Joseph Audio RM33si Signature loudspeaker JA Follow-up

John Atkinson wrote a Follow-up in January 2002 (Vol.25 No.1)

When Chip Stern reviewed the floorstanding, $7499/pair RM33si loudspeaker from Joseph Audio in October 2001 (pp.181-187), he was about as enthusiastic as it is possible to get: "Jeff Joseph and Richard Modaferri have achieved something special with the RM33si—a design that might in time prove to be a standard by which other speakers are compared." Yet when I performed my usual set of measurements, I was bothered by finding the Joseph's voltage sensitivity much lower than the 88dB/2.83V/m specified. I was also concerned by a measured lack of integration between the midrange unit and the woofer.

I wondered at the time if Chip had damaged something when he clipped his Musical Fidelity amplifier on Telarc's infamous 1812 cannons. Jeff Joseph therefore brought over a second pair of RM33sis that he could vouch for performing as specified, so that I could do a Follow-Up. (The serial numbers of Chip's samples were 33-006L & R; of my samples, 33-014L & R.)

The impedance curves of the two samples (fig.1 is one of Chip's samples, fig.2 is one of the replacements) are broadly similar, but differ in details. In particular, the Chip sample dips to a minimum value of 6 ohms at 450Hz, while the new sample's minimum is a significantly higher 7.8 ohms at 300Hz. The tuning of the base-mounted port seemed identical at 23Hz, but again the impedance at the tuning frequency is higher: 14.2 ohms vs 11.9 ohms. Perhaps headbanger Chip had indeed damaged a component. Nevertheless, the speaker's sensitivity still measured a low 81.5dB(B)/2.83V/m, 1dB lower than the BBC LS3/5A minimonitor. This speaker needs a muscular amplifier if it is to be driven to musically appropriate levels.

Fig.1 Joseph RM33si, sample 1, electrical impedance (solid) and phase (dashed). (2 ohms/vertical div.)

Fig.2 Joseph RM33si, sample 2, electrical impedance (solid) and phase (dashed). (2 ohms/vertical div.)

Turning next to the integration between the drive-units, fig.3 shows the individual responses of the sample I measured for Chip's review, taken in the nearfield using the tiny Mitey Mike II microphone and DRA Labs MLSSA system. (The speaker was raised from the floor on 1" cones for the measurements, and the relative levels in this graph have been adjusted in the ratio of the radiating diameters, as recommended by DRA Labs' Doug Rife.)

Fig.3 Joseph RM33si, sample 1, nearfield midrange, woofer, and port responses.

As I noted in October, the port output is somewhat suppressed, and extends quite high in frequency. The midrange unit rolls off with a steep slope below a small peak at 125Hz but with a slight shelf centered on 55Hz, the frequency of a low-level cabinet resonance. The side-mounted woofer peaks a little between 40Hz and 50Hz, before starting a very steep rolloff above the lower crossover frequency of 125Hz. The outputs of all three radiators overlap between 75Hz and 200Hz, which will make the speaker's behavior complex in this range.

Fig.4 shows how these individual responses add, taking into account both acoustic phase and the difference in path length from each drive-unit to a nominal farfield listening position. Above 500Hz, this complex sum is spliced to the true farfield response, averaged across a 30 degrees horizontal window on the tweeter axis. As I noted in October, the RM33si's response in the midrange and treble is extraordinarily smooth, though with a slight rising trend apparent. In the bass, however, there appears to be a lack of energy in the transition region between the woofer and midrange, and an uneven character in the woofer's and port's passbands before the speaker rolls off to reach -12dB at 20Hz.

Fig.4 Joseph RM33si, sample 1, anechoic response on tweeter axis at 50", averaged across 30 degrees horizontal window and corrected for microphone response, with complex sum of nearfield midrange, woofer, and port responses plotted below 350Hz.