Patients with severely impaired high-frequency hearing and sufficient residual low-frequency hearing can be provided with a cochlear implant (CI), thereby facilitating ipsilateral electric and acoustic stimulation with established advantages over electric stimulation alone. However, partial or complete hearing loss often occurred after implantation due to, inter alia, acute mechanical trauma to cochlear structures during electrode insertion. Possibilities of intraoperative monitoring using electrocochleography (ECochG) have recently been studied in CI patients, primarily using the ongoing response to low-frequency tone bursts consisting of the cochlear microphonic (CM) and the auditory nerve neurophonic. By contrast, the transient neural response to tone bursts, that is, compound action potential (CAP), was generally less detectable or less sensitive as a monitoring measure, thus falling short of providing useful contribution to electrocochleography analysis. In this study, we investigate using chirps to evoke more robust CAP responses in a limited frequency band by synchronizing neural firing, and thereby improving CAP sensitivity to mechanical trauma in a guinea pig model of cochlear implantation. Stimuli were band-limited between 100 Hz and 10 kHz to investigate their frequency range selectivity as a preliminary model for low-frequency hearing. They were constructed by adding a harmonic series either with zero phase delay (click) or by adjusting the phase delay at a rate that is inversely related to a traveling wave delay model (chirp), with three different parameters to examine level-dependent delay compression. The amplitude spectrum was thus identical between stimuli with differences only in phase. In Experiment 1, we compared input-output functions recorded at the round window in normal-hearing guinea pigs and implemented a high-pass noise masking paradigm to infer neural contribution to the CAP. In Experiment 2, guinea pigs were implanted with a custom-built CI electrode using a motorized micromanipulator. Acute mechanical trauma was simulated during the electrode insertion. At each insertion step, CAP and CM responses were measured at the round window for the following stimuli: broad-band click, band-limited click, and band-limited chirps (3 parameters), and tone bursts at frequencies 1, 2, 4, and 8 kHz. Chirps compared with the equal-band click showed significantly lower thresholds and steeper slopes of sigmoid-fitted input-output functions. The shorter chirp evoked significantly larger amplitudes than click when compared at equal sensation level. However, the click evoked larger amplitudes than chirps at higher levels and correspondingly achieved larger saturation amplitudes. The results of the high-pass noise masking paradigm suggest that chirps could efficiently synchronize neural firing in their targeted frequency band, while the click recruited more basal fibers outside its limited band. Finally, monitoring sensitivity during electrode insertion, defined as relative amplitude change per unit distance, was higher for chirp-evoked CAP and tone burst-evoked CM, but smaller for CAP responses evoked by clicks or tone bursts. The chirp was shown to be an efficient stimulus in synchronizing neural firing for a limited frequency band in the guinea pig model. This study provides a proof of principle for using chirp-evoked CAP as a comprehensive neural measure in CI patients with residual hearing.