Application of nuclear densometer to compaction compliance test 2

[Robert168]

Dear Members:

The nuclear densometer is regarded as as a fast and accurate tool to test the density and water content of compacted soils.

For sure it is fast as it takes only 15 seconds to do the measurement. However, is it accurate? If the readings are weird, what might be the cause or causes?

Thanks in advance.

 

[cap4000]

Assuming your using it properly,I hope its not leaking radiation. Send it back to the manufacturer or get it tested ASAP. Good Luck

 

[GeoPaveTraffic]

In general the nuclear desometers that I have used are accurate. However, like any tool they can be misused. If you suspect that the readings are not accurate, do a nuke test and then a sand cone/drive tube/etc. test at the same location. Do this four or five times and you will have your answer. The COE requires just this on some projects.

 

[BigRedGeo]

A lot of things can throw a nuclear gage off, such as atmospheric moisture conditions, or certain elements in the soils. I believe the moisture readings are the most sensitive. Make sure you standardize at least daily, and take trench offsets if you are testing in a trench. If, after doing this, and checking your results with a sand cone/drive ring/rubber balloon, your results are still questionable, either something is wrong with your nuke, or there is something in the soil that is throwing the readings off. There are certain elements in certain soils that can lead to erroneous readings.....I don't remember what the elements are, but I remember, for instance, that micaceous clays will give erroneous readings. When I was in grad school, I tried using a nuke on hydrated fly ash......and it didn't work, giving results all over the place (density was close, but moisture readings were unpredictable). After looking into this a little more, and talking with some people on the ASTM committe, that certain things will throw off the readings, and nukes shouldn't be used in certain soil types.

 

[Grouser]

It is important that the proximity of dense and/or hydrogen containing objects be accounted for in the pre-test calibration. It may be necessary to calibrate for each test.
Nuclear densometers require a minimum of 75% intimate contact with the soil. All test sites should be levelled and voids, grooves, etc. filled with soil of approximately the same composition and moisture content. Small amounts of sand can be used to fill the small voids in cohesive soils.
You should be able to see the entire footprint of the gauge in the soil after it is removed from the test site. The gauge should not be able to be rocked when on the test site.
It doesn't hurt to recover one or two moisture samples from each soil being tested on a daily basis and run an oven-dry on them to establish a running correlation for the moisture correction for each machine.
A good dose of common sense doesn't hurt. Test results lower than 90% and higher than 105% should be rerun to indicate repeatability of the equipment and operator.
Soils failing by visual inspection are failing regardless what the machine says except those soils that "set up" and are often compacted with an excess of moisture.

 

[BigH]

Good points above - I started using the Nuke back in 1975 - and this was when we had the charts for converting the count readings into wet weight and water weight. At that time, when we used the nukes, if a contractor questioned the readings, then the "ultimate" test would be the sand cone or the Washington Densometer. He'd believe the sand method - it is what "he" grew up with. Back in 1979, I had a lot of problems getting compaction using crushed stone - I then took samples at each test location and did a lab moisture on them. I found out that the lab moisture was 70% of the nuke moisture. This sure helped out on compaction!! - about 2% better compaction. So, what this says is that there are differences and calibrations can be made to the sand cone, etc. to see what offsets are needed. I understand today that the nukes have offsets that can be entered into the nuke and these take account of a particular machine's "tricky personality". This goes beyond the normal factory calibration. Also, one other point - here in India, when you use the nuke, you need to double the tests required compared to a sand cone - in their statistical approach. In China, they had the nukes - no one wanted to use them as they were afraid of not being able to have (or make) babies (an important thing in the one kid family in China). Here in India, at the beginning, they were all wearing lead vests! Had one test running so I went up to the machine - took a look, then ran away fast - you've never seen so many scampering in such discombobulated directions. Thanks for letting me reminsice.

 

[dirtsqueezer]

Wow- a huge question. I also agree with all of the above, with more of a sway toward the human error. The Troxler I use has built-in tests that run before each time you turn it on. Problems for me come when it just plain doesn't work, and there you just send it in for repairs. If you could get more specific, I could narrow it down, but here are a few some common mistakes I try to watch for when I test:


'Measuring air'- the gauge has to be flush with the ground. This is a challenge when testing oversize material- it's sometimes hard to get a 'straight' hole. You will find one side of the gauge up off the ground. It just means the inspector has to pound another hole.

Native soil- variability. The problem here is usually when you have multiple inspectors. Different inspectors have different learning curves/interpretations, and this can get convoluted. My company has been fairly responsive to requests for a single resident inspector. You just need to make the call.

Thin lifts- make sure you're not punching through to another layer. Simple mistake, simple solution- just ask before you test.

Oversize material- will lead to erratically high tests. You just need to take more tests, so you have a reference if a few tests come up unusually high. Maybe two tests for every one you'd normally take.

Crushed rock- I agree. It always comes up, especially with the larger material- 1 1/4". The more you water it, the more fines wash out, the lower your numbers will be (because the proctor's actually lowered). I'd stick w/ a 5/8" CSTC if possible- more fines and always much easier to test.

