Questionable Modified Proctor test

[sdh7711]

I had a contractor using a backfill material that was tested using the modified proctor method yielding an optimum water content of 6.8% and max dry unit weight of 21.28kN/m3. They recently "retested" the backfill material getting a new corrected wc of 10.7% and max dry unit weight of 19.94 kN/m3. Thats a pretty drastic change in optimum wc and a drop in the max dry unit weight on the same material in my opinion. I am now getitng compaction test results with many coming above 100%, which itself not a huge deal, but several have been above 102% and even 103%. I am questioning these results now and while I am not concerned with the values around 100% I am worried that any test that resulted in a 95% to 97% percent compaction are not "true" values and may be lower than the specification requirement of 95% compaction. Thoughts?

 

[jgailla]

What type of soil is it? There are soils with double highs, although I haven't seen any.

 

[EireChch]

If you feel that way, you could re check with a Nuclear Densometer.

Thats of course if it is possible, maybe concrete has been poured on top or whatever. If that is the case you might be best letting sleeping dogs ly!

If additional fill has been placed, then excavating down to maybe 1.5m and preforming some Nukes might be an option.

Even if the "true" values were 90% they may have gained density over time due to additional fill on top providing a surcharge or dissipation of porewater pressure (if wetter than optimum).

If more than 1.5m of fill has been placed, again let sleeping dogs ly....Its only going to be marginally below the true 95% value (hopefully) and will that have much of an effect.

 

[sdh7711]

Its structural fill gravel with sand and some silt present. Previous and current modified proctor tests showed standard bell curves no double dips. 90% passing 2" sieve, 82% passing 1.5", 66% passing 1", 44% passing No 4, 11.5% passing the 200 sieve. LL was 28 and PL was 21 with a PI of 7. They had been having field labor remove "plus size" aggregate during compaction as I already brought that up as an issue and now are getintg a screen on site to do it more thoroughly.

All the compaction test results were recorded wiht the nuclear denso.

 

[dik]

With the discrepancy, I'd redo the testing; you should not have that variance.

Dik

 

[oldestguy]

What size lab compaction molds and what "gravel correction"? I'd do some double checking with the large size sand cone method. How was the nuc calibrated?

 

[sdh7711]

The mold volume was 2131.1 cubic cm or 130 cubic inch. Done to ASTM D1557-12 and procedure C was used. Didnt see what if any gravel correction on the data sheet. Dont have the calibration of the Nuc readily available but when they were doing compaction with the prior testing densities were in the high 80s (required recompaction) up to 96 and 97% with only two tests getting over 100% none over 102%. While this is no subsititue for seeing the claibration method and results I dont think the nuc is off any.

 

[Ron]

Getting compaction results of over 100 percent is not unusual if using a well graded material, good compaction methods and the moisture content is at or very near optimum. That doesn't concern me unless the results are routinely in the 104-105% range.

Soils as you have described can vary in their Proctor value rather easily because of variations in the particle sizes at any location.

Depending on the area to be tested, multiple Proctor tests might be required.

 

[Mad Mike]

My chime is that too often a variable back-fill material is classified based on 1 density test, which is then carried over to every density check conducted. Silt is relatively light, stone (gravel) can be relatively dense and the various ratios of these particles can throw your density testing off.

My first thought would be that the Mod. density of the material varies more than was anticipated, be it due to variable gravel content or otherwise. I would get a couple of extra tests run and use an average Mod. if possible.

All the best,
Mike

 

[Okiryu]

Are these 2 proctor results from the same lab? You may need to use an independent 3rd party lab to validate the results.

 

[darthsoilsguy2]

mod proctors are harder to check-point in the field. i don't like mods specified for structural fill.

 

[Ron]

dsg2...depends on the soil type. In the coastal plains soils that I deal with, check points usually work fine, particularly if the fines content is a bit higher than in clean fine sands. Our sands are poorly graded (SP) and can sometimes yield wacky Proctor results, but overall pretty good. We don't typically use the standard Proctor except in non-structural embankment soils.

Assuming you work mostly in glacial till and residual soils, you probably have sharply peaked Modified Proctor curves, high unit weights and little room for moisture ranges. Let me know if I'm off base....I'm doing some writing and still have time to change a few things if necessary!

 

[darthsoilsguy2]

Ron, i was thinking more from the practical perspective of what the lab is set up for. i worked for a few different labs that did mostly commercial and university projects (not as much DOT stuff) where std still gets specified more..... and we stocked 2 or 3 mod proctor kits and those mod kits would spend the time on the lab floor or the lab closet and would mostly get used for road gravel testing in lab. We had all the field tech trucks stocked up with personal std proctor kits and all that equipment takes up space with the concrete stuff too.... if we did have a mod proctor job, it would be an opportunity for error if the tech runs a std proctor and an opportunity for error if the tech wasn't prepared for the job to skip the check point and rely solely on the preliminary lab testing. Also, our trucks were stocked with concrete reaction blocks to beat it against since there is no concrete slab in the field. With the longer drop height of the mod hammer and the reaction block height, the overall apparatus gets to be taller and unwieldy so the technician's technique could be picking up a little more noise on the precision and bias side too..... maybe the mod proctor is more common elsewhere, but when i see it i just think 'ugh, what a pain in the butt' cause it would be a little bit of a curveball for most of the jobs we would do.

 

[Ron]

dsg2...that's interesting. I spent about 30 years of my career thus far in testing lab work....18 years with a major international geotech and materials firm, 5 years with medium size regional firms and 7 years owning my own testing and consulting firm.....ours setups were exactly the opposite! Our work was predominately Modified with the little standard Proctor hammer gathering dust in the cabinet. Our other offices in the southeast Piedmont regions and in the Ky/Tn region were more like yours.

 

[darthsoilsguy2]

Haha Ron, i've got 10 cumulative years in testing labs (when i wasn't in design) and those labs were based in North Carolina and would cross over to VA and SC occasionally. As i was writing my last post and got to the end, i put a ".... maybe the mod proctor is more common elsewhere" right about that time i was thinking to myself 'Hmm... maybe NC is a weirdo outlier on this thing.... that kinda sounds Right. Wouldn't be the 1st time, at all'