand that a lot of skill is needed. I just can't figure out what "skills" I need to develop. The density values I obtain seem too low based on the way the backfill behaves. I mean, this stuff is compacted to the MAX! And when I have someone with a nuke on site, I do Sand Cone tests exactly where he tested and my values come up 10% lower than his every time. All of the errors I can think of (caving, sand bridging) would result in calculated hole volumes being smaller than actual, and the resulting density appearing to be higher than actual. Can anyone help me identify some sources of error?

 

[oldestguy]

Several possibilities or combination of them, in order of importance or likelihood:

Not correctly calibrated sand (calibrate over known volume, such as large compaction mold); not correct calibration of what is in the cone; dug hole too small (use at least 8" cone and one gallon jug) the larger the better, starting surface is not even and not of same density as below; disturbance by compaction on walls of dug hole; use most of the sand in the jug, not only small part of it; scale not accurate; moisture determination on non-representative material (dry the whole sample as a check)

Is the fill quite full of obvious voids (like single sized rock) where sand can be lost?

If your outfit is using a quart jar and a small cone, forget the results as meaningful. It takes a lot of sand. On some sites I have gone to a very large hole, like 18" x 18" x 8" and spilled in the sand as it also was done for calibration.

finally recognize the nuke may be inaccurate. Was it calibrated for the site soils? I'd suspect that as a starter. Calibrating in the lab is not sufficient.

 

[Ron]

OG is exactly right. Other sources of error are:

vibration during testing
high moisture content bleeding into calibrated sand

OG has given you the primary issue...make sure your sand and system are calibrated together and use a large quantity of sand...dig the hole to conform to the template and then make sure the bottom of the hole is bowl shaped....if it is too square, sand will bridge.

 

[BigH]

Interesting as most would consider the sand cone method to be much more reliable than the nuclear method (that's my experience at least).

 

[fattdad]

When doing a sand cone test, you release falling sand into a hole. There is this metal plate that sits over the hole. You need to make sure you know how much sand is lost to the metal plate and the cone, as this sand has nothing to do with the hole that you dug.

So, you need calibration on the density of the sand. You also need calibration to know the sand lost to the equipment and not lost to the hole.

I'd agree, sand cone test more trustworthy. Then again, I'm old, but not the oldest - ha.

f-d

¡papá gordo ain’t no madre flaca!

 

[TDAA]

There is also potential to loose soil as it is dug out of the hole due to your mechanical action or wind.

Based on being 10% low (at least 10pcf for a lot of soils), I would suggest you check the sand unit weight calibration, and your cone calibration. These two would likely be the main items on your end, while vibration my be causing it from the contractors end.

Also, if you have the equipment, try checking the density with other methods such as a drive cylinder, or push a shelby tube. Se if they are closer to backing you up or the nuke gauge.

When you compare the densities from the Cone and the nuke, look at the wet density. That will take out the moisture content variability.

 

[Ron]

The soil dug out of the hole is irrelevant...it is only used for the moisture content.

I agree that the SC method is more reliable and has long been used as the field correlation standard for other methods, particularly nuclear.

The calibration, as previously noted, should be a "system calibration" as F-D noted to include the plate, the cone and particularly don't forget that amount of sand that fills the template area just under the cone. The sand should be calibrated routinely if you are in a humid area.

Calibrate the cone, jug, and template on a glass plate.

 

[BigH]

One source of SC error would occur if when digging the hole they overdug the sides and the top of the hole (in the ground) was bigger than the hole in the plate - then you may not get the proper filling of "all" the hole you dug - but this would/should be small.

 

[TDAA]

The soil dug out of the hole is irrelevant...it is only used for the moisture content.

I am sorry Ron, I think that is the first way-off-base post of yours I have read, out of the many good posts covering a variety of topics.

Since the sand is only used to figure out the volume of the hole, what do you use to figure out how much the soil in the hole weighed?

See, I usually take the irrelevant weight of the soil from the hole and divide by the volume of said hole to determine the moist unit weight.

 

[Ron]

TDAA...you are correct...I had a lapse of thought there! Thank you for you kind remarks! It has been at least 20 years since I personally did a sandcone density. I stand corrected!!

 

[BigH]

Senior moment . . .

 

[Ron]

Geez BigH...you coulda gone all day without sayin' that!!

 

[BigH]

Not really, it's what I have to say - a lot!

 

[oldestguy]

A final point. I started out doing these tests with Ottawa (Illinois) sand. That is a good material to use, if available. Then, due to costs, the firm I was with switched to a sand also coming from the St. Peter Formation in Wisconsin, being a finer gradation. The finer stuff would jam up the valve (Oh yes, you have to take into account what is in the valve, or not) and technicians tended to vibrate the cone while "closing" a jammed valve. Not always jammed, so some variations can come between the field and the lab calibration on that score.

Also, what were you taught as to how to do the test? I once observed a well-known and research paper writer oriented university professor do that test in a legal case. I had never before seen a smaller cone or jug!!! Was he your prof.??

 

[BigH]

Ya know - don't send a boy to do a man's job! We got some Ottawa sand a few years back and it cost me $10/kg here in Indonesia! We only use it for cement cube "referee tests".

 

[Ron]

20-30 sandblasting sand works well and flows nicely. As BigH noted, Ottawa sand is way too expensive.

 

[lviescas]

I want to thank all you bright, busy professionals for dropping what you were doing to help me with this problem. I have implemented every suggestion and I will get back to you with the results. I hope I can be as helpful someday when I've been at this a little longer.

 

[msesi]

In principle sand cone replacement is more accurate than troxler when carryout field density tests.However sand replacement method is slow as it involve more activities.In order to use troxler you are required to calibrate the troxler on site in line with the sand replacement method.First step carryout a sand replacement and the same time just short from hole carryout field density test by the use of troxler.Continue with the same procedure say twenty or thirty stations.Then plot the test results obtained by the use of troxler and by the use of sand replacement on the same scale.It will be obvious that the troxler test results obtained will be higher compared to test results obtained by the use of sand replacement method.Then find the coorelation and find the correction factor.Then you can choose to use troxler for it is faster and then multiplied by obtained test results by the factor obtained.