Also, if you have any doubts, just talk with your inspector . Call him! Or just go to the site. If you have questions about the test, he should be able to give you more specific answers- if not, then maybe just ask for another inspector, for a reference. It is, after all, just business.

 

[iandig]

Just a few points to add to the comments above.
The gauge uses different sources to measure the density and the water content. The different sources penetrate the ground to different depths, neutrons are used to assess the moisture and they don't have the same penetration as the gamma photons for the density. We normally only expect the mosture 'assessment' to be limited to the upper 50mm of the layer under investigation, whereas the density is taken accross the full depth of the extended source.
Different minerology in the ground can have different effects on the 'assessment' of the density. PFA [FLY ASH]can dramatically effect the moisture reading, typically twice that of the lab determined MC. Granite has been shown to effect the density reading, but not the moisture 'assessment'. Strange readings can be that the density doesn't match the moisture, if this is the case it may be the material as much as the machine.
Please note the use of the word 'assessment', the gauge tells you what it 'thinks' it is, a calibration between the readings and the actual density should always be used. The ways in which the calibration can be carried out can vary depending upon the material, but in cases of doubt, undertaking a series of readings and actual measurements will help in the statistical analysis of the results.
One of the most common problems we have are flat batteries, especially re-chargable ones which after overuse can drain very rapidly. Moisture in the electrics is also a problem, as can be dust. If you intend to 'clean' the machine, dry air used for computers is very useful.

 

[dirtsqueezer]

I would just like to add a note on nuke gauges in general. I've been using different models of Troxler for 4 years. In the vast majority of projects I've been on, they have been more dependable than the vehicles I used to get me there. I've never had or heard of a radiation leak, or for that matter trouble with densities, with a few notable exceptions. If you have a specific problem, we could discuss it, but against a contractor with deleterious morals, I would consider the nuke gauge a better tool than possibly a lawer. I think the short discussion here reflects this. Best of luck.

 

[Focht3]

If someone else noted this, I apologize for the duplication - I scanned this thread since I'm short of time today.

[blue]Robert168[/blue]:
You can improve the accuracy of the measurements by running the test for a longer period of time. Personally, I don't like 15 second tests and usually push for two minute readings unless we are taking lots of densities on site and the results are almost all passing.

But even a two minute test contains some potential for error. That's why I always ask my technician, "How did the pin drive?" when they call in about an unexpected low test result. Ultimately, engineering judgment should come into play when unusual results are encountered, regardless of the method used to take the measurement.



Please see FAQ731-376 for great suggestions on how to make the best use of Eng-Tips Fora. See faq158-922 for recommendations regarding the question, "How Do You Evaluate Fill Settlement Beneath Structures?"

 

[Grouser]

Focht has hit the pin on the head!!!
The test is run when the pin is driven into the soil. The nuclear gauge gives numbers to record.
Here's what the inspector should be paying attention to:

1. Does the pin drive with the same resistance for the entire length(depth) of insertion?
Lower resistance with depth means the compactive effort is insufficient and/or the lift thickness is excessive. Increasing resistance with depth means sufficient or excessive compactive effort, or a change in (denser)material.

2. Does the pin drive with a similar resistance as the same material elsewhere on site?
Uniformity of resistance means consistant material and effort.

Is pin resistance similar for similar materials on different sites?
Again, uniformity of materials and effort for those of us blessed with processed materials used on a majority of sites in the area.

Density measurements include about 2 to 3 inches below the bottom of the probe, depending on the density. Looser or denser soil imediately below the probe can affect the reading. Hence a backscatter is inappropriate for lifts exceeding 2 to 4 inches. Try to have the test depth 2 inches less than the lift thickness.

Big vibratory compactors get maximum densification 2 to 3 feet below the surface, therefore it's not a bad idea to dig for tests. Also, have the vibration turned off for final passes on the top lift (static rolling) otherwise the top portion of the lift is loosened by the vibration (more prevalent on finer cohesionless material). Well graded material sticks together somewhat better.

Who will do a research project correlating vibration frequency with grain size or finess modulus?

Focht3
I'e found that density readings are not appreciably affected by test duration. Have you found otherwise? I believe it is the moisture reading that is more time dependent. The solution is to correct as compared to oven-dry as discussed in my previous post. The majority of soils tested in my area are lime or silica sands or soft limestone prepared aggregates. I have less recent experience with clays which appear to be more prevalent in your area. Maybe we need to start a new topic?

 

[Robert168]

Thank you all for your useful replies. Now I know that as long as the nuke is calibrated and used correctly, it is very "objective" on that shoot point.

Erroneous readings must have resulted from the difference between the compacted soil tested on and the soil used for proctor test. For example, when the soil contains more gravel at or near test depth, the reading will become higher as pointed out by dirtsqueezer.

By reading the answers above, I feel like I am an experienced inspector now. Thanks again.

