nuclear density gauges are not accurate 3

They are better than sand replacements in my own experience. I too have over 26 years experience, although we did not get gauges until later in the 80's. When I get the gauges calibrated (usually 2 or 3 on a job at one time in case 1 goes down) I always ask for the R² value to be included with the calibration. Where this is outside of 95%, re-calibrate the gauge, get a different one and/or look at the material under assessment and see what could be causing the big errors in repetitive readings. (PFA, Ash, slag, organics, some rock types like granites etc…)
Always have at least one 'other' density measurement technique done per day of earthworks filling, and normally with clays this would be a core cutter every 5 to 10 NDG readings and run it on a rolling calibration. More difficult with granular soils ( I have very little faith in the SRD test), so would normally have a spec requirement for calibration boxes to be done on a regular basis. Where stabilised soils are being tested, this would probably be one per day. Then use the data from the calibration box as a rolling assessment. I also plot the results of the data on a rolling basis to see if there are any trends with the data which could indicate the gauge is giving erroneous readings.
One big point to note is that there are a LOT of older gauges out there at the moment due to the cost of de-commissioning them. From my own experience lots are still in circulation that should not be and a number have been ‘given’ away to other test houses who have less scruples. So when I start to doubt the readings one of the first questions is “how old is the gauge’ and can I look at the ‘daily checks’.
What you do get with a gauge is a lot of data in a short space of time. How much reliance you place on an individual reading will be a reflection of your own confidence in the kit, technician, material, and how critical the data is.
When I write specs, I always ask for additional in-situ assessments to be done at each and every NDG reading, which is normally either a set of hand vanes or mexe probes so I can make a judgement as to the value of the data. I would never rely on an uncorrected gauge in isolation (just writing a review on someone else's work where they have done this, and the numbers just don't stack up!!).

The best use for a nuke gauge is that a bureaucrat can have a sheet of paper with a number in it to put in their file.
If it has the magic number on it, then nothing can be wrong.
I get test reports back sometimes with the compaction at 110% Modified Proctor and other nonsense, but hey, it passed, so no worries.

I think that Ian made the right point - the guages should be calibrated for a particular material - as one might do for site calibration of STP N value vs shear vane Su. I, too, found problems with the guages I used in the late 70s. Was not "getting" compaction but when i compared moisture content (lab) with moisture content (nuke), we found that the lab value was 70% of the nuke value - this makes a difference of about 2-3% which make our compaction "good". They are a tool - as with any tool, one needs to know the applicability and limitations. On our job here we will be using a double probe nuke - so I will want the contractor to calibrate vs other method (likely dug hole lined with film and volume determined by water).

Let us not forget that the gauge and any other testing method is a tool to support observations. Terzaghi and Peck stressed the observational method - these tools we have are straying us from the basics.

What an argument!! If I've heard this one once in the last 30 years, I've heard it a hundred times.

You have to calibrate the Nuke to the material for cross reference as both iandig and BigH noted. Back when Nuke gages were becoming popular, we calibrated them to sand cone density tests. Not a bad idea, but a direct volume comparison is even better (drive sleeve)...problem was that you can't use drive sleeve in all materials (material > 1/4"). Comparing several methods, you could get a good idea of the accuracy of the Nuke gage. Keep in mind that the gage calibration is to a standardized density block...not soil or asphalt. You still need to compare it to conventional methods of soil/asphalt density.

As for moisture, the gages have gotten better over the years. They used to be awful for moisture content. Almost completely unreliable in any material that had a reasonably high calcium content. Not great in most other materials either. We found that a Calcium Carbide moisture test was more accurate than the Nuke in most cases.

Almost any electronic or nondestructive test method must be calibrate to physical test measurements. At the least, just to give a bit of "peace of mind".

Good comments.

brownbag-

I would agree with your statement...but also think "some" of the floating numbers begin in between sampling and the lab work done. I do however think the guage requires the contractor to provide good compactive effort at a reasonably close moisture content, when used correctly.

What about the nuke gauge operators? In my experience, incorrect proctor selection is more responsible for bad compaction test results than the nuke data. How many times have you heard a technician say they used curve X because the numbers seemed reasonable, and make no mention of soil type?

how do you calabrate your nukes. I have always used a sandcone, but out of 22 tech in the office, maybe three knows how to run a sandcone. I am properly the only one that ran one in the last 15 years.

what other methods for comparing a nuke.

brownbagg....Calibration is done at the company's calibration facility. What you do in the field is a correlation check. Sand Cone was generally the accepted method and was the method prescribed by the Corps of Engineers for correlation. I consider the sand cone to be an indirect correlation since the sand cone method is a secondary measure of volume. The drive sleeve method is a direct volumetric measure, and in sands, clayey sands, clay, and silty sands is a more accurate representation of volume and, therefore, in-place density. If you have pieces of stone or concrete in the material to be tested, or if the material is a graded aggregate, then the sand cone is the preferred method.

I would encourage you to teach the other techs how to do the sand cone test. No, you won't use it very often in practice, but if they learn to run the test, they'll have a better understanding of soil density and compaction. One of the problems with electronic "black box" methods of testing is that the original basis of the testing is lost. That leads the technicians to believe anything the "black box" tells them. Understanding "why" is much more important than understand "how". Teach them "why".

The drive tube or shelby tube, I feel is less accurate because you can over drive a sample and compact the sample or the sample does not complete feel the tube and breaks off. Now using three or four different method to prove would be better. Sand cone, drive tube, what else.

as far as teaching the crew, ex waffle house cooks just care about the hours, push the button, write it down, when quitting time.


This is not an argument just trying to pick everybody brain on proving nuke accurate.

brownbagg...I agree, bad techniques can be used with any method to get inaccurate results. If done properly, the drive sleeve method is a good, direct volume measurement.

I would not want to use a testing laboratory that doesn't train its technicians.

it really not the lab, its mainly the labor force, how many people you know want to grow up to make cylinders. It not just labs, its everywhere. Concrete driver dont care about the product etc. I just feel that I provide a service I need to be able to prove ten different ways to the client. Once the nuke leave the factory you are stuck not knowing right or wrong. We both know one little mistake with sandcone or drive tube ruin that method too. you cant make money running sandcones so they dont use them.

I like to add my two cents. I may not have much experience as the others, (11 years), but I agree that you should have more than one way to prove the soil is compacted. I prefer to us the proof roll over the nuke gauge if at all possible. I know engineers like numebrs because they are defendable in court. Here is an example were a nuke guage comes out OK and the proof roll fail. I ran a density test on some soils and obtained 100% comapction at very close to optimum moisture by laboratory proctor. The specification also called for a proof roll. When I proofed rolled the site very close to where I took the compaction test, the truck buried itself to the axles. Once we removed the truck we found that there was 2 feet of silty clay material ontop of 3 to 5 feet of peat. I know this is an usual story. Just thought I share.

My 2 cents. I grew up using sand cone and also used some of the first experimental nuke gages. My grad work was measuring moisture variations with a probe down a tube and a laboratory Geiger counter scaler powered by a generator. Even then I noticed a lot to be desired about the method, but it worked for my purposes.

Many a time later on when a contractor would call and say he was being penalized unfairly by a competitor inspector using the nuke, it was pretty easy to show the nuke was wrong. That helped bring business our way, so maybe it was not so bad after all.

Anyhow, as for sand cone, tests have been done to check its accuracy and it has been found that how you leave the sides of the hole plays a big part in accuracy achieved. Slight disturbance makes big difference. Of course other things like sand being kept dry, etc. are important.

Then comes a soils professor out doing some, on the side, density testing work for a client. I happened to be on the job for the competitor, just watching. His sand jar was a canning jar. Cone diameter was maybe 4 inches. Can you imagine? Yet that was sold as one form of test gear. Makes one wonder if some of us use our heads out there. Maybe the profs should stick to the books.

No one so far has mentioned the size of the hole, cone size, etc. Maybe those having little faith in it use canning jars?

No one has mentioned rubber balloon either?

rockiologist-what you described is a stability failure, not a density failure. Density does not imply stability unless you correlate it to a CBR or some other stability test.

Proofrolling is done to show areas of instability and should not be a final measure of density, unless of course you can correlate number of passes to achieved compaction.

I can show you equally many instances of proofrolling that show no instability but did not achieve compaction. One common example of such material is slightly silty, slightly clayey fine sand that is very hard on the dry side of optimum but often compaction is not achieved on the dry side. The result, when surface or groundwater moves in, is a significant loss of stability because of the lack of compaction.

You can achieve compaction without stability, but you cannot achieve true stability without compaction.

I,m still looking for more ways to prove a nuke accurate. even maybe a test block that I can weigh

If you can get hold of a copy of BS 1377-9, whihc covers in-situ testing of soils, there are a number of 'standard' methodologies for calibrating a density gauge. Depending on what the end-use of the readings is, there are different 'reccomended/specified' methods for calibrating the gauge against the soil or material under test.
These methods do inlcude comparision between the guge and other in-situ methods, as well as the use of the calibration box.
I don't have a copy of this to hand, but will speak to a couple of labs I know and see if I can 'source' the exact methodologies for you if that will help. You do have to bear in mind that the way in which most ofthe BS's are written, it does assume you know what you are doing, following a recipe does not mean you are a good chef!

oldestguy - I, though not as old, have used the rubber balloon! Lots easier to lug around that a cone with a lot of sand . . . Nukes are fast - good for a green or red light answer; sand cones, I feel would be considered more dependable (but, hey, I'm almost old). All the discussion boils down to having a good understanding of what you are doing and doing so in a reasonable fashion - I don't think that lab calibrations should be considered as accurate as site specific material calibrations (but with the latter, you will have to calibrate many many types of soils). As for rockiologist's comments - one must remember that the cone - or the nuke - only determines the near surface density and doesn't "know" the underlying causes. In his case, the proofrolling was critical but the soils encountered should have been known and accounted for in design. So it shouldn't have been a surprise.

jgailla...I hope you are rejecting such test results. Your firm's liability increases if you don't. Even if you are not the "approving authority", send a memo commenting on the results. At the least, the bureaucrat will have to close the loop on an open comment, and accept or reject the results....lowering your liability, and maybe even getting a better job out of the testing lab.

In your experience with a testing lab, you saw the variability of the quality of work of individuals. It's the same in most labs. The key is the quality of the review and unfortunately, some labs send out results with no review.