Summary We intend (1) to show that formation characteristics affect the cement bond log (CBL) signal in most wells, (2) to prove that high-amplitude CBL's obtained in concentric casing-string configurations are often an artifact resulting paradoxically from the excellent paradoxically from the excellent quality of the cement job, and (3) to propose procedures for running CBL's propose procedures for running CBL's in concentric casings to eliminate this artifact. Experiments performed to study the influence of the cement/ formation or cement/external-casing interface show that, in all cases, these interfaces induce perturbations on the received waveform. The strength of these perturbations depends mainly on the impedance contrast at the interface. Field logs confirm the laboratory results and clearly show that high CBL amplitudes in well-cemented concentric casings are an artifact that could easily be eliminated by appropriate setting of the measurement windows. Introduction Primary cementing is one of the most critical and most difficult operations Primary cementing is one of the most critical and most difficult operations in the life of an oil well. Considerable effort has been and is still being devoted to improving all aspects of cementing, including postjob evaluation. Early postjob evaluation methods consisted exclusively of either pressure testing the casing or locating the top of cement behind the pressure testing the casing or locating the top of cement behind the casing. Since the development of CBL's in the early 1960'S, it has been possible to quantify the results of a cement job, within certain possible to quantify the results of a cement job, within certain measurement limitations, throughout the cemented interval. Recent experimental work proved that CBL attenuation rate is related more directly to the cement acoustic impedance than to the cement compressive strength. Classic CBL interpretation requires that the cement-sheath thickness be sufficient to ensure that no energy reflected from the cement external interface is superimposed on the measured Peak El amplitude. Until recently, 19 mm [0.75 in.] was considered a sufficient cement-sheath thickness; however, Ref. 6 showed that the necessary minimum thickness is a function of the acoustic properties of the cement. True values of minimum cement-sheath thickness can vary from 25.4 to 76 mm [1 to 3 in.]. Because these values are much larger than the average annular gap observed on most production strings, it is clear that in most cases soundenergy reflections at the cement external interface affect the CBL signal. This paper presents the results of extensive experimental work performed with realistic annuli. The influence of the cement external performed with realistic annuli. The influence of the cement external interface is investigated as a function of tool spacing, cement, and formation characteristics. Formations were simulated with cement formulations of known acoustic properties. In the case of concentric casings, experimental results are confirmed by field logs. From an analysis of these results, a simple modification of the CBL running procedure is proposed. This procedure effectively eliminates problems procedure is proposed. This procedure effectively eliminates problems related to narrow cement-sheath thickness in concentric casings. Experimental Setup Test Slurry. Cemoil API Class G cement was mixed at 1. 9 kg/L [15.8 lbm/gal]. The mixing procedure was as described in Ref. 6; a high-shear paddle was used to mix the slurry for 25 minutes before pouring the paddle was used to mix the slurry for 25 minutes before pouring the slurry into the annular gap. A sample was kept for pulse velocity measurements. The thickening time of this cement slurry was around 6 hours, estimated from the temperature increase inside the casing. During setting, pulse velocity measurements of the slurry were found unreliable, as shown in Ref. 7. For this reason, cement acoustic impedance is given only for times exceeding 8 hours. Synthetic Formations. These were simulated with special cement slurry designs. Three different slurries were designed and prepared to have lower or higher acoustic impedance than the test cement slurry. This was achieved by using an appropriate water/cement ratio combined with the proper amount of weighting agent or extender. After mixing, the slurry was proper amount of weighting agent or extender. After mixing, the slurry was poured into the annular space between a 160-mm [6.3-in.]-OD PVC pipe and a poured into the annular space between a 160-mm [6.3-in.]-OD PVC pipe and a 310-mm [12.2-in.] -ID steel pipe. A sample of each slurry was kept for acoustic impedance measurements. When the cement was set, the PVC pipe was removed, leaving a hollow cylinder of synthetic formation 75 mm [2.95 in.] thick. The complete assembly was then kept saturated with water for at least 3 months. This ensured that the synthetic formation material was fully hydrated with stable properties for the final tests. Table 1 gives the relevant properties of each synthetic formation at the time of the tests. The weakest formation (Test F3) had an acoustic impedance of 3.45 Mrayl (see SI Metric Conversion Factors), similar to some unconsolidated shales. The remaining two formations (Tests F1 and F2) had acoustic impedances of 8.3 and 9.5 Mrayl, respectively, comparable with a reservoir rock with 25% porosity. Test Cells. These consisted of casing sections 1.67 m [5.5 ft] long immersed in a water bath 1.52 m [5 ft] deep. The water bath can hold up to seven test cells at a temperature controllable between 20 and 70 degrees C [68 and 158 degrees F]. JPT P. 1158