The only time that we ever use anything over 40 db is for flooded members in the water, everything else that we do is around the low 20's.  Not sure why he's going that high for, but everyone learns a different way.

Charles

Better check that guy papers, might be a fake.

I do know one thing - it is a serious matter to challenge the owner's UT tech.  It's even worse if it turns out he was right ("egg-on-face" that is).

Since the owner hired the tech, he might not have to produce documentation for your satisfaction, nor show that he has a procedure.  Afterall, he is satisfying his client and not you.  The same party that you need to satisfy.

Having said that, I think your best bet is to diplomatically present your case to the owner. 
You have consulted your Level III and his recommended Level II scanned the work with acceptable results.  Your guy with 10+ years of UT experience found acceptable results. Something is not quite right.  "Naturally, you want to produce a good product but you don't want to chase issues that might not be valid."

It would be great if you can have your Level III check it himself.

If the suggestion to remove the backing and re-scan is practical for you, I would do so.  Then if you get good results, you can chalk that up to good faith effort to resolve the problem - good teamwork on your part.  But if you have a lot more CJP joints to do, and the owner then wants all backing removed, naturally there would be a change-order to cover your additional costs.

My USN58 is set with the reference level at 64dB set at 50% FSH, and I scan according to the chart at the bottom of the page of Table 6.2 which is dependant on the sound path length.....through 2.5 inches = 14db above my 64db ref.

My older USN50 had a ref level of 47dB at 50% FSH.....

So to answer your question, yes, I have my newer machine set a bit higher than your TPI.

Maybe get your UT guy and the TPI to set up(calibrate) off the same IIW block, I mean check both verify the BIP, verify the Angle, verify the Distance, and verify the Sensitivity and then see where each machine ends up....note the differences and go inspect the joint again.

edit: Someone else mentioned that he might be calling the backing bar...maybe...so use the sound path and some trig(just bypass the trig in the machine for this situation)...map it out and see where the sound is going. I use ACAD when in doubt, but graph paper will work as long as you keep everything to scale.

Based on what your stating here, my money is on the owners TPI findings. Scanning at 24DB??? Having said that, the TPI may be closer to the truth, but something smells fishy about them as well. In all, it may behoove you to contact another independent group to assess the group you have been using. I have to call BS on 24DB scanning level. Not to mention this statement "plus 14 Db is too high". The code is an equal opportunity offender, and not subject to technician alteration due to opinion. It can only change if both parties agree to it in which case its not 'D1.1' UT.

Just wondering if you have resolved this issue yet, and what you discovered?

I would have to agree that the 24dB is pretty low for scanning. Although technically... you should be able to scan at reference level as long as you are slow and careful.  For what it's worth, I do have some analog scopes that come in lower than the digitals, and I have a couple of transducers from NDT Instruments that are hotter than my KB ones.... so equipment does make a difference.

Besides accurate plotting, it might be a good idea to size down the transducer and bump the Mhz in order to evaluate the indication better. the Code will allow this, with the Level III's blessing.  Let's face it, a 5/8 X 5/8 probe is a big beam for 3/8" material.... Maybe a 1/4" or 3/8" 5 MHz would be a wiser selection for evaluation.

~thirdeye~

Where is the written procedure?

Since you posted in the D1 section and you reference D1.1, I assume you are testing to the requirements of clause 6. The procedures to be followed are delineated in D1.1 with directions on the calibration of the test instrument, frequency of calibration, etc. The scanning technique is spelled out as well as the db to be used for scanning once the reference level has been determined while setting the instrument up on the 1/16 inch diameter side drilled hole. Setting sensitivity on the reference hole will determine the reference level to which all reflectors are compared. Each machine is different and the type of transducer and wedge will influence the reference level.

The rating system is used to either accept or reject the indication based on the amplitude of the signal relative to reference. If the technician cannot show you how he calculated the rating, it is time to find another technician. Make sure the technician is using the correct table for the weld service class, i.e., tension, compression, or cyclic loading.

I map out indications when there is any doubt as to what is causing the reflection. I use my AutoCad program on my laptop to map exactly what the source of the signal is. It takes time, but when there is a question regarding the source of a reflector, my time is less expensive than the cost of cutting out a "good" weld and making a new weld. It provides the client with the warm fuzzy feeling that I have half an idea of what I'm doing. It helps to build my client's confidence that I am not calling good welds bad and inccuring needless additional expense and lost production time.  The technician should be able to map the location of the reflector if he is payng attention to his technique and plots the surface distance from the transducer as well as the depth. That should provide a good idea of whether the reflector is the backing or something other than the backing.

Best regards  - Al

I agree with Joe. Some of the old USK-7 machines would ref. at around 24 or so, but the scan level would be at least 38 depending on the sound path. Most of the newer machines will ref. at around 45 to 65 depending on the machine, transducer, cable, tempature, couplant, ect. Either way you have to scan at least 14db over your ref. level. The excuse of too much grass is BS. Once a signal is detected, the gain will be turned back down to evaluate the indication based on the set ref. level. Indications are easily skipped when the scan level is too low. The owners guy  sounds like he may be closer to being right, however, he should always have a copy of his procedures in the truck with him. All NDT labs should have an SOP (Standard Operating Procedure) manual which each inspector should have a copy of. If he can't at least produce his SOP within a day or two, something is wrong about that lab. Labs should have an SOP which contains a written practice outlining their training and certification procedures. All of this is required to be performing NDT in accordance with ASNT SNT-TC-1A. If the guy is rejecting your work, you have every right to ask for a copy of his certifications, procedures, and written practice.
I have heard the backing bar thing many times. If the flaw is detected in the first leg, it is not the backing bar. If the flaw is detected in the first part of the second leg, the inspector should back up to see if it can be seen in the third leg. If it can be seen in the third leg, it is not the backing bar. Also, the weld can be checked from face A and face B, or the weld can be ground smooth to see if the indication can be seen in the first leg.

I would suggest that all parties involved sit down with the code book and look at what it says. It really does not matter where you start with (your reference db) what matters is correctly adding the amout of gain (db) to do your scanning.  When you find an indication it is compared to the reference db, at that point scanning db is a moot point.  After the indication is brought to reference, that number becomes what you subtract reference from, and then attenuation from, to achieve your rating. (a-b-c= D) If the A-B-C = D thing throws you, those letters come from the D1.1/D1.5 report form.
The main thing is that when one tech says accept and another says reject, its time to sit down and figure out if everyone is on the same page.
Hope this helps.
D