Now I would like to put forward a "real" situation to see how you use your skills to evaluate the compaction by percentage. Hope I am not demanding too much.

A utilities trench was compacted with silty clay, traces of gravel with a max. dry density of 1780 kg/m3 at 14.2% of moist content. The specified compaction is 95%. 4 shoots were done and the densometer readings(dry density) were 1817, 1711, 1605 and 1724 kg/m3. How do you assess the compaction done on this trench?

 

[Focht3]

[blue]Grouser[/blue]:

I would generally agree, but that may have more to do with the problems with obtaining a "representative" moisture content. The statistics are about the same for both moisture and density when used with a standard count block.

I can't find the reference now, but Humboldt used to provide a table showing the error band for various count periods. It's worth reviewing. As I recall, a 15 second count could be off by more than 1 pcf and 0.5% - a significant amount if the test result was marginal. Jumping to a 2 minute reading improved things a good bit.

Someone check those numbers - I haven't looked at those materials in over 8 years, and can't seem to find them...



Please see FAQ731-376 for great suggestions on how to make the best use of Eng-Tips Fora. See faq158-922 for recommendations regarding the question, "How Do You Evaluate Fill Settlement Beneath Structures?"

 

[BigH]

I'd love to pass on a picture of the "get-up" they use here when doing a nuke test - but, alas, no pics . . .

 

[dirtsqueezer]

Robert168,

That is a fun question. I'll put on my blindfold and take a shot.

There are major problems with the results, the first of which is that one of your readings is greater than the max dry density! They call it 'max' for a reason... This tells me the proctor is not representative of your trench material- there is more oversize in the field than what you brought into the lab, or it's just a different material.

The second thing is the erratic compaction results. Given that and the result 'over' the max compaction, I'd say it's possible you just have too much cobble-sized material for a good test. If that's the case, I can tell you that the Federal Highways Administration, as hard-nosed as they are, do not test compaction on fill that has too much large rock (cobbles), as it has a tendancy to segregate , rendering representative testing impossible. Their solution is a 'method based' approach, as I've heard Focht3 talk about, in their case a minimum number of passes with specified compaction equipment. I could quote a some specific wording, if you'd like. If that's not the case, well do I get points for originality?

More information could render more guesses. What were the moisture readings for those tests? Material type? Depth of trench? Compaction equipment? Character of contractor? Don't worry about being demanding, this is all- for fun!

 

[BigH]

Dirtsqueezer - maximum dry density is a maximum??? Not really. Robert168 might be giving ASTM 698 - standard proctor. This is a maximum. But, if he had used ASTM 1558 - modified proctor - then his 100% standard would definitely had been exceeded by the 100%MDD of Modified Proctor. Then, if he pounds and pounds with heavy compaction equipment (not the trench scenario, I agree), then even his 100% MDD Modified might be exceeded - the higher the energy input, the higher the MDD and lower the OMC - right???

You are correct to a degree that the one "high" result might be due to different material, etc. but the other three are below MDD (which one???). I doubt that the gravel content is that critical in that he indicates only a "trace" gravel - trace implies less than 10% (at least where I come from). Silty clay is always difficult to compact - especially in trenches using light equipment. Need to knead the silty clay and most small equipment is vibration based.

I agree that for larger materials, the performance criteria is best - I always get a kick out of those who believe the numbers in compaction of a sandy gravel!!

In my years in the business, it is always interesting to see how much everyone knows about compaction. The client might be a chemical engineer - but he is an expert in compacation. Same for a client who is an accountant. And yet, we, in these forums show this isn't the case. Further, to back this up, if you pick up Fang's Foundation Handbook (chapter by Hilf) - you find that the chapter on compaction is the largest (perhaps second largest after site investigations) in the book. If compaction was so "easy" why is the compaction chapter so large and why do we spend so much time talking about compaction in these forums? There is always something "up" with compaction!! along with interesting insights of those in the field.

 

[Focht3]

In my years in the business, it is always interesting to see how much everyone knows about compaction.

Ain't that the truth!



Please see FAQ731-376 for great suggestions on how to make the best use of Eng-Tips Fora. See faq158-922 for recommendations regarding the question, "How Do You Evaluate Fill Settlement Beneath Structures?"

 

[Robert168]

By reading your kind replies, I am more "experienced" than before. Thanks.

 

[gbowen]

Big H

I am always amazed that people assume that anything over 100% compaction (modified or standard proctor)is impossible. I have clients that when given results exceeding 100% insist that we must have done something wrong.

My reply when I am up to explaining it is to use the example of a standard proctor and hypothetically double the number of lifts and hammer blows. Given the same material this hypothetical scenario should far exceed the maximum dry density of the standard proctor.

I also use a zero air voids curve to demonstrate the theorethical maximum.

 

[Focht3]

Agreed!



Please see FAQ731-376 for great suggestions on how to make the best use of Eng-Tips Fora. See faq158-922 for recommendations regarding the question, "How Do You Evaluate Fill Settlement Beneath Structures